ASP.NET Core Documentation<a class="headerlink" href="#asp-net-core-documentation" title="Permalink to this headline">¶

This tutorial will teach you the basics of building an ASP.NET Core MVC web app using Visual Studio 2015.

Install Visual Studio and .NET Core¶

Install Visual Studio Community 2015. Select the Community download and the default installation. Skip this step if you have Visual Studio 2015 installed.
- Visual Studio 2015 Home page installer
Install .NET Core + Visual Studio tooling

Create a web app¶

From the Visual Studio Start page, tap New Project.

Alternatively, you can use the menus to create a new project. Tap File > New > Project.

Complete the New Project dialog:

In the left pane, tap Web
In the center pane, tap ASP.NET Core Web Application (.NET Core)
Name the project “MvcMovie” (It’s important to name the project “MvcMovie” so when you copy code, the namespace will match. )
Tap OK

Complete the New ASP.NET Core Web Application - MvcMovie dialog:

Tap Web Application
Clear Host in the cloud
Tap OK.

Visual Studio used a default template for the MVC project you just created, so you have a working app right now by entering a project name and selecting a few options. This is a simple “Hello World!” project, and it’s a good place to start,

Tap F5 to run the app in debug mode or Ctl-F5 in non-debug mode.

Visual Studio starts IIS Express and runs your app. Notice that the address bar shows localhost:port# and not something like example.com. That’s because localhost always points to your own local computer, which in this case is running the app you just created. When Visual Studio creates a web project, a random port is used for the web server. In the image above, the port number is 1234. When you run the app, you’ll see a different port number.
Launching the app with Ctrl+F5 (non-debug mode) allows you to make code changes, save the file, refresh the browser, and see the code changes. Many developers prefer to use non-debug mode to quickly launch the app and view changes.
You can launch the app in debug or non-debug mode from the Debug menu item:

You can debug the app by tapping the IIS Express button

The default template gives you working Home, Contact, About, Register and Log in links. The browser image above doesn’t show these links. Depending on the size of your browser, you might need to click the navigation icon to show them.

In the next part of this tutorial, we’ll learn about MVC and start writing some code.

Adding a controller¶

The Model-View-Controller (MVC) architectural pattern separates an app into three main components: the Model, the View, and the Controller. The MVC pattern helps you create apps that are testable and easier to maintain and update than traditional monolithic apps. MVC-based apps contain:

Models: Classes that represent the data of the app and that use validation logic to enforce business rules for that data. Typically, model objects retrieve and store model state in a database. In this tutorial, a Movie model retrieves movie data from a database, provides it to the view or updates it. Updated data is written to a SQL Server database.
Views: Views are the components that display the app’s user interface (UI). Generally, this UI displays the model data.
Controllers: Classes that handle browser requests, retrieve model data, and then specify view templates that return a response to the browser. In an MVC app, the view only displays information; the controller handles and responds to user input and interaction. For example, the controller handles route data and query-string values, and passes these values to the model. The model might use these values to query the database.

The MVC pattern helps you create apps that separate the different aspects of the app (input logic, business logic, and UI logic), while providing a loose coupling between these elements. The pattern specifies where each kind of logic should be located in the app. The UI logic belongs in the view. Input logic belongs in the controller. Business logic belongs in the model. This separation helps you manage complexity when you build an app, because it enables you to work on one aspect of the implementation at a time without impacting the code of another. For example, you can work on the view code without depending on the business logic code.

We’ll be covering all these concepts in this tutorial series and show you how to use them to build a simple movie app. The following image shows the Models, Views and Controllers folders in the MVC project.

In Solution Explorer, right-click Controllers > Add > New Item... > MVC Controller Class

In the Add New Item dialog, enter HelloWorldController.

Replace the contents of Controllers/HelloWorldController.cs with the following:

using Microsoft.AspNetCore.Mvc;
using System.Text.Encodings.Web;

namespace MvcMovie.Controllers
{
    public class HelloWorldController : Controller
    {
        // 
        // GET: /HelloWorld/

        public string Index()
        {
            return "This is my default action...";
        }

        // 
        // GET: /HelloWorld/Welcome/ 

        public string Welcome()
        {
            return "This is the Welcome action method...";
        }
    }
}

Every public method in a controller is callable as an HTTP endpoint. In the sample above, both methods return a string. Note the comments preceding each method:

public class HelloWorldController : Controller
{
    // 
    // GET: /HelloWorld/

    public string Index()
    {
        return "This is my default action...";
    }

    // 
    // GET: /HelloWorld/Welcome/ 

    public string Welcome()
    {
        return "This is the Welcome action method...";
    }
}

The first comment states this is an HTTP GET method that is invoked by appending “/HelloWorld/” to the base URL. The second comment specifies an HTTP GET method that is invoked by appending “/HelloWorld/Welcome/” to the URL. Later on in the tutorial we’ll use the scaffolding engine to generate HTTP POST methods.

Run the app in non-debug mode (press Ctrl+F5) and append “HelloWorld” to the path in the address bar. (In the image below, http://localhost:1234/HelloWorld is used, but you’ll have to replace 1234 with the port number of your app.) The Index method returns a string. You told the system to return some HTML, and it did!

MVC invokes controller classes (and the action methods within them) depending on the incoming URL. The default URL routing logic used by MVC uses a format like this to determine what code to invoke:

/[Controller]/[ActionName]/[Parameters]

You set the format for routing in the Startup.cs file.

    app.UseMvc(routes =>
    {
        routes.MapRoute(
            name: "default",
            template: "{controller=Home}/{action=Index}/{id?}");
    });
}

When you run the app and don’t supply any URL segments, it defaults to the “Home” controller and the “Index” method specified in the template line highlighted above.

The first URL segment determines the controller class to run. So localhost:xxxx/HelloWorld maps to the HelloWorldController class. The second part of the URL segment determines the action method on the class. So localhost:xxxx/HelloWorld/Index would cause the Index method of the HelloWorldController class to run. Notice that we only had to browse to localhost:xxxx/HelloWorld and the Index method was called by default. This is because Index is the default method that will be called on a controller if a method name is not explicitly specified. The third part of the URL segment ( id) is for route data. We’ll see route data later on in this tutorial.

Browse to http://localhost:xxxx/HelloWorld/Welcome. The Welcome method runs and returns the string “This is the Welcome action method...”. For this URL, the controller is HelloWorld and Welcome is the action method. We haven’t used the [Parameters] part of the URL yet.

Let’s modify the example slightly so that you can pass some parameter information from the URL to the controller (for example, /HelloWorld/Welcome?name=Scott&numtimes=4). Change the Welcome method to include two parameters as shown below. Note that the code uses the C# optional-parameter feature to indicate that the numTimes parameter defaults to 1 if no value is passed for that parameter.

public string Welcome(string name, int numTimes = 1)
{
    return HtmlEncoder.Default.Encode($"Hello {name}, numTimes: {numTimes}");
}

Note

The code above uses HtmlEncoder.Default.Encode to protect the app from malicious input (namely JavaScript). It also uses Interpolated Strings.

Note

In Visual Studio 2015, when you are running in IIS Express without debugging (Ctl+F5), you don’t need to build the app after changing the code. Just save the file, refresh your browser and you can see the changes.

Run your app and browse to:

http://localhost:xxxx/HelloWorld/Welcome?name=Rick&numtimes=4

(Replace xxxx with your port number.) You can try different values for name and numtimes in the URL. The MVC model binding system automatically maps the named parameters from the query string in the address bar to parameters in your method. See Model Binding for more information.

In the sample above, the URL segment (Parameters) is not used, the name and numTimes parameters are passed as query strings. The ? (question mark) in the above URL is a separator, and the query strings follow. The & character separates query strings.

Replace the Welcome method with the following code:

public string Welcome(string name, int ID = 1)
{
    return HtmlEncoder.Default.Encode($"Hello {name}, id: {ID}");
}

Run the app and enter the following URL: http://localhost:xxx/HelloWorld/Welcome/3?name=Rick

This time the third URL segment matched the route parameter id. The Welcome method contains a parameter id that matched the URL template in the MapRoute method. The trailing ? (in id?) indicates the id parameter is optional.

app.UseMvc(routes =>
{
    routes.MapRoute(
        name: "default",
        template: "{controller=Home}/{action=Index}/{id?}");
});

In these examples the controller has been doing the “VC” portion of MVC - that is, the view and controller work. The controller is returning HTML directly. Generally you don’t want controllers returning HTML directly, since that becomes very cumbersome to code and maintain. Instead we’ll typically use a separate Razor view template file to help generate the HTML response. We’ll do that in the next tutorial.

Adding a view¶

In this section you’re going to modify the HelloWorldController class to use Razor view template files to cleanly encapsulate the process of generating HTML responses to a client.

You’ll create a view template file using Razor. Razor-based view templates have a .cshtml file extension, and provide an elegant way to create HTML output using C#. Razor seemlessly blends C# and HTML, minimizing the number of characters and keystrokes required when writing a view template, and enables a fast, fluid coding workflow.

Currently the Index method returns a string with a message that is hard-coded in the controller class. Change the Index method to return a View object, as shown in the following code:

public IActionResult Index()
{
    return View();
}

The Index method above uses a view template to generate an HTML response to the browser. Controller methods (also known as action methods) such as the Index method above, generally return an IActionResult (or a class derived from ActionResult), not primitive types like string.

Right click on the Views folder, and then Add > New Folder and name the folder HelloWorld.
Right click on the Views/HelloWorld folder, and then Add > New Item.
In the Add New Item - MvcMovie dialog
- In the search box in the upper-right, enter view
- Tap MVC View Page
- In the Name box, keep the default Index.cshtml
- Tap Add

Replace the contents of the Views/HelloWorld/Index.cshtml Razor view file with the following:

@{
    ViewData["Title"] = "Index";
}

<h2>Index</h2>

<p>Hello from our View Template!</p>

Navigate to http://localhost:xxxx/HelloWorld. The Index method in the HelloWorldController didn’t do much work; it simply ran the statement return View();, which specified that the method should use a view template file to render a response to the browser. Because you didn’t explicitly specify the name of the view template file to use, MVC defaulted to using the Index.cshtml view file in the /Views/HelloWorld folder. The image below shows the string “Hello from our View Template!” hard-coded in the view.

If your browser window is small (for example on a mobile device), you might need to toggle (tap) the Bootstrap navigation button in the upper right to see the to the Home, About, and Contact links.

Changing views and layout pages¶

Tap on the menu links (MvcMovie, Home, About). Each page shows the same menu layout. The menu layout is implemented in the Views/Shared/_Layout.cshtml file. Open the Views/Shared/_Layout.cshtml file.

Layout templates allow you to specify the HTML container layout of your site in one place and then apply it across multiple pages in your site. Find the @RenderBody() line. RenderBody is a placeholder where all the view-specific pages you create show up, “wrapped” in the layout page. For example, if you select the About link, the Views/Home/About.cshtml view is rendered inside the RenderBody method.

Change the title and menu link in the layout file¶

Change the contents of the title element. Change the anchor text in the layout template to “MVC Movie” and the controller from Home to Movies as highlighted below:

<!DOCTYPE html>
<html>
<head>
    <meta charset="utf-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>@ViewData["Title"] - Movie App </title>

    <environment names="Development">
        <link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />
        <link rel="stylesheet" href="~/css/site.css" />
    </environment>
    <environment names="Staging,Production">
        <link rel="stylesheet" href="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.6/css/bootstrap.min.css"
              asp-fallback-href="~/lib/bootstrap/dist/css/bootstrap.min.css"
              asp-fallback-test-class="sr-only" asp-fallback-test-property="position" asp-fallback-test-value="absolute" />
        <link rel="stylesheet" href="~/css/site.min.css" asp-append-version="true" />
    </environment>
</head>
<body>
    <div class="navbar navbar-inverse navbar-fixed-top">
        <div class="container">
            <div class="navbar-header">
                <button type="button" class="navbar-toggle" data-toggle="collapse" data-target=".navbar-collapse">
                    <span class="sr-only">Toggle navigation</span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                </button>
                <a asp-area="" asp-controller="Movies" asp-action="Index" class="navbar-brand">MvcMovie</a>
            </div>
            <div class="navbar-collapse collapse">
                <ul class="nav navbar-nav">
                    <li><a asp-area="" asp-controller="Home" asp-action="Index">Home</a></li>
                    <li><a asp-area="" asp-controller="Home" asp-action="About">About</a></li>
                    <li><a asp-area="" asp-controller="Home" asp-action="Contact">Contact</a></li>
                </ul>
            </div>
        </div>
    </div>
    <div class="container body-content">
        @RenderBody()
        <hr />
        <footer>
            <p>&copy; 2016 - MvcMovie</p>
        </footer>
    </div>

Warning

We haven’t implemented the Movies controller yet, so if you click on that link, you’ll get a 404 (Not found) error.

Save your changes and tap the About link. Notice how each page displays the Mvc Movie link. We were able to make the change once in the layout template and have all pages on the site reflect the new link text and new title.

Examine the Views/_ViewStart.cshtml file:

@{
    Layout = "_Layout";
}

The Views/_ViewStart.cshtml file brings in the Views/Shared/_Layout.cshtml file to each view. You can use the Layout property to set a different layout view, or set it to null so no layout file will be used.

Now, let’s change the title of the Index view.

Open Views/HelloWorld/Index.cshtml. There are two places to make a change:

The text that appears in the title of the browser

The secondary header (<h2> element).

You’ll make them slightly different so you can see which bit of code changes which part of the app.

@{
    ViewData["Title"] = "Movie List";
}

<h2>My Movie List</h2>

<p>Hello from our View Template!</p>

ViewData["Title"] = "Movie List"; in the code above sets the Title property of the ViewDataDictionary to “Movie List”. The Title property is used in the <title> HTML element in the layout page:

<title>@ViewData["Title"] - Movie App</title>

Save your change and refresh the page. Notice that the browser title, the primary heading, and the secondary headings have changed. (If you don’t see changes in the browser, you might be viewing cached content. Press Ctrl+F5 in your browser to force the response from the server to be loaded.) The browser title is created with ViewData["Title"] we set in the Index.cshtml view template and the additional “- Movie App” added in the layout file.

Also notice how the content in the Index.cshtml view template was merged with the Views/Shared/_Layout.cshtml view template and a single HTML response was sent to the browser. Layout templates make it really easy to make changes that apply across all of the pages in your application. To learn more see Layout.

Our little bit of “data” (in this case the “Hello from our View Template!” message) is hard-coded, though. The MVC application has a “V” (view) and you’ve got a “C” (controller), but no “M” (model) yet. Shortly, we’ll walk through how to create a database and retrieve model data from it.

Passing Data from the Controller to the View¶

Before we go to a database and talk about models, though, let’s first talk about passing information from the controller to a view. Controller actions are invoked in response to an incoming URL request. A controller class is where you write the code that handles the incoming browser requests, retrieves data from a database, and ultimately decides what type of response to send back to the browser. View templates can then be used from a controller to generate and format an HTML response to the browser.

Controllers are responsible for providing whatever data or objects are required in order for a view template to render a response to the browser. A best practice: A view template should never perform business logic or interact with a database directly. Instead, a view template should work only with the data that’s provided to it by the controller. Maintaining this “separation of concerns” helps keep your code clean, testable and more maintainable.

Currently, the Welcome method in the HelloWorldController class takes a name and a ID parameter and then outputs the values directly to the browser. Rather than have the controller render this response as a string, let’s change the controller to use a view template instead. The view template will generate a dynamic response, which means that you need to pass appropriate bits of data from the controller to the view in order to generate the response. You can do this by having the controller put the dynamic data (parameters) that the view template needs in a ViewData dictionary that the view template can then access.

Return to the HelloWorldController.cs file and change the Welcome method to add a Message and NumTimes value to the ViewData dictionary. The ViewData dictionary is a dynamic object, which means you can put whatever you want in to it; the ViewData object has no defined properties until you put something inside it. The MVC model binding system automatically maps the named parameters (name and numTimes) from the query string in the address bar to parameters in your method. The complete HelloWorldController.cs file looks like this:

using Microsoft.AspNetCore.Mvc;
using System.Text.Encodings.Web;

namespace MvcMovie.Controllers
{
    public class HelloWorldController : Controller
    {
        public IActionResult Index()
        {
            return View();
        }

        public IActionResult Welcome(string name, int numTimes = 1)
        {
            ViewData["Message"] = "Hello " + name;
            ViewData["NumTimes"] = numTimes;

            return View();
        }
    }
}

The ViewData dictionary object contains data that will be passed to the view. Next, you need a Welcome view template.

Right click on the Views/HelloWorld folder, and then Add > New Item.
In the Add New Item - MvcMovie dialog
- In the search box in the upper-right, enter view
- Tap MVC View Page
- In the Name box, enter Welcome.cshtml
- Tap Add

You’ll create a loop in the Welcome.cshtml view template that displays “Hello” NumTimes. Replace the contents of Views/HelloWorld/Welcome.cshtml with the following:

@{
    ViewData["Title"] = "About";
}

<h2>Welcome</h2>

<ul>
    @for (int i = 0; i < (int)ViewData["NumTimes"]; i++)
    {
        <li>@ViewData["Message"]</li>
    }
</ul>

Save your changes and browse to the following URL:

http://localhost:xxxx/HelloWorld/Welcome?name=Rick&numtimes=4

Data is taken from the URL and passed to the controller using the MVC model binder . The controller packages the data into a ViewData dictionary and passes that object to the view. The view then renders the data as HTML to the browser.

In the sample above, we used the ViewData dictionary to pass data from the controller to a view. Later in the tutorial, we will use a view model to pass data from a controller to a view. The view model approach to passing data is generally much preferred over the ViewData dictionary approach.

Well, that was a kind of an “M” for model, but not the database kind. Let’s take what we’ve learned and create a database of movies.

Adding a model¶

In this section you’ll add some classes for managing movies in a database. These classes will be the “Model” part of the MVC app.

You’ll use a .NET Framework data-access technology known as the Entity Framework Core to define and work with these data model classes. Entity Framework Core (often referred to as EF Core) features a development paradigm called Code First. You write the code first, and the database tables are created from this code. Code First allows you to create data model objects by writing simple classes. (These are also known as POCO classes, from “plain-old CLR objects.”) The database is created from your classes. If you are required to create the database first, you can still follow this tutorial to learn about MVC and EF app development.

Create a new project with individual user accounts¶

In the current version of the ASP.NET Core MVC tools for Visual Studio, scaffolding a model is only supported when you create a new project with individual user accounts. We hope to have this fixed in the next tooling update. Until that’s fixed, you’ll need to create a new project with the same name. Because the project has the same name, you’ll need to create it in another directory.

From the Visual Studio Start page, tap New Project.

Alternatively, you can use the menus to create a new project. Tap File > New > Project.

Complete the New Project dialog:

In the left pane, tap Web
In the center pane, tap ASP.NET Core Web Application (.NET Core)
Change the location to a different directory from the previous project you created or you’ll get an error
Name the project “MvcMovie” (It’s important to name the project “MvcMovie” so when you copy code, the namespace will match.)
Tap OK

Warning

You must have the Authentication set to Individual User Accounts in this release for the scaffolding engine to work.

In the New ASP.NET Core Web Application - MvcMovie dialog:

tap Web Application
tap the Change Authentication button and change the authentication to Individual User Accounts and tap OK

Follow the instructions in Change the title and menu link in the layout file so you can tap the MvcMovie link to invoke the Movie controller. We’ll scaffold the movies controller in this tutorial.

Adding data model classes¶

In Solution Explorer, right click the Models folder > Add > Class. Name the class Movie and add the following properties:

using System;

namespace MvcMovie.Models
{
    public class Movie
    {
        public int ID { get; set; }
        public string Title { get; set; }
        public DateTime ReleaseDate { get; set; }
        public string Genre { get; set; }
        public decimal Price { get; set; }
    }
}

In addition to the properties you’d expect to model a movie, the ID field is required by the DB for the primary key. Build the project. If you don’t build the app, you’ll get an error in the next section. We’ve finally added a Model to our MVC app.

Scaffolding a controller¶

In Solution Explorer, right-click the Controllers folder > Add > Controller.

In the Add Scaffold dialog, tap MVC Controller with views, using Entity Framework > Add.

Complete the Add Controller dialog

Model class: Movie(MvcMovie.Models)
Data context class: ApplicationDbContext(MvcMovie.Models)
Views:: Keep the default of each option checked
Controller name: Keep the default MoviesController
Tap Add

The Visual Studio scaffolding engine creates the following:

A movies controller (Controllers/MoviesController.cs)
Create, Delete, Details, Edit and Index Razor view files (Views/Movies)

Visual Studio automatically created the CRUD (create, read, update, and delete) action methods and views for you (the automatic creation of CRUD action methods and views is known as scaffolding). You’ll soon have a fully functional web application that lets you create, list, edit, and delete movie entries.

If you run the app and click on the Mvc Movie link, you’ll get the following errors:

We’ll follow those instructions to get the database ready for our Movie app.

Update the database¶

Warning

You must stop IIS Express before you update the database.

To Stop IIS Express:¶

Right click the IIS Express system tray icon in the notification area

Tap Exit or Stop Site

Alternatively, you can exit and restart Visual Studio

Open a command prompt in the project directory (MvcMovie/src/MvcMovie). Follow these instructions for a quick way to open a folder in the project directory.
- Open a file in the root of the project (for this example, use Startup.cs.)
- Right click on Startup.cs > Open Containing Folder.

Shift + right click a folder > Open command window here

Run cd .. to move back up to the project directory

Run the following commands in the command prompt:

dotnet ef migrations add Initial
dotnet ef database update

Note

If IIS-Express is running, you’ll get the error CS2012: Cannot open ‘MvcMovie/bin/Debug/netcoreapp1.0/MvcMovie.dll’ for writing – ‘The process cannot access the file ‘MvcMovie/bin/Debug/netcoreapp1.0/MvcMovie.dll’ because it is being used by another process.’

dotnet ef commands¶

dotnet (.NET Core) is a cross-platform implementation of .NET. You can read about it here
dotnet ef migrations add Initial Runs the Entity Framework .NET Core CLI migrations command and creates the initial migration. The parameter “Initial” is arbitrary, but customary for the first (initial) database migration. This operation creates the Data/Migrations/<date-time>_Initial.cs file containing the migration commands to add (or drop) the Movie table to the database
dotnet ef database update Updates the database with the migration we just created

Test the app¶

Note

If your browser is unable to connect to the movie app you might need to wait for IIS Express to load the app. It can sometimes take up to 30 seconds to build the app and have it ready to respond to requests.

Run the app and tap the Mvc Movie link
Tap the Create New link and create a movie

Note

You may not be able to enter decimal points or commas in the Price field. To support jQuery validation for non-English locales that use a comma (”,”) for a decimal point, and non US-English date formats, you must take steps to globalize your app. See Additional resources for more information. For now, just enter whole numbers like 10.

Note

In some locales you’ll need to specify the date format. See the highlighted code below.

using System;
using System.ComponentModel.DataAnnotations;

namespace MvcMovie.Models
{
    public class Movie
    {
        public int ID { get; set; }
        public string Title { get; set; }
        [DisplayFormat(DataFormatString = "{0:yyyy-MM-dd}", ApplyFormatInEditMode = true)]
        public DateTime ReleaseDate { get; set; }
        public string Genre { get; set; }
        public decimal Price { get; set; }
    }
}

Tapping Create causes the form to be posted to the server, where the movie information is saved in a database. You are then redirected to the /Movies URL, where you can see the newly created movie in the listing.

Create a couple more movie entries. Try the Edit, Details, and Delete links, which are all functional.

Examining the Generated Code¶

Open the Controllers/MoviesController.cs file and examine the generated Index method. A portion of the movie controller with the Index method is shown below:

public class MoviesController : Controller
{
    private readonly ApplicationDbContext _context;

    public MoviesController(ApplicationDbContext context)
    {
        _context = context;
    }

    // GET: Movies
    public async Task<IActionResult> Index()
    {
        return View(await _context.Movie.ToListAsync());
    }

The constructor uses Dependency Injection to inject the database context into the controller. The database context is used in each of the CRUD methods in the controller.

A request to the Movies controller returns all the entries in the Movies table and then passes the data to the Index view.

Strongly typed models and the @model keyword¶

Earlier in this tutorial, you saw how a controller can pass data or objects to a view using the ViewData dictionary. The ViewData dictionary is a dynamic object that provides a convenient late-bound way to pass information to a view.

MVC also provides the ability to pass strongly typed objects to a view. This strongly typed approach enables better compile-time checking of your code and richer IntelliSense in Visual Studio (VS). The scaffolding mechanism in VS used this approach (that is, passing a strongly typed model) with the MoviesController class and views when it created the methods and views.

Examine the generated Details method in the Controllers/MoviesController.cs file:

public async Task<IActionResult> Details(int? id)
{
    if (id == null)
    {
        return NotFound();
    }

    var movie = await _context.Movie.SingleOrDefaultAsync(m => m.ID == id);
    if (movie == null)
    {
        return NotFound();
    }

    return View(movie);
}

The id parameter is generally passed as route data, for example http://localhost:1234/movies/details/1 sets:

The controller to the movies controller (the first URL segment)
The action to details (the second URL segment)
The id to 1 (the last URL segment)

You could also pass in the id with a query string as follows:

http://localhost:1234/movies/details?id=1

If a Movie is found, an instance of the Movie model is passed to the Details view:

return View(movie);

Examine the contents of the Views/Movies/Details.cshtml file:

@model MvcMovie.Models.Movie

@{
    ViewData["Title"] = "Details";
}

<h2>Details</h2>

<div>
    <h4>Movie</h4>
    <hr />
    <dl class="dl-horizontal">
        <dt>
            @Html.DisplayNameFor(model => model.Genre)
        </dt>
        <dd>
            @Html.DisplayFor(model => model.Genre)
        </dd>
        <dt>
            @Html.DisplayNameFor(model => model.Price)
        </dt>
        <dd>
            @Html.DisplayFor(model => model.Price)
        </dd>
        <dt>
            @Html.DisplayNameFor(model => model.ReleaseDate)
        </dt>
        <dd>
            @Html.DisplayFor(model => model.ReleaseDate)
        </dd>
        <dt>
            @Html.DisplayNameFor(model => model.Title)
        </dt>
        <dd>
            @Html.DisplayFor(model => model.Title)
        </dd>
    </dl>
</div>
<div>
    <a asp-action="Edit" asp-route-id="@Model.ID">Edit</a> |
    <a asp-action="Index">Back to List</a>
</div>

By including a @model statement at the top of the view file, you can specify the type of object that the view expects. When you created the movie controller, Visual Studio automatically included the following @model statement at the top of the Details.cshtml file:

@model MvcMovie.Models.Movie

This @model directive allows you to access the movie that the controller passed to the view by using a Model object that’s strongly typed. For example, in the Details.cshtml view, the code passes each movie field to the DisplayNameFor and DisplayFor HTML Helpers with the strongly typed Model object. The Create and Edit methods and views also pass a Movie model object.

Examine the Index.cshtml view and the Index method in the Movies controller. Notice how the code creates a List object when it calls the View method. The code passes this Movies list from the Index action method to the view:

public async Task<IActionResult> Index()
{
    return View(await _context.Movie.ToListAsync());
}

When you created the movies controller, Visual Studio automatically included the following @model statement at the top of the Index.cshtml file:

@model IEnumerable<MvcMovie.Models.Movie>

The @model directive allows you to access the list of movies that the controller passed to the view by using a Model object that’s strongly typed. For example, in the Index.cshtml view, the code loops through the movies with a foreach statement over the strongly typed Model object:

@model IEnumerable<MvcMovie.Models.Movie>

@{
    ViewData["Title"] = "Index";
}

<h2>Index</h2>

<p>
    <a asp-action="Create">Create New</a>
</p>
<table class="table">
    <thead>
        <tr>
            <th>
                @Html.DisplayNameFor(model => model.Genre)
            </th>
            <th>
                @Html.DisplayNameFor(model => model.Price)
            </th>
            <th>
                @Html.DisplayNameFor(model => model.ReleaseDate)
            </th>
            <th>
                @Html.DisplayNameFor(model => model.Title)
            </th>
            <th></th>
        </tr>
    </thead>
    <tbody>
@foreach (var item in Model) {
        <tr>
            <td>
                @Html.DisplayFor(modelItem => item.Genre)
            </td>
            <td>
                @Html.DisplayFor(modelItem => item.Price)
            </td>
            <td>
                @Html.DisplayFor(modelItem => item.ReleaseDate)
            </td>
            <td>
                @Html.DisplayFor(modelItem => item.Title)
            </td>
            <td>
                <a asp-action="Edit" asp-route-id="@item.ID">Edit</a> |
                <a asp-action="Details" asp-route-id="@item.ID">Details</a> |
                <a asp-action="Delete" asp-route-id="@item.ID">Delete</a>
            </td>
        </tr>
}
    </tbody>
</table>

Because the Model object is strongly typed (as an IEnumerable<Movie> object), each item in the loop is typed as Movie. Among other benefits, this means that you get compile-time checking of the code and full IntelliSense support in the code editor:

You now have a database and pages to display, edit, update and delete data. In the next tutorial, we’ll work with the database.

Additional resources¶

Working with SQL Server LocalDB¶

The ApplicationDbContext class handles the task of connecting to the database and mapping Movie objects to database records. The database context is registered with the Dependency Injection container in the ConfigureServices method in the Startup.cs file:

public void ConfigureServices(IServiceCollection services)
{
    // Add framework services.
    services.AddDbContext<ApplicationDbContext>(options =>
        options.UseSqlServer(Configuration.GetConnectionString("DefaultConnection")));

The ASP.NET Core Configuration system reads the ConnectionString. For local development, it gets the connection string from the appsettings.json file:

{
  "ConnectionStrings": {
    "DefaultConnection": "Server=(localdb)\\mssqllocaldb;Database=aspnet-MvcMovie-4ae3798a;Trusted_Connection=True;MultipleActiveResultSets=true"
  },
  "Logging": {
    "IncludeScopes": false,

When you deploy the app to a test or production server, you can use an environment variable or another approach to set the connection string to a real SQL Server. See Configuration .

SQL Server Express LocalDB¶

LocalDB is a lightweight version of the SQL Server Express Database Engine that is targeted for program development. LocalDB starts on demand and runs in user mode, so there is no complex configuration. By default, LocalDB database creates “*.mdf” files in the C:/Users/<user> directory.

From the View menu, open SQL Server Object Explorer (SSOX).

Right click on the Movie table > View Designer

Note the key icon next to ID. By default, EF will make a property named ID the primary key.

Right click on the Movie table > View Data

Seed the database¶

Create a new class named SeedData in the Models folder. Replace the generated code with the following:

using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.DependencyInjection;
using MvcMovie.Data;
using System;
using System.Linq;

namespace MvcMovie.Models
{
    public static class SeedData
    {
        public static void Initialize(IServiceProvider serviceProvider)
        {
            using (var context = new ApplicationDbContext(
                serviceProvider.GetRequiredService<DbContextOptions<ApplicationDbContext>>()))
            {
                // Look for any movies.
                if (context.Movie.Any())
                {
                    return;   // DB has been seeded
                }

                context.Movie.AddRange(
                     new Movie
                     {
                         Title = "When Harry Met Sally",
                         ReleaseDate = DateTime.Parse("1989-1-11"),
                         Genre = "Romantic Comedy",
                         Price = 7.99M
                     },

                     new Movie
                     {
                         Title = "Ghostbusters ",
                         ReleaseDate = DateTime.Parse("1984-3-13"),
                         Genre = "Comedy",
                         Price = 8.99M
                     },

                     new Movie
                     {
                         Title = "Ghostbusters 2",
                         ReleaseDate = DateTime.Parse("1986-2-23"),
                         Genre = "Comedy",
                         Price = 9.99M
                     },

                   new Movie
                   {
                       Title = "Rio Bravo",
                       ReleaseDate = DateTime.Parse("1959-4-15"),
                       Genre = "Western",
                       Price = 3.99M
                   }
                );
                context.SaveChanges();
            }
        }
    }
}

Notice if there are any movies in the DB, the seed initializer returns.

if (context.Movie.Any())
{
    return;   // DB has been seeded
}

Add the seed initializer to the end of the Configure method in the Startup.cs file:

    app.UseMvc(routes =>
    {
        routes.MapRoute(
            name: "default",
            template: "{controller=Home}/{action=Index}/{id?}");
    });

    SeedData.Initialize(app.ApplicationServices);
}

Test the app

Delete all the records in the DB. You can do this with the delete links in the browser or from SSOX.
Force the app to initialize (call the methods in the Startup class) so the seed method runs. To force initialization, IIS Express must be stopped and restarted. You can do this with any of the following approaches:
- Right click the IIS Express system tray icon in the notification area and tap Exit or Stop Site

If you were running VS in non-debug mode, press F5 to run in debug mode

If you were running VS in debug mode, stop the debugger and press ^F5

Note

If the database doesn’t initialize, put a break point on the line if (context.Movie.Any()) and start debugging.

The app shows the seeded data.

Controller methods and views¶

We have a good start to the movie app, but the presentation is not ideal. We don’t want to see the time (12:00:00 AM in the image below) and ReleaseDate should be two words.

Open the Models/Movie.cs file and add the highlighted lines shown below:

public class Movie
{
    public int ID { get; set; }
    public string Title { get; set; }

    [Display(Name = "Release Date")]
    [DataType(DataType.Date)]
    public DateTime ReleaseDate { get; set; }
    public string Genre { get; set; }
    public decimal Price { get; set; }
}

Right click on a red squiggly line > Quick Actions.

Tap using System.ComponentModel.DataAnnotations;

Visual studio adds using System.ComponentModel.DataAnnotations;.

Let’s remove the using statements that are not needed. They show up by default in a light grey font. Right click anywhere in the Movie.cs file > Organize Usings > Remove Unnecessary Usings.

The updated code:

using System;
using System.ComponentModel.DataAnnotations;

namespace MvcMovie.Models
{
    public class Movie
    {
        public int ID { get; set; }
        public string Title { get; set; }

        [Display(Name = "Release Date")]
        [DataType(DataType.Date)]
        public DateTime ReleaseDate { get; set; }
        public string Genre { get; set; }
        public decimal Price { get; set; }
    }
}

We’ll cover DataAnnotations in the next tutorial. The Display attribute specifies what to display for the name of a field (in this case “Release Date” instead of “ReleaseDate”). The DataType attribute specifies the type of the data, in this case it’s a date, so the time information stored in the field is not displayed.

Browse to the Movies controller and hold the mouse pointer over an Edit link to see the target URL.

The Edit, Details, and Delete links are generated by the MVC Core Anchor Tag Helper in the Views/Movies/Index.cshtml file.

<td>
    <a asp-action="Edit" asp-route-id="@item.ID">Edit</a> |
    <a asp-action="Details" asp-route-id="@item.ID">Details</a> |
    <a asp-action="Delete" asp-route-id="@item.ID">Delete</a>
</td>

Tag Helpers enable server-side code to participate in creating and rendering HTML elements in Razor files. In the code above, the AnchorTagHelper dynamically generates the HTML href attribute value from the controller action method and route id. You use View Source from your favorite browser or use the F12 tools to examine the generated markup. The F12 tools are shown below.

Recall the format for routing set in the Startup.cs file.

app.UseMvc(routes =>
{
    routes.MapRoute(
        name: "default",
        template: "{controller=Home}/{action=Index}/{id?}");
});

ASP.NET Core translates http://localhost:1234/Movies/Edit/4 into a request to the Edit action method of the Movies controller with the parameter ID of 4. (Controller methods are also known as action methods.)

Tag Helpers are one of the most popular new features in ASP.NET Core. See Additional resources for more information.

Open the Movies controller and examine the two Edit action methods:

public async Task<IActionResult> Edit(int? id)
{
    if (id == null)
    {
        return NotFound();
    }

    var movie = await _context.Movie.SingleOrDefaultAsync(m => m.ID == id);
    if (movie == null)
    {
        return NotFound();
    }
    return View(movie);
}

[HttpPost]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Edit(int id, [Bind("ID,Genre,Price,ReleaseDate,Title")] Movie movie)
{
    if (id != movie.ID)
    {
        return NotFound();
    }

    if (ModelState.IsValid)
    {
        try
        {
            _context.Update(movie);
            await _context.SaveChangesAsync();
        }
        catch (DbUpdateConcurrencyException)
        {
            if (!MovieExists(movie.ID))
            {
                return NotFound();
            }
            else
            {
                throw;
            }
        }
        return RedirectToAction("Index");
    }
    return View(movie);
}

The [Bind] attribute is one way to protect against over-posting. You should only include properties in the [Bind] attribute that you want to change. See Protect your controller from over-posting for more information. ViewModels provide an alternative approach to prevent over-posting.

Notice the second Edit action method is preceded by the [HttpPost] attribute.

[HttpPost]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Edit(int id, [Bind("ID,Genre,Price,ReleaseDate,Title")] Movie movie)
{
    if (id != movie.ID)
    {
        return NotFound();
    }

    if (ModelState.IsValid)
    {
        try
        {
            _context.Update(movie);
            await _context.SaveChangesAsync();
        }
        catch (DbUpdateConcurrencyException)
        {
            if (!MovieExists(movie.ID))
            {
                return NotFound();
            }
            else
            {
                throw;
            }
        }
        return RedirectToAction("Index");
    }
    return View(movie);
}

The HttpPostAttribute attribute specifies that this Edit method can be invoked only for POST requests. You could apply the [HttpGet] attribute to the first edit method, but that’s not necessary because [HttpGet] is the default.

The ValidateAntiForgeryTokenAttribute attribute is used to prevent forgery of a request and is paired up with an anti-forgery token generated in the edit view file (Views/Movies/Edit.cshtml). The edit view file generates the anti-forgery token with the Form Tag Helper.

<form asp-action="Edit">

The Form Tag Helper generates a hidden anti-forgery token that must match the [ValidateAntiForgeryToken] generated anti-forgery token in the Edit method of the Movies controller. For more information, see 🔧 Anti-Request Forgery.

The HttpGet Edit method takes the movie ID parameter, looks up the movie using the Entity Framework SingleOrDefaultAsync method, and returns the selected movie to the Edit view. If a movie cannot be found, NotFound (HTTP 404) is returned.

public async Task<IActionResult> Edit(int? id)
{
    if (id == null)
    {
        return NotFound();
    }

    var movie = await _context.Movie.SingleOrDefaultAsync(m => m.ID == id);
    if (movie == null)
    {
        return NotFound();
    }
    return View(movie);
}

When the scaffolding system created the Edit view, it examined the Movie class and created code to render <label> and <input> elements for each property of the class. The following example shows the Edit view that was generated by the visual studio scaffolding system:

@model MvcMovie.Models.Movie

@{
    ViewData["Title"] = "Edit";
}

<h2>Edit</h2>

<form asp-action="Edit">
    <div class="form-horizontal">
        <h4>Movie</h4>
        <hr />
        <div asp-validation-summary="ModelOnly" class="text-danger"></div>
    <input type="hidden" asp-for="ID" />
        <div class="form-group">
            <label asp-for="Genre" class="col-md-2 control-label"></label>
            <div class="col-md-10">
                <input asp-for="Genre" class="form-control" />
                <span asp-validation-for="Genre" class="text-danger"></span>
            </div>
        </div>
        <div class="form-group">
            <label asp-for="Price" class="col-md-2 control-label"></label>
            <div class="col-md-10">
                <input asp-for="Price" class="form-control" />
                <span asp-validation-for="Price" class="text-danger"></span>
            </div>
        </div>
        <div class="form-group">
            <label asp-for="ReleaseDate" class="col-md-2 control-label"></label>
            <div class="col-md-10">
                <input asp-for="ReleaseDate" class="form-control" />
                <span asp-validation-for="ReleaseDate" class="text-danger"></span>
            </div>
        </div>
        <div class="form-group">
            <label asp-for="Title" class="col-md-2 control-label"></label>
            <div class="col-md-10">
                <input asp-for="Title" class="form-control" />
                <span asp-validation-for="Title" class="text-danger"></span>
            </div>
        </div>
        <div class="form-group">
            <div class="col-md-offset-2 col-md-10">
                <input type="submit" value="Save" class="btn btn-default" />
            </div>
        </div>
    </div>
</form>

<div>
    <a asp-action="Index">Back to List</a>
</div>

@section Scripts {
    @{await Html.RenderPartialAsync("_ValidationScriptsPartial");}
}

Notice how the view template has a @model MvcMovie.Models.Movie statement at the top of the file — this specifies that the view expects the model for the view template to be of type Movie.

The scaffolded code uses several Tag Helper methods to streamline the HTML markup. The - Label Tag Helper displays the name of the field (“Title”, “ReleaseDate”, “Genre”, or “Price”). The Input Tag Helper renders an HTML <input> element. The Validation Tag Helper displays any validation messages associated with that property.

Run the application and navigate to the /Movies URL. Click an Edit link. In the browser, view the source for the page. The generated HTML for the <form> element is shown below.

<form action="/Movies/Edit/7" method="post">
    <div class="form-horizontal">
        <h4>Movie</h4>
        <hr />
        <div class="text-danger" />
        <input type="hidden" data-val="true" data-val-required="The ID field is required." id="ID" name="ID" value="7" />
        <div class="form-group">
            <label class="control-label col-md-2" for="Genre" />
            <div class="col-md-10">
                <input class="form-control" type="text" id="Genre" name="Genre" value="Western" />
                <span class="text-danger field-validation-valid" data-valmsg-for="Genre" data-valmsg-replace="true"></span>
            </div>
        </div>
        <div class="form-group">
            <label class="control-label col-md-2" for="Price" />
            <div class="col-md-10">
                <input class="form-control" type="text" data-val="true" data-val-number="The field Price must be a number." data-val-required="The Price field is required." id="Price" name="Price" value="3.99" />
                <span class="text-danger field-validation-valid" data-valmsg-for="Price" data-valmsg-replace="true"></span>
            </div>
        </div>
        <!-- Markup removed for brevity -->
        <div class="form-group">
            <div class="col-md-offset-2 col-md-10">
                <input type="submit" value="Save" class="btn btn-default" />
            </div>
        </div>
    </div>
    <input name="__RequestVerificationToken" type="hidden" value="CfDJ8Inyxgp63fRFqUePGvuI5jGZsloJu1L7X9le1gy7NCIlSduCRx9jDQClrV9pOTTmqUyXnJBXhmrjcUVDJyDUMm7-MF_9rK8aAZdRdlOri7FmKVkRe_2v5LIHGKFcTjPrWPYnc9AdSbomkiOSaTEg7RU" />
</form>

The <input> elements are in an HTML <form> element whose action attribute is set to post to the /Movies/Edit/id URL. The form data will be posted to the server when the Save button is clicked. The last line before the closing </form> element shows the hidden XSRF token generated by the Form Tag Helper.

Processing the POST Request¶

The following listing shows the [HttpPost] version of the Edit action method.

[HttpPost]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Edit(int id, [Bind("ID,Genre,Price,ReleaseDate,Title")] Movie movie)
{
    if (id != movie.ID)
    {
        return NotFound();
    }

    if (ModelState.IsValid)
    {
        try
        {
            _context.Update(movie);
            await _context.SaveChangesAsync();
        }
        catch (DbUpdateConcurrencyException)
        {
            if (!MovieExists(movie.ID))
            {
                return NotFound();
            }
            else
            {
                throw;
            }
        }
        return RedirectToAction("Index");
    }
    return View(movie);
}

The [ValidateAntiForgeryToken] attribute validates the hidden XSRF token generated by the anti-forgery token generator in the Form Tag Helper

The model binding system takes the posted form values and creates a Movie object that’s passed as the movie parameter. The ModelState.IsValid method verifies that the data submitted in the form can be used to modify (edit or update) a Movie object. If the data is valid it’s saved. The updated (edited) movie data is saved to the database by calling the SaveChangesAsync method of database context. After saving the data, the code redirects the user to the Index action method of the MoviesController class, which displays the movie collection, including the changes just made.

Before the form is posted to the server, client side validation checks any validation rules on the fields. If there are any validation errors, an error message is displayed and the form is not posted. If JavaScript is disabled, you won’t have client side validation but the server will detect the posted values that are not valid, and the form values will be redisplayed with error messages. Later in the tutorial we examine Model Validation validation in more detail. The Validation Tag Helper in the Views/Book/Edit.cshtml view template takes care of displaying appropriate error messages.

All the HttpGet methods in the movie controller follow a similar pattern. They get a movie object (or list of objects, in the case of Index), and pass the object (model) to the view. The Create method passes an empty movie object to the Create view. All the methods that create, edit, delete, or otherwise modify data do so in the [HttpPost] overload of the method. Modifying data in an HTTP GET method is a security risk, as in ASP.NET MVC Tip #46 – Don’t use Delete Links because they create Security Holes. Modifying data in a HTTP GET method also violates HTTP best practices and the architectural REST pattern, which specifies that GET requests should not change the state of your application. In other words, performing a GET operation should be a safe operation that has no side effects and doesn’t modify your persisted data.

Additional resources¶

Globalization and localization
Introduction to Tag Helpers
Authoring Tag Helpers
🔧 Anti-Request Forgery
Protect your controller from over-posting
ViewModels
Form Tag Helper
Input Tag Helper
Label Tag Helper
Select Tag Helper
Validation Tag Helper

Adding Search¶

In this section you’ll add search capability to the Index action method that lets you search movies by genre or name.

Update the Index action method to enable search:

public async Task<IActionResult> Index(string searchString)
{
    var movies = from m in _context.Movie
                 select m;

    if (!String.IsNullOrEmpty(searchString))
    {
        movies = movies.Where(s => s.Title.Contains(searchString));
    }

    return View(await movies.ToListAsync());
}

The first line of the Index action method creates a LINQ query to select the movies:

var movies = from m in _context.Movie
             select m;

The query is only defined at this point, it has not been run against the database.

If the searchString parameter contains a string, the movies query is modified to filter on the value of the search string, using the following code:

   if (!String.IsNullOrEmpty(searchString))
   {
       movies = movies.Where(s => s.Title.Contains(searchString));
   }

The s => s.Title.Contains() code above is a Lambda Expression. Lambdas are used in method-based LINQ queries as arguments to standard query operator methods such as the Where method or Contains used in the code above. LINQ queries are not executed when they are defined or when they are modified by calling a method such as Where, Contains or OrderBy. Instead, query execution is deferred, which means that the evaluation of an expression is delayed until its realized value is actually iterated over or the ToListAsync method is called. For more information about deferred query execution, see Query Execution.

Note

The Contains method is run on the database, not the c# code above. On the database, Contains maps to SQL LIKE, which is case insensitive.

Navigate to /Movies/Index. Append a query string such as ?searchString=ghost to the URL. The filtered movies are displayed.

If you change the signature of the Index method to have a parameter named id, the id parameter will match the optional {id} placeholder for the default routes set in Startup.cs.

app.UseMvc(routes =>
{
    routes.MapRoute(
        name: "default",
        template: "{controller=Home}/{action=Index}/{id?}");
});

You can quickly rename the searchString parameter to id with the rename command. Right click on searchString > Rename.

The rename targets are highlighted.

Change the parameter to id and all occurrences of searchString change to id.

The previous Index method:

public async Task<IActionResult> Index(string searchString)
{
    var movies = from m in _context.Movie
                 select m;

    if (!String.IsNullOrEmpty(searchString))
    {
        movies = movies.Where(s => s.Title.Contains(searchString));
    }

    return View(await movies.ToListAsync());
}

The updated Index method:

public async Task<IActionResult> Index(string id)
{
    var movies = from m in _context.Movie
                 select m;

    if (!String.IsNullOrEmpty(id))
    {
        movies = movies.Where(s => s.Title.Contains(id));
    }

    return View(await movies.ToListAsync());
}

You can now pass the search title as route data (a URL segment) instead of as a query string value.

However, you can’t expect users to modify the URL every time they want to search for a movie. So now you’ll add UI to help them filter movies. If you changed the signature of the Index method to test how to pass the route-bound ID parameter, change it back so that it takes a parameter named searchString:

public async Task<IActionResult> Index(string searchString)
{
    var movies = from m in _context.Movie
                 select m;

    if (!String.IsNullOrEmpty(searchString))
    {
        movies = movies.Where(s => s.Title.Contains(searchString));
    }

    return View(await movies.ToListAsync());
}

Open the Views/Movies/Index.cshtml file, and add the <form> markup highlighted below:

@{
    ViewData["Title"] = "Index";
}

<h2>Index</h2>

<p>
    <a asp-action="Create">Create New</a>
</p>

<form asp-controller="Movies" asp-action="Index">
    <p>
        Title: <input type="text" name="SearchString">
        <input type="submit" value="Filter" />
    </p>
</form>

<table class="table">

The HTML <form> tag uses the Form Tag Helper, so when you submit the form, the filter string is posted to the Index action of the movies controller. Save your changes and then test the filter.

There’s no [HttpPost] overload of the Index method as you might expect. You don’t need it, because the method isn’t changing the state of the app, just filtering data.

You could add the following [HttpPost] Index method.

[HttpPost]
public string Index(string searchString, bool notUsed)
{
    return "From [HttpPost]Index: filter on " + searchString;
}

The notUsed parameter is used to create an overload for the Index method. We’ll talk about that later in the tutorial.

If you add this method, the action invoker would match the [HttpPost] Index method, and the [HttpPost] Index method would run as shown in the image below.

However, even if you add this [HttpPost] version of the Index method, there’s a limitation in how this has all been implemented. Imagine that you want to bookmark a particular search or you want to send a link to friends that they can click in order to see the same filtered list of movies. Notice that the URL for the HTTP POST request is the same as the URL for the GET request (localhost:xxxxx/Movies/Index) – there’s no search information in the URL. The search string information is sent to the server as a form field value. You can verify that with the F12 Developer tools or the excellent Fiddler tool. Start the F12 tool:

Tap the http://localhost:xxx/Movies HTTP POST 200 line and then tap Body > Request Body.

You can see the search parameter and XSRF token in the request body. Note, as mentioned in the previous tutorial, the Form Tag Helper generates an XSRF anti-forgery token. We’re not modifying data, so we don’t need to validate the token in the controller method.

Because the search parameter is in the request body and not the URL, you can’t capture that search information to bookmark or share with others. We’ll fix this by specifying the request should be HTTP GET. Notice how intelliSense helps us update the markup.

Notice the distinctive font in the <form> tag. That distinctive font indicates the tag is supported by Tag Helpers.

Now when you submit a search, the URL contains the search query string. Searching will also go to the HttpGet Index action method, even if you have a HttpPost Index method.

The following markup shows the change to the form tag:

<form asp-controller="Movies" asp-action="Index" method="get">

Adding Search by Genre¶

Add the following MovieGenreViewModel class to the Models folder:

using Microsoft.AspNetCore.Mvc.Rendering;
using System.Collections.Generic;

namespace MvcMovie.Models
{
    public class MovieGenreViewModel
    {
        public List<Movie> movies;
        public SelectList genres;
        public string movieGenre { get; set; }
    }
}

The move-genre view model will contain:

a list of movies

a SelectList containing the list of genres. This will allow the user to select a genre from the list.

movieGenre, which contains the selected genre

Replace the Index method with the following code:

public async Task<IActionResult> Index(string movieGenre, string searchString)
{
    // Use LINQ to get list of genre's.
    IQueryable<string> genreQuery = from m in _context.Movie
                                    orderby m.Genre
                                    select m.Genre;

    var movies = from m in _context.Movie
                 select m;

    if (!String.IsNullOrEmpty(searchString))
    {
        movies = movies.Where(s => s.Title.Contains(searchString));
    }

    if (!String.IsNullOrEmpty(movieGenre))
    {
        movies = movies.Where(x => x.Genre == movieGenre);
    }

    var movieGenreVM = new MovieGenreViewModel();
    movieGenreVM.genres = new SelectList(await genreQuery.Distinct().ToListAsync());
    movieGenreVM.movies = await movies.ToListAsync();

    return View(movieGenreVM);
}

The following code is a LINQ query that retrieves all the genres from the database.

IQueryable<string> genreQuery = from m in _context.Movie
                                orderby m.Genre
                                select m.Genre;

The SelectList of genres is created by projecting the distinct genres (we don’t want our select list to have duplicate genres).

movieGenreVM.genres = new SelectList(await genreQuery.Distinct().ToListAsync())

Adding search by genre to the Index view¶

@model MovieGenreViewModel

@{
    ViewData["Title"] = "Index";
}

<h2>Index</h2>

<p>
    <a asp-action="Create">Create New</a>
</p>

<form asp-controller="Movies" asp-action="Index" method="get">
    <p>
        <select asp-for="movieGenre" asp-items="Model.genres">
            <option value="">All</option>
        </select>

        Title: <input type="text" name="SearchString">
        <input type="submit" value="Filter" />
    </p>
</form>

<table class="table">
    <tr>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].Genre)
        </th>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].Price)
        </th>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].ReleaseDate)
        </th>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].Title)
        </th>
        <th></th>
    </tr>
    <tbody>
        @foreach (var item in Model.movies)
        {
            <tr>
                <td>
                    @Html.DisplayFor(modelItem => item.Genre)
                </td>
                <td>
                    @Html.DisplayFor(modelItem => item.Price)
                </td>
                <td>
                    @Html.DisplayFor(modelItem => item.ReleaseDate)
                </td>
                <td>
                    @Html.DisplayFor(modelItem => item.Title)
                </td>
                <td>
                    <a asp-action="Edit" asp-route-id="@item.ID">Edit</a> |
                    <a asp-action="Details" asp-route-id="@item.ID">Details</a> |
                    <a asp-action="Delete" asp-route-id="@item.ID">Delete</a>
                </td>
            </tr>
        }
    </tbody>
</table>

Test the app by searching by genre, by movie title, and by both.

Adding a New Field¶

In this section you’ll use Entity Framework Code First Migrations to add a new field to the model and migrate that change to the database.

When you use EF Code First to automatically create a database, Code First adds a table to the database to help track whether the schema of the database is in sync with the model classes it was generated from. If they aren’t in sync, EF throws an exception. This makes it easier to track down issues at development time that you might otherwise only find (by obscure errors) at run time.

Adding a Rating Property to the Movie Model¶

Open the Models/Movie.cs file and add a Rating property:

public class Movie
{
    public int ID { get; set; }
    public string Title { get; set; }

    [Display(Name = "Release Date")]
    [DataType(DataType.Date)]
    public DateTime ReleaseDate { get; set; }
    public string Genre { get; set; }
    public decimal Price { get; set; }
    public string Rating { get; set; }
}

Build the app (Ctrl+Shift+B).

Because you’ve added a new field to the Movie class, you also need to update the binding white list so this new property will be included. Update the [Bind] attribute for Create and Edit action methods to include the Rating property:

[Bind("ID,Title,ReleaseDate,Genre,Price,Rating")]

You also need to update the view templates in order to display, create and edit the new Rating property in the browser view.

Edit the /Views/Movies/Index.cshtml file and add a Rating field:

<table class="table">
    <tr>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].Genre)
        </th>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].Price)
        </th>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].ReleaseDate)
        </th>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].Title)
        </th>
        <th>
            @Html.DisplayNameFor(model => model.movies[0].Rating)
        </th>
        <th></th>
    </tr>
    <tbody>
        @foreach (var item in Model.movies)
        {
            <tr>
                <td>
                    @Html.DisplayFor(modelItem => item.Genre)
                </td>
                <td>
                    @Html.DisplayFor(modelItem => item.Price)
                </td>
                <td>
                    @Html.DisplayFor(modelItem => item.ReleaseDate)
                </td>
                <td>
                    @Html.DisplayFor(modelItem => item.Title)
                </td>
                <td>
                    @Html.DisplayFor(modelItem => item.Rating)
                </td>

Update the /Views/Movies/Create.cshtml with a Rating field. You can copy/paste the previous “form group” and let intelliSense help you update the fields. IntelliSense works with Tag Helpers.

The app won’t work until we update the DB to include the new field. If you run it now, you’ll get the following SqlException:

You’re seeing this error because the updated Movie model class is different than the schema of the Movie table of the existing database. (There’s no Rating column in the database table.)

There are a few approaches to resolving the error:

Have the Entity Framework automatically drop and re-create the database based on the new model class schema. This approach is very convenient early in the development cycle when you are doing active development on a test database; it allows you to quickly evolve the model and database schema together. The downside, though, is that you lose existing data in the database — so you don’t want to use this approach on a production database! Using an initializer to automatically seed a database with test data is often a productive way to develop an application.
Explicitly modify the schema of the existing database so that it matches the model classes. The advantage of this approach is that you keep your data. You can make this change either manually or by creating a database change script.
Use Code First Migrations to update the database schema.

For this tutorial, we’ll use Code First Migrations.

Update the SeedData class so that it provides a value for the new column. A sample change is shown below, but you’ll want to make this change for each new Movie.

     new Movie
     {
         Title = "When Harry Met Sally",
         ReleaseDate = DateTime.Parse("1989-1-11"),
         Genre = "Romantic Comedy",
         Rating = "R",
         Price = 7.99M
     },

Warning

You must stop IIS Express before you run the dotnet ef commands. See To Stop IIS Express:

Build the solution then open a command prompt. Enter the following commands:

dotnet ef migrations add Rating
dotnet ef database update

The migrations add command tells the migration framework to examine the current Movie model with the current Movie DB schema and create the necessary code to migrate the DB to the new model. The name “Rating” is arbitrary and is used to name the migration file. It’s helpful to use a meaningful name for the migration step.

If you delete all the records in the DB, the initialize will seed the DB and include the Rating field. You can do this with the delete links in the browser or from SSOX.

Run the app and verify you can create/edit/display movies with a Rating field. You should also add the Rating field to the Edit, Details, and Delete view templates.

Adding Validation¶

In this section you’ll add validation logic to the Movie model, and you’ll ensure that the validation rules are enforced any time a user attempts to create or edit a movie.

Keeping things DRY¶

One of the design tenets of MVC is DRY (“Don’t Repeat Yourself”). ASP.NET MVC encourages you to specify functionality or behavior only once, and then have it be reflected everywhere in an app. This reduces the amount of code you need to write and makes the code you do write less error prone, easier to test, and easier to maintain.

The validation support provided by MVC and Entity Framework Core Code First is a great example of the DRY principle in action. You can declaratively specify validation rules in one place (in the model class) and the rules are enforced everywhere in the app.

Let’s look at how you can take advantage of this validation support in the movie app.

Adding validation rules to the movie model¶

Open the Movie.cs file. DataAnnotations provides a built-in set of validation attributes that you apply declaratively to any class or property. (It also contains formatting attributes like DataType that help with formatting and don’t provide any validation.)

Update the Movie class to take advantage of the built-in Required, StringLength, RegularExpression, and Range validation attributes.

public class Movie
{
    public int ID { get; set; }

    [StringLength(60, MinimumLength = 3)]
    public string Title { get; set; }

    [Display(Name = "Release Date")]
    [DataType(DataType.Date)]
    public DateTime ReleaseDate { get; set; }

    [RegularExpression(@"^[A-Z]+[a-zA-Z''-'\s]*$")]
    [Required]
    [StringLength(30)]
    public string Genre { get; set; }

    [Range(1, 100)]
    [DataType(DataType.Currency)]
    public decimal Price { get; set; }

    [RegularExpression(@"^[A-Z]+[a-zA-Z''-'\s]*$")]
    [StringLength(5)]
    public string Rating { get; set; }
}

The validation attributes specify behavior that you want to enforce on the model properties they are applied to. The Required and MinimumLength attributes indicates that a property must have a value; but nothing prevents a user from entering white space to satisfy this validation. The RegularExpression attribute is used to limit what characters can be input. In the code above, Genre and Rating must use only letters (white space, numbers and special characters are not allowed). The Range attribute constrains a value to within a specified range. The StringLength attribute lets you set the maximum length of a string property, and optionally its minimum length. Value types (such as decimal, int, float, DateTime) are inherently required and don’t need the [Required] attribute.

Having validation rules automatically enforced by ASP.NET helps make your app more robust. It also ensures that you can’t forget to validate something and inadvertently let bad data into the database.

Validation Error UI in MVC¶

Run the app and navigate to the Movies controller.

Tap the Create New link to add a new movie. Fill out the form with some invalid values. As soon as jQuery client side validation detects the error, it displays an error message.

Note

You may not be able to enter decimal points or commas in the Price field. To support jQuery validation for non-English locales that use a comma (”,”) for a decimal point, and non US-English date formats, you must take steps to globalize your app. See Additional resources for more information. For now, just enter whole numbers like 10.

Notice how the form has automatically rendered an appropriate validation error message in each field containing an invalid value. The errors are enforced both client-side (using JavaScript and jQuery) and server-side (in case a user has JavaScript disabled).

A significant benefit is that you didn’t need to change a single line of code in the MoviesController class or in the Create.cshtml view in order to enable this validation UI. The controller and views you created earlier in this tutorial automatically picked up the validation rules that you specified by using validation attributes on the properties of the Movie model class. Test validation using the Edit action method, and the same validation is applied.

The form data is not sent to the server until there are no client side validation errors. You can verify this by putting a break point in the HTTP Post method, by using the Fiddler tool , or the F12 Developer tools.

How Validation Occurs in the Create View and Create Action Method¶

You might wonder how the validation UI was generated without any updates to the code in the controller or views. The next listing shows the two Create methods.

public IActionResult Create()
{
    return View();
}

// POST: Movies/Create
// To protect from overposting attacks, please enable the specific properties you want to bind to, for 
// more details see http://go.microsoft.com/fwlink/?LinkId=317598.
[HttpPost]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Create([Bind("ID,Genre,Price,ReleaseDate,Title,Rating")] Movie movie)
{
    if (ModelState.IsValid)
    {
        _context.Add(movie);
        await _context.SaveChangesAsync();
        return RedirectToAction("Index");
    }
    return View(movie);
}

The first (HTTP GET) Create action method displays the initial Create form. The second ([HttpPost]) version handles the form post. The second Create method (The [HttpPost] version) calls ModelState.IsValid to check whether the movie has any validation errors. Calling this method evaluates any validation attributes that have been applied to the object. If the object has validation errors, the Create method re-displays the form. If there are no errors, the method saves the new movie in the database. In our movie example, the form is not posted to the server when there are validation errors detected on the client side; the second Create method is never called when there are client side validation errors. If you disable JavaScript in your browser, client validation is disabled and you can test the HTTP POST Create method ModelState.IsValid detecting any validation errors.

You can set a break point in the [HttpPost] Create method and verify the method is never called, client side validation will not submit the form data when validation errors are detected. If you disable JavaScript in your browser, then submit the form with errors, the break point will be hit. You still get full validation without JavaScript. The following image shows how to disable JavaScript in Internet Explorer.

The following image shows how to disable JavaScript in the FireFox browser.

The following image shows how to disable JavaScript in the Chrome browser.

After you disable JavaScript, post invalid data and step through the debugger.

Below is portion of the Create.cshtml view template that you scaffolded earlier in the tutorial. It’s used by the action methods shown above both to display the initial form and to redisplay it in the event of an error.

<form asp-action="Create">
    <div class="form-horizontal">
        <h4>Movie</h4>
        <hr />
        <div asp-validation-summary="ModelOnly" class="text-danger"></div>
        <div class="form-group">
            <label asp-for="Genre" class="col-md-2 control-label"></label>
            <div class="col-md-10">
                <input asp-for="Genre" class="form-control" />
                <span asp-validation-for="Genre" class="text-danger"></span>
            </div>
        </div>
        @*Markup removed for brevity.*@
        <div class="form-group">
            <label asp-for="Rating" class="col-md-2 control-label"></label>
            <div class="col-md-10">
                <input asp-for="Rating" class="form-control" />
                <span asp-validation-for="Rating" class="text-danger"></span>
            </div>
        </div>
        <div class="form-group">
            <div class="col-md-offset-2 col-md-10">
                <input type="submit" value="Create" class="btn btn-default" />
            </div>
        </div>
    </div>
</form>

The Input Tag Helper consumes the DataAnnotations attributes and produces HTML attributes needed for jQuery Validation on the client side. The Validation Tag Helper displays a validation errors. See Validation for more information.

What’s really nice about this approach is that neither the controller nor the Create view template knows anything about the actual validation rules being enforced or about the specific error messages displayed. The validation rules and the error strings are specified only in the Movie class. These same validation rules are automatically applied to the Edit view and any other views templates you might create that edit your model.

When you need to change validation logic, you can do so in exactly one place by adding validation attributes to the model (in this example, the Movie class). You won’t have to worry about different parts of the application being inconsistent with how the rules are enforced — all validation logic will be defined in one place and used everywhere. This keeps the code very clean, and makes it easy to maintain and evolve. And it means that that you’ll be fully honoring the DRY principle.

Using DataType Attributes¶

Open the Movie.cs file and examine the Movie class. The System.ComponentModel.DataAnnotations namespace provides formatting attributes in addition to the built-in set of validation attributes. We’ve already applied a DataType enumeration value to the release date and to the price fields. The following code shows the ReleaseDate and Price properties with the appropriate DataType attribute.

[Display(Name = "Release Date")]
[DataType(DataType.Date)]
public DateTime ReleaseDate { get; set; }

[Range(1, 100)]
[DataType(DataType.Currency)]
public decimal Price { get; set; }

The DataType attributes only provide hints for the view engine to format the data (and supply attributes such as <a> for URL’s and <a href="mailto:EmailAddress.com"> for email. You can use the RegularExpression attribute to validate the format of the data. The DataType attribute is used to specify a data type that is more specific than the database intrinsic type, they are not validation attributes. In this case we only want to keep track of the date, not the time. The DataType Enumeration provides for many data types, such as Date, Time, PhoneNumber, Currency, EmailAddress and more. The DataType attribute can also enable the application to automatically provide type-specific features. For example, a mailto: link can be created for DataType.EmailAddress, and a date selector can be provided for DataType.Date in browsers that support HTML5. The DataType attributes emits HTML 5 data- (pronounced data dash) attributes that HTML 5 browsers can understand. The DataType attributes do not provide any validation.

DataType.Date does not specify the format of the date that is displayed. By default, the data field is displayed according to the default formats based on the server’s CultureInfo.

The DisplayFormat attribute is used to explicitly specify the date format:

[DisplayFormat(DataFormatString = "{0:yyyy-MM-dd}", ApplyFormatInEditMode = true)]
public DateTime ReleaseDate { get; set; }

The ApplyFormatInEditMode setting specifies that the formatting should also be applied when the value is displayed in a text box for editing. (You might not want that for some fields — for example, for currency values, you probably do not want the currency symbol in the text box for editing.)

You can use the DisplayFormat attribute by itself, but it’s generally a good idea to use the DataType attribute. The DataType attribute conveys the semantics of the data as opposed to how to render it on a screen, and provides the following benefits that you don’t get with DisplayFormat:

The browser can enable HTML5 features (for example to show a calendar control, the locale-appropriate currency symbol, email links, etc.)
By default, the browser will render data using the correct format based on your locale
The DataType attribute can enable MVC to choose the right field template to render the data (the DisplayFormat if used by itself uses the string template).

Note

jQuery validation does not work with the Range attribute and DateTime. For example, the following code will always display a client side validation error, even when the date is in the specified range:

[Range(typeof(DateTime), "1/1/1966", "1/1/2020")]

You will need to disable jQuery date validation to use the Range attribute with DateTime. It’s generally not a good practice to compile hard dates in your models, so using the Range attribute and DateTime is discouraged.

The following code shows combining attributes on one line:

public class Movie
{
    public int ID { get; set; }

    [StringLength(60, MinimumLength = 3)]
    public string Title { get; set; }

    [Display(Name = "Release Date"), DataType(DataType.Date)]
    public DateTime ReleaseDate { get; set; }

    [RegularExpression(@"^[A-Z]+[a-zA-Z''-'\s]*$"), Required, StringLength(30)]
    public string Genre { get; set; }

    [Range(1, 100), DataType(DataType.Currency)]
    public decimal Price { get; set; }

    [RegularExpression(@"^[A-Z]+[a-zA-Z''-'\s]*$"), StringLength(5)]
    public string Rating { get; set; }
}

In the next part of the series, we’ll review the application and make some improvements to the automatically generated Details and Delete methods.

Additional resources¶

Examining the Details and Delete methods¶

Open the Movie controller and examine the Details method:

public async Task<IActionResult> Details(int? id)
{
    if (id == null)
    {
        return NotFound();
    }

    var movie = await _context.Movie.SingleOrDefaultAsync(m => m.ID == id);
    if (movie == null)
    {
        return NotFound();
    }

    return View(movie);
}

The MVC scaffolding engine that created this action method adds a comment showing a HTTP request that invokes the method. In this case it’s a GET request with three URL segments, the Movies controller, the Details method and a id value. Recall these segments are defined in Startup.

app.UseMvc(routes =>
{
    routes.MapRoute(
        name: "default",
        template: "{controller=Home}/{action=Index}/{id?}");
});

Code First makes it easy to search for data using the SingleOrDefaultAsync method. An important security feature built into the method is that the code verifies that the search method has found a movie before the code tries to do anything with it. For example, a hacker could introduce errors into the site by changing the URL created by the links from http://localhost:xxxx/Movies/Details/1 to something like http://localhost:xxxx/Movies/Details/12345 (or some other value that doesn’t represent an actual movie). If you did not check for a null movie, the app would throw an exception.

Examine the Delete and DeleteConfirmed methods.

public async Task<IActionResult> Delete(int? id)
{
    if (id == null)
    {
        return NotFound();
    }

    var movie = await _context.Movie.SingleOrDefaultAsync(m => m.ID == id);
    if (movie == null)
    {
        return NotFound();
    }

    return View(movie);
}

// POST: Movies/Delete/5
[HttpPost, ActionName("Delete")]
[ValidateAntiForgeryToken]
public async Task<IActionResult> DeleteConfirmed(int id)
{
    var movie = await _context.Movie.SingleOrDefaultAsync(m => m.ID == id);
    _context.Movie.Remove(movie);
    await _context.SaveChangesAsync();
    return RedirectToAction("Index");
}

Note that the HTTP GET Delete method doesn’t delete the specified movie, it returns a view of the movie where you can submit (HttpPost) the deletion. Performing a delete operation in response to a GET request (or for that matter, performing an edit operation, create operation, or any other operation that changes data) opens up a security hole.

The [HttpPost] method that deletes the data is named DeleteConfirmed to give the HTTP POST method a unique signature or name. The two method signatures are shown below:

// GET: Movies/Delete/5
public async Task<IActionResult> Delete(int? id)

// POST: Movies/Delete/
[HttpPost, ActionName("Delete")]
[ValidateAntiForgeryToken]
public async Task<IActionResult> DeleteConfirmed(int id)

The common language runtime (CLR) requires overloaded methods to have a unique parameter signature (same method name but different list of parameters). However, here you need two Delete methods – one for GET and one for POST – that both have the same parameter signature. (They both need to accept a single integer as a parameter.)

There are two approaches to this problem, one is to give the methods different names. That’s what the scaffolding mechanism did in the preceding example. However, this introduces a small problem: ASP.NET maps segments of a URL to action methods by name, and if you rename a method, routing normally wouldn’t be able to find that method. The solution is what you see in the example, which is to add the ActionName("Delete") attribute to the DeleteConfirmed method. That attribute performs mapping for the routing system so that a URL that includes /Delete/ for a POST request will find the DeleteConfirmed method.

Another common work around for methods that have identical names and signatures is to artificially change the signature of the POST method to include an extra (unused) parameter. That’s what we did in a previous post when we added the notUsed parameter. You could do the same thing here for the [HttpPost] Delete method:

[ValidateAntiForgeryToken]
public async Task<IActionResult> Delete(int id, bool notUsed)
{
    var movie = await _context.Movie.SingleOrDefaultAsync(m => m.ID == id);
    _context.Movie.Remove(movie);
    await _context.SaveChangesAsync();
    return RedirectToAction("Index");
}

ASP.NET Core on Nano Server¶

Attention

This tutorial uses a pre-release version of the Nano Server installation option of Windows Server Technical Preview 5. You may use the software in the virtual hard disk image only to internally demonstrate and evaluate it. You may not use the software in a live operating environment. Please see https://go.microsoft.com/fwlink/?LinkId=624232 for specific information about the end date for the preview.

In this tutorial, you’ll take an existing ASP.NET Core app and deploy it to a Nano Server instance running IIS.

Sections:

Introduction
Setting up the Nano Server Instance
Creating a file share
Open port in the Firewall
Installing IIS
Installing the ASP.NET Core Module (ANCM)
Installing .NET Core Framework
Publishing the application
Known issue running .NET Core CLI on Nano Server and Workaround
Running the Application

Introduction ¶

Nano Server is an installation option in Windows Server 2016, offering a tiny footprint, better security and better servicing than Server Core or full Server. Please consult the official Nano Server documentation for more details. There are 3 ways for you try out Nano Server for yourself:

You can download the Windows Server 2016 Technical Preview 5 ISO file, and build a Nano Server image
Download the Nano Server developer VHD
Create a VM in Azure using the Nano Server image in the Azure Gallery. If you don’t have an Azure account, you can get a free 30-day trial

In this tutorial, we will be using the pre-built Nano Server Developer VHD from Windows Server Technical Preview 5.

Before proceeding with this tutorial, you will need the published output of an existing ASP.NET Core application. Ensure your application is built to run in a 64-bit process.

Setting up the Nano Server Instance ¶

Create a new Virtual Machine using Hyper-V on your development machine using the previously downloaded VHD. The machine will require you to set an administator password before logging on. At the VM console, press F11 to set the password before the first logon.

After setting the local password, you will manage Nano Server using PowerShell remoting.

Connecting to your Nano Server Instance using PowerShell Remoting¶

Open an elevated PowerShell window to add your remote Nano Server instance to your TrustedHosts list.

$nanoServerIpAddress = "10.83.181.14"
Set-Item WSMan:\localhost\Client\TrustedHosts "$nanoServerIpAddress" -Concatenate -Force

NOTE: Replace the variable $nanoServerIpAddress with the correct IP address.

Once you have added your Nano Server instance to your TrustedHosts, you can connect to it using PowerShell remoting

$nanoServerSession = New-PSSession -ComputerName $nanoServerIpAddress -Credential ~\Administrator
Enter-PSSession $nanoServerSession

A successful connection results in a prompt with a format looking like: [10.83.181.14]: PS C:\Users\Administrator\Documents>

Open port in the Firewall ¶

Run the following commands in the remote session to open up a port in the firewall to listen for TCP traffic.

New-NetFirewallRule -Name "AspNet5 IIS" -DisplayName "Allow HTTP on TCP/8000" -Protocol TCP -LocalPort 8000 -Action Allow -Enabled True

Installing IIS ¶

Add the NanoServerPackage provider from the PowerShell gallery. Once the provider is installed and imported, you can install Windows packages.

Run the following commands in the PowerShell session that was created earlier:

Install-PackageProvider NanoServerPackage
Import-PackageProvider NanoServerPackage
Install-NanoServerPackage -Name Microsoft-NanoServer-IIS-Package

To quickly verify if IIS is setup correctly, you can visit the url http://<nanoserver-ip-address>/ and should see a welcome page. When IIS is installed, by default a web site called Default Web Site listening on port 80 is created.

Installing the ASP.NET Core Module (ANCM)¶

The ASP.NET Core Module is an IIS 7.5+ module which is responsible for process management of ASP.NET Core HTTP listeners and to proxy requests to processes that it manages. At the moment, the process to install the ASP.NET Core Module for IIS is manual. You will need to install the version of the .NET Core Windows Server Hosting bundle on a regular (not Nano) machine. After installing the bundle on a regular machine, you will need to copy the following files to the file share that we created earlier.

On a regular (not Nano) machine run the following copy commands:

copy C:\windows\system32\inetsrv\aspnetcore.dll ``\\<nanoserver-ip-address>\AspNetCoreSampleForNano``
copy C:\windows\system32\inetsrv\config\schema\aspnetcore_schema.xml ``\\<nanoserver-ip-address>\AspNetCoreSampleForNano``

On a Nano machine, you will need to copy the following files from the file share that we created earlier to the valid locations. So, run the following copy commands:

copy C:\PublishedApps\AspNetCoreSampleForNano\aspnetcore.dll C:\windows\system32\inetsrv\
copy C:\PublishedApps\AspNetCoreSampleForNano\aspnetcore_schema.xml C:\windows\system32\inetsrv\config\schema\

Run the following script in the remote session:

# Backup existing applicationHost.config
copy C:\Windows\System32\inetsrv\config\applicationHost.config C:\Windows\System32\inetsrv\config\applicationHost_BeforeInstallingANCM.config

Import-Module IISAdministration

# Initialize variables
$aspNetCoreHandlerFilePath="C:\windows\system32\inetsrv\aspnetcore.dll"
Reset-IISServerManager -confirm:$false
$sm = Get-IISServerManager

# Add AppSettings section 
$sm.GetApplicationHostConfiguration().RootSectionGroup.Sections.Add("appSettings")

# Set Allow for handlers section
$appHostconfig = $sm.GetApplicationHostConfiguration()
$section = $appHostconfig.GetSection("system.webServer/handlers")
$section.OverrideMode="Allow"

# Add aspNetCore section to system.webServer
$sectionaspNetCore = $appHostConfig.RootSectionGroup.SectionGroups["system.webServer"].Sections.Add("aspNetCore")
$sectionaspNetCore.OverrideModeDefault = "Allow"
$sm.CommitChanges()

# Configure globalModule
Reset-IISServerManager -confirm:$false
$globalModules = Get-IISConfigSection "system.webServer/globalModules" | Get-IISConfigCollection
New-IISConfigCollectionElement $globalModules -ConfigAttribute @{"name"="AspNetCoreModule";"image"=$aspNetCoreHandlerFilePath}

# Configure module
$modules = Get-IISConfigSection "system.webServer/modules" | Get-IISConfigCollection
New-IISConfigCollectionElement $modules -ConfigAttribute @{"name"="AspNetCoreModule"}

# Backup existing applicationHost.config
copy C:\Windows\System32\inetsrv\config\applicationHost.config C:\Windows\System32\inetsrv\config\applicationHost_AfterInstallingANCM.config

NOTE: Delete the files aspnetcore.dll and aspnetcore_schema.xml from the share after the above step.

Installing .NET Core Framework ¶

If you published a portable app, .NET Core must be installed on the target machine. Execute the following Powershell script in a remote Powershell session to install the .NET Framework on your Nano Server.

$SourcePath = "https://go.microsoft.com/fwlink/?LinkID=809115"
$DestinationPath = "C:\dotnet"

$EditionId = (Get-ItemProperty -Path 'HKLM:\SOFTWARE\Microsoft\Windows NT\CurrentVersion' -Name 'EditionID').EditionId

if (($EditionId -eq "ServerStandardNano") -or
    ($EditionId -eq "ServerDataCenterNano") -or
    ($EditionId -eq "NanoServer") -or
    ($EditionId -eq "ServerTuva")) {

    $TempPath = [System.IO.Path]::GetTempFileName()
    if (($SourcePath -as [System.URI]).AbsoluteURI -ne $null)
    {
        $handler = New-Object System.Net.Http.HttpClientHandler
        $client = New-Object System.Net.Http.HttpClient($handler)
        $client.Timeout = New-Object System.TimeSpan(0, 30, 0)
        $cancelTokenSource = [System.Threading.CancellationTokenSource]::new()
        $responseMsg = $client.GetAsync([System.Uri]::new($SourcePath), $cancelTokenSource.Token)
        $responseMsg.Wait()
        if (!$responseMsg.IsCanceled)
        {
            $response = $responseMsg.Result
            if ($response.IsSuccessStatusCode)
            {
                $downloadedFileStream = [System.IO.FileStream]::new($TempPath, [System.IO.FileMode]::Create, [System.IO.FileAccess]::Write)
                $copyStreamOp = $response.Content.CopyToAsync($downloadedFileStream)
                $copyStreamOp.Wait()
                $downloadedFileStream.Close()
                if ($copyStreamOp.Exception -ne $null)
                {
                    throw $copyStreamOp.Exception
                }
            }
        }
    }
    else
    {
        throw "Cannot copy from $SourcePath"
    }
    [System.IO.Compression.ZipFile]::ExtractToDirectory($TempPath, $DestinationPath)
    Remove-Item $TempPath
}

Publishing the application ¶

Copy over the published output of your existing application to the file share.

You may need to make changes to your web.config to point to where you extracted dotnet.exe. Alternatively, you can add dotnet.exe to your path.

Example of how a web.config might look like if dotnet.exe was not on the path:

<?xml version="1.0" encoding="utf-8"?>
<configuration>
  <system.webServer>
    <handlers>
      <add name="aspNetCore" path="*" verb="*" modules="AspNetCoreModule" resourceType="Unspecified" />
    </handlers>
    <aspNetCore processPath="C:\dotnet\dotnet.exe" arguments=".\AspNetCoreSampleForNano.dll" stdoutLogEnabled="false" stdoutLogFile=".\logs\stdout" forwardWindowsAuthToken="true" />
  </system.webServer>
</configuration>

Run the following commands in the remote session to create a new site in IIS for the published app. This script uses the DefaultAppPool for simplicity. For more considerations on running under an application pool, see Application Pools.

Import-module IISAdministration
New-IISSite -Name "AspNetCore" -PhysicalPath c:\PublishedApps\AspNetCoreSampleForNano -BindingInformation "*:8000:"

Known issue running .NET Core CLI on Nano Server and Workaround ¶

If you’re using Nano Server Technical Preview 5 with .NET Core CLI, you will need to copy all DLL files from c:\windows\system32\forwarders to c:\windows\system32, due to a bug that has since been fixed in later releases.

If you use dotnet publish, make sure to copy all DLL files from c:\windows\system32\forwarders to your publish directory as well.

If your Nano Server Technical Preview 5 build is updated or serviced, please make sure to repeat this process, in case any of the DLLs have been updated as well.

Running the Application ¶

The published web app should be accessible in browser at http://<nanoserver-ip-address>:8000. If you have set up logging as described in Log creation and redirection, you should be able to view your logs at C:\PublishedApps\AspNetCoreSampleForNano\logs.

🔧 Creating Backend Services for Native Mobile Applications¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Developing ASP.NET Core applications using dotnet watch¶

By Victor Hurdugaci

Introduction¶

dotnet watch is a development time tool that runs a dotnet command when source files change. It can be used to compile, run tests, or publish when code changes.

In this tutorial we’ll use an existing WebApi application that calculates the sum and product of two numbers to demonstrate the use cases of dotnet watch. The sample application contains an intentional bug that we’ll fix as part of this tutorial.

Getting started¶

Start by downloading the sample application. It contains two projects, WebApp (a web application) and WebAppTests (unit tests for the web application)

In a console, open the folder where you downloaded the sample application and run:

dotnet restore
cd WebApp
dotnet run

The console output will show messages similar to the ones below, indicating that the application is now running and waiting for requests:

$ dotnet run
Project WebApp (.NETCoreApp,Version=v1.0) will be compiled because inputs were modified
Compiling WebApp for .NETCoreApp,Version=v1.0

Compilation succeeded.
  0 Warning(s)
  0 Error(s)

Time elapsed 00:00:02.6049991

Hosting environment: Production
Content root path: /Users/user/dev/aspnet/Docs/aspnet/tutorials/dotnet-watch/sample/WebApp
Now listening on: http://localhost:5000
Application started. Press Ctrl+C to shut down.

In a web browser, navigate to http://localhost:5000/api/math/sum?a=4&b=5 and you should see the result 9.

If you navigate to http://localhost:5000/api/math/product?a=4&b=5 instead, you’d expect to get the result 20. Instead, you get 9 again.

We’ll fix that.

Adding `dotnet watch` to a project¶

Add Microsoft.DotNet.Watcher.Tools to the tools section of the WebApp/project.json file as in the example below:

"tools": {
  "Microsoft.DotNet.Watcher.Tools": "1.0.0-preview2-final"
},

Run dotnet restore.

The console output will show messages similar to the ones below:

log  : Restoring packages for /Users/user/dev/aspnet/Docs/aspnet/tutorials/dotnet-watch/sample/WebApp/project.json...
log  : Restoring packages for tool 'Microsoft.DotNet.Watcher.Tools' in /Users/user/dev/aspnet/Docs/aspnet/tutorials/dotnet-watch/sample/WebApp/project.json...
log  : Installing Microsoft.DotNet.Watcher.Core 1.0.0-preview2-final.
log  : Installing Microsoft.DotNet.Watcher.Tools 1.0.0-preview2-final.

Running `dotnet` commands using `dotnet watch`¶

Any dotnet command can be run with dotnet watch: For example:

Command	Command with watch
`dotnet run`	`dotnet watch run`
`dotnet run -f net451`	`dotnet watch run -f net451`
`dotnet run -f net451 -- --arg1`	`dotnet watch run -f net451 -- --arg1`
`dotnet test`	`dotnet watch test`

To run WebApp using the watcher, run dotnet watch run in the WebApp folder. The console output will show messages similar to the ones below, indicating that dotnet watch is now watching code files:

user$ dotnet watch run
[DotNetWatcher] info: Running dotnet with the following arguments: run
[DotNetWatcher] info: dotnet process id: 39746
Project WebApp (.NETCoreApp,Version=v1.0) was previously compiled. Skipping compilation.
Hosting environment: Production
Content root path: /Users/user/dev/aspnet/Docs/aspnet/tutorials/dotnet-watch/sample/WebApp
Now listening on: http://localhost:5000
Application started. Press Ctrl+C to shut down.

Making changes with `dotnet watch`¶

Make sure dotnet watch is running.

Let’s fix the bug that we discovered when we tried to compute the product of two number.

Open WebApp/Controllers/MathController.cs.

We’ve intentionally introduced a bug in the code.

    public static int Product(int a, int b)
    {
        // We have an intentional bug here
        // + should be *
        return a + b; 
    }

Fix the code by replacing a + b with a * b.

Save the file. The console output will show messages similar to the ones below, indicating that dotnet watch detected a file change and restarted the application.

[DotNetWatcher] info: File changed: /Users/user/dev/aspnet/Docs/aspnet/tutorials/dotnet-watch/sample/WebApp/Controllers/MathController.cs
[DotNetWatcher] info: Running dotnet with the following arguments: run
[DotNetWatcher] info: dotnet process id: 39940
Project WebApp (.NETCoreApp,Version=v1.0) will be compiled because inputs were modified
Compiling WebApp for .NETCoreApp,Version=v1.0
Compilation succeeded.
  0 Warning(s)
  0 Error(s)
Time elapsed 00:00:03.3312829

Hosting environment: Production
Content root path: /Users/user/dev/aspnet/Docs/aspnet/tutorials/dotnet-watch/sample/WebApp
Now listening on: http://localhost:5000
Application started. Press Ctrl+C to shut down.

Verify http://localhost:5000/api/math/product?a=4&b=5 returns the correct result.

Running tests using `dotnet watch`¶

The file watcher can run other dotnet commands like test or publish.

Open the WebAppTests folder that already has dotnet watch in project.json.
Run dotnet watch test.

If you previously fixed the bug in the MathController then you’ll see an output similar to the one below, otherwise you’ll see a test failure:

WebAppTests user$ dotnet watch test
[DotNetWatcher] info: Running dotnet with the following arguments: test
[DotNetWatcher] info: dotnet process id: 40193
Project WebApp (.NETCoreApp,Version=v1.0) was previously compiled. Skipping compilation.
Project WebAppTests (.NETCoreApp,Version=v1.0) was previously compiled. Skipping compilation.
xUnit.net .NET CLI test runner (64-bit .NET Core osx.10.11-x64)
  Discovering: WebAppTests
  Discovered:  WebAppTests
  Starting:    WebAppTests
  Finished:    WebAppTests
=== TEST EXECUTION SUMMARY ===
   WebAppTests  Total: 2, Errors: 0, Failed: 0, Skipped: 0, Time: 0.259s
SUMMARY: Total: 1 targets, Passed: 1, Failed: 0.
[DotNetWatcher] info: dotnet exit code: 0
[DotNetWatcher] info: Waiting for a file to change before restarting dotnet...

Once all the tests run, the watcher will indicate that it’s waiting for a file to change before restarting dotnet test.

Open the controller file in WebApp/Controllers/MathController.cs and change some code. If you haven’t fixed the product bug, do it now. Save the file.

dotnet watch will detect the file change and rerun the tests. The console output will show messages similar to the one below:

[DotNetWatcher] info: File changed: /Users/user/dev/aspnet/Docs/aspnet/tutorials/dotnet-watch/sample/WebApp/Controllers/MathController.cs
[DotNetWatcher] info: Running dotnet with the following arguments: test
[DotNetWatcher] info: dotnet process id: 40233
Project WebApp (.NETCoreApp,Version=v1.0) will be compiled because inputs were modified
Compiling WebApp for .NETCoreApp,Version=v1.0
Compilation succeeded.
  0 Warning(s)
  0 Error(s)
Time elapsed 00:00:03.2127590
Project WebAppTests (.NETCoreApp,Version=v1.0) will be compiled because dependencies changed
Compiling WebAppTests for .NETCoreApp,Version=v1.0
Compilation succeeded.
  0 Warning(s)
  0 Error(s)
Time elapsed 00:00:02.1204052

xUnit.net .NET CLI test runner (64-bit .NET Core osx.10.11-x64)
  Discovering: WebAppTests
  Discovered:  WebAppTests
  Starting:    WebAppTests
  Finished:    WebAppTests
=== TEST EXECUTION SUMMARY ===
   WebAppTests  Total: 2, Errors: 0, Failed: 0, Skipped: 0, Time: 0.260s
SUMMARY: Total: 1 targets, Passed: 1, Failed: 0.
[DotNetWatcher] info: dotnet exit code: 0

[DotNetWatcher] info: Waiting for a file to change before restarting dotnet...

Fundamentals¶

Application Startup¶

ASP.NET Core provides complete control of how individual requests are handled by your application. The Startup class is the entry point to the application, setting up configuration and wiring up services the application will use. Developers configure a request pipeline in the Startup class that is used to handle all requests made to the application.

Sections:

The Startup class
The Configure method
The ConfigureServices method
Services Available in Startup
Additional Resources

The Startup class ¶

In ASP.NET Core, the Startup class provides the entry point for an application, and is required for all applications. It’s possible to have environment-specific startup classes and methods (see Working with Multiple Environments 다중 환경에서 작업하기), but regardless, one Startup class will serve as the entry point for the application. ASP.NET searches the primary assembly for a class named Startup (in any namespace). You can specify a different assembly to search using the Hosting:Application configuration key. It doesn’t matter whether the class is defined as public; ASP.NET will still load it if it conforms to the naming convention. If there are multiple Startup classes, this will not trigger an exception. ASP.NET will select one based on its namespace (matching the project’s root namespace first, otherwise using the class in the alphabetically first namespace).

The Startup class can optionally accept dependencies in its constructor that are provided through dependency injection. Typically, the way an application will be configured is defined within its Startup class’s constructor (see Configuration). The Startup class must define a Configure method, and may optionally also define a ConfigureServices method, which will be called when the application is started.

The Configure method ¶

The Configure method is used to specify how the ASP.NET application will respond to individual HTTP requests. At its simplest, you can configure every request to receive the same response. However, most real-world applications require more functionality than this. More complex sets of pipeline configuration can be encapsulated in middleware and added using extension methods on IApplicationBuilder.

Your Configure method must accept an IApplicationBuilder parameter. Additional services, like IHostingEnvironment and ILoggerFactory may also be specified, in which case these services will be injected by the server if they are available. In the following example from the default web site template, you can see several extension methods are used to configure the pipeline with support for BrowserLink, error pages, static files, ASP.NET MVC, and Identity.

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    loggerFactory.AddConsole(Configuration.GetSection("Logging"));
    loggerFactory.AddDebug();

    if (env.IsDevelopment())
    {
        app.UseDeveloperExceptionPage();
        app.UseDatabaseErrorPage();
        app.UseBrowserLink();
    }
    else
    {
        app.UseExceptionHandler("/Home/Error");
    }

    app.UseStaticFiles();

    app.UseIdentity();

    // Add external authentication middleware below. To configure them please see http://go.microsoft.com/fwlink/?LinkID=532715

    app.UseMvc(routes =>
    {
        routes.MapRoute(
            name: "default",
            template: "{controller=Home}/{action=Index}/{id?}");
    });
}

Each Use extension method adds middleware to the request pipeline. For instance, the UseMvc extension method adds the routing middleware to the request pipeline and configures MVC as the default handler.

You can learn all about middleware and using IApplicationBuilder to define your request pipeline in the Middleware topic.

The ConfigureServices method ¶

Your Startup class can optionally include a ConfigureServices method for configuring services that are used by your application. The ConfigureServices method is a public method on your Startup class that takes an IServiceCollection instance as a parameter and optionally returns an IServiceProvider. The ConfigureServices method is called before Configure. This is important, because some features like ASP.NET MVC require certain services to be added in ConfigureServices before they can be wired up to the request pipeline.

Just as with Configure, it is recommended that features that require substantial setup within ConfigureServices be wrapped up in extension methods on IServiceCollection. You can see in this example from the default web site template that several Add[Something] extension methods are used to configure the app to use services from Entity Framework, Identity, and MVC:

public void ConfigureServices(IServiceCollection services)
{
    // Add framework services.
    services.AddDbContext<ApplicationDbContext>(options =>
        options.UseSqlServer(Configuration.GetConnectionString("DefaultConnection")));

    services.AddIdentity<ApplicationUser, IdentityRole>()
        .AddEntityFrameworkStores<ApplicationDbContext>()
        .AddDefaultTokenProviders();

    services.AddMvc();

    // Add application services.
    services.AddTransient<IEmailSender, AuthMessageSender>();
    services.AddTransient<ISmsSender, AuthMessageSender>();
}

Adding services to the services container makes them available within your application via dependency injection. Just as the Startup class is able to specify dependencies its methods require as parameters, rather than hard-coding to a specific implementation, so too can your middleware, MVC controllers and other classes in your application.

The ConfigureServices method is also where you should add configuration option classes that you would like to have available in your application. See the Configuration topic to learn more about configuring options.

Services Available in Startup ¶

ASP.NET Core provides certain application services and objects during your application’s startup. You can request certain sets of these services by simply including the appropriate interface as a parameter on your Startup class’s constructor or one of its Configure or ConfigureServices methods. The services available to each method in the Startup class are described below. The framework services and objects include:

IApplicationBuilder: Used to build the application request pipeline. Available only to the Configure method in Startup. Learn more about Request Features.
IHostingEnvironment: Provides the current EnvironmentName, ContentRootPath, WebRootPath, and web root file provider. Available to the Startup constructor and Configure method.
ILoggerFactory: Provides a mechanism for creating loggers. Available to the Startup constructor and Configure method. Learn more about Logging.
IServiceCollection: The current set of services configured in the container. Available only to the ConfigureServices method, and used by that method to configure the services available to an application.

Looking at each method in the Startup class in the order in which they are called, the following services may be requested as parameters:

Startup Constructor - IApplicationEnvironment - IHostingEnvironment - ILoggerFactory

ConfigureServices - IServiceCollection

Configure - IApplicationBuilder - IApplicationEnvironment - IHostingEnvironment - ILoggerFactory

Note

Although ILoggerFactory is available in the constructor, it is typically configured in the Configure method. Learn more about Logging.

Additional Resources ¶

Middleware¶

By Steve Smith and Rick Anderson

Sections:

What is middleware
Creating a middleware pipeline with IApplicationBuilder
Built-in middleware
Writing middleware
Additional Resources

What is middleware ¶

Middleware are software components that are assembled into an application pipeline to handle requests and responses. Each component chooses whether to pass the request on to the next component in the pipeline, and can perform certain actions before and after the next component is invoked in the pipeline. Request delegates are used to build the request pipeline. The request delegates handle each HTTP request.

Request delegates are configured using Run, Map, and Use extension methods on the IApplicationBuilder type that is passed into the Configure method in the Startup class. An individual request delegate can be specified in-line as an anonymous method, or it can be defined in a reusable class. These reusable classes are middleware, or middleware components. Each middleware component in the request pipeline is responsible for invoking the next component in the pipeline, or short-circuiting the chain if appropriate.

Migrating HTTP Modules to Middleware explains the difference between request pipelines in ASP.NET Core and the previous versions and provides more middleware samples.

Creating a middleware pipeline with IApplicationBuilder ¶

The ASP.NET request pipeline consists of a sequence of request delegates, called one after the next, as this diagram shows (the thread of execution follows the black arrows):

Each delegate has the opportunity to perform operations before and after the next delegate. Any delegate can choose to stop passing the request on to the next delegate, and instead handle the request itself. This is referred to as short-circuiting the request pipeline, and is desirable because it allows unnecessary work to be avoided. For example, an authorization middleware might only call the next delegate if the request is authenticated; otherwise it could short-circuit the pipeline and return a “Not Authorized” response. Exception handling delegates need to be called early on in the pipeline, so they are able to catch exceptions that occur in deeper calls within the pipeline.

You can see an example of setting up the request pipeline in the default web site template that ships with Visual Studio 2015. The Configure method adds the following middleware components:

Error handling (for both development and non-development environments)
Static file server
Authentication
MVC

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    loggerFactory.AddConsole(Configuration.GetSection("Logging"));
    loggerFactory.AddDebug();

    if (env.IsDevelopment())
    {
        app.UseDeveloperExceptionPage();
        app.UseDatabaseErrorPage();
        app.UseBrowserLink();
    }
    else
    {
        app.UseExceptionHandler("/Home/Error");
    }

    app.UseStaticFiles();

    app.UseIdentity();

    // Add external authentication middleware below. To configure them please see http://go.microsoft.com/fwlink/?LinkID=532715

    app.UseMvc(routes =>
    {
        routes.MapRoute(
            name: "default",
            template: "{controller=Home}/{action=Index}/{id?}");
    });
}

In the code above (in non-development environments), UseExceptionHandler is the first middleware added to the pipeline, therefore will catch any exceptions that occur in later calls.

The static file module provides no authorization checks. Any files served by it, including those under wwwroot are publicly available. If you want to serve files based on authorization:

Store them outside of wwwroot and any directory accessible to the static file middleware.
Deliver them through a controller action, returning a FileResult where authorization is applied.

A request that is handled by the static file module will short circuit the pipeline. (see Working with Static Files.) If the request is not handled by the static file module, it’s passed on to the Identity module, which performs authentication. If the request is not authenticated, the pipeline is short circuited. If the request does not fail authentication, the last stage of this pipeline is called, which is the MVC framework.

Note

The order in which you add middleware components is generally the order in which they take effect on the request, and then in reverse for the response. This can be critical to your app’s security, performance and functionality. In the code above, the static file middleware is called early in the pipeline so it can handle requests and short circuit without going through unnecessary components. The authentication middleware is added to the pipeline before anything that handles requests that need to be authenticated. Exception handling must be registered before other middleware components in order to catch exceptions thrown by those components.

The simplest possible ASP.NET application sets up a single request delegate that handles all requests. In this case, there isn’t really a request “pipeline”, so much as a single anonymous function that is called in response to every HTTP request.

app.Run(async context =>
{
    await context.Response.WriteAsync("Hello, World!");
});

The first App.Run delegate terminates the pipeline. In the following example, only the first delegate (“Hello, World!”) will run.

public void Configure(IApplicationBuilder app)
{
    app.Run(async context =>
    {
        await context.Response.WriteAsync("Hello, World!");
    });

    app.Run(async context =>
    {
        await context.Response.WriteAsync("Hello, World, Again!");
    });

You chain multiple request delegates together; the next parameter represents the next delegate in the pipeline. You can terminate (short-circuit) the pipeline by not calling the next parameter. You can typically perform actions both before and after the next delegate, as this example demonstrates:

public void ConfigureLogInline(IApplicationBuilder app, ILoggerFactory loggerfactory)
{
    loggerfactory.AddConsole(minLevel: LogLevel.Information);
    var logger = loggerfactory.CreateLogger(_environment);
    app.Use(async (context, next) =>
    {
        logger.LogInformation("Handling request.");
        await next.Invoke();
        logger.LogInformation("Finished handling request.");
    });

    app.Run(async context =>
    {
        await context.Response.WriteAsync("Hello from " + _environment);
    });
}

Warning

Avoid modifying HttpResponse after invoking next, one of the next components in the pipeline may have written to the response, causing it to be sent to the client.

Note

This ConfigureLogInline method is called when the application is run with an environment set to LogInline. Learn more about Working with Multiple Environments 다중 환경에서 작업하기. We will be using variations of Configure[Environment] to show different options in the rest of this article. The easiest way to run the samples in Visual Studio is with the web command, which is configured in project.json. See also Application Startup.

In the above example, the call to await next.Invoke() will call into the next delegate await context.Response.WriteAsync("Hello from " + _environment);. The client will receive the expected response (“Hello from LogInline”), and the server’s console output includes both the before and after messages:

Run, Map, and Use¶

You configure the HTTP pipeline using Run, Map, and Use. The Run method short circuits the pipeline (that is, it will not call a next request delegate). Thus, Run should only be called at the end of your pipeline. Run is a convention, and some middleware components may expose their own Run[Middleware] methods that should only run at the end of the pipeline. The following two middleware are equivalent as the Use version doesn’t use the next parameter:

public void ConfigureEnvironmentOne(IApplicationBuilder app)
{
    app.Run(async context =>
    {
        await context.Response.WriteAsync("Hello from " + _environment);
    });
}

public void ConfigureEnvironmentTwo(IApplicationBuilder app)
{
    app.Use(async (context, next) =>
    {
        await context.Response.WriteAsync("Hello from " + _environment);
    });
}

Note

The IApplicationBuilder interface exposes a single Use method, so technically they’re not all extension methods.

We’ve already seen several examples of how to build a request pipeline with Use. Map* extensions are used as a convention for branching the pipeline. The current implementation supports branching based on the request’s path, or using a predicate. The Map extension method is used to match request delegates based on a request’s path. Map simply accepts a path and a function that configures a separate middleware pipeline. In the following example, any request with the base path of /maptest will be handled by the pipeline configured in the HandleMapTest method.

private static void HandleMapTest(IApplicationBuilder app)
{
    app.Run(async context =>
    {
        await context.Response.WriteAsync("Map Test Successful");
    });
}

public void ConfigureMapping(IApplicationBuilder app)
{
    app.Map("/maptest", HandleMapTest);

}

Note

When Map is used, the matched path segment(s) are removed from HttpRequest.Path and appended to HttpRequest.PathBase for each request.

In addition to path-based mapping, the MapWhen method supports predicate-based middleware branching, allowing separate pipelines to be constructed in a very flexible fashion. Any predicate of type Func<HttpContext, bool> can be used to map requests to a new branch of the pipeline. In the following example, a simple predicate is used to detect the presence of a query string variable branch:

private static void HandleBranch(IApplicationBuilder app)
{
    app.Run(async context =>
    {
        await context.Response.WriteAsync("Branch used.");
    });
}

public void ConfigureMapWhen(IApplicationBuilder app)
{
    app.MapWhen(context => {
        return context.Request.Query.ContainsKey("branch");
    }, HandleBranch);

    app.Run(async context =>
    {
        await context.Response.WriteAsync("Hello from " + _environment);
    });
}

Using the configuration shown above, any request that includes a query string value for branch will use the pipeline defined in the HandleBranch method (in this case, a response of “Branch used.”). All other requests (that do not define a query string value for branch) will be handled by the delegate defined on line 17.

You can also nest Maps:

app.Map("/level1", level1App => {
    level1App.Map("/level2a", level2AApp => {
        // "/level1/level2a"
        //...
    });
    level1App.Map("/level2b", level2BApp => {
        // "/level1/level2b"
        //...
    });
});

Built-in middleware ¶

ASP.NET ships with the following middleware components:

Middleware¶
Middleware	Description
Authentication	Provides authentication support.
CORS	Configures Cross-Origin Resource Sharing.
Routing	Define and constrain request routes.
Session	Provides support for managing user sessions.
Static Files	Provides support for serving static files, and directory browsing.

Writing middleware ¶

The CodeLabs middleware tutorial provides a good introduction to writing middleware.

For more complex request handling functionality, the ASP.NET team recommends implementing the middleware in its own class, and exposing an IApplicationBuilder extension method that can be called from the Configure method. The simple logging middleware shown in the previous example can be converted into a middleware class that takes in the next RequestDelegate in its constructor and supports an Invoke method as shown:

RequestLoggerMiddleware.cs¶

using System.Threading.Tasks;
using Microsoft.AspNetCore.Http;
using Microsoft.Extensions.Logging;

namespace MiddlewareSample
{
    public class RequestLoggerMiddleware
    {
        private readonly RequestDelegate _next;
        private readonly ILogger _logger;

        public RequestLoggerMiddleware(RequestDelegate next, ILoggerFactory loggerFactory)
        {
            _next = next;
            _logger = loggerFactory.CreateLogger<RequestLoggerMiddleware>();
        }

        public async Task Invoke(HttpContext context)
        {
            _logger.LogInformation("Handling request: " + context.Request.Path);
            await _next.Invoke(context);
            _logger.LogInformation("Finished handling request.");
        }
    }
}

The middleware follows the Explicit Dependencies Principle and exposes all of its dependencies in its constructor. Middleware can take advantage of the UseMiddleware<T> extension to inject services directly into their constructors, as shown in the example below. Dependency injected services are automatically filled, and the extension takes a params array of arguments to be used for non-injected parameters.

RequestLoggerExtensions.cs¶

public static class RequestLoggerExtensions
{
    public static IApplicationBuilder UseRequestLogger(this IApplicationBuilder builder)
    {
        return builder.UseMiddleware<RequestLoggerMiddleware>();
    }
}

Using the extension method and associated middleware class, the Configure method becomes very simple and readable.

public void ConfigureLogMiddleware(IApplicationBuilder app,
    ILoggerFactory loggerfactory)
{
    loggerfactory.AddConsole(minLevel: LogLevel.Information);

    app.UseRequestLogger();

    app.Run(async context =>
    {
        await context.Response.WriteAsync("Hello from " + _environment);
    });
}

Although RequestLoggerMiddleware requires an ILoggerFactory parameter in its constructor, neither the Startup class nor the UseRequestLogger extension method need to explicitly supply it. Instead, it is automatically provided through dependency injection performed within UseMiddleware<T>.

Testing the middleware (by setting the Hosting:Environment environment variable to LogMiddleware) should result in output like the following (when using WebListener):

Note

The UseStaticFiles extension method (which creates the StaticFileMiddleware) also uses UseMiddleware<T>. In this case, the StaticFileOptions parameter is passed in, but other constructor parameters are supplied by UseMiddleware<T> and dependency injection.

Additional Resources ¶

Working with Static Files¶

Static files, such as HTML, CSS, image, and JavaScript, are assets that an ASP.NET Core app can serve directly to clients.

Sections

Serving static files
Static file authorization
Enabling directory browsing
Serving a default document
UseFileServer
Non-standard content types
Additional Resources

Serving static files ¶

Static files are typically located in the web root (<content-root>/wwwroot) folder. See Content root and Web root in Introduction to ASP.NET Core for more information. You generally set the content root to be the current directory so that your project’s web root will be found while in development.

public static void Main(string[] args)
{
    var host = new WebHostBuilder()
        .UseKestrel()
        .UseContentRoot(Directory.GetCurrentDirectory())
        .UseIISIntegration()
        .UseStartup<Startup>()
        .Build();

    host.Run();
}

Static files can be stored in any folder under the web root and accessed with a relative path to that root. For example, when you create a default Web application project using Visual Studio, there are several folders created within the wwwroot folder - css, images, and js. The URI to access an image in the images subfolder:

http://<app>/images/<imageFileName>
http://localhost:9189/images/banner3.svg

In order for static files to be served, you must configure the Middleware to add static files to the pipeline. The static file middleware can be configured by adding a dependency on the Microsoft.AspNetCore.StaticFiles package to your project and then calling the UseStaticFiles extension method from Startup.Configure:

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    app.UseStaticFiles();
}

app.UseStaticFiles(); makes the files in web root (wwwroot by default) servable. Later I’ll show how to make other directory contents servable with UseStaticFiles.

You must include “Microsoft.AspNetCore.StaticFiles” in the project.json file.

Note

web root defaults to the wwwroot directory, but you can set the web root directory with UseWebRoot. See Introduction to ASP.NET Core for more information.

Suppose you have a project hierarchy where the static files you wish to serve are outside the web root. For example:

wwwroot

css

images

...

MyStaticFiles

test.png

For a request to access test.png, configure the static files middleware as follows:

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    app.UseStaticFiles();

    app.UseStaticFiles(new StaticFileOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"MyStaticFiles")),
        RequestPath = new PathString("/StaticFiles")
    });
}

A request to http://<app>/StaticFiles/test.png will serve the test.png file.

Static file authorization ¶

The static file module provides no authorization checks. Any files served by it, including those under wwwroot are publicly available. To serve files based on authorization:

Store them outside of wwwroot and any directory accessible to the static file middleware and
Serve them through a controller action, returning a FileResult where authorization is applied

Enabling directory browsing ¶

Directory browsing allows the user of your web app to see a list of directories and files within a specified directory. Directory browsing is disabled by default for security reasons (see Considerations). To enable directory browsing, call the UseDirectoryBrowser extension method from Startup.Configure:

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    app.UseStaticFiles(); // For the wwwroot folder

    app.UseStaticFiles(new StaticFileOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"wwwroot\images")),
        RequestPath = new PathString("/MyImages")
    });

    app.UseDirectoryBrowser(new DirectoryBrowserOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"wwwroot\images")),
        RequestPath = new PathString("/MyImages")
    });
}

And add required services by calling AddDirectoryBrowser extension method from Startup.ConfigureServices:

public void ConfigureServices(IServiceCollection services)
{
    services.AddDirectoryBrowser();
}

The code above allows directory browsing of the wwwroot/images folder using the URL http://<app>/MyImages, with links to each file and folder:

See Considerations on the security risks when enabling browsing.

Note the two app.UseStaticFiles calls. The first one is required to serve the CSS, images and JavaScript in the wwwroot folder, and the second call for directory browsing of the wwwroot/images folder using the URL http://<app>/MyImages:

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    app.UseStaticFiles(); // For the wwwroot folder

    app.UseStaticFiles(new StaticFileOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"wwwroot\images")),
        RequestPath = new PathString("/MyImages")
    });

    app.UseDirectoryBrowser(new DirectoryBrowserOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"wwwroot\images")),
        RequestPath = new PathString("/MyImages")
    });
}

Serving a default document ¶

Setting a default home page gives site visitors a place to start when visiting your site. In order for your Web app to serve a default page without the user having to fully qualify the URI, call the UseDefaultFiles extension method from Startup.Configure as follows.

public void Configure(IApplicationBuilder app)
{
    app.UseDefaultFiles();
    app.UseStaticFiles();
}

Note

UseDefaultFiles must be called before UseStaticFiles to serve the default file. UseDefaultFiles is a URL re-writer that doesn’t actually serve the file. You must enable the static file middleware (UseStaticFiles) to serve the file.

With UseDefaultFiles, requests to a folder will search for:

default.htm

default.html

index.htm

index.html

The first file found from the list will be served as if the request was the fully qualified URI (although the browser URL will continue to show the URI requested).

The following code shows how to change the default file name to mydefault.html.

public void Configure(IApplicationBuilder app)
{
    // Serve my app-specific default file, if present.
    DefaultFilesOptions options = new DefaultFilesOptions();
    options.DefaultFileNames.Clear();
    options.DefaultFileNames.Add("mydefault.html");
    app.UseDefaultFiles(options);
    app.UseStaticFiles();
}

UseFileServer ¶

UseFileServer combines the functionality of UseStaticFiles, UseDefaultFiles, and UseDirectoryBrowser.

The following code enables static files and the default file to be served, but does not allow directory browsing:

app.UseFileServer();

The following code enables static files, default files and directory browsing:

app.UseFileServer(enableDirectoryBrowsing: true);

See Considerations on the security risks when enabling browsing. As with UseStaticFiles, UseDefaultFiles, and UseDirectoryBrowser, if you wish to serve files that exist outside the web root, you instantiate and configure an FileServerOptions object that you pass as a parameter to UseFileServer. For example, given the following directory hierarchy in your Web app:

wwwroot
- css
- images
- ...
MyStaticFiles
- test.png
- default.html

Using the hierarchy example above, you might want to enable static files, default files, and browsing for the MyStaticFiles directory. In the following code snippet, that is accomplished with a single call to FileServerOptions.

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    app.UseStaticFiles();

    app.UseFileServer(new FileServerOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"MyStaticFiles")),
        RequestPath = new PathString("/StaticFiles"),
        EnableDirectoryBrowsing = true
    });
}

If enableDirectoryBrowsing is set to true you are required to call AddDirectoryBrowser extension method from Startup.ConfigureServices:

public void ConfigureServices(IServiceCollection services)
{
    services.AddDirectoryBrowser();
}

Using the file hierarchy and code above:

URI	Response
http://<app>/StaticFiles/test.png	MyStaticFiles/test.png
http://<app>/StaticFiles	MyStaticFiles/default.html

If no default named files are in the MyStaticFiles directory, http://<app>/StaticFiles returns the directory listing with clickable links:

Note

UseDefaultFiles and UseDirectoryBrowser will take the url http://<app>/StaticFiles without the trailing slash and cause a client side redirect to http://<app>/StaticFiles/ (adding the trailing slash). Without the trailing slash relative URLs within the documents would be incorrect.

FileExtensionContentTypeProvider¶

The FileExtensionContentTypeProvider class contains a collection that maps file extensions to MIME content types. In the following sample, several file extensions are registered to known MIME types, the ”.rtf” is replaced, and ”.mp4” is removed.

public void Configure(IApplicationBuilder app)
{
    // Set up custom content types -associating file extension to MIME type
    var provider = new FileExtensionContentTypeProvider();
    // Add new mappings
    provider.Mappings[".myapp"] = "application/x-msdownload";
    provider.Mappings[".htm3"] = "text/html";
    provider.Mappings[".image"] = "image/png";
    // Replace an existing mapping
    provider.Mappings[".rtf"] = "application/x-msdownload";
    // Remove MP4 videos.
    provider.Mappings.Remove(".mp4");

    app.UseStaticFiles(new StaticFileOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"wwwroot\images")),
        RequestPath = new PathString("/MyImages"),
        ContentTypeProvider = provider
    });

    app.UseDirectoryBrowser(new DirectoryBrowserOptions()
    {
        FileProvider = new PhysicalFileProvider(
            Path.Combine(Directory.GetCurrentDirectory(), @"wwwroot\images")),
        RequestPath = new PathString("/MyImages")
    });
}

See MIME content types.

Non-standard content types ¶

The ASP.NET static file middleware understands almost 400 known file content types. If the user requests a file of an unknown file type, the static file middleware returns a HTTP 404 (Not found) response. If directory browsing is enabled, a link to the file will be displayed, but the URI will return an HTTP 404 error.

The following code enables serving unknown types and will render the unknown file as an image.

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    app.UseStaticFiles(new StaticFileOptions
    {
        ServeUnknownFileTypes = true,
        DefaultContentType = "image/png"
    });
}

With the code above, a request for a file with an unknown content type will be returned as an image.

Warning

Enabling ServeUnknownFileTypes is a security risk and using it is discouraged. FileExtensionContentTypeProvider (explained below) provides a safer alternative to serving files with non-standard extensions.

Considerations¶

Warning

UseDirectoryBrowser and UseStaticFiles can leak secrets. We recommend that you not enable directory browsing in production. Be careful about which directories you enable with UseStaticFiles or UseDirectoryBrowser as the entire directory and all sub-directories will be accessible. We recommend keeping public content in its own directory such as <content root>/wwwroot, away from application views, configuration files, etc.

The URLs for content exposed with UseDirectoryBrowser and UseStaticFiles are subject to the case sensitivity and character restrictions of their underlying file system. For example, Windows is case insensitive, but Mac and Linux are not.
ASP.NET Core applications hosted in IIS use the ASP.NET Core Module to forward all requests to the application including requests for static files. The IIS static file handler is not used because it doesn’t get a chance to handle requests before they are handled by the ASP.NET Core Module.
To remove the IIS static file handler (at the server or website level):
- Navigate to the Modules feature
- Select StaticFileModule in the list
- Tap Remove in the Actions sidebar

Warning

If the IIS static file handler is enabled and the ASP.NET Core Module (ANCM) is not correctly configured (for example if web.config was not deployed), static files will be served.

Code files (including c# and Razor) should be placed outside of the app project’s web root (wwwroot by default). This creates a clean separation between your app’s client side content and server side source code, which prevents server side code from being leaked.

Additional Resources ¶

Routing¶

By Ryan Nowak, Steve Smith, and Rick Anderson

Routing is used to map requests to route handlers. Routes are configured when the application starts up, and can extract values from the URL that will be used for request processing. Routing functionality is also responsible for generating links using the defined routes in ASP.NET apps.

This document covers the low level ASP.NET Core routing. For ASP.NET Core MVC routing, see 🔧 Routing to Controller Actions

Sections

Routing basics
Using Routing Middleware
Route Template Reference
Route Constraint Reference
URL Generation Reference

Routing basics ¶

Routing uses routes (implementations of IRouter) to:

map incoming requests to route handlers
generate URLs used in responses

Generally an app has a single collection of routes. The route collection is processed in order. Requests look for a match in the route collection by URL matching. Responses use routing to genenerate URLs.

Routing is connected to the middleware pipeline by the RouterMiddleware class. ASP.NET MVC adds routing to the middleware pipeline as part of its configuration. To learn about using routing as a standalone component, see using-routing-middleware.

URL matching¶

URL matching is the process by which routing dispatches an incoming request to a handler. This process is generally based on data in the URL path, but can be extended to consider any data in the request. The ability to dispatch requests to separate handlers is key to scaling the size and complexity of an application.

Incoming requests enter the RouterMiddleware which calls the RouteAsync method on each route in sequence. The IRouter instance chooses whether to handle the request by setting the RouteContext Handler to a non-null RequestDelegate. If a handler is set a route, it will be invoked to process the request and no further routes will be processed. If all routes are executed, and no handler is found for a request, the middleware calls next and the next middleware in the request pipeline is invoked.

The primary input to RouteAsync is the RouteContext HttpContext associated with the current request. The RouteContext.Handler and RouteContext RouteData are outputs that will be set after a successful match.

A successful match during RouteAsync also will set the properties of the RouteContext.RouteData to appropriate values based on the request processing that was done. The RouteContext.RouteData contains important state information about the result of a route when it successfully matches a request.

RouteData Values is a dictionary of route values produced from the route. These values are usually determined by tokenizing the URL, and can be used to accept user input, or to make further dispatching decisions inside the application.

RouteData DataTokens is a property bag of additional data related to the matched route. DataTokens are provided to support associating state data with each route so the application can make decisions later based on which route matched. These values are developer-defined and do not affect the behavior of routing in any way. Additionally, values stashed in data tokens can be of any type, in contrast to route values which must be easily convertable to and from strings.

RouteData Routers is a list of the routes that took part in successfully matching the request. Routes can be nested inside one another, and the Routers property reflects the path through the logical tree of routes that resulted in a match. Generally the first item in Routers is the route collection, and should be used for URL generation. The last item in Routers is the route that matched.

URL generation¶

URL generation is the process by which routing can create a URL path based on a set of route values. This allows for a logical separation between your handlers and the URLs that access them.

URL generation follows a similar iterative process, but starts with user or framework code calling into the GetVirtualPath method of the route collection. Each route will then have its GetVirtualPath method called in sequence until until a non-null VirtualPathData is returned.

The primary inputs to GetVirtualPath are:

Routes primarily use the route values provided by the Values and AmbientValues to decide where it is possible to generate a URL and what values to include. The AmbientValues are the set of route values that were produced from matching the current request with the routing system. In contrast, Values are the route values that specify how to generate the desired URL for the current operation. The HttpContext is provided in case a route needs to get services or additional data associated with the current context.

Tip

Think of Values as being a set of overrides for the AmbientValues. URL generation tries to reuse route values from the current request to make it easy to generate URLs for links using the same route or route values.

The output of GetVirtualPath is a VirtualPathData. VirtualPathData is a parallel of RouteData; it contains the VirtualPath for the output URL as well as the some additional properties that should be set by the route.

The VirtualPathData VirtualPath property contains the virtual path produced by the route. Depending on your needs you may need to process the path further. For instance, if you want to render the generated URL in HTML you need to prepend the base path of the application.

The VirtualPathData Router is a reference to the route that successfully generated the URL.

The VirtualPathData DataTokens properties is a dictionary of additional data related to the route that generated the URL. This is the parallel of RouteData.DataTokens.

Creating routes¶

Routing provides the Route class as the standard implementation of IRouter. Route uses the route template syntax to define patterns that will match against the URL path when RouteAsync is called. Route will use the same route template to generate a URL when GetVirtualPath is called.

Most applications will create routes by calling MapRoute or one of the similar extension methods defined on IRouteBuilder. All of these methods will create an instance of Route and add it to the route collection.

Note

MapRoute doesn’t take a route handler parameter - it only adds routes that will be handled by the DefaultHandler. Since the default handler is an IRouter, it may decide not to handle the request. For example, ASP.NET MVC is typically configured as a default handler that only handles requests that match an available controller and action. To learn more about routing to MVC, see 🔧 Routing to Controller Actions.

This is an example of a MapRoute call used by a typical ASP.NET MVC route definition:

routes.MapRoute(
    name: "default",
    template: "{controller=Home}/{action=Index}/{id?}");

This template will match a URL path like /Products/Details/17 and extract the route values { controller = Products, action = Details, id = 17 }. The route values are determined by splitting the URL path into segments, and matching each segment with the route parameter name in the route template. Route parameters are named. They are defined by enclosing the parameter name in braces { }.

The template above could also match the URL path / and would produce the values { controller = Home, action = Index }. This happens because the {controller} and {action} route parameters have default values, and the id route parameter is optional. An equals = sign followed by a value after the route parameter name defines a default value for the parameter. A question mark ? after the route parameter name defines the parameter as optional. Route parameters with a default value always produce a route value when the route matches - optional parameters will not produce a route value if there was no corresponding URL path segment.

See route-template-reference for a thorough description of route template features and syntax.

This example includes a route constraint:

routes.MapRoute(
    name: "default",
    template: "{controller=Home}/{action=Index}/{id:int}");

This template will match a URL path like /Products/Details/17, but not /Products/Details/Apples. The route parameter definition {id:int} defines a route constraint for the id route parameter. Route constraints implement IRouteConstraint and inspect route values to verify them. In this example the route value id must be convertable to an integer. See route-constraint-reference for a more detailed explaination of route constraints that are provided by the framework.

Additional overloads of MapRoute accept values for constraints, dataTokens, and defaults. These additional parameters of MapRoute are defined as type object. The typical usage of these parameters is to pass an anonymously typed object, where the property names of the anonymous type match route parameter names.

The following two examples create equivalent routes:

routes.MapRoute(
    name: "default_route",
    template: "{controller}/{action}/{id?}",
    defaults: new { controller = "Home", action = "Index" });

routes.MapRoute(
    name: "default_route",
    template: "{controller=Home}/{action=Index}/{id?}");

Tip

The inline syntax for defining constraints and defaults can be more convenient for simple routes. However, there are features such as data tokens which are not supported by inline syntax.

This example demonstrates a few more features:

routes.MapRoute(
  name: "blog",
  template: "Blog/{*article}",
  defaults: new { controller = "Blog", action = "ReadArticle" });

This template will match a URL path like /Blog/All-About-Routing/Introduction and will extract the values { controller = Blog, action = ReadArticle, article = All-About-Routing/Introduction }. The default route values for controller and action are produced by the route even though there are no corresponding route parameters in the template. Default values can be specified in the route template. The article route parameter is defined as a catch-all by the appearance of an asterix * before the route parameter name. Catch-all route parameters capture the remainder of the URL path, and can also match the empty string.

This example adds route constraints and data tokens:

routes.MapRoute(
    name: "us_english_products",
    template: "en-US/Products/{id}",
    defaults: new { controller = "Products", action = "Details" },
    constraints: new { id = new IntRouteConstraint() },
    dataTokens: new { locale = "en-US" });

This template will match a URL path like /en-US/Products/5 and will extract the values { controller = Products, action = Details, id = 5 } and the data tokens { locale = en-US }.

URL generation¶

The Route class can also perform URL generation by combining a set of route values with its route template. This is logically the reverse process of matching the URL path.

Tip

To better understand URL generation, imagine what URL you want to generate and then think about how a route template would match that URL. What values would be produced? This is the rough equivalent of how URL generation works in the Route class.

This example uses a basic ASP.NET MVC style route:

routes.MapRoute(
    name: "default",
    template: "{controller=Home}/{action=Index}/{id?}");

With the route values { controller = Products, action = List }, this route will generate the URL /Products/List. The route values are substituted for the corresponding route parameters to form the URL path. Since id is an optional route parameter, it’s no problem that it doesn’t have a value.

With the route values { controller = Home, action = Index }, this route will generate the URL /. The route values that were provided match the default values so the segments corresponding to those values can be safely omitted. Note that both URLs generated would round-trip with this route definition and produce the same route values that were used to generate the URL.

Tip

An app using ASP.NET MVC should use UrlHelper to generate URLs instead of calling into routing directly.

For more details about the URL generation process, see url-generation-reference.

Using Routing Middleware ¶

To use routing middleware, add it to the dependencies in project.json:

"Microsoft.AspNetCore.Routing": <current version>

Add routing to the service container in Startup.cs:

public void ConfigureServices(IServiceCollection services)
{
    services.AddRouting();
}

Routes must configured in the Configure method in the Startup class. The sample below uses these APIs:

RouteBuilder
Build
MapGet Matches only HTTP GET requests
UseRouter

public void Configure(IApplicationBuilder app, ILoggerFactory loggerFactory)
{
    var trackPackageRouteHandler = new RouteHandler(context =>
    {
        var routeValues = context.GetRouteData().Values;
        return context.Response.WriteAsync(
            $"Hello! Route values: {string.Join(", ", routeValues)}");
    });

    var routeBuilder = new RouteBuilder(app, trackPackageRouteHandler);

    routeBuilder.MapRoute(
        "Track Package Route",
        "package/{operation:regex(track|create|detonate)}/{id:int}");

    routeBuilder.MapGet("hello/{name}", context =>
    {
        var name = context.GetRouteValue("name");
        // This is the route handler when HTTP GET "hello/<anything>"  matches
        // To match HTTP GET "hello/<anything>/<anything>, 
        // use routeBuilder.MapGet("hello/{*name}"
        return context.Response.WriteAsync($"Hi, {name}!");
    });            

    var routes = routeBuilder.Build();
    app.UseRouter(routes);

The table below shows the resposes with the given URIs.

URI	Response
/package/create/3	Hello! Route values: [operation, create], [id, 3]
/package/track/-3	Hello! Route values: [operation, track], [id, -3]
/package/track/-3/	Hello! Route values: [operation, track], [id, -3]
/package/track/	<Fall through, no match>
GET /hello/Joe	Hi, Joe!
POST /hello/Joe	<Fall through, matches HTTP GET only>
GET /hello/Joe/Smith	<Fall through, no match>

If you are configuring a single route, call app.UseRouter passing in an IRouter instance. You won’t need to call RouteBuilder.

The framework provides a set of extension methods for creating routes such as:

Some of these methods such as MapGet require a RequestDelegate to be provided. The RequestDelegate will be used as the route handler when the route matches. Other methods in this family allow configuring a middleware pipeline which will be used as the route handler. If the Map method doesn’t accept a handler, such as MapRoute, then it will use the DefaultHandler.

The Map[Verb] methods use constraints to limit the route to the HTTP Verb in the method name. For example, see MapGet and MapVerb.

Route Template Reference ¶

Tokens within curly braces ({ }) define route parameters which will be bound if the route is matched. You can define more than one route parameter in a route segment, but they must be separated by a literal value. For example {controller=Home}{action=Index} would not be a valid route, since there is no literal value between {controller} and {action}. These route parameters must have a name, and may have additional attributes specified.

Literal text other than route parameters (for example, {id}) and the path separator / must match the text in the URL. Text matching is case-insensitive and based on the decoded representation of the URLs path. To match the literal route parameter delimiter { or }, escape it by repeating the character ({{ or }}).

URL patterns that attempt to capture a filename with an optional file extension have additional considerations. For example, using the template files/{filename}.{ext?} - When both filename and ext exist, both values will be populated. If only filename exists in the URL, the route matches because the trailing period . is optional. The following URLs would match this route:

/files/myFile.txt
/files/myFile.
/files/myFile

You can use the * character as a prefix to a route parameter to bind to the rest of the URI - this is called a catch-all parameter. For example, blog/{*slug} would match any URI that started with /blog and had any value following it (which would be assigned to the slug route value). Catch-all parameters can also match the empty string.

Route parameters may have default values, designated by specifying the default after the parameter name, separated by an =. For example, {controller=Home} would define Home as the default value for controller. The default value is used if no value is present in the URL for the parameter. In addition to default values, route parameters may be optional (specified by appending a ? to the end of the parameter name, as in id?). The difference between optional and “has default” is that a route parameter with a default value always produces a value; an optional parameter has a vaule only when one is provided.

Route parameters may also have constraints, which must match the route value bound from the URL. Adding a colon : and constraint name after the route parameter name specifies an inline constraint on a route parameter. If the constraint requires arguments those are provided enclosed in parentheses ( ) after the constraint name. Multiple inline constraints can be specified by appending another colon : and constraint name. The constraint name is passed to the IInlineConstraintResolver service to create an instance of IRouteConstraint to use in URL processing. For example, the route template blog/{article:minlength(10)} specifies the minlength constraint with the argument 10. For more description route constraints, and a listing of the constraints provided by the framework, see route-constraint-reference.

The following table demonstrates some route templates and their behavior.

Route Template	Example Matching URL	Notes
hello	/hello	Only matches the single path ‘/hello’
hello	/hello	Only matches the single path ‘/hello’
{Page=Home}	/	Matches and sets `Page` to `Home`
{Page=Home}	/Contact	Matches and sets `Page` to `Contact`
{controller}/{action}/{id?}	/Products/List	Maps to `Products` controller and `List` action
{controller}/{action}/{id?}	/Products/Details/123	Maps to `Products` controller and `Details` action. `id` set to 123
{controller=Home}/ {action=Index}/{id?}	/	Maps to `Home` controller and `Index` method; `id` is ignored.

Using a template is generally the simplest approach to routing. Constraints and defaults can also be specified outside the route template.

Tip

Enable Logging to see how the built in routing implementations, such as Route, match requests.

Route Constraint Reference ¶

Route constraints execute when a Route has matched the syntax of the incoming URL and tokenized the URL path into route values. Route constraints generally inspect the route value associated via the route template and make a simple yes/no decision about whether or not the value is acceptable. Some route constraints use data outside the route value to consider whether the request can be routed. For example, the HttpMethodRouteConstraint can accept or reject a request based on its HTTP verb.

Warning

Avoid using constraints for input validation, because doing so means that invalid input will result in a 404 (Not Found) instead of a 400 with an appropriate error message. Route constraints should be used to disambiguate between similar routes, not to validate the inputs for a particular route.

The following table demonstrates some route constraints and their expected behavior.

Inline Route Constraints¶
constraint	Example	Example Match	Notes
`int`	{id:int}	123	Matches any integer
`bool`	{active:bool}	true	Matches `true` or `false`
`datetime`	{dob:datetime}	2016-01-01	Matches a valid `DateTime` value (in the invariant culture - see options)
`decimal`	{price:decimal}	49.99	Matches a valid `decimal` value
`double`	{weight:double}	4.234	Matches a valid `double` value
`float`	{weight:float}	3.14	Matches a valid `float` value
`guid`	{id:guid}	7342570B-<snip>	Matches a valid `Guid` value
`long`	{ticks:long}	123456789	Matches a valid `long` value
`minlength(value)`	{username:minlength(5)}	steve	String must be at least 5 characters long.
`maxlength(value)`	{filename:maxlength(8)}	somefile	String must be no more than 8 characters long.
`length(min,max)`	{filename:length(4,16)}	Somefile.txt	String must be at least 8 and no more than 16 characters long.
`min(value)`	{age:min(18)}	19	Value must be at least 18.
`max(value)`	{age:max(120)}	91	Value must be no more than 120.
`range(min,max)`	{age:range(18,120)}	91	Value must be at least 18 but no more than 120.
`alpha`	{name:alpha}	Steve	String must consist of alphabetical characters.
`regex(expression)`	{ssn:regex(d{3}-d{2}-d{4})}	123-45-6789	String must match the provided regular expression.
`required`	{name:required}	Steve	Used to enforce that a non-parameter value is present during during URL generation.

Warning

Route constraints that verify the URL can be converted to a CLR type (such as int or DateTime) always use the invariant culture - they assume the URL is non-localizable. The framework-provided route constraints do not modify the values stored in route values. All route values parsed from the URL will be stored as strings. For example, the Float route constraint will attempt to convert the route value to a float, but the converted value is used only to verify it can be converted to a float.

Tip

To constrain a parameter to a known set of possible values, you can use a regular expression ( for example {action:regex(list|get|create)}. This would only match the action route value to list, get, or create. If passed into the constraints dictionary, the string “list|get|create” would be equivalent. Constraints that are passed in the constraints dictionary (not inline within a template) that don’t match one of the known constraints are also treated as regular expressions.

URL Generation Reference ¶

The example below shows how to generate a link to a route given a dictionary of route values and a RouteCollection.

app.Run(async (context) =>
{
    var dictionary = new RouteValueDictionary
    {
        { "operation", "create" },
        { "id", 123}
    };

    var vpc = new VirtualPathContext(context, null, dictionary, "Track Package Route");
    var path = routes.GetVirtualPath(vpc).VirtualPath;

    context.Response.ContentType = "text/html";
    await context.Response.WriteAsync("Menu<hr/>");
    await context.Response.WriteAsync($"<a href='{path}'>Create Package 123</a><br/>");
});

The VirtualPath generated at the end of the sample above is /package/create/123.

The second parameter to the VirtualPathContext constructor is a collection of ambient values. Ambient values provide convenience by limiting the number of values a developer must specify within a certain request context. The current route values of the current request are considered ambient values for link generation. For example, in an ASP.NET MVC app if you are in the About action of the HomeController, you don’t need to specify the controller route value to link to the Index action (the ambient value of Home will be used).

Ambient values that don’t match a parameter are ignored, and ambient values are also ignored when an explicitly-provided value overrides it, going from left to right in the URL.

Values that are explicitly provided but which don’t match anything are added to the query string. The following table shows the result when using the route template {controller}/{action}/{id?}.

Generating links with `{controller}/{action}/{id?}` template¶
Ambient Values	Explicit Values	Result
controller=”Home”	action=”About”	`/Home/About`
controller=”Home”	controller=”Order”,action=”About”	`/Order/About`
controller=”Home”,color=”Red”	action=”About”	`/Home/About`
controller=”Home”	action=”About”,color=”Red”	`/Home/About?color=Red`

If a route has a default value that doesn’t correspond to a parameter and that value is explicitly provided, it must match the default value. For example:

routes.MapRoute("blog_route", "blog/{*slug}",
  defaults: new { controller = "Blog", action = "ReadPost" });

Link generation would only generate a link for this route when the matching values for controller and action are provided.

Error Handling¶

When errors occur in your ASP.NET app, you can handle them in a variety of ways, as described in this article.

Sections

Configuring an Exception Handling Page
Using the Developer Exception Page
Configuring Status Code Pages
Limitations of Exception Handling During Client-Server Interaction
Server Exception Handling
Startup Exception Handling
ASP.NET MVC Error Handling

Configuring an Exception Handling Page ¶

You configure the pipeline for each request in the Startup class’s Configure() method (learn more about Application Startup). You can add a simple exception page, meant only for use during development, very easily. All that’s required is to add a dependency on Microsoft.AspNetCore.Diagnostics to the project and then add one line to Configure() in Startup.cs:

public void Configure(IApplicationBuilder app, 
    IHostingEnvironment env)
{
    app.UseIISPlatformHandler();

    if (env.IsDevelopment())
    {
        app.UseDeveloperExceptionPage();
    }

The above code includes a check to ensure the environment is development before adding the call to UseDeveloperExceptionPage. This is a good practice, since you typically do not want to share detailed exception information about your application publicly while it is in production. Learn more about configuring environments.

The sample application includes a simple mechanism for creating an exception:

public static void HomePage(IApplicationBuilder app)
{
    app.Run(async (context) =>
    {
        if (context.Request.Query.ContainsKey("throw"))
        {
            throw new Exception("Exception triggered!");
        }
        var builder = new StringBuilder();
        builder.AppendLine("<html><body>Hello World!");
        builder.AppendLine("<ul>");
        builder.AppendLine("<li><a href=\"/?throw=true\">Throw Exception</a></li>");
        builder.AppendLine("<li><a href=\"/missingpage\">Missing Page</a></li>");
        builder.AppendLine("</ul>");
        builder.AppendLine("</body></html>");

        context.Response.ContentType = "text/html";
        await context.Response.WriteAsync(builder.ToString());
    });
}

If a request includes a non-empty querystring parameter for the variable throw (e.g. a path of /?throw=true), an exception will be thrown. If the environment is set to Development, the developer exception page is displayed:

When not in development, it’s a good idea to configure an exception handler path using the UseExceptionHandler middleware:

app.UseExceptionHandler("/Error");

Using the Developer Exception Page ¶

The developer exception page displays useful diagnostics information when an unhandled exception occurs within the web processing pipeline. The page includes several tabs with information about the exception that was triggered and the request that was made. The first tab includes a stack trace:

The next tab shows the query string parameters, if any:

_images/developer-exception-page-query.png

In this case, you can see the value of the throw parameter that was passed to this request. This request didn’t have any cookies, but if it did, they would appear on the Cookies tab. You can see the headers that were passed in the last tab:

_images/developer-exception-page-headers.png

Configuring Status Code Pages ¶

By default, your app will not provide a rich status code page for HTTP status codes such as 500 (Internal Server Error) or 404 (Not Found). You can configure the StatusCodePagesMiddleware adding this line to the Configure method:

app.UseStatusCodePages();

By default, this middleware adds very simple, text-only handlers for common status codes. For example, the following is the result of a 404 Not Found status code:

The middleware supports several different extension methods. You can pass it a custom lamdba expression:

app.UseStatusCodePages(context =>
  context.HttpContext.Response.SendAsync("Handler, status code: " +
  context.HttpContext.Response.StatusCode, "text/plain"));

Alternately, you can simply pass it a content type and a format string:

app.UseStatusCodePages("text/plain", "Response, status code: {0}");

The middleware can handle redirects (with either relative or absolute URL paths), passing the status code as part of the URL:

app.UseStatusCodePagesWithRedirects("~/errors/{0}");

In the above case, the client browser will see a 302 / Found status and will redirect to the URL provided.

Alternately, the middleware can re-execute the request from a new path format string:

app.UseStatusCodePagesWithReExecute("/errors/{0}");

The UseStatusCodePagesWithReExecute method will still return the original status code to the browser, but will also execute the handler given at the path specified.

If you need to disable status code pages for certain requests, you can do so using the following code:

var statusCodePagesFeature = context.Features.Get<IStatusCodePagesFeature>();
if (statusCodePagesFeature != null)
{
  statusCodePagesFeature.Enabled = false;
}

Limitations of Exception Handling During Client-Server Interaction ¶

Web apps have certain limitations to their exception handling capabilities because of the nature of disconnected HTTP requests and responses. Keep these in mind as you design your app’s exception handling behavior.

Once the headers for a response have been sent, you cannot change the response’s status code, nor can any exception pages or handlers run. The response must be completed or the connection aborted.
If the client disconnects mid-response, you cannot send them the rest of the content of that response.
There is always the possibility of an exception occuring one layer below your exception handling layer.
Don’t forget, exception handling pages can have exceptions, too. It’s often a good idea for production error pages to consist of purely static content.

Following the above recommendations will help ensure your app remains responsive and is able to gracefully handle exceptions that may occur.

Server Exception Handling ¶

In addition to the exception handling logic in your app, the server hosting your app will perform some exception handling. If the server catches an exception before the headers have been sent it will send a 500 Internal Server Error response with no body. If it catches an exception after the headers have been sent it must close the connection. Requests that are not handled by your app will be handled by the server, and any exception that occurs will be handled by the server’s exception handling. Any custom error pages or exception handling middleware or filters you have configured for your app will not affect this behavior.

Startup Exception Handling ¶

One of the trickiest places to handle exceptions in your app is during its startup. Only the hosting layer can handle exceptions that take place during app startup. Exceptions that occur in your app’s startup can also impact server behavior. For example, to enable SSL in Kestrel, one must configure the server with KestrelServerOptions.UseHttps(). If an exception happens before this line in Startup, then by default hosting will catch the exception, start the server, and display an error page on the non-SSL port. If an exception happens after that line executes, then the error page will be served over HTTPS instead.

ASP.NET MVC Error Handling ¶

MVC apps have some additional options when it comes to handling errors, such as configuring exception filters and performing model validation.

Exception Filters¶

Exception filters can be configured globally or on a per-controller or per-action basis in an MVC app. These filters handle any unhandled exception that occurs during the execution of a controller action or another filter, and are not called otherwise. Exception filters are detailed in filters.

Tip

Exception filters are good for trapping exceptions that occur within MVC actions, but they’re not as flexible as error handling middleware. Prefer middleware for the general case, and use filters only where you need to do error handling differently based on which MVC action was chosen.

Handling Model State Errors¶

Model validation occurs prior to each controller action being invoked, and it is the action method’s responsibility to inspect ModelState.IsValid and react appropriately. In many cases, the appropriate reaction is to return some kind of error response, ideally detailing the reason why model validation failed.

Some apps will choose to follow a standard convention for dealing with model validation errors, in which case a filter may be an appropriate place to implement such a policy. You should test how your actions behave with valid and invalid model states (learn more about testing controller logic).

Globalization and localization¶

Rick Anderson, Damien Bowden, Bart Calixto, Nadeem Afana

Creating a multilingual website with ASP.NET Core will allow your site to reach a wider audience. ASP.NET Core provides services and middleware for localizing into different languages and cultures.

Internationalization involves Globalization and Localization. Globalization is the process of designing apps that support different cultures. Globalization adds support for input, display, and output of a defined set of language scripts that relate to specific geographic areas.

Localization is the process of adapting a globalized app, which you have already processed for localizability, to a particular culture/locale. For more information see Globalization and localization terms near the end of this document.

App localization involves the following:

Make the app’s content localizable
Provide localized resources for the languages and cultures you support
Implement a strategy to select the language/culture for each request

Sections:

Make the app’s content localizable
View localization
DataAnnotations localization
Provide localized resources for the languages and cultures you support
Working with resource files
Implement a strategy to select the language/culture for each request
Resource file naming
Globalization and localization terms
Additional Resources

Make the app’s content localizable ¶

Introduced in ASP.NET Core, IStringLocalizer and IStringLocalizer<T> were architected to improve productivity when developing localized apps. IStringLocalizer uses the ResourceManager and ResourceReader to provide culture-specific resources at run time. The simple interface has an indexer and an IEnumerable for returning localized strings. IStringLocalizer doesn’t require you to store the default language strings in a resource file. You can develop an app targeted for localization and not need to create resource files early in development. The code below shows how to wrap the string “About Title” for localization.

using Microsoft.AspNet.Mvc;
using Microsoft.Extensions.Localization;

namespace Localization.StarterWeb.Controllers
{
    [Route("api/[controller]")]
    public class AboutController : Controller
    {
        private readonly IStringLocalizer<AboutController> _localizer;

        public AboutController(IStringLocalizer<AboutController> localizer)
        {
            _localizer = localizer;
        }

        [HttpGet]
        public string Get()
        {
            return _localizer["About Title"];
        }
    }
}

In the code above, the IStringLocalizer<T> implementation comes from Dependency Injection. If the localized value of “About Title” is not found, then the indexer key is returned, that is, the string “About Title”. You can leave the default language literal strings in the app and wrap them in the localizer, so that you can focus on developing the app. You develop your app with your default language and prepare it for the localization step without first creating a default resource file. Alternatively, you can use the traditional approach and provide a key to retrieve the default language string. For many developers the new workflow of not having a default language .resx file and simply wrapping the string literals can reduce the overhead of localizing an app. Other developers will prefer the traditional work flow as it can make it easier to work with longer string literals and make it easier to update localized strings.

Use the IHtmlLocalizer<T> implementation for resources that contain HTML. IHtmlLocalizer HTML encodes arguments that are formatted in the resource string, but not the resource string. In the sample highlighted below, only the value of name parameter is HTML encoded.

using System;
using Microsoft.AspNet.Http;
using Microsoft.AspNet.Localization;
using Microsoft.AspNet.Mvc;
using Microsoft.AspNet.Mvc.Localization;

namespace Localization.StarterWeb.Controllers
{
    public class BookController : Controller
    {
        private readonly IHtmlLocalizer<BookController> _localizer;

        public BookController(IHtmlLocalizer<BookController> localizer)
        {
            _localizer = localizer;
        }

        public IActionResult Hello(string name)
        {
            ViewData["Message"] = _localizer["<b>Hello</b><i> {0}</i>", name];

            return View();
        }

Note:	You generally want to only localize text and not HTML.

At the lowest level, you can get IStringLocalizerFactory out of Dependency Injection:

 public class TestController : Controller
 {
     private readonly IStringLocalizer _localizer;
     private readonly IStringLocalizer _localizer2;

     public TestController(IStringLocalizerFactory factory)
     {
         _localizer = factory.Create(typeof(SharedResource));
         _localizer2 = factory.Create("SharedResource", location: null);
     }       

     public IActionResult About()
     {
         ViewData["Message"] = _localizer["Your application description page."] 
             + " loc 2: " + _localizer2["Your application description page."];

         return View();
     }        

The code above demonstrates each of the two factory create methods.

You can partition your localized strings by controller, area, or have just one container. In the sample app, a dummy class named SharedResource is used for shared resources.

// Dummy class to group shared resources

namespace Localization.StarterWeb
{
    public class SharedResource
    {
    }
}

Some developers use the Startup class to contain global or shared strings. In the sample below, the InfoController and the SharedResource localizers are used:

 public class InfoController : Controller
 {
     private readonly IStringLocalizer<InfoController> _localizer;
     private readonly IStringLocalizer<SharedResource> _sharedLocalizer;

     public InfoController(IStringLocalizer<InfoController> localizer,
                    IStringLocalizer<SharedResource> sharedLocalizer)
     {
         _localizer = localizer;
         _sharedLocalizer = sharedLocalizer;
     }

     public string TestLoc()
     {
         string msg = "Shared resx: " + _sharedLocalizer["Hello!"] +
                      " Info resx " + _localizer["Hello!"];
         return msg;
     }

View localization ¶

The IViewLocalizer service provides localized strings for a view. The ViewLocalizer class implements this interface and finds the resource location from the view file path. The following code shows how to use the default implementation of IViewLocalizer:

@using Microsoft.AspNet.Mvc.Localization

@inject IViewLocalizer Localizer

@{
    ViewData["Title"] = Localizer["About"];
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<p>@Localizer["Use this area to provide additional information."]</p>

The default implementation of IViewLocalizer finds the resource file based on the view’s file name. There is no option to use a global shared resource file. ViewLocalizer implements the localizer using IHtmlLocalizer, so Razor doesn’t HTML encode the localized string. You can parameterize resource strings and IViewLocalizer will HTML encode the parameters, but not the resource string. Consider the following Razor markup:

@Localizer["<i>Hello</i> <b>{0}!</b>", UserManager.GetUserName(User)]

A French resource file could contain the following:

Key	Value
<i>Hello</i> <b>{0}!</b>	<i>Bonjour</i> <b>{0}!</b>

The rendered view would contain the HTML markup from the resource file.

Note:	You generally want to only localize text and not HTML.

To use a shared resource file in a view, inject IHtmlLocalizer<T>:

@using Microsoft.AspNet.Mvc.Localization
@using Localization.StarterWeb.Services

@inject IViewLocalizer Localizer
@inject IHtmlLocalizer<SharedResource> SharedLocalizer

@{
    ViewData["Title"] = Localizer["About"];
}
<h2>@ViewData["Title"].</h2>

<h1>@SharedLocalizer["Hello!"]</h1>

DataAnnotations localization ¶

DataAnnotations error messages are localized with IStringLocalizer<T>. Using the option ResourcesPath = "Resources", the error messages in RegisterViewModel can be stored in either of the following paths:

Resources/ViewModels.Account.RegisterViewModel.fr.resx
Resources/ViewModels/Account/RegisterViewModel.fr.resx

 public class RegisterViewModel
 {
     [Required(ErrorMessage = "The Email field is required.")]
     [EmailAddress(ErrorMessage = "The Email field is not a valid e-mail address.")]
     [Display(Name = "Email")]
     public string Email { get; set; }

     [Required(ErrorMessage = "The Password field is required.")]
     [StringLength(8, ErrorMessage = "The {0} must be at least {2} characters long.", MinimumLength = 6)]
     [DataType(DataType.Password)]
     [Display(Name = "Password")]
     public string Password { get; set; }

     [DataType(DataType.Password)]
     [Display(Name = "Confirm password")]
     [Compare("Password", ErrorMessage = "The password and confirmation password do not match.")]
     public string ConfirmPassword { get; set; }
 }

The runtime doesn’t look up localized strings for non-validation attributes. In the code above, “Email” (from [Display(Name = "Email")]) will not be localized.

Provide localized resources for the languages and cultures you support ¶

SupportedCultures and SupportedUICultures¶

ASP.NET Core allows you to specify two culture values, SupportedCultures and SupportedUICultures. The CultureInfo object for SupportedCultures determines the results of culture-dependent functions, such as date, time, number, and currency formatting. SupportedCultures also determines the sorting order of text, casing conventions, and string comparisons. See CultureInfo.CurrentCulture for more info on how the server gets the Culture. The SupportedUICultures determines which translates strings (from .resx files) are looked up by the ResourceManager. The ResourceManager simply looks up culture-specific strings that is determined by CurrentUICulture. Every thread in .NET has CurrentCulture and CurrentUICulture objects. ASP.NET Core inspects these values when rendering culture-dependent functions. For example, if the current thread’s culture is set to “en-US” (English, United States), DateTime.Now.ToLongDateString() displays “Thursday, February 18, 2016”, but if CurrentCulture is set to “es-ES” (Spanish, Spain) the output will be “jueves, 18 de febrero de 2016”.

Working with resource files ¶

A resource file is a useful mechanism for separating localizable strings from code. Translated strings for the non-default language are isolated .resx resource files. For example, you might want to create Spanish resource file named Welcome.es.resx containing translated strings. “es” is the language code for Spanish. To create this resource file in Visual Studio:

In Solution Explorer, right click on the folder which will contain the resource file > Add > New Item.

In the Search installed templates box, enter “resource” and name the file.

Enter the key value (native string) in the Name column and the translated string in the Value column.

Visual Studio shows the Welcome.es.resx file.

Generating resource files with Visual Studio¶

If you create a resource file in Visual Studio without a culture in the file name (for example, Welcome.resx), Visual Studio will create a C# class with a property for each string. That’s usually not what you want with ASP.NET Core; you typically don’t have a default .resx resource file (A .resx file without the culture name). We suggest you create the .resx file with a culture name (for example Welcome.fr.resx). When you create a .resx file with a culture name, Visual Studio will not generate the class file. We anticipate that many developers will not create a default language resource file.

Adding Other Cultures¶

Each language and culture combination (other than the default language) requires a unique resource file. You can create resource files for different cultures and locales by creating new resource files in which the ISO language codes are part of the file name (for example, en-us, fr-ca, and en-gb). These ISO codes are placed between the file name and the .resx file name extension, as in Welcome.es-MX.resx (Spanish/Mexico). To specify a culturally neutral language, you would eliminate the country code, such as Welcome.fr.resx for the French language.

Implement a strategy to select the language/culture for each request ¶

Configuring localization¶

Localization is configured in the ConfigureServices method:

  public void ConfigureServices(IServiceCollection services)
  {

      services.AddLocalization(options => options.ResourcesPath = "Resources");

      services.AddMvc()
        .AddViewLocalization(LanguageViewLocationExpanderFormat.Suffix)
        .AddDataAnnotationsLocalization();

AddLocalization Adds the localization services to the services container. The code above also sets the resources path to “Resources”.
AddViewLocalization Adds support for localized view files. In this sample view localization is based on the view file suffix. For example “fr” in the Index.fr.cshtml file.
AddDataAnnotationsLocalization Adds support for localized DataAnnotations validation messages through IStringLocalizer abstractions.

Localization middleware¶

The current culture on a request is set in the localization Middleware. The localization middleware is enabled in the Configure method of Startup.cs file.

  public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)

      var supportedCultures = new[]
      {
          new CultureInfo("en-US"),
          new CultureInfo("en-AU"),
          new CultureInfo("en-GB"),
          new CultureInfo("en"),
          new CultureInfo("es-ES"),
          new CultureInfo("es-MX"),
          new CultureInfo("es"),
          new CultureInfo("fr-FR"),
          new CultureInfo("fr"),
      };

      app.UseRequestLocalization(new RequestLocalizationOptions
      {
          DefaultRequestCulture = new RequestCulture("en-US"),
          // Formatting numbers, dates, etc.
          SupportedCultures = supportedCultures,
          // UI strings that we have localized.
          SupportedUICultures = supportedCultures
      });

      // Remaining code omitted for brevity.

UseRequestLocalization initializes a RequestLocalizationOptions object. On every request the list of RequestCultureProvider in the RequestLocalizationOptions is enumerated and the first provider that can successfully determine the request culture is used. The default providers come from the RequestLocalizationOptions class:

The default list goes from most specific to least specific. Later in the article we’ll see how you can change the order and even add a custom culture provider. If none of the providers can determine the request culture, the DefaultRequestCulture is used.

QueryStringRequestCultureProvider¶

Some apps will use a query string to set the culture and UI culture. For apps that use the cookie or Accept-Language header approach, adding a query string to the URL is useful for debugging and testing code. By default, the QueryStringRequestCultureProvider is registered as the first localization provider in the RequestCultureProvider list. You pass the query string parameters culture and ui-culture. The following example sets the specific culture (language and region) to Spanish/Mexico:

http://localhost:5000/?culture=es-MX&ui-culture=es-MX

If you only pass in one of the two (culture or ui-culture), the query string provider will set both values using the one you passed in. For example, setting just the culture will set both the Culture and the UICulture:

http://localhost:5000/?culture=es-MX

CookieRequestCultureProvider¶

Production apps will often provide a mechanism to set the culture with the ASP.NET Core culture cookie. Use the MakeCookieValue method to create a cookie.

The CookieRequestCultureProvider DefaultCookieName returns the default cookie name used to track the user’s preferred culture information. The default cookie name is ”.AspNetCore.Culture”.

The cookie format is c=%LANGCODE%|uic=%LANGCODE%, where c is Culture and uic is UICulture, for example:

c=’en-UK’|uic=’en-US’

If you only specify one of culture info and UI culture, the specified culture will be used for both culture info and UI culture.

The Accept-Language HTTP header¶

The Accept-Language header is settable in most browsers and was originally intended to specify the user’s language. This setting indicates what the browser has been set to send or has inherited from the underlying operating system. The Accept-Language HTTP header from a browser request is not an infallible way to detect the user’s preferred language (see Setting language preferences in a browser). A production app should include a way for a user to customize their choice of culture.

Setting the Accept-Language HTTP header in IE¶

From the gear icon, tap Internet Options.
Tap Languages.

Tap Set Language Preferences.
Tap Add a language.
Add the language.
Tap the language, then tap Move Up.

Using a custom provider¶

Suppose you want to let your customers store their language and culture in your databases. You could write a provider to look up these values for the user. The following code shows how to add a custom provider:

services.Configure<RequestLocalizationOptions>(options =>
{
    var supportedCultures = new[]
    {
        new CultureInfo("en-US"),
        new CultureInfo("fr")
    };

    options.DefaultRequestCulture = new RequestCulture(culture: "en-US", uiCulture: "en-US");
    options.SupportedCultures = supportedCultures;
    options.SupportedUICultures = supportedCultures;

    options.RequestCultureProviders.Insert(0, new CustomRequestCultureProvider(async context =>
    {
       // My custom request culture logic
      return new ProviderCultureResult("en");
    }));
});

Use RequestLocalizationOptions to add or remove localization providers.

Resource file naming ¶

Resources are named for the type of their class minus the default namespace (which is also the name of the assembly). For example, a French resource in the LocalizationWebsite.Web project for the class LocalizationWebsite.Web.Startup would be named Startup.fr.resx. The class LocalizationWebsite.Web.Controllers.HomeController would be Controllers.HomeController.fr.resx. If for some reason your targeted class is not in the base namespace you will need the full type name. For example, in the sample project a type ExtraNamespace.Tools would be ExtraNamespace.Tools.fr.resx.

In the sample project, the ConfigureServices method sets the ResourcesPath to “Resources”, so the project relative path for the home controller’s French resource file is Resources/Controllers.HomeController.fr.resx. Alternatively, you can use folders to organize resource files. For the home controller, the path would be Resources/Controllers/HomeController.fr.resx. If you don’t use the ResourcesPath option, the .resx file would go in the project base directory. The resource file for HomeController would be named Controllers.HomeController.fr.resx. The choice of using the dot or path naming convention depends on how you want to organize your resource files.

Resource name	Dot or path naming
Resources/Controllers.HomeController.fr.resx	Dot
Resources/Controllers/HomeController.fr.resx	Path

Resource files using @inject IViewLocalizer in Razor views follow a similar pattern. The resource file for a view can be named using either dot naming or path naming. Razor view resource files mimic the path of their associated view file. Assuming we set the ResourcesPath to “Resources”, the French resource file associated with the Views/Book/About.cshtml view could be either of the following:

Resources/Views/Home/About.fr.resx
Resources/Views.Home.About.fr.resx

If you don’t use the ResourcesPath option, the .resx file for a view would be located in the same folder as the view.

If you remove the ”.fr” culture designator AND you have the culture set to French (via cookie or other mechanism), the default resource file is read and strings are localized. The Resource manager designates a default or fallback resource, when nothing meets your requested culture you’re served the *.resx file without a culture designator. If you want to just return the key when missing a resource for the requested culture you must not have a default resource file.

Setting the culture programmatically¶

This sample Localization.StarterWeb project on GitHub contains UI to set the Culture. The Views/Shared/_SelectLanguagePartial.cshtml file allows you to select the culture from the list of supported cultures:

@using Microsoft.AspNet.Builder
@using Microsoft.AspNet.Http.Features
@using Microsoft.AspNet.Localization
@using Microsoft.AspNet.Mvc.Localization
@using Microsoft.Extensions.Options

@inject IViewLocalizer Localizer
@inject IOptions<RequestLocalizationOptions> LocOptions

@{
    var requestCulture = Context.Features.Get<IRequestCultureFeature>();
    var cultureItems = LocOptions.Value.SupportedUICultures
        .Select(c => new SelectListItem { Value = c.Name, Text = c.DisplayName })
        .ToList();
}

<div title="@Localizer["Request culture provider:"] @requestCulture?.Provider?.GetType().Name">
    <form id="selectLanguage" asp-controller="Home" 
          asp-action="SetLanguage" asp-route-returnUrl="@Context.Request.Path" 
          method="post" class="form-horizontal" role="form">
        @Localizer["Language:"] <select name="culture"
          asp-for="@requestCulture.RequestCulture.UICulture.Name" asp-items="cultureItems">
        </select>
    </form>
</div>

The Views/Shared/_SelectLanguagePartial.cshtml file is added to the footer section of the layout file so it will be available to all views:

  <div class="container body-content">
      @RenderBody()
      <hr />
      <footer>
          <div class="row">
              <div class="col-md-6">
                  <p>&copy; 2015 - Localization.StarterWeb</p>
              </div>
              <div class="col-md-6 text-right">
                  @await Html.PartialAsync("_SelectLanguagePartial")
              </div>
          </div>
      </footer>
  </div>

The SetLanguage method sets the culture cookie.

  [HttpPost]
  public IActionResult SetLanguage(string culture, string returnUrl)
  {
      Response.Cookies.Append(
          CookieRequestCultureProvider.DefaultCookieName,
          CookieRequestCultureProvider.MakeCookieValue(new RequestCulture(culture)),
          new CookieOptions { Expires = DateTimeOffset.UtcNow.AddYears(1) }
      );

      return LocalRedirect(returnUrl);
  }

You can’t simply plug in the _SelectLanguagePartial.cshtml to sample code for this project. The Localization.StarterWeb project on GitHub has code to flow the RequestLocalizationOptions to a Razor partial through the Dependency Injection container.

Globalization and localization terms ¶

The process of localizing your app also requires a basic understanding of relevant character sets commonly used in modern software development and an understanding of the issues associated with them. Although all computers store text as numbers (codes), different systems store the same text using different numbers. The localization process refers to translating the app user interface (UI) for a specific culture/locale.

Localizability is an intermediate process for verifying that a globalized app is ready for localization.

The RFC 4646 format for the culture name is “<languagecode2>-<country/regioncode2>”, where <languagecode2> is the language code and <country/regioncode2> is the subculture code. For example, es-CL for Spanish (Chile), en-US for English (United States), and en-AU for English (Australia). RFC 4646 is a combination of an ISO 639 two-letter lowercase culture code associated with a language and an ISO 3166 two-letter uppercase subculture code associated with a country or region. See Language Culture Name.

Internationalization is often abbreviated to “I18N”. The abbreviation takes the first and last letters and the number of letters between them, so 18 stands for the number of letters between the first “I” and the last “N”. The same applies to Globalization (G11N), and Localization (L10N).

Terms:

Globalization (G11N): The process of making an app support different languages and regions.
Localization (L10N): The process of customizing an app for a given language and region.
Internationalization (I18N): Describes both globalization and localization.
Culture: It is a language and, optionally, a region.
Neutral culture: A culture that has a specified language, but not a region. (for example “en”, “es”)
Specific culture: A culture that has a specified language and region. (for example “en-US”, “en-GB”, “es-CL”)
Locale: A locale is the same as a culture.

Additional Resources ¶

Localization.StarterWeb project used in the article.
Resource Files in Visual Studio
Resources in .resx Files

Configuration¶

Steve Smith, Daniel Roth

ASP.NET Core supports a variety of different configuration options. Application configuration data can come from files using built-in support for JSON, XML, and INI formats, as well as from environment variables, command line arguments or an in-memory collection. You can also write your own custom configuration provider.

Sections:

Getting and setting configuration settings
Using the built-in sources
Using Options and configuration objects
Writing custom providers
Summary

Getting and setting configuration settings ¶

ASP.NET Core’s configuration system has been re-architected from previous versions of ASP.NET, which relied on System.Configuration and XML configuration files like web.config. The new configuration model provides streamlined access to key/value based settings that can be retrieved from a variety of sources. Applications and frameworks can then access configured settings in a strongly typed fashion using the new Options pattern.

To work with settings in your ASP.NET application, it is recommended that you only instantiate a Configuration in your application’s Startup class. Then, use the Options pattern to access individual settings.

At its simplest, Configuration is just a collection of sources, which provide the ability to read and write name/value pairs. If a name/value pair is written to Configuration, it is not persisted. This means that the written value will be lost when the sources are read again.

You must configure at least one source in order for Configuration to function correctly. The following sample shows how to test working with Configuration as a key/value store:

var builder = new ConfigurationBuilder();
builder.AddInMemoryCollection();
var config = builder.Build();
config["somekey"] = "somevalue";

// do some other work

var setting = config["somekey"]; // also returns "somevalue"

Note

You must set at least one configuration source.

It’s not unusual to store configuration values in a hierarchical structure, especially when using external files (e.g. JSON, XML, INI). In this case, configuration values can be retrieved using a : separated key, starting from the root of the hierarchy. For example, consider the following appsettings.json file:

{
  "ConnectionStrings": {
    "DefaultConnection": "Server=(localdb)\\mssqllocaldb;Database=aspnet-WebApplication1-26e8893e-d7c0-4fc6-8aab-29b59971d622;Trusted_Connection=True;MultipleActiveResultSets=true"
  },
  "Logging": {
    "IncludeScopes": false,
    "LogLevel": {
      "Default": "Debug",
      "System": "Information",
      "Microsoft": "Information"
    }
  }
}

The application uses configuration to configure the right connection string. Access to the DefaultConnection setting is achieved through this key: ConnectionStrings:DefaultConnection, or by using the GetConnectionString extension method and passing in "DefaultConnection".

The settings required by your application and the mechanism used to specify those settings (configuration being one example) can be decoupled using the options pattern. To use the options pattern you create your own options class (probably several different classes, corresponding to different cohesive groups of settings) that you can inject into your application using an options service. You can then specify your settings using configuration or whatever mechanism you choose.

Note

You could store your Configuration instance as a service, but this would unnecessarily couple your application to a single configuration system and specific configuration keys. Instead, you can use the Options pattern to avoid these issues.

Using the built-in sources ¶

The configuration framework has built-in support for JSON, XML, and INI configuration files, as well as support for in-memory configuration (directly setting values in code) and the ability to pull configuration from environment variables and command line parameters. Developers are not limited to using a single configuration source. In fact several may be set up together such that a default configuration is overridden by settings from another source if they are present.

Adding support for additional configuration sources is accomplished through extension methods. These methods can be called on a ConfigurationBuilder instance in a standalone fashion, or chained together as a fluent API. Both of these approaches are demonstrated in the sample below.

// work with with a builder using multiple calls
var builder = new ConfigurationBuilder();
builder.SetBasePath(Directory.GetCurrentDirectory());
builder.AddJsonFile("appsettings.json");
var connectionStringConfig = builder.Build();

// chain calls together as a fluent API
var config = new ConfigurationBuilder()
    .SetBasePath(Directory.GetCurrentDirectory())
    .AddJsonFile("appsettings.json")
    .AddEntityFrameworkConfig(options =>
        options.UseSqlServer(connectionStringConfig.GetConnectionString("DefaultConnection"))
    )
    .Build();

The order in which configuration sources are specified is important, as this establishes the precedence with which settings will be applied if they exist in multiple locations. In the example below, if the same setting exists in both appsettings.json and in an environment variable, the setting from the environment variable will be the one that is used. The last configuration source specified “wins” if a setting exists in more than one location. The ASP.NET team recommends specifying environment variables last, so that the local environment can override anything set in deployed configuration files.

Note

To override nested keys through environment variables in shells that don’t support : in variable names, replace them with __ (double underscore).

It can be useful to have environment-specific configuration files. This can be achieved using the following:

public Startup(IHostingEnvironment env)
{
    var builder = new ConfigurationBuilder()
        .SetBasePath(env.ContentRootPath)
        .AddJsonFile("appsettings.json", optional: true, reloadOnChange: true)
        .AddJsonFile($"appsettings.{env.EnvironmentName}.json", optional: true);

    if (env.IsDevelopment())
    {
        // For more details on using the user secret store see http://go.microsoft.com/fwlink/?LinkID=532709
        builder.AddUserSecrets();
    }

    builder.AddEnvironmentVariables();
    Configuration = builder.Build();
}

The IHostingEnvironment service is used to get the current environment. In the Development environment, the highlighted line of code above would look for a file named appsettings.Development.json and use its values, overriding any other values, if it’s present. Learn more about Working with Multiple Environments 다중 환경에서 작업하기.

When specifying files as configuration sources, you can optionally specify whether changes to the file should result in the settings being reloaded. This is configured by passing in a true value for the reloadOnChange parameter when calling AddJsonFile or similar file-based extension methods.

Warning

You should never store passwords or other sensitive data in configuration provider code or in plain text configuration files. You also shouldn’t use production secrets in your development or test environments. Instead, such secrets should be specified outside the project tree, so they cannot be accidentally committed into the configuration provider repository. Learn more about Working with Multiple Environments 다중 환경에서 작업하기 and managing Safe storage of app secrets during development.

One way to leverage the order precedence of Configuration is to specify default values, which can be overridden. In the console application below, a default value for the username setting is specified in an in-memory collection, but this is overridden if a command line argument for username is passed to the application. You can see in the output how many different configuration sources are configured in the application at each stage of its execution.

using System;
using System.Collections.Generic;
using System.Linq;
using Microsoft.Extensions.Configuration;

namespace ConfigConsole
{
    public static class Program
    {
        public static void Main(string[] args)
        {
            var builder = new ConfigurationBuilder();
            Console.WriteLine("Initial Config Sources: " + builder.Sources.Count());

            builder.AddInMemoryCollection(new Dictionary<string, string>
            {
                { "username", "Guest" }
            });

            Console.WriteLine("Added Memory Source. Sources: " + builder.Sources.Count());

            builder.AddCommandLine(args);
            Console.WriteLine("Added Command Line Source. Sources: " + builder.Sources.Count());

            var config = builder.Build();
            string username = config["username"];

            Console.WriteLine($"Hello, {username}!");
        }
    }
}

When run, the program will display the default value unless a command line parameter overrides it.

Using Options and configuration objects ¶

The options pattern enables using custom options classes to represent a group of related settings. A class needs to have a public read-write property for each setting and a constructor that does not take any parameters (e.g. a default constructor) in order to be used as an options class.

It’s recommended that you create well-factored settings objects that correspond to certain features within your application, thus following the Interface Segregation Principle (ISP) (classes depend only on the configuration settings they use) as well as Separation of Concerns (settings for disparate parts of your app are managed separately, and thus are less likely to negatively impact one another).

A simple MyOptions class is shown here:

public class MyOptions
{
    public string Option1 { get; set; }
    public int Option2 { get; set; }
}

Options can be injected into your application using the IOptions<TOptions> accessor service. For example, the following controller uses IOptions<MyOptions> to access the settings it needs to render the Index view:

public class HomeController : Controller
{
    private readonly IOptions<MyOptions> _optionsAccessor;

    public HomeController(IOptions<MyOptions> optionsAccessor)
    {
        _optionsAccessor = optionsAccessor;
    }

    // GET: /<controller>/
    public IActionResult Index() => View(_optionsAccessor.Value);
}

Tip

Learn more about Dependency Injection.

To setup the IOptions<TOptions> service you call the AddOptions extension method during startup in your ConfigureServices method:

public void ConfigureServices(IServiceCollection services)
{
    // Setup options with DI
    services.AddOptions();

The Index view displays the configured options:

You configure options using the Configure<TOptions> extension method. You can configure options using a delegate or by binding your options to configuration:

public void ConfigureServices(IServiceCollection services)
{
    // Setup options with DI
    services.AddOptions();

    // Configure MyOptions using config by installing Microsoft.Extensions.Options.ConfigurationExtensions
    services.Configure<MyOptions>(Configuration);

    // Configure MyOptions using code
    services.Configure<MyOptions>(myOptions =>
    {
        myOptions.Option1 = "value1_from_action";
    });

    // Configure MySubOptions using a sub-section of the appsettings.json file
    services.Configure<MySubOptions>(Configuration.GetSection("subsection"));

    // Add framework services.
    services.AddMvc();
}

When you bind options to configuration, each property in your options type is bound to a configuration key of the form property:subproperty:.... For example, the MyOptions.Option1 property is bound to the key Option1, which is read from the option1 property in appsettings.json. Note that configuration keys are case insensitive.

Each call to Configure<TOptions> adds an IConfigureOptions<TOptions> service to the service container that is used by the IOptions<TOptions> service to provide the configured options to the application or framework. If you want to configure your options using objects that must be obtained from the service container (for example, to read settings from a database) you can use the AddSingleton<IConfigureOptions<TOptions>> extension method to register a custom IConfigureOptions<TOptions> service.

You can have multiple IConfigureOptions<TOptions> services for the same option type and they are all applied in order. In the example above, the values of Option1 and Option2 are both specified in appsettings.json, but the value of Option1 is overridden by the configured delegate with the value “value1_from_action”.

Writing custom providers ¶

In addition to using the built-in configuration providers, you can also write your own. To do so, you simply implement the IConfigurationSource interface, which exposes a Build method. The build method configures and returns an IConfigurationProvider.

Example: Entity Framework Settings¶

You may wish to store some of your application’s settings in a database, and access them using Entity Framework Core (EF). There are many ways in which you could choose to store such values, ranging from a simple table with a column for the setting name and another column for the setting value, to having separate columns for each setting value. In this example, we’re going to create a simple configuration provider that reads name-value pairs from a database using EF.

To start off we’ll define a simple ConfigurationValue entity for storing configuration values in the database:

public class ConfigurationValue
{
    public string Id { get; set; }
    public string Value { get; set; }
}

You need a ConfigurationContext to store and access the configured values using EF:

public class ConfigurationContext : DbContext
{
    public ConfigurationContext(DbContextOptions options) : base(options)
    {
    }

    public DbSet<ConfigurationValue> Values { get; set; }
}

Create an EntityFrameworkConfigurationSource that inherits from IConfigurationSource:

using System;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;

namespace CustomConfigurationProvider
{
    public class EntityFrameworkConfigurationSource : IConfigurationSource
    {
        private readonly Action<DbContextOptionsBuilder> _optionsAction;

        public EntityFrameworkConfigurationSource(Action<DbContextOptionsBuilder> optionsAction)
        {
            _optionsAction = optionsAction;
        }

        public IConfigurationProvider Build(IConfigurationBuilder builder)
        {
            return new EntityFrameworkConfigurationProvider(_optionsAction);
        }
    }
}

Next, create the custom configuration provider by inheriting from ConfigurationProvider. The configuration data is loaded by overriding the Load method, which reads in all of the configuration data from the configured database. For demonstration purposes, the configuration provider also takes care of initializing the database if it hasn’t already been created and populated:

using System;
using System.Collections.Generic;
using System.Linq;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;

namespace CustomConfigurationProvider
{
    public class EntityFrameworkConfigurationProvider : ConfigurationProvider
    {
        public EntityFrameworkConfigurationProvider(Action<DbContextOptionsBuilder> optionsAction)
        {
            OptionsAction = optionsAction;
        }

        Action<DbContextOptionsBuilder> OptionsAction { get; }

        public override void Load()
        {
            var builder = new DbContextOptionsBuilder<ConfigurationContext>();
            OptionsAction(builder);

            using (var dbContext = new ConfigurationContext(builder.Options))
            {
                dbContext.Database.EnsureCreated();
                Data = !dbContext.Values.Any()
                    ? CreateAndSaveDefaultValues(dbContext)
                    : dbContext.Values.ToDictionary(c => c.Id, c => c.Value);
            }
        }

        private static IDictionary<string, string> CreateAndSaveDefaultValues(
            ConfigurationContext dbContext)
        {
            var configValues = new Dictionary<string, string>
                {
                    { "key1", "value_from_ef_1" },
                    { "key2", "value_from_ef_2" }
                };
            dbContext.Values.AddRange(configValues
                .Select(kvp => new ConfigurationValue { Id = kvp.Key, Value = kvp.Value })
                .ToArray());
            dbContext.SaveChanges();
            return configValues;
        }
    }
}

Note the values that are being stored in the database (“value_from_ef_1” and “value_from_ef_2”); these are displayed in the sample below to demonstrate the configuration is reading values from the database properly.

By convention you can also add an AddEntityFrameworkConfiguration extension method for adding the configuration source:

using System;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;

namespace CustomConfigurationProvider
{
    public static class EntityFrameworkExtensions
    {
        public static IConfigurationBuilder AddEntityFrameworkConfig(
            this IConfigurationBuilder builder, Action<DbContextOptionsBuilder> setup)
        {
            return builder.Add(new EntityFrameworkConfigurationSource(setup));
        }
    }
}

You can see an example of how to use this custom configuration provider in your application in the following example. Create a new ConfigurationBuilder to set up your configuration sources. To add the EntityFrameworkConfigurationProvider, you first need to specify the EF data provider and connection string. How should you configure the connection string? Using configuration of course! Add an appsettings.json file as a configuration source to bootstrap setting up the EntityFrameworkConfigurationProvider. By adding the database settings to an existing configuration with other sources specified, any settings specified in the database will override settings specified in appsettings.json:

using System;
using System.IO;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;

namespace CustomConfigurationProvider
{
    public static class Program
    {
        public static void Main()
        {
            // work with with a builder using multiple calls
            var builder = new ConfigurationBuilder();
            builder.SetBasePath(Directory.GetCurrentDirectory());
            builder.AddJsonFile("appsettings.json");
            var connectionStringConfig = builder.Build();

            // chain calls together as a fluent API
            var config = new ConfigurationBuilder()
                .SetBasePath(Directory.GetCurrentDirectory())
                .AddJsonFile("appsettings.json")
                .AddEntityFrameworkConfig(options =>
                    options.UseSqlServer(connectionStringConfig.GetConnectionString("DefaultConnection"))
                )
                .Build();

            Console.WriteLine("key1={0}", config["key1"]);
            Console.WriteLine("key2={0}", config["key2"]);
            Console.WriteLine("key3={0}", config["key3"]);
        }
    }
}

Run the application to see the configured values:

Summary ¶

ASP.NET Core provides a very flexible configuration model that supports a number of different file-based options, as well as command-line, in-memory, and environment variables. It works seamlessly with the options model so that you can inject strongly typed settings into your application or framework. You can create your own custom configuration providers as well, which can work with or replace the built-in providers, allowing for extreme flexibility.

Logging¶

ASP.NET Core has built-in support for logging, and allows developers to easily leverage their preferred logging framework’s functionality as well. Implementing logging in your application requires a minimal amount of setup code. Once this is in place, logging can be added wherever it is desired.

Sections:

Implementing Logging in your Application
Configuring Logging in your Application
Logging Recommendations
Summary

Implementing Logging in your Application ¶

Adding logging to a component in your application is done by requesting either an ILoggerFactory or an ILogger<T> via Dependency Injection. If an ILoggerFactory is requested, a logger must be created using its CreateLogger method. The following example shows how to do this:

var logger = loggerFactory.CreateLogger("Catchall Endpoint");
logger.LogInformation("No endpoint found for request {path}", context.Request.Path);

When a logger is created, a category name must be provided. The category name specifies the source of the logging events. By convention this string is hierarchical, with categories separated by dot (.) characters. Some logging providers have filtering support that leverages this convention, making it easier to locate logging output of interest. In this article’s sample application, logging is configured to use the built-in ConsoleLogger (see Configuring Logging in your Application below). To see the console logger in action, run the sample application using the dotnet run command, and make a request to configured URL (localhost:5000). You should see output similar to the following:

You may see more than one log statement per web request you make in your browser, since most browsers will make multiple requests (i.e. for the favicon file) when attempting to load a page. Note that the console logger displayed the log level (info in the image above) followed by the category ([Catchall Endpoint]), and then the message that was logged.

The call to the log method can utilize a format string with named placeholders (like {path}). These placeholders are populated in the order in which they appear by the args values passed into the method call. Some logging providers will store these names along with their mapped values in a dictionary that can later be queried. In the example below, the request path is passed in as a named placeholder:

logger.LogInformation("No endpoint found for request {path}", context.Request.Path);

In your real world applications, you will want to add logging based on application-level, not framework-level, events. For instance, if you have created a Web API application for managing To-Do Items (see Building Your First Web API with ASP.NET Core MVC and Visual Studio), you might add logging around the various operations that can be performed on these items.

The logic for the API is contained within the TodoController, which uses Dependency Injection to request the services it requires via its constructor. Ideally, classes should follow this example and use their constructor to define their dependencies explicitly as parameters. Rather than requesting an ILoggerFactory and creating an instance of ILogger explicitly, TodoController demonstrates another way to work with loggers in your application - you can request an ILogger<T> (where T is the class requesting the logger).

[Route("api/[controller]")]
public class TodoController : Controller
{
    private readonly ITodoRepository _todoRepository;
    private readonly ILogger<TodoController> _logger;

    public TodoController(ITodoRepository todoRepository, 
        ILogger<TodoController> logger)
    {
        _todoRepository = todoRepository;
        _logger = logger;
    }

    [HttpGet]
    public IEnumerable<TodoItem> GetAll()
    {
        _logger.LogInformation(LoggingEvents.LIST_ITEMS, "Listing all items");
        EnsureItems();
        return _todoRepository.GetAll();
    }

Within each controller action, logging is done through the use of the local field, _logger, as shown on line 17, above. This technique is not limited to controllers, but can be utilized by any of your application services that utilize Dependency Injection.

Working with ILogger<T>¶

As we have just seen, your application can request an instance of ILogger<T> as a dependency in a class’s constructor, where T is the type performing logging. The TodoController shows an example of this approach. When this technique is used, the logger will automatically use the type’s name as its category name. By requesting an instance of ILogger<T>, your class doesn’t need to create an instance of a logger via ILoggerFactory. You can use this approach anywhere you don’t need the additional functionality offered by ILoggerFactory.

Logging Verbosity Levels¶

When adding logging statements to your application, you must specify a LogLevel. The LogLevel allows you to control the verbosity of the logging output from your application, as well as the ability to pipe different kinds of log messages to different loggers. For example, you may wish to log debug messages to a local file, but log errors to the machine’s event log or a database.

ASP.NET Core defines six levels of logging verbosity, ordered by increasing importance or severity:

Trace: Used for the most detailed log messages, typically only valuable to a developer debugging an issue. These messages may contain sensitive application data and so should not be enabled in a production environment. Disabled by default. Example: Credentials: {"User":"someuser", "Password":"P@ssword"}
Debug: These messages have short-term usefulness during development. They contain information that may be useful for debugging, but have no long-term value. This is the default most verbose level of logging. Example: Entering method Configure with flag set to true
Information: These messages are used to track the general flow of the application. These logs should have some long term value, as opposed to Verbose level messages, which do not. Example: Request received for path /foo
Warning: The Warning level should be used for abnormal or unexpected events in the application flow. These may include errors or other conditions that do not cause the application to stop, but which may need to be investigated in the future. Handled exceptions are a common place to use the Warning log level. Examples: Login failed for IP 127.0.0.1 or FileNotFoundException for file foo.txt
Error: An error should be logged when the current flow of the application must stop due to some failure, such as an exception that cannot be handled or recovered from. These messages should indicate a failure in the current activity or operation (such as the current HTTP request), not an application-wide failure. Example: Cannot insert record due to duplicate key violation
Critical: A critical log level should be reserved for unrecoverable application or system crashes, or catastrophic failure that requires immediate attention. Examples: data loss scenarios, out of disk space

The Logging package provides helper extension methods for each LogLevel value, allowing you to call, for example, LogInformation, rather than the more verbose Log(LogLevel.Information, ...) method. Each of the LogLevel-specific extension methods has several overloads, allowing you to pass in some or all of the following parameters:

string data: The message to log.
EventId eventId: A numeric id to associate with the log, which can be used to associate a series of logged events with one another. Event IDs should be static and specific to a particular kind of event that is being logged. For instance, you might associate adding an item to a shopping cart as event id 1000 and completing a purchase as event id 1001. This allows intelligent filtering and processing of log statements.
string format: A format string for the log message.
object[] args: An array of objects to format.
Exception error: An exception instance to log.

Note

The EventId type can be implicitly casted to int, so you can just pass an int to this argument.

Note

Some loggers, such as the built-in ConsoleLogger used in this article, will ignore the eventId parameter. If you need to display it, you can include it in the message string. This is done in the following sample so you can easily see the eventId associated with each message, but in practice you would not typically include it in the log message.

In the TodoController example, event id constants are defined for each event, and log statements are configured at the appropriate verbosity level based on the success of the operation. In this case, successful operations log as Information and not found results are logged as Warning (error handling is not shown).

[HttpGet]
public IEnumerable<TodoItem> GetAll()
{
    _logger.LogInformation(LoggingEvents.LIST_ITEMS, "Listing all items");
    EnsureItems();
    return _todoRepository.GetAll();
}

[HttpGet("{id}", Name = "GetTodo")]
public IActionResult GetById(string id)
{
    _logger.LogInformation(LoggingEvents.GET_ITEM, "Getting item {0}", id);
    var item = _todoRepository.Find(id);
    if (item == null)
    {
        _logger.LogWarning(LoggingEvents.GET_ITEM_NOTFOUND, "GetById({0}) NOT FOUND", id);
        return NotFound();
    }
    return new ObjectResult(item);
}

Note

It is recommended that you perform application logging at the level of your application and its APIs, not at the level of the framework. The framework already has logging built in which can be enabled simply by setting the appropriate logging verbosity level.

To see more detailed logging at the framework level, you can adjust the LogLevel specified to your logging provider to something more verbose (like Debug or Trace). For example, if you modify the AddConsole call in the Configure method to use LogLevel.Trace and run the application, the result shows much more framework-level detail about each request:

The console logger prefixes debug output with “dbug: ”; there is no trace level debugging enabled by the framework by default. Each log level has a corresponding four character prefix that is used, so that log messages are consistently aligned.

Log Level	Prefix
Critical	crit
Error	fail
Warning	warn
Information	info
Debug	dbug
Trace	trce

Scopes¶

In the course of logging information within your application, you can group a set of logical operations within a scope. A scope is an IDisposable type returned by calling the ILogger.BeginScope<TState> method, which lasts from the moment it is created until it is disposed. The built-in TraceSource logger returns a scope instance that is responsible for starting and stopping tracing operations. Any logging state, such as a transaction id, is attached to the scope when it is created.

Scopes are not required, and should be used sparingly, if at all. They’re best used for operations that have a distinct beginning and end, such as a transaction involving multiple resources.

Configuring Logging in your Application ¶

To configure logging in your ASP.NET Core application, you should resolve ILoggerFactory in the Configure method of your Startup class. ASP.NET Core will automatically provide an instance of ILoggerFactory using Dependency Injection when you add a parameter of this type to the Configure method.

    public void Configure(IApplicationBuilder app,
        IHostingEnvironment env,
        ILoggerFactory loggerFactory)

Once you’ve added ILoggerFactory as a parameter, you configure loggers within the Configure method by calling methods (or extension methods) on the logger factory. We have already seen an example of this configuration at the beginning of this article, when we added console logging by calling loggerFactory.AddConsole.

Each logger provides its own set of extension methods to ILoggerFactory. The console, debug, and event log loggers allow you to specify the minimum logging level at which those loggers should write log messages. The console and debug loggers provide extension methods accepting a function to filter log messages according to their logging level and/or category (for example, logLevel => logLevel >= LogLevel.Warning or (category, loglevel) => category.Contains("MyController") && loglevel >= LogLevel.Trace). The event log logger provides a similar overload that takes an EventLogSettings instance as argument, which may contain a filtering function in its Filter property. The TraceSource logger does not provide any of those overloads, since its logging level and other parameters are based on the SourceSwitch and TraceListener it uses.

A LoggerFactory instance can optionally be configured with custom FilterLoggerSettings. The example below configures custom log levels for different scopes, limiting system and Microsoft built-in logging to warnings while allowing the app to log at debug level by default. The WithFilter method returns a new ILoggerFactory that will filter the log messages passed to all logger providers registered with it. It does not affect any other ILoggerFactory instances, including the original ILoggerFactory instance.

loggerFactory
    .WithFilter(new FilterLoggerSettings
    {
        { "Microsoft", LogLevel.Warning },
        { "System", LogLevel.Warning },
        { "ToDoApi", LogLevel.Debug }
    })
    .AddConsole();

Configuring TraceSource Logging¶

When running on the full .NET Framework you can configuring logging to use the existing System.Diagnostics.TraceSource libraries and providers, including easy access to the Windows event log. TraceSource allows you to route messages to a variety of listeners and is already in use by many organizations.

First, be sure to add the Microsoft.Extensions.Logging.TraceSource package to your project (in project.json), along with any specific trace source packages you’ll be using (in this case, TextWriterTraceListener):

  "Microsoft.AspNetCore.Mvc": "1.0.0",
  "Microsoft.AspNetCore.Server.Kestrel": "1.0.0",
  "Microsoft.AspNetCore.Server.IISIntegration": "1.0.0",
  "Microsoft.AspNetCore.StaticFiles": "1.0.0",
  "Microsoft.Extensions.Logging": "1.0.0",
  "Microsoft.Extensions.Logging.Console": "1.0.0",
  "Microsoft.Extensions.Logging.Filter": "1.0.0",
  "Microsoft.Extensions.Logging.TraceSource": "1.0.0"
},
"tools": {
  "Microsoft.AspNetCore.Server.IISIntegration.Tools": {

The following example demonstrates how to configure a TraceSourceLogger instance for an application, logging only Warning or higher priority messages. Each call to AddTraceSource takes a TraceListener. The call configures a TextWriterTraceListener to write to the console window. This log output will be in addition to the console logger that was already added to this sample, but its behavior is slightly different.

// add Trace Source logging
var testSwitch = new SourceSwitch("sourceSwitch", "Logging Sample");
testSwitch.Level = SourceLevels.Warning;
loggerFactory.AddTraceSource(testSwitch,
    new TextWriterTraceListener(writer: Console.Out));

The sourceSwitch is configured to use SourceLevels.Warning, so only Warning (or higher) log messages are picked up by the TraceListener instance.

The API action below logs a warning when the specified id is not found:

[HttpGet("{id}", Name = "GetTodo")]
public IActionResult GetById(string id)
{
    _logger.LogInformation(LoggingEvents.GET_ITEM, "Getting item {0}", id);
    var item = _todoRepository.Find(id);
    if (item == null)
    {
        _logger.LogWarning(LoggingEvents.GET_ITEM_NOTFOUND, "GetById({0}) NOT FOUND", id);
        return NotFound();
    }
    return new ObjectResult(item);
}

To test out this code, you can trigger logging a warning by running the app from the console and navigating to http://localhost:5000/api/Todo/0. You should see output similar to the following:

The yellow line with the “warn: ” prefix, along with the following line, is output by the ConsoleLogger. The next line, beginning with “TodoApi.Controllers.TodoController”, is output from the TraceSource logger. There are many other TraceSource listeners available, and the TextWriterTraceListener can be configured to use any TextWriter instance, making this a very flexible option for logging.

Configuring Other Providers¶

In addition to the built-in loggers, you can configure logging to use other providers. Add the appropriate package to your project.json file, and then configure it just like any other provider. Typically, these packages include extension methods on ILoggerFactory to make it easy to add them.

elmah.io - provider for the elmah.io service

Loggr - provider for the Loggr service

NLog - provider for the NLog library

Serilog - provider for the Serilog library

You can create your own custom providers as well, to support other logging frameworks or your own internal logging requirements.

Logging Recommendations ¶

The following are some recommendations you may find helpful when implementing logging in your ASP.NET Core applications.

Log using the correct LogLevel. This will allow you to consume and route logging output appropriately based on the importance of the messages.
Log information that will enable errors to be identified quickly. Avoid logging irrelevant or redundant information.
Keep log messages concise without sacrificing important information.
Although loggers will not log if disabled, consider adding code guards around logging methods to prevent extra method calls and log message setup overhead, especially within loops and performance critical methods.
Name your loggers with a distinct prefix so they can easily be filtered or disabled. Remember the Create<T> extension will create loggers named with the full name of the class.
Use Scopes sparingly, and only for actions with a bounded start and end. For example, the framework provides a scope around MVC actions. Avoid nesting many scopes within one another.
Application logging code should be related to the business concerns of the application. Increase the logging verbosity to reveal additional framework-related concerns, rather than implementing yourself.

Summary ¶

ASP.NET Core provides built-in support for logging, which can easily be configured within the Startup class and used throughout the application. Logging verbosity can be configured globally and per logging provider to ensure actionable information is logged appropriately. Built-in providers for console and trace source logging are included in the framework; other logging frameworks can easily be configured as well.

🔧 File Providers¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Dependency Injection¶

Steve Smith

ASP.NET Core is designed from the ground up to support and leverage dependency injection. ASP.NET Core applications can leverage built-in framework services by having them injected into methods in the Startup class, and application services can be configured for injection as well. The default services container provided by ASP.NET Core provides a minimal feature set and is not intended to replace other containers.

Sections:

What is Dependency Injection?
Using Framework-Provided Services
Registering Your Own Services
Service Lifetimes and Registration Options
Request Services
Designing Your Services For Dependency Injection
Replacing the default services container
Recommendations
Additional Resources

What is Dependency Injection?¶

Dependency injection (DI) is a technique for achieving loose coupling between objects and their collaborators, or dependencies. Rather than directly instantiating collaborators, or using static references, the objects a class needs in order to perform its actions are provided to the class in some fashion. Most often, classes will declare their dependencies via their constructor, allowing them to follow the Explicit Dependencies Principle. This approach is known as “constructor injection”.

When classes are designed with DI in mind, they are more loosely coupled because they do not have direct, hard-coded dependencies on their collaborators. This follows the Dependency Inversion Principle, which states that “high level modules should not depend on low level modules; both should depend on abstractions.” Instead of referencing specific implementations, classes, request abstractions (typically interfaces) which are provided to them when they are constructed. Extracting dependencies into interfaces and providing implementations of these interfaces as parameters is also an example of the Strategy design pattern.

When a system is designed to use DI, with many classes requesting their dependencies via their constructor (or properties), it’s helpful to have a class dedicated to creating these classes with their associated dependencies. These classes are referred to as containers, or more specifically, Inversion of Control (IoC) containers or Dependency Injection (DI) containers. A container is essentially a factory that is responsible for providing instances of types that are requested from it. If a given type has declared that it has dependencies, and the container has been configured to provide the dependency types, it will create the dependencies as part of creating the requested instance. In this way, complex dependency graphs can be provided to classes without the need for any hard-coded object construction. In addition to creating objects with their dependencies, containers typically manage object lifetimes within the application.

ASP.NET Core includes a simple built-in container (represented by the IServiceProvider interface) that supports constructor injection by default, and ASP.NET makes certain services available through DI. ASP.NET’s container refers to the types it manages as services. Throughout the rest of this article, services will refer to types that are managed by ASP.NET Core’s IoC container. You configure the built-in container’s services in the ConfigureServices method in your application’s Startup class.

Note

Martin Fowler has written an extensive article on Inversion of Control Containers and the Dependency Injection Pattern. Microsoft Patterns and Practices also has a great description of Dependency Injection.

Note

This article covers Dependency Injection as it applies to all ASP.NET applications. Dependency Injection within MVC controllers is covered in Dependency Injection and Controllers.

Using Framework-Provided Services ¶

The ConfigureServices method in the Startup class is responsible for defining the services the application will use, including platform features like Entity Framework Core and ASP.NET Core MVC. Initially, the IServiceCollection provided to ConfigureServices has just a handful of services defined. Below is an example of how to add additional services to the container using a number of extensions methods like AddDbContext, AddIdentity, and AddMvc.

// This method gets called by the runtime. Use this method to add services to the container.
public void ConfigureServices(IServiceCollection services)
{
    // Add framework services.
    services.AddDbContext<ApplicationDbContext>(options =>
        options.UseSqlServer(Configuration.GetConnectionString("DefaultConnection")));

    services.AddIdentity<ApplicationUser, IdentityRole>()
        .AddEntityFrameworkStores<ApplicationDbContext>()
        .AddDefaultTokenProviders();

    services.AddMvc();

    // Add application services.
    services.AddTransient<IEmailSender, AuthMessageSender>();
    services.AddTransient<ISmsSender, AuthMessageSender>();
}

The features and middleware provided by ASP.NET, such as MVC, follow a convention of using a single AddService extension method to register all of the services required by that feature.

Tip

You can request certain framework-provided services within Startup methods through their parameter lists - see Application Startup for more details.

Of course, in addition to configuring the application to take advantage of various framework features, you can also use ConfigureServices to configure your own application services.

Registering Your Own Services ¶

You can register your own application services as follows. The first generic type represents the type (typically an interface) that will be requested from the container. The second generic type represents the concrete type that will be instantiated by the container and used to fulfill such requests.

services.AddTransient<IEmailSender, AuthMessageSender>();
services.AddTransient<ISmsSender, AuthMessageSender>();

Note

Each services.Add<service> calls adds (and potentially configures) services. For example, services.AddMvc() adds the services MVC requires.

The AddTransient method is used to map abstract types to concrete services that are instantiated separately for every object that requires it. This is known as the service’s lifetime, and additional lifetime options are described below. It is important to choose an appropriate lifetime for each of the services you register. Should a new instance of the service be provided to each class that requests it? Should one instance be used throughout a given web request? Or should a single instance be used for the lifetime of the application?

In the sample for this article, there is a simple controller that displays character names, called CharactersController. Its Index method displays the current list of characters that have been stored in the application, and initializes the collection with a handful of characters if none exist. Note that although this application uses Entity Framework Core and the ApplicationDbContext class for its persistence, none of that is apparent in the controller. Instead, the specific data access mechanism has been abstracted behind an interface, ICharacterRepository, which follows the repository pattern. An instance of ICharacterRepository is requested via the constructor and assigned to a private field, which is then used to access characters as necessary.

public class CharactersController : Controller
{
    private readonly ICharacterRepository _characterRepository;

    public CharactersController(ICharacterRepository characterRepository)
    {
        _characterRepository = characterRepository;
    }

    // GET: /characters/
    public IActionResult Index()
    {
        PopulateCharactersIfNoneExist();
        var characters = _characterRepository.ListAll();

        return View(characters);
    }
    
    private void PopulateCharactersIfNoneExist()
    {
        if (!_characterRepository.ListAll().Any())
        {
            _characterRepository.Add(new Character("Darth Maul"));
            _characterRepository.Add(new Character("Darth Vader"));
            _characterRepository.Add(new Character("Yoda"));
            _characterRepository.Add(new Character("Mace Windu"));
        }
    }
}

The ICharacterRepository simply defines the two methods the controller needs to work with Character instances.

using System.Collections.Generic;
using DependencyInjectionSample.Models;

namespace DependencyInjectionSample.Interfaces
{
    public interface ICharacterRepository
    {
        IEnumerable<Character> ListAll();
        void Add(Character character);
    }
}

This interface is in turn implemented by a concrete type, CharacterRepository, that is used at runtime.

Note

The way DI is used with the CharacterRepository class is a general model you can follow for all of your application services, not just in “repositories” or data access classes.

using System.Collections.Generic;
using System.Linq;
using DependencyInjectionSample.Interfaces;

namespace DependencyInjectionSample.Models
{
    public class CharacterRepository : ICharacterRepository
    {
        private readonly ApplicationDbContext _dbContext;

        public CharacterRepository(ApplicationDbContext dbContext)
        {
            _dbContext = dbContext;
        }

        public IEnumerable<Character> ListAll()
        {
            return _dbContext.Characters.AsEnumerable();
        }

        public void Add(Character character)
        {
            _dbContext.Characters.Add(character);
            _dbContext.SaveChanges();
        }
    }
}

Note that CharacterRepository requests an ApplicationDbContext in its constructor. It is not unusual for dependency injection to be used in a chained fashion like this, with each requested dependency in turn requesting its own dependencies. The container is responsible for resolving all of the dependencies in the graph and returning the fully resolved service.

Note

Creating the requested object, and all of the objects it requires, and all of the objects those require, is sometimes referred to as an object graph. Likewise, the collective set of dependencies that must be resolved is typically referred to as a dependency tree or dependency graph.

In this case, both ICharacterRepository and in turn ApplicationDbContext must be registered with the services container in ConfigureServices in Startup. ApplicationDbContext is configured with the call to the extension method AddDbContext<T>. The following code shows the registration of the CharacterRepository type.

public void ConfigureServices(IServiceCollection services)
{
    services.AddDbContext<ApplicationDbContext>(options =>
        options.UseInMemoryDatabase()
    );

    // Add framework services.
    services.AddMvc();

    // Register application services.
    services.AddScoped<ICharacterRepository, CharacterRepository>();
    services.AddTransient<IOperationTransient, Operation>();
    services.AddScoped<IOperationScoped, Operation>();
    services.AddSingleton<IOperationSingleton, Operation>();
    services.AddSingleton<IOperationSingletonInstance>(new Operation(Guid.Empty));
    services.AddTransient<OperationService, OperationService>();
}

Entity Framework contexts should be added to the services container using the Scoped lifetime. This is taken care of automatically if you use the helper methods as shown above. Repositories that will make use of Entity Framework should use the same lifetime.

Warning

The main danger to be wary of is resolving a Scoped service from a singleton. It’s likely in such a case that the service will have incorrect state when processing subsequent requests.

Service Lifetimes and Registration Options ¶

ASP.NET services can be configured with the following lifetimes:

Transient: Transient lifetime services are created each time they are requested. This lifetime works best for lightweight, stateless services.
Scoped: Scoped lifetime services are created once per request.
Singleton: Singleton lifetime services are created the first time they are requested (or when ConfigureServices is run if you specify an instance there) and then every subsequent request will use the same instance. If your application requires singleton behavior, allowing the services container to manage the service’s lifetime is recommended instead of implementing the singleton design pattern and managing your object’s lifetime in the class yourself.

Services can be registered with the container in several ways. We have already seen how to register a service implementation with a given type by specifying the concrete type to use. In addition, a factory can be specified, which will then be used to create the instance on demand. The third approach is to directly specify the instance of the type to use, in which case the container will never attempt to create an instance.

To demonstrate the difference between these lifetime and registration options, consider a simple interface that represents one or more tasks as an operation with a unique identifier, OperationId. Depending on how we configure the lifetime for this service, the container will provide either the same or different instances of the service to the requesting class. To make it clear which lifetime is being requested, we will create one type per lifetime option:

using System;

namespace DependencyInjectionSample.Interfaces
{
    public interface IOperation
    {
        Guid OperationId { get; }
    }

    public interface IOperationTransient : IOperation
    {
    }
    public interface IOperationScoped : IOperation
    {
    }
    public interface IOperationSingleton : IOperation
    {
    }
    public interface IOperationSingletonInstance : IOperation
    {
    }
}

We implement these interfaces using a single class, Operation, that accepts a Guid in its constructor, or uses a new Guid if none is provided.

Next, in ConfigureServices, each type is added to the container according to its named lifetime:

services.AddTransient<IOperationTransient, Operation>();
services.AddScoped<IOperationScoped, Operation>();
services.AddSingleton<IOperationSingleton, Operation>();
services.AddSingleton<IOperationSingletonInstance>(new Operation(Guid.Empty));
services.AddTransient<OperationService, OperationService>();

Note that the IOperationSingletonInstance service is using a specific instance with a known ID of Guid.Empty so it will be clear when this type is in use. We have also registered an OperationService that depends on each of the other Operation types, so that it will be clear within a request whether this service is getting the same instance as the controller, or a new one, for each operation type. All this service does is expose its dependencies as properties, so they can be displayed in the view.

using DependencyInjectionSample.Interfaces;

namespace DependencyInjectionSample.Services
{
    public class OperationService
    {
        public IOperationTransient TransientOperation { get; }
        public IOperationScoped ScopedOperation { get; }
        public IOperationSingleton SingletonOperation { get; }
        public IOperationSingletonInstance SingletonInstanceOperation { get; }

        public OperationService(IOperationTransient transientOperation,
            IOperationScoped scopedOperation,
            IOperationSingleton singletonOperation,
            IOperationSingletonInstance instanceOperation)
        {
            TransientOperation = transientOperation;
            ScopedOperation = scopedOperation;
            SingletonOperation = singletonOperation;
            SingletonInstanceOperation = instanceOperation;
        }
    }
}

To demonstrate the object lifetimes within and between separate individual requests to the application, the sample includes an OperationsController that requests each kind of IOperation type as well as an OperationService. The Index action then displays all of the controller’s and service’s OperationId values.

using DependencyInjectionSample.Interfaces;
using DependencyInjectionSample.Services;
using Microsoft.AspNetCore.Mvc;

namespace DependencyInjectionSample.Controllers
{
    public class OperationsController : Controller
    {
        private readonly OperationService _operationService;
        private readonly IOperationTransient _transientOperation;
        private readonly IOperationScoped _scopedOperation;
        private readonly IOperationSingleton _singletonOperation;
        private readonly IOperationSingletonInstance _singletonInstanceOperation;

        public OperationsController(OperationService operationService,
            IOperationTransient transientOperation,
            IOperationScoped scopedOperation,
            IOperationSingleton singletonOperation,
            IOperationSingletonInstance singletonInstanceOperation)
        {
            _operationService = operationService;
            _transientOperation = transientOperation;
            _scopedOperation = scopedOperation;
            _singletonOperation = singletonOperation;
            _singletonInstanceOperation = singletonInstanceOperation;
        }

        public IActionResult Index()
        {
            // viewbag contains controller-requested services
            ViewBag.Transient = _transientOperation;
            ViewBag.Scoped = _scopedOperation;
            ViewBag.Singleton = _singletonOperation;
            ViewBag.SingletonInstance = _singletonInstanceOperation;
            
            // operation service has its own requested services
            ViewBag.Service = _operationService;
            return View();
        }
    }
}

Now two separate requests are made to this controller action:

Observe which of the OperationId values varies within a request, and between requests.

Transient objects are always different; a new instance is provided to every controller and every service.
Scoped objects are the same within a request, but different across different requests
Singleton objects are the same for every object and every request (regardless of whether an instance is provided in ConfigureServices)

Request Services ¶

The services available within an ASP.NET request from HttpContext are exposed through the RequestServices collection.

Request Services represent the services you configure and request as part of your application. When your objects specify dependencies, these are satisfied by the types found in RequestServices, not ApplicationServices.

Generally, you shouldn’t use these properties directly, preferring instead to request the types your classes you require via your class’s constructor, and letting the framework inject these dependencies. This yields classes that are easier to test (see Testing) and are more loosely coupled.

Note

Prefer requesting dependencies as constructor parameters to accessing the RequestServices collection.

Designing Your Services For Dependency Injection ¶

You should design your services to use dependency injection to get their collaborators. This means avoiding the use of stateful static method calls (which result in a code smell known as static cling) and the direct instantiation of dependent classes within your services. It may help to remember the phrase, New is Glue, when choosing whether to instantiate a type or to request it via dependency injection. By following the SOLID Principles of Object Oriented Design, your classes will naturally tend to be small, well-factored, and easily tested.

What if you find that your classes tend to have way too many dependencies being injected? This is generally a sign that your class is trying to do too much, and is probably violating SRP - the Single Responsibility Principle. See if you can refactor the class by moving some of its responsibilities into a new class. Keep in mind that your Controller classes should be focused on UI concerns, so business rules and data access implementation details should be kept in classes appropriate to these separate concerns.

With regard to data access specifically, you can easily inject Entity Framework DbContext types into your controllers, assuming you’ve configured EF in your Startup class. However, it is best to avoid depending directly on DbContext in your UI project. Instead, depend on an abstraction (like a Repository interface), and restrict knowledge of EF (or any other specific data access technology) to the implementation of this interface. This will reduce the coupling between your application and a particular data access strategy, and will make testing your application code much easier.

Replacing the default services container ¶

The built-in services container is mean to serve the basic needs of the framework and most consumer applications built on it. However, developers who wish to replace the built-in container with their preferred container can easily do so. The ConfigureServices method typically returns void, but if its signature is changed to return IServiceProvider, a different container can be configured and returned. There are many IOC containers available for .NET. In this example, the Autofac package is used.

First, add the appropriate container package(s) to the dependencies property in project.json:

"dependencies" : {
  "Autofac": "4.0.0-rc2-237",
  "Autofac.Extensions.DependencyInjection": "4.0.0-rc2-200"
},

Next, configure the container in ConfigureServices and return an IServiceProvider:

public IServiceProvider ConfigureServices(IServiceCollection services)
{
  services.AddMvc();
  // add other framework services

  // Add Autofac
  var containerBuilder = new ContainerBuilder();
  containerBuilder.RegisterModule<DefaultModule>();
  containerBuilder.Populate(services);
  var container = containerBuilder.Build();
  return container.Resolve<IServiceProvider>();
}

Note

When using a third-party DI container, you must change ConfigureServices so that it returns IServiceProvider instead of void.

Finally, configure Autofac as normal in DefaultModule:

public class DefaultModule : Module
{
  protected override void Load(ContainerBuilder builder)
  {
    builder.RegisterType<CharacterRepository>().As<ICharacterRepository>();
  }
}

At runtime, Autofac will be used to resolve types and inject dependencies. Learn more about using Autofac and ASP.NET Core

Recommendations ¶

When working with dependency injection, keep the following recommendations in mind:

DI is for objects that have complex dependencies. Controllers, services, adapters, and repositories are all examples of objects that might be added to DI.
Avoid storing data and configuration directly in DI. For example, a user’s shopping cart shouldn’t typically be added to the services container. Configuration should use the Options Model. Similarly, avoid “data holder” objects that only exist to allow access to some other object. It’s better to request the actual item needed via DI, if possible.
Avoid static access to services.
Avoid service location in your application code.
Avoid static access to HttpContext.

Note

Like all sets of recommendations, you may encounter situations where ignoring one is required. We have found exceptions to be rare – mostly very special cases within the framework itself.

Remember, dependency injection is an alternative to static/global object access patterns. You will not be able to realize the benefits of DI if you mix it with static object access.

Additional Resources ¶

Working with Multiple Environments 다중 환경에서 작업하기¶

ASP.NET Core introduces improved support for controlling application behavior across multiple environments, such as development, staging, and production. Environment variables are used to indicate which environment the application is running in, allowing the app to be configured appropriately.

ASP.NET Core는 개발, 스테이징, 운영서버등이 존재하는 다중 환경에서 애플리케이션의 행동을 환경에 맞게 조정하기 위해 개선된 방안을 소개하고 있다. 환경 변수(environment variables)에 정보를 기록한 후, 앱이 환경 변수를 참조하여 스스로 환경에 맞는 앱을 구성할 수 있도록 한 것이 그 개념이다.

Sections:

Development, Staging, Production
Determining the environment at runtime 실행 시점에 환경 결정하기
Startup conventions Startup 컨벤션
Summary 정리
Additional Resources 추가 자료

예제 다운로드

Development, Staging, Production ¶

(이 괄호 블럭은 읽고 지워주세요. 제목을 한글로 개발, 스테이징, 운영 이라고 하고 싶었으나 의미보다는 환경변수에 설정하는 값으로 문맥에서 사용되기도 하기에 영문을 유지하는게 낫다고 판단했습니다)

ASP.NET Core references a particular environment variable, ASPNETCORE_ENVIRONMENT to describe the environment the application is currently running in. This variable can be set to any value you like, but three values are used by convention: Development, Staging, and Production. You will find these values used in the samples and templates provided with ASP.NET Core.

ASP.NET Core는 특정 환경 변수, ASPNETCORE_ENVIRONMENT 를 참조하여 애플리케이션이 현재 운영되고 있는 환경을 알아낸다. 환경 변수를 통해 운영환경을 애플리케이션에게 설명해 준다고 할 수 있다. 이 변수는 여러분이 원하는 값을 설정할 수 있지만 컨벤션에 의해 Development, Staging, Production 라는 세 가지 값을 사용한다. 이 값들은 ASP.NET Core와 함께 제공되는 예제와 템플릿 등에 사용되므로 쉽게 찾아볼 수 있다.

The current environment setting can be detected programmatically from within your application. In addition, you can use the Environment tag helper to include certain sections in your view based on the current application environment.

애플리케이션에서는 프로그램적으로 현재 환경 설정을 탐지할 수 있다. 게다가 Envrionment 태그 헬퍼 를 사용하면 현재 애플리케이션 환경에만 사용할 컨텐츠를 뷰 에 포함시킬 수도 있다.

Note

The specified environment name is case insensitive. Whether you set the variable to Development or development or DEVELOPMENT the results will be the same.

Note

환경 변수의 값으로 사용되는 환경 이름은 대소문자를 가리지 않는다. 변수를 어떻게 설정하든 - Development or development or DEVELOPMENT - 결과는 같다.

Development¶

(참고하고 지워주세요. 제목을 개발환경이라고 하고 싶었으나 Development가 문맥에서 환경 변수의 값이므로 영문을 유지했습니다. 이전과 동일)

This should be the environment used when developing an application. When using Visual Studio, this setting can be specified in your project’s debug profiles, such as for IIS Express, shown here:

Development는 애플리케이션을 개발할 때 사용되는 환경이 되어야 한다. 비주얼 스튜디오를 사용할 때, 이 설정은 프로젝트의 디버그 프로파일에 명시된다. 아래 그림과 같이 이 설정은 개발 머신의 IIS Express를 위한 것이다.

When you modify the default settings created with the project, your changes are persisted in launchSettings.json in the Properties folder. This file holds settings specific to each profile Visual Studio is configured to use to launch the application, including any environment variables that should be used. (Debug profiles are discussed in more detail in Servers). For example, after adding another profile configured to use IIS Express, but using an ASPNETCORE_ENVIRONMENT value of Staging, the launchSettings.json file in our sample project is shown below:

프로젝트의 기본 설정을 수정하면 그 내용은 Properties 폴더의 launchSettings.json 파일에 저장된다. 이 파일은 비주얼 스튜디오가 애플리케이션을 구동하도록 설정되는 것과 관련하여, 어떤 환경 변수가 사용되어야 하는지를 포함한 모든 프로파일 설정을 갖고 있다. (디버그 프로파일은 Servers 라는 문서에서 보다 자세한 내용을 다룬다). 예를 들어, IIS Express 가 다른 프로파일을 사용하도록 프로파일을 추가하고자 하는데, Staging 이라는 값을 ASPNETCORE_ENVIRONMENT 환경 변수로 사용하고 싶다면 아래의 예제 파일 launchSettings.json 과 같이 설정한다.

launchSettings.json¶

{
  "iisSettings": {
    "windowsAuthentication": false,
    "anonymousAuthentication": true,
    "iisExpress": {
      "applicationUrl": "http://localhost:40088/",
      "sslPort": 0
    }
  },
  "profiles": {
    "IIS Express": {
      "commandName": "IISExpress",
      "launchBrowser": true,
      "environmentVariables": {
        "ASPNETCORE_ENVIRONMENT": "Development"
      }
    },
    "IIS Express (Staging)": {
      "commandName": "IISExpress",
      "launchBrowser": true,
      "environmentVariables": {
        "ASPNETCORE_ENVIRONMENT": "Staging"
      }
    }
  }
}

Note

Changes made to project profiles or to launchSettings.json directly may not take effect until the web server used is restarted (in particular, Kestrel must be restarted before it will detect changes made to its environment).

Note

프로젝트 프로파일 또는 launchSettings.json 파일에서 변경된 내용은 웹 서버가 재시작되기 전까지 반영되지 않을 수 있다. (특히 Kestrel 서버는 반드시 재시작되어야만 환경의 변화를 감지할 수 있다).

You can create multiple different launch profiles for various different configurations of your application, including those that require other environment variables.

필요에 따라 여러 개의 구동 프로파일(launch profile)을 만들어 애플리케이션에서 다양한 구성을 할 수 있다. 다른 환경 변수들을 필요로 하는 프로파일을 작성하는 것도 포함된다.

Warning

Environment variables stored in launchSettings.json are not secured in any way and will be part of the source code repository for your project, if you use one. Never store credentials or other secret data in this file. If you need a place to store such data, use the Secret Manager tool described in Safe storage of app secrets during development.

Warning

launchSettings.json 파일에 저장되는 환경 변수들은 사실 안전하지 않다. 그리고, 프로젝트 소스 코드 저장소의 일부가 될 것이다. 신원정보 또는 보안정보를 절대로 이 파일에 저장해서는 안된다. 그런 데이터를 저장하려는 장소가 필요하다면 Safe storage of app secrets during development 문서에 설명된 Secret Manager 도구를 사용하도록 한다.

Staging¶

By convention, a Staging environment is a pre-production environment used for final testing before deployment to production. Ideally, its physical characteristics should mirror that of production, so that any issues that may arise in production occur first in the staging environment, where they can be addressed without impact to users.

컨벤션에 따르면 Staging 환경은 운영에 가기 바로전 최종 테스트를 진행하는 사전 운영환경이다. 이 환경은 운영 환경의 물리적 특성을 동일하게 갖추는 것이 이상적이다. 그래서, 운영 환경에서 혹여나 발생할 만한 이슈를 Staging 환경에서 미리 포착하여 사용자에게 어떤 영향도 미치지 않고 해결할 수 있다.

Production¶

The Production environment is the environment in which the application runs when it is live and being used by end users. This environment should be configured to maximize security, performance, and application robustness. Some common settings that a production environment might have that would differ from development include:

Turn on caching
Ensure all client-side resources are bundled, minified, and potentially served from a CDN
Turn off diagnostic ErrorPages
Turn on friendly error pages
Enable production logging and monitoring (for example, Application Insights)

Production 환경은 애플리케이션이 라이브 상태가 된 후 최종 사용자에 의해 사용될 때, 애플리케이션이 동작하는 환경이다. 이 환경은 보안, 성능, 견고함을 최대로 구성해야 한다. 개발 환경과 달리 운영 환경이 가져야 할 일반적인 설정은 다음과 같다.

캐싱 사용하기
모든 클라이언트 단 리소스를 같이 묶고(bundle), 최소화(minify)한다. CDN을 통해 클라이언트 단 파일을 서비스하는 것도 잠재적인 옵션이다
오류를 진단하기 위해 사용하는 오류 페이지를 사용하지 않는다
사용자 친화적인 오류 페이지를 사용한다
로깅, 모니터링을 활성화한다. (예, Application Insights)

This is by no means meant to be a complete list. It’s best to avoid scattering environment checks in many parts of your application. Instead, the recommended approach is to perform such checks within the application’s Startup class(es) wherever possible

이 리스트는 단지 예제일뿐 운영을 위한 완벽한 리스트가 아니다. 애플리케이션에서 환경 관련한 확인이 필요하다면 여기 저기에서 비일관적으로 하는 것을 피하고 Startup 클래스와 같이 한 곳의 제한된 곳에서 확인하도록 하자.

Determining the environment at runtime 실행 시점에 환경 결정하기 ¶

The IHostingEnvironment service provides the core abstraction for working with environments. This service is provided by the ASP.NET hosting layer, and can be injected into your startup logic via Dependency Injection. The ASP.NET Core web site template in Visual Studio uses this approach to load environment-specific configuration files (if present) and to customize the app’s error handling settings. In both cases, this behavior is achieved by referring to the currently specified environment by calling EnvironmentName or IsEnvironment on the instance of IHostingEnvironment passed into the appropriate method.

IHostingEnvironment 서비스는 운영 환경과 관련된 핵심적인 추상화를 제공한다. 이 서비스는 ASP.NET 호스팅 단에서 제공되는데 Dependency Injection 을 사용해 startup 로직에서 주입될 수도 있다. 비주얼 스튜디오의 ASP.NET Core 웹 사이트 템플릿이 이 접근법을 사용하는데, 특정 환경에 대한 구성 파일이 있다면 이 파일을 로드하고 오류 처리를 환경에 따라 다르게 하도록 구성한다. IHostingEnvironment 가 주입되던 그렇지 않던 환경에 따라 이런 식으로 처리할 수 있는데 IHostingEnvironment 인스턴스의 EnvironmentName 속성 또는 IsEnvironment 메서드를 사용해서 가능하다.

Note

If you need to check whether the application is running in a particular environment, use env.IsEnvironment("environmentname") since it will correctly ignore case (instead of checking if env.EnvironmentName == "Development" for example).

Note

애플리케이션이 특정 환경에서 동작하는지 확인할 필요가 있다면 env.IsEnvironment("environmentname") 을 사용하자. 이 문장은 env.EnvironmentName == "Development" 문장과 달리 대소문자를 구분하지 않는다.

For example, you can use the following code in your Configure method to setup environment specific error handling:

예를 들어, 특정 환경에서의 오류 처리를 설정하려면 다음의 코드를 Configure 메서드에서 사용할 수 있다.

public void Configure(IApplicationBuilder app, IHostingEnvironment env)
{
    if (env.IsDevelopment())
    {
        app.UseDeveloperExceptionPage();
        app.UseDatabaseErrorPage();
        app.UseBrowserLink();
    }
    else
    {
        app.UseExceptionHandler("/Home/Error");
    }
    // ...

If the app is running in a Development environment, then it enables the runtime support necessary to use the “BrowserLink” feature in Visual Studio, development-specific error pages (which typically should not be run in production) and special database error pages (which provide a way to apply migrations and should therefore only be used in development). Otherwise, if the app is not running in a development environment, a standard error handling page is configured to be displayed in response to any unhandled exceptions.

Development 환경에서 앱을 실행한다면, 비주얼 스튜디오에서 “BrowserLink” 기능을 사용하기 위해 런타임 지원 필요(runtime support necessary)가 활성화되는데, 이 기능은 개발용 오류 화면과 특별한 데이터베이스 오류 화면도 같이 지원한다. 개발용 오류 화면은 통상적으로 운영 환경에서 사용해서는 안되고, 데이터베이스 특정 오류 화면은 마이그레이션을 적용하는 방법을 제공하기 때문에 개발 환경에서만 사용되어야 한다. 반면, 앱이 개발 환경에서 실행되지 않을 경우에는, 표준 오류 화면을 통해 처리되지 않은 예외를 보여준다.

You may need to determine which content to send to the client at runtime, depending on the current environment. For example, in a development environment you generally serve non-minimized scripts and style sheets, which makes debugging easier. Production and test environments should serve the minified versions and generally from a CDN. You can do this using the Environment tag helper. The Environment tag helper will only render its contents if the current environment matches one of the environments specified using the names attribute.

실행환경에 따라 어떤 컨텐트를 클라이언트에 보내야 할 지 결정해야 할 상황이 있을 수 있다. 예를 들면, 개발 환경에서는 디버깅 편의를 위해서 압축되지 않은 스크립트와 스타일 시트를 사용지만, 운영 환경과 테스트 환경에서는 응답 성능 향상을 위해 압축된 버전을 사용한다. 이 최소화된 버전은 일반적으로 CDN을 통해 받는다. 태그 헬퍼 를 사용하여 이런 작업을 할 수 있는데, Environment 태그 헬퍼를 사용하면 현재 환경이 태그의 name 속성에 기재된 환경과 일치할 때만 해당 컨텐츠를 출력한다.

<environment names="Development">
    <link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />
    <link rel="stylesheet" href="~/css/site.css" />
</environment>
<environment names="Staging,Production">
    <link rel="stylesheet" href="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.6/css/bootstrap.min.css"
          asp-fallback-href="~/lib/bootstrap/dist/css/bootstrap.min.css"
          asp-fallback-test-class="sr-only" asp-fallback-test-property="position" asp-fallback-test-value="absolute" />
    <link rel="stylesheet" href="~/css/site.min.css" asp-append-version="true" />
</environment>

To get started with using tag helpers in your application see Introduction to Tag Helpers.

여러분의 애플리케이션에서 태그 헬퍼 사용을 고려한다면 Introduction to Tag Helpers 문서를 참조하자.

Startup conventions Startup 컨벤션 ¶

ASP.NET Core supports a convention-based approach to configuring an application’s startup based on the current environment. You can also programmatically control how your application behaves according to which environment it is in, allowing you to create and manage your own conventions.

ASP.NET Core는 현재 환경에 따라 애플리케이션 구동방식을 구성할 수 있는 컨벤션 기반의 접근법을 지원한다. 여러분은 또한 프로그램적으로 애플리케이션의 행동 방식을 환경에 따라 제어할 수 있기 때문에 자신만의 컨벤션을 만들고 관리할 수도 있다.

When an ASP.NET Core application starts, the Startup class is used to bootstrap the application, load its configuration settings, etc. (learn more about ASP.NET startup). However, if a class exists named Startup{EnvironmentName} (for example StartupDevelopment), and the ASPNETCORE_ENVIRONMENT environment variable matches that name, then that Startup class is used instead. Thus, you could configure Startup for development, but have a separate StartupProduction that would be used when the app is run in production. Or vice versa.

ASP.NET Core 애플리케이션이 시작될 때, Startup 클래스가 애플리케이션을 구동하고, 구성 설정등을 불러오는 등의 일을 한다 (learn more about ASP.NET startup). 그러나, Startup{EnvironmentName} 처럼, 환경 이름을 포함한 Startup 클래스를 정의하고 (예, StartupDevelopment), ASPNETCORE_ENVIRONMENT 변수 값이 환경 이름과 일치한다면, 환경 특화된 클래스가 사용될 것이다. 따라서, 개발환경에 사용될 Startup 만을 구성할 수도 있겠지만 운영환경을 고려하여 StartupProduction 클래스를 별도로 가져갈 수도 있다. 혹은 그 반대로, Startup 을 운영환경에 사용하고 개발환경을 위한 StartupDevelopment 를 별도로 정의할 수도 있다.

In addition to using an entirely separate Startup class based on the current environment, you can also make adjustments to how the application is configured within a Startup class. The Configure() and ConfigureServices() methods support environment-specific versions similar to the Startup class itself, of the form Configure{EnvironmentName}() and Configure{EnvironmentName}Services(). If you define a method ConfigureDevelopment() it will be called instead of Configure() when the environment is set to development. Likewise, ConfigureDevelopmentServices() would be called instead of ConfigureServices() in the same environment.

환경에 기초해 완벽하게 분리된 Startup 클래스를 운영하는 것과 더불어, Startup 클래스 내부에서 애플리케이션을 어떻게 구성할지 조정하는 것도 가능하다. Startup 클래스 자체가 환경에 특화된 별도 클래스를 가질 수 있는 것처럼 Configure() 와 ConfigureService() 메서드도 Configure{EnvironmentName}() 와 Configure{EnvironmentName}Services() 의 형태로 환경 특화된 메서드를 지원한다. 환경이 development로 지정되었을 때, ConfigureDevelopment() 메서드를 정의했다면 이 것이 Configure() 메서드 대신 호출될 것이다. 같은 환경이라면 앞서 방식과 동일하게 ConfigureServices() 메서드 대신 ConfigureDevelopmentServices() 메서드가 호출될 것이다.

Summary 정리 ¶

ASP.NET Core provides a number of features and conventions that allow developers to easily control how their applications behave in different environments. When publishing an application from development to staging to production, environment variables set appropriately for the environment allow for optimization of the application for debugging, testing, or production use, as appropriate.

ASP.NET Core 는 개발자들이 애플리케이션 배포 환경에 맞추어 관리하려고 할 때 필요한 많은 기능과 컨벤션을 제공한다. 개발환경에서부터 애플리케이션을 스테이징, 운영으로 배포할 때, 환경에 맞추어 설정된 환경 변수들은 애플리케이션을 디버깅, 테스팅 또는 운영 모드로 최적화시키는 역할을 한다.

Additional Resources 추가 자료 ¶

Managing Application State¶

In ASP.NET Core, application state can be managed in a variety of ways, depending on when and how the state is to be retrieved. This article provides a brief overview of several options, and focuses on installing and configuring Session state support in ASP.NET Core applications.

Sections

Application State Options
Working with HttpContext.Items
Installing and Configuring Session
A Working Sample Using Session

Application State Options ¶

Application state refers to any data that is used to represent the current representation of the application. This includes both global and user-specific data. Previous versions of ASP.NET (and even ASP) have had built-in support for global Application and Session state stores, as well as a variety of other options.

Note

The Application store had the same characteristics as the ASP.NET Cache, with fewer capabilities. In ASP.NET Core, Application no longer exists; applications written for previous versions of ASP.NET that are migrating to ASP.NET Core replace Application with a Caching implementation.

Application developers are free to use different state storage providers depending on a variety of factors:

How long does the data need to persist?
How large is the data?
What format is the data?
Can it be serialized?
How sensitive was the data? Could it be stored on the client?

Based on answers to these questions, application state in ASP.NET Core apps can be stored or managed in a variety of ways.

HttpContext.Items¶

The Items collection is the best location to store data that is only needed while processing a given request. Its contents are discarded after each request. It is best used as a means of communicating between components or middleware that operate at different points in time during a request, and have no direct relationship with one another through which to pass parameters or return values. See Working with HttpContext.Items, below.

Querystring and Post¶

State from one request can be provided to another request by adding values to the new request’s querystring or by POSTing the data. These techniques should not be used with sensitive data, because these techniques require that the data be sent to the client and then sent back to the server. It is also best used with small amounts of data. Querystrings are especially useful for capturing state in a persistent manner, allowing links with embedded state to be created and sent via email or social networks, for use potentially far into the future. However, no assumption can be made about the user making the request, since URLs with querystrings can easily be shared, and care must also be taken to avoid Cross-Site Request Forgery (CSRF) attacks (for instance, even assuming only authenticated users are able to perform actions using querystring-based URLs, an attacker could trick a user into visiting such a URL while already authenticated).

Cookies¶

Very small pieces of state-related data can be stored in Cookies. These are sent with every request, and so the size should be kept to a minimum. Ideally, only an identifier should be used, with the actual data stored somewhere on the server, keyed to the identifier.

Session¶

Session storage relies on a cookie-based identifier to access data related to a given browser session (a series of requests from a particular browser and machine). You can’t necessarily assume that a session is restricted to a single user, so be careful what kind of information you store in Session. It is a good place to store application state that is specific to a particular session but which doesn’t need to be persisted permanently (or which can be reproduced as needed from a persistent store). See Installing and Configuring Session, below for more details.

Cache¶

Caching provides a means of storing and efficiently retrieving arbitrary application data based on developer-defined keys. It provides rules for expiring cached items based on time and other considerations. Learn more about Caching.

Configuration¶

Configuration can be thought of as another form of application state storage, though typically it is read-only while the application is running. Learn more about Configuration.

Other Persistence¶

Any other form of persistent storage, whether using Entity Framework and a database or something like Azure Table Storage, can also be used to store application state, but these fall outside of what ASP.NET supports directly.

Working with HttpContext.Items ¶

The HttpContext abstraction provides support for a simple dictionary collection of type IDictionary<object, object>, called Items. This collection is available from the start of an HttpRequest` and is discarded at the end of each request. You can access it by simply assigning a value to a keyed entry, or by requesting the value for a given key.

For example, some simple Middleware could add something to the Items collection:

app.Use(async (context, next) =>
{
    // perform some verification
    context.Items["isVerified"] = true;
    await next.Invoke();
});

and later in the pipeline, another piece of middleware could access it:

app.Run(async (context) =>
{
    await context.Response.WriteAsync("Verified request? " + context.Items["isVerified"]);
});

Note

Since keys into Items are simple strings, if you are developing middleware that needs to work across many applications, you may wish to prefix your keys with a unique identifier to avoid key collisions (e.g. “MyComponent.isVerified” instead of just “isVerified”).

Installing and Configuring Session ¶

ASP.NET Core ships a session package that provides middleware for managing session state. You can install it by including a reference to the Microsoft.AspNetCore.Session package in your project.json file.

Once the package is installed, Session must be configured in your application’s Startup class. Session is built on top of IDistributedCache, so you must configure this as well, otherwise you will receive an error.

Note

If you do not configure at least one IDistributedCache implementation, you will get an exception stating “Unable to resolve service for type ‘Microsoft.Extensions.Caching.Distributed.IDistributedCache’ while attempting to activate ‘Microsoft.AspNetCore.Session.DistributedSessionStore’.”

ASP.NET ships with several implementations of IDistributedCache, including an in-memory option (to be used during development and testing only). To configure session using this in-memory option add the Microsoft.Extensions.Caching.Memory package in your project.json file and then add the following to ConfigureServices:

services.AddDistributedMemoryCache();
services.AddSession();

Then, add the following to Configure and you’re ready to use session in your application code:

app.UseSession();

You can reference Session from HttpContext once it is installed and configured.

Note

If you attempt to access Session before UseSession has been called, you will get an InvalidOperationException exception stating that “Session has not been configured for this application or request.”

Warning

If you attempt to create a new Session (i.e. no session cookie has been created yet) after you have already begun writing to the Response stream, you will get an InvalidOperationException as well, stating that “The session cannot be established after the response has started”. This exception may not be displayed in the browser; you may need to view the web server log to discover it, as shown below:

_images/session-after-response-error.png

Implementation Details¶

Session uses a cookie to track and disambiguate between requests from different browsers. By default this cookie is named ”.AspNet.Session” and uses a path of “/”. Further, by default this cookie does not specify a domain, and is not made available to client-side script on the page (because CookieHttpOnly defaults to true).

These defaults, as well as the default IdleTimeout (used on the server independent from the cookie), can be overridden when configuring Session by using SessionOptions as shown here:

services.AddSession(options =>
{
  options.CookieName = ".AdventureWorks.Session";
  options.IdleTimeout = TimeSpan.FromSeconds(10);
});

The IdleTimeout is used by the server to determine how long a session can be idle before its contents are abandoned. Each request made to the site that passes through the Session middleware (regardless of whether Session is read from or written to within that middleware) will reset the timeout. Note that this is independent of the cookie’s expiration.

Note

Session is non-locking, so if two requests both attempt to modify the contents of session, the last one will win. Further, Session is implemented as a coherent session, which means that all of the contents are stored together. This means that if two requests are modifying different parts of the session (different keys), they may still impact each other.

ISession¶

Once session is installed and configured, you refer to it via HttpContext, which exposes a property called Session of type ISession. You can use this interface to get and set values in Session, such as byte[].

public interface ISession
{
    bool IsAvailable { get; }
    string Id { get; }
    IEnumerable<string> Keys { get; }
    Task LoadAsync();
    Task CommitAsync();
    bool TryGetValue(string key, out byte[] value);
    void Set(string key, byte[] value);
    void Remove(string key);
    void Clear();
}

Because Session is built on top of IDistributedCache, you must always serialize the object instances being stored. Thus, the interface works with byte[] not simply object. However, there are extension methods that make working with simple types such as String and Int32 easier, as well as making it easier to get a byte[] value from session.

// session extension usage examples
context.Session.SetInt32("key1", 123);
int? val = context.Session.GetInt32("key1");
context.Session.SetString("key2", "value");
string stringVal = context.Session.GetString("key2");
byte[] result = context.Session.Get("key3");

If you’re storing more complex objects, you will need to serialize the object to a byte[] in order to store them, and then deserialize them from byte[] when retrieving them.

A Working Sample Using Session ¶

The associated sample application demonstrates how to work with Session, including storing and retrieving simple types as well as custom objects. In order to see what happens when session expires, the sample has configured sessions to last just 10 seconds:

{
    services.AddDistributedMemoryCache();

    services.AddSession(options =>
    {
        options.IdleTimeout = TimeSpan.FromSeconds(10);
    });
}

When you first navigate to the web server, it displays a screen indicating that no session has yet been established:

This default behavior is produced by the following middleware in Startup.cs, which runs when requests are made that do not already have an established session (note the highlighted sections):

// main catchall middleware
app.Run(async context =>
{
    RequestEntryCollection collection = GetOrCreateEntries(context);

    if (collection.TotalCount() == 0)
    {
        await context.Response.WriteAsync("<html><body>");
        await context.Response.WriteAsync("Your session has not been established.<br>");
        await context.Response.WriteAsync(DateTime.Now.ToString() + "<br>");
        await context.Response.WriteAsync("<a href=\"/session\">Establish session</a>.<br>");
    }
    else
    {
        collection.RecordRequest(context.Request.PathBase + context.Request.Path);
        SaveEntries(context, collection);

        // Note: it's best to consistently perform all session access before writing anything to response
        await context.Response.WriteAsync("<html><body>");
        await context.Response.WriteAsync("Session Established At: " + context.Session.GetString("StartTime") + "<br>");
        foreach (var entry in collection.Entries)
        {
            await context.Response.WriteAsync("Request: " + entry.Path + " was requested " + entry.Count + " times.<br />");
        }

        await context.Response.WriteAsync("Your session was located, you've visited the site this many times: " + collection.TotalCount() + "<br />");
    }
    await context.Response.WriteAsync("<a href=\"/untracked\">Visit untracked part of application</a>.<br>");
    await context.Response.WriteAsync("</body></html>");
});

GetOrCreateEntries is a helper method that will retrieve a RequestEntryCollection instance from Session if it exists; otherwise, it creates the empty collection and returns that. The collection holds RequestEntry instances, which keep track of the different requests the user has made during the current session, and how many requests they’ve made for each path.

public class RequestEntry
{
    public string Path { get; set; }
    public int Count { get; set; }
}

public class RequestEntryCollection
{
    public List<RequestEntry> Entries { get; set; } = new List<RequestEntry>();

    public void RecordRequest(string requestPath)
    {
        var existingEntry = Entries.FirstOrDefault(e => e.Path == requestPath);
        if (existingEntry != null) { existingEntry.Count++; return; }

        var newEntry = new RequestEntry()
        {
            Path = requestPath,
            Count = 1
        };
        Entries.Add(newEntry);
    }

    public int TotalCount()
    {
        return Entries.Sum(e => e.Count);
    }
}

Fetching the current instance of RequestEntryCollection is done via the GetOrCreateEntries helper method:

private RequestEntryCollection GetOrCreateEntries(HttpContext context)
{
    RequestEntryCollection collection = null;
    byte[] requestEntriesBytes = context.Session.Get("RequestEntries");

    if (requestEntriesBytes != null && requestEntriesBytes.Length > 0)
    {
        string json = System.Text.Encoding.UTF8.GetString(requestEntriesBytes);
        return JsonConvert.DeserializeObject<RequestEntryCollection>(json);
    }
    if (collection == null)
    {
        collection = new RequestEntryCollection();
    }
    return collection;
}

When the entry for the object exists in Session, it is retrieved as a byte[] type, and then deserialized using a MemoryStream and a BinaryFormatter, as shown above. If the object isn’t in Session, the method returns a new instance of the RequestEntryCollection.

In the browser, clicking the Establish session hyperlink makes a request to the path “/session”, and returns this result:

Refreshing the page results in the count incrementing; returning to the root of the site (after making a few more requests) results in this display, summarizing all of the requests that were made during the current session:

_images/session-established-with-request-counts.png

Establishing the session is done in the middleware that handles requests to “/session”:

// establish session
app.Map("/session", subApp =>
{
    subApp.Run(async context =>
    {
        // uncomment the following line and delete session coookie to generate an error due to session access after response has begun
        // await context.Response.WriteAsync("some content");
        RequestEntryCollection collection = GetOrCreateEntries(context);
        collection.RecordRequest(context.Request.PathBase + context.Request.Path);
        SaveEntries(context, collection);
        if (context.Session.GetString("StartTime") == null)
        {
            context.Session.SetString("StartTime", DateTime.Now.ToString());
        }

        await context.Response.WriteAsync("<html><body>");
        await context.Response.WriteAsync($"Counting: You have made {collection.TotalCount()} requests to this application.<br><a href=\"/\">Return</a>");
        await context.Response.WriteAsync("</body></html>");
    });
});

Requests to this path will get or create a RequestEntryCollection, will add the current path to it, and then will store it in session using the helper method SaveEntries, shown below:

private void SaveEntries(HttpContext context, RequestEntryCollection collection)
{
    string json = JsonConvert.SerializeObject(collection);
    byte[] serializedResult = System.Text.Encoding.UTF8.GetBytes(json);

    context.Session.Set("RequestEntries", serializedResult);
}

SaveEntries demonstrates how to serialize a custom object into a byte[] for storage in Session using a MemoryStream and a BinaryFormatter.

The sample includes one more piece of middleware worth mentioning, which is mapped to the “/untracked” path. You can see its configuration here:

// example middleware that does not reference session at all and is configured before app.UseSession()
app.Map("/untracked", subApp =>
{
    subApp.Run(async context =>
    {
        await context.Response.WriteAsync("<html><body>");
        await context.Response.WriteAsync("Requested at: " + DateTime.Now.ToString("hh:mm:ss.ffff") + "<br>");
        await context.Response.WriteAsync("This part of the application isn't referencing Session...<br><a href=\"/\">Return</a>");
        await context.Response.WriteAsync("</body></html>");
    });
});

app.UseSession();

Note that this middleware is configured before the call to app.UseSession() is made (on line 13). Thus, the Session feature is not available to this middleware, and requests made to it do not reset the session IdleTimeout. You can confirm this behavior in the sample application by refreshing the untracked path several times within 10 seconds, and then return to the application root. You will find that your session has expired, despite no more than 10 seconds having passed between your requests to the application.

Servers¶

ASP.NET Core is completely decoupled from the web server environment that hosts the application. ASP.NET Core supports hosting in IIS and IIS Express, and self-hosting scenarios using the Kestrel and WebListener HTTP servers. Additionally, developers and third party software vendors can create custom servers to host their ASP.NET Core apps.

Sections:

Servers and commands
Supported Features by Server
IIS and IIS Express
WebListener
Kestrel
Choosing a server
Custom Servers
Additional Reading

Servers and commands ¶

ASP.NET Core was designed to decouple web applications from the underlying HTTP server. Traditionally, ASP.NET apps have been windows-only hosted on Internet Information Server (IIS). The recommended way to run ASP.NET Core applications on Windows is using IIS as a reverse-proxy server. The HttpPlatformHandler module in IIS manages and proxies requests to an HTTP server hosted out-of-process. ASP.NET Core ships with two different HTTP servers:

Microsoft.AspNetCore.Server.Kestrel (AKA Kestrel, cross-platform)
Microsoft.AspNetCore.Server.WebListener (AKA WebListener, Windows-only, preview)

ASP.NET Core does not directly listen for requests, but instead relies on the HTTP server implementation to surface the request to the application as a set of feature interfaces composed into an HttpContext. While WebListener is Windows-only, Kestrel is designed to run cross-platform. You can configure your application to be hosted by any or all of these servers by specifying commands in your project.json file. You can even specify an application entry point for your application, and run it as an executable (using dotnet run) rather than hosting it in a separate process.

The default web host for ASP.NET apps developed using Visual Studio is IIS Express functioning as a reverse proxy server for Kestrel. The “Microsoft.AspNetCore.Server.Kestrel” and “Microsoft.AspNetCore.Server.IISIntegration” dependencies are included in project.json by default, even with the Empty web site template. Visual Studio provides support for multiple profiles, associated with IIS Express. You can manage these profiles and their settings in the Debug tab of your web application project’s Properties menu or from the launchSettings.json file.

The sample project for this article is configured to support each server option in the project.json file:

project.json (truncated)¶

{
  "webroot": "wwwroot",
  "version": "1.0.0-*",

  "dependencies": {
    "Microsoft.AspNet.Server.Kestrel": "1.0.0-rc1-final",
    "Microsoft.AspNet.Server.WebListener": "1.0.0-rc1-final"
  },

  "commands": {
    "run": "run server.urls=http://localhost:5003",
    "web": "Microsoft.AspNet.Hosting --server Microsoft.AspNet.Server.Kestrel --server.urls http://localhost:5000",
    "weblistener": "Microsoft.AspNet.Hosting --server WebListener --server.urls http://localhost:5004"
  },

  "frameworks": {
    "dnx451": { },

The run command will launch the application from the void main method. The run command configures and starts an instance of Kestrel.

program.cs¶

using System;
using System.Threading.Tasks;
using Microsoft.AspNet.Hosting;
using Microsoft.Extensions.Configuration;
using Microsoft.AspNet.Builder;
using Microsoft.Extensions.Logging;
using Microsoft.AspNet.Server.Kestrel;

namespace ServersDemo
{
    /// <summary>
    /// This demonstrates how the application can be launched in a console application. 
    /// Executing the "dnx run" command in the application folder will run this app.
    /// </summary>
    public class Program
    {
        private readonly IServiceProvider _serviceProvider;

        public Program(IServiceProvider serviceProvider)
        {
            _serviceProvider = serviceProvider;
        }

        public Task<int> Main(string[] args)
        {
            //Add command line configuration source to read command line parameters.
            var builder = new ConfigurationBuilder();
            builder.AddCommandLine(args);
            var config = builder.Build();

            using (new WebHostBuilder(config)
                .UseServer("Microsoft.AspNet.Server.Kestrel")
                .Build()
                .Start())
            {
                Console.WriteLine("Started the server..");
                Console.WriteLine("Press any key to stop the server");
                Console.ReadLine();
            }
            return Task.FromResult(0);
        }
    }
}

Supported Features by Server ¶

ASP.NET defines a number of Request Features. The following table lists the WebListener and Kestrel support for request features.

Feature	WebListener	Kestrel
IHttpRequestFeature	Yes	Yes
IHttpResponseFeature	Yes	Yes
IHttpAuthenticationFeature	Yes	No
IHttpUpgradeFeature	Yes (with limits)	Yes
IHttpBufferingFeature	Yes	No
IHttpConnectionFeature	Yes	Yes
IHttpRequestLifetimeFeature	Yes	Yes
IHttpSendFileFeature	Yes	No
IHttpWebSocketFeature	No*	No*
IRequestIdentifierFeature	Yes	No
ITlsConnectionFeature	Yes	Yes
ITlsTokenBindingFeature	Yes	No

Configuration options¶

You can provide configuration options (by command line parameters or a configuration file) that are read on server startup.

The Microsoft.AspNetCore.Hosting command supports server parameters (such as Kestrel or WebListener) and a server.urls configuration key. The server.urls configuration key is a semicolon-separated list of URL prefixes that the server should handle.

The project.json file shown above demonstrates how to pass the server.urls parameter directly:

"web": "Microsoft.AspNetCore.Kestrel --server.urls http://localhost:5004"

Alternately, a JSON configuration file can be used,

"kestrel": "Microsoft.AspNetCore.Hosting"

The hosting.json can include the settings the server will use (including the server parameter, as well):

{
  "server": "Microsoft.AspNetCore.Server.Kestrel",
  "server.urls": "http://localhost:5004/"
}

Programmatic configuration¶

The server hosting the application can be referenced programmatically via the IApplicationBuilder interface, available in the Configure method in Startup. IApplicationBuilder exposes Server Features of type IFeatureCollection. IServerAddressesFeature only expose a Addresses property, but different server implementations may expose additional functionality. For instance, WebListener exposes AuthenticationManager that can be used to configure the server’s authentication:

public void Configure(IApplicationBuilder app, IApplicationLifetime lifetime, ILoggerFactory loggerFactory)
{
    var webListenerInfo = app.ServerFeatures.Get<WebListener>();
    if (webListenerInfo != null)
    {
        webListenerInfo.AuthenticationManager.AuthenticationSchemes =
            AuthenticationSchemes.AllowAnonymous;
    }

    var serverAddress = app.ServerFeatures.Get<IServerAddressesFeature>()?.Addresses.FirstOrDefault();

    app.Run(async (context) =>
    {
        var message = String.Format("Hello World from {0}",
                                serverAddress);
        await context.Response.WriteAsync(message);
    });
}

IIS and IIS Express ¶

IIS is the most feature rich server, and includes IIS management functionality and access to other IIS modules. Hosting ASP.NET Core no longer uses the System.Web infrastructure used by prior versions of ASP.NET.

ASP.NET Core Module¶

In ASP.NET Core on Windows, the web application is hosted by an external process outside of IIS. The ASP.NET Core Module is a native IIS module that is used to proxy requests to external processes that it manages. See ASP.NET Core Module Configuration Reference for more details.

WebListener ¶

WebListener is a Windows-only HTTP server for ASP.NET Core. It runs directly on the Http.Sys kernel driver, and has very little overhead.

You can add support for WebListener to your ASP.NET application by adding the “Microsoft.AspNetCore.Server.WebListener” dependency in project.json and the following command:

"web": "Microsoft.AspNetCore.Hosting --server Microsoft.AspNetCore.Server.WebListener --server.urls http://localhost:5000"

Note

WebListener is currently still in preview.

Kestrel ¶

Kestrel is a cross-platform web server based on libuv, a cross-platform asynchronous I/O library. You add support for Kestrel by including Microsoft.AspNetCore.Server.Kestrel in your project’s dependencies listed in project.json.

Learn more about working with Kestrel to create Your First ASP.NET Core Application on a Mac Using Visual Studio Code.

Note

Kestrel is designed to be run behind a proxy (for example IIS or Nginx) and should not be deployed directly facing the Internet.

Choosing a server ¶

If you intend to deploy your application on a Windows server, you should run IIS as a reverse proxy server that manages and proxies requests to Kestrel. If deploying on Linux, you should run a comparable reverse proxy server such as Apache or Nginx to proxy requests to Kestrel (see Publish to a Linux Production Environment).

Custom Servers ¶

You can create your own server in which to host ASP.NET apps, or use other open source servers. When implementing your own server, you’re free to implement just the feature interfaces your application needs, though at a minimum you must support IHttpRequestFeature and IHttpResponseFeature.

Since Kestrel is open source, it makes an excellent starting point if you need to implement your own custom server. Like all of ASP.NET Core, you’re welcome to contribute any improvements you make back to the project.

Kestrel currently supports a limited number of feature interfaces, but additional features will be added in the future.

Additional Reading ¶

Request Features

Request Features¶

Individual web server features related to how HTTP requests and responses are handled have been factored into separate interfaces. These abstractions are used by individual server implementations and middleware to create and modify the application’s hosting pipeline.

Sections:

Feature interfaces
Feature collections
Middleware and request features
Summary
Additional Resources

Feature interfaces ¶

ASP.NET Core defines a number of HTTP feature interfaces in Microsoft.AspNetCore.Http.Features which are used by servers to identify the features they support. The following feature interfaces handle requests and return responses:

IHttpRequestFeature: Defines the structure of an HTTP request, including the protocol, path, query string, headers, and body.
IHttpResponseFeature: Defines the structure of an HTTP response, including the status code, headers, and body of the response.
IHttpAuthenticationFeature: Defines support for identifying users based on a ClaimsPrincipal and specifying an authentication handler.
IHttpUpgradeFeature: Defines support for HTTP Upgrades, which allow the client to specify which additional protocols it would like to use if the server wishes to switch protocols.
IHttpBufferingFeature: Defines methods for disabling buffering of requests and/or responses.
IHttpConnectionFeature: Defines properties for local and remote addresses and ports.
IHttpRequestLifetimeFeature: Defines support for aborting connections, or detecting if a request has been terminated prematurely, such as by a client disconnect.
IHttpSendFileFeature: Defines a method for sending files asynchronously.
IHttpWebSocketFeature: Defines an API for supporting web sockets.
IHttpRequestIdentifierFeature: Adds a property that can be implemented to uniquely identify requests.
ISessionFeature: Defines ISessionFactory and ISession abstractions for supporting user sessions.
ITlsConnectionFeature: Defines an API for retrieving client certificates.
ITlsTokenBindingFeature: Defines methods for working with TLS token binding parameters.

Note

ISessionFeature is not a server feature, but is implemented by the SessionMiddleware (see Managing Application State).

Feature collections ¶

The Features property of HttpContext provides an interface for getting and setting the available HTTP features for the current request. Since the feature collection is mutable even within the context of a request, middleware can be used to modify the collection and add support for additional features.

Middleware and request features ¶

While servers are responsible for creating the feature collection, middleware can both add to this collection and consume features from the collection. For example, the StaticFileMiddleware accesses the IHttpSendFileFeature feature. If the feature exists, it is used to send the requested static file from its physical path. Otherwise, a slower alternative method is used to send the file. When available, the IHttpSendFileFeature allows the operating system to open the file and perform a direct kernel mode copy to the network card.

Additionally, middleware can add to the feature collection established by the server. Existing features can even be replaced by middleware, allowing the middleware to augment the functionality of the server. Features added to the collection are available immediately to other middleware or the underlying application itself later in the request pipeline.

By combining custom server implementations and specific middleware enhancements, the precise set of features an application requires can be constructed. This allows missing features to be added without requiring a change in server, and ensures only the minimal amount of features are exposed, thus limiting attack surface area and improving performance.

Summary ¶

Feature interfaces define specific HTTP features that a given request may support. Servers define collections of features, and the initial set of features supported by that server, but middleware can be used to enhance these features.

Additional Resources ¶

Open Web Interface for .NET (OWIN)¶

By Steve Smith and Rick Anderson

ASP.NET Core supports the Open Web Interface for .NET (OWIN). OWIN allows web apps to be decoupled from web servers. It defines a standard way for middleware to be used in a pipeline to handle requests and associated responses. ASP.NET Core applications and middleware can interoperate with OWIN-based applications, servers, and middleware.

Sections:

Running OWIN middleware in the ASP.NET pipeline
Using ASP.NET Hosting on an OWIN-based server
Run ASP.NET Core on an OWIN-based server and use its WebSockets support
OWIN keys
Additional Resources

Running OWIN middleware in the ASP.NET pipeline ¶

ASP.NET Core’s OWIN support is deployed as part of the Microsoft.AspNetCore.Owin package. You can import OWIN support into your project by adding this package as a dependency in your project.json file:

  "dependencies": {
    "Microsoft.AspNetCore.Server.IISIntegration": "1.0.0",
    "Microsoft.AspNetCore.Server.Kestrel": "1.0.0",
    "Microsoft.AspNetCore.Owin": "1.0.0"
  },

OWIN middleware conforms to the OWIN specification, which requires a Func<IDictionary<string, object>, Task> interface, and specific keys be set (such as owin.ResponseBody). The following simple OWIN middleware displays “Hello World”:

public Task OwinHello(IDictionary<string, object> environment)
{
    string responseText = "Hello World via OWIN";
    byte[] responseBytes = Encoding.UTF8.GetBytes(responseText);

    // OWIN Environment Keys: http://owin.org/spec/spec/owin-1.0.0.html
    var responseStream = (Stream)environment["owin.ResponseBody"];
    var responseHeaders = (IDictionary<string, string[]>)environment["owin.ResponseHeaders"];

    responseHeaders["Content-Length"] = new string[] { responseBytes.Length.ToString(CultureInfo.InvariantCulture) };
    responseHeaders["Content-Type"] = new string[] { "text/plain" };

    return responseStream.WriteAsync(responseBytes, 0, responseBytes.Length);
}

The sample signature returns a Task and accepts an IDictionary<string, object> as required by OWIN.

The following code shows how to add the OwinHello middleware (shown above) to the ASP.NET pipeline with the UseOwin extension method.

public void Configure(IApplicationBuilder app)
{
    app.UseOwin(pipeline =>
    {
        pipeline(next => OwinHello);
    });
}

You can configure other actions to take place within the OWIN pipeline.

Note

Response headers should only be modified prior to the first write to the response stream.

Note

Multiple calls to UseOwin is discouraged for performance reasons. OWIN components will operate best if grouped together.

app.UseOwin(pipeline =>
{
    pipeline(next =>
    {
        // do something before
        return OwinHello;
        // do something after
    });
});

Using ASP.NET Hosting on an OWIN-based server ¶

OWIN-based servers can host ASP.NET applications. One such server is Nowin, a .NET OWIN web server. In the sample for this article, I’ve included a project that references Nowin and uses it to create an IServer capable of self-hosting ASP.NET Core.

using System;
using System.Collections.Generic;
using System.Linq;
using System.Net;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Hosting.Server;
using Microsoft.AspNetCore.Hosting.Server.Features;
using Microsoft.AspNetCore.Http.Features;
using Microsoft.AspNetCore.Owin;
using Microsoft.Extensions.Options;
using Nowin;

namespace NowinSample
{
    public class NowinServer : IServer
    {
        private INowinServer _nowinServer;
        private ServerBuilder _builder;

        public IFeatureCollection Features { get; } = new FeatureCollection();

        public NowinServer(IOptions<ServerBuilder> options)
        {
            Features.Set<IServerAddressesFeature>(new ServerAddressesFeature());
            _builder = options.Value;
        }

        public void Start<TContext>(IHttpApplication<TContext> application)
        {
            // Note that this example does not take into account of Nowin's "server.OnSendingHeaders" callback.
            // Ideally we should ensure this method is fired before disposing the context. 
            Func<IDictionary<string, object>, Task> appFunc = async env =>
            {
                // The reason for 2 level of wrapping is because the OwinFeatureCollection isn't mutable
                // so features can't be added
                var features = new FeatureCollection(new OwinFeatureCollection(env));

                var context = application.CreateContext(features);
                try
                {
                    await application.ProcessRequestAsync(context);
                }
                catch (Exception ex)
                {
                    application.DisposeContext(context, ex);
                    throw;
                }

                application.DisposeContext(context, null);
            };

            // Add the web socket adapter so we can turn OWIN websockets into ASP.NET Core compatible web sockets.
            // The calling pattern is a bit different
            appFunc = OwinWebSocketAcceptAdapter.AdaptWebSockets(appFunc);

            // Get the server addresses
            var address = Features.Get<IServerAddressesFeature>().Addresses.First();

            var uri = new Uri(address);
            var port = uri.Port;
            IPAddress ip;
            if (!IPAddress.TryParse(uri.Host, out ip))
            {
                ip = IPAddress.Loopback;
            }

            _nowinServer = _builder.SetAddress(ip)
                                    .SetPort(port)
                                    .SetOwinApp(appFunc)
                                    .Build();
            _nowinServer.Start();
        }

        public void Dispose()
        {
            _nowinServer?.Dispose();
        }
    }
}

IServer is an interface that requires an Features property and a Start method.

Start is responsible for configuring and starting the server, which in this case is done through a series of fluent API calls that set addresses parsed from the IServerAddressesFeature. Note that the fluent configuration of the _builder variable specifies that requests will be handled by the appFunc defined earlier in the method. This Func is called on each request to process incoming requests.

We’ll also add an IWebHostBuilder extension to make it easy to add and configure the Nowin server.

using System;
using Microsoft.AspNetCore.Hosting.Server;
using Microsoft.Extensions.DependencyInjection;
using Nowin;
using NowinSample;

namespace Microsoft.AspNetCore.Hosting
{
    public static class NowinWebHostBuilderExtensions
    {
        public static IWebHostBuilder UseNowin(this IWebHostBuilder builder)
        {
            return builder.ConfigureServices(services =>
            {
                services.AddSingleton<IServer, NowinServer>();
            });
        }

        public static IWebHostBuilder UseNowin(this IWebHostBuilder builder, Action<ServerBuilder> configure)
        {
            builder.ConfigureServices(services =>
            {
                services.Configure(configure);
            });
            return builder.UseNowin();
        }
    }
}

With this in place, all that’s required to run an ASP.NET application using this custom server to call the extension in Program.cs:

using System;
using System.Collections.Generic;
using System.IO;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Hosting;

namespace NowinSample
{
    public class Program
    {
        public static void Main(string[] args)
        {
            var host = new WebHostBuilder()
                .UseNowin()
                .UseContentRoot(Directory.GetCurrentDirectory())
                .UseIISIntegration()
                .UseStartup<Startup>()
                .Build();

            host.Run();
        }
    }
}

Learn more about ASP.NET Servers.

Run ASP.NET Core on an OWIN-based server and use its WebSockets support ¶

Another example of how OWIN-based servers’ features can be leveraged by ASP.NET Core is access to features like WebSockets. The .NET OWIN web server used in the previous example has support for Web Sockets built in, which can be leveraged by an ASP.NET Core application. The example below shows a simple web app that supports Web Sockets and echoes back everything sent to the server through WebSockets.

    public class Startup
    {
        public void Configure(IApplicationBuilder app)
        {
            app.Use(async (context, next) =>
            {
                if (context.WebSockets.IsWebSocketRequest)
                {
                    WebSocket webSocket = await context.WebSockets.AcceptWebSocketAsync();
                    await EchoWebSocket(webSocket);
                }
                else
                {
                    await next();
                }
            });

            app.Run(context =>
            {
                return context.Response.WriteAsync("Hello World");
            });
        }

        private async Task EchoWebSocket(WebSocket webSocket)
        {
            byte[] buffer = new byte[1024];
            WebSocketReceiveResult received = await webSocket.ReceiveAsync(
                new ArraySegment<byte>(buffer), CancellationToken.None);

            while (!webSocket.CloseStatus.HasValue)
            {
                // Echo anything we receive
                await webSocket.SendAsync(new ArraySegment<byte>(buffer, 0, received.Count), 
                    received.MessageType, received.EndOfMessage, CancellationToken.None);

                received = await webSocket.ReceiveAsync(new ArraySegment<byte>(buffer), 
                    CancellationToken.None);
            }

            await webSocket.CloseAsync(webSocket.CloseStatus.Value, 
                webSocket.CloseStatusDescription, CancellationToken.None);
        }
    }
}

This sample is configured using the same NowinServer as the previous one - the only difference is in how the application is configured in its Configure method. A test using a simple websocket client demonstrates the application:

OWIN keys ¶

OWIN depends on an IDictionary<string,object> object to communicate information throughout an HTTP Request/Response exchange. ASP.NET Core implements the keys listed below. See the primary specification, extensions, and OWIN Key Guidelines and Common Keys.

Request Data (OWIN v1.0.0)¶

Key	Value (type)	Description
owin.RequestScheme	`String`
owin.RequestMethod	`String`
owin.RequestPathBase	`String`
owin.RequestPath	`String`
owin.RequestQueryString	`String`
owin.RequestProtocol	`String`
owin.RequestHeaders	`IDictionary<string,string[]>`
owin.RequestBody	`Stream`

Request Data (OWIN v1.1.0)¶

Key	Value (type)	Description
owin.RequestId	`String`	Optional

Response Data (OWIN v1.0.0)¶

Key	Value (type)	Description
owin.ResponseStatusCode	`int`	Optional
owin.ResponseReasonPhrase	`String`	Optional
owin.ResponseHeaders	`IDictionary<string,string[]>`
owin.ResponseBody	`Stream`

Other Data (OWIN v1.0.0)¶

Key	Value (type)	Description
owin.CallCancelled	`CancellationToken`
owin.Version	`String`

Common Keys¶

Key	Value (type)	Description
ssl.ClientCertificate	`X509Certificate`
ssl.LoadClientCertAsync	`Func<Task>`
server.RemoteIpAddress	`String`
server.RemotePort	`String`
server.LocalIpAddress	`String`
server.LocalPort	`String`
server.IsLocal	`bool`
server.OnSendingHeaders	`Action<Action<object>,object>`

SendFiles v0.3.0¶

Key	Value (type)	Description
sendfile.SendAsync	See delegate signature	Per Request

Opaque v0.3.0¶

Key	Value (type)	Description
opaque.Version	`String`
opaque.Upgrade	`OpaqueUpgrade`	See delegate signature
opaque.Stream	`Stream`
opaque.CallCancelled	`CancellationToken`

WebSocket v0.3.0¶

Key	Value (type)	Description
websocket.Version	`String`
websocket.Accept	`WebSocketAccept`	See delegate signature.
websocket.AcceptAlt		Non-spec
websocket.SubProtocol	`String`	See RFC6455 Section 4.2.2 Step 5.5
websocket.SendAsync	`WebSocketSendAsync`	See delegate signature.
websocket.ReceiveAsync	`WebSocketReceiveAsync`	See delegate signature.
websocket.CloseAsync	`WebSocketCloseAsync`	See delegate signature.
websocket.CallCancelled	`CancellationToken`
websocket.ClientCloseStatus	`int`	Optional
websocket.ClientCloseDescription	`String`	Optional

Additional Resources ¶

Choosing the Right .NET For You on the Server¶

By Daniel Roth

ASP.NET Core is based on the .NET Core project model, which supports building applications that can run cross-platform on Windows, Mac and Linux. When building a .NET Core project you also have a choice of which .NET flavor to target your application at: .NET Framework (CLR), .NET Core (CoreCLR) or Mono. Which .NET flavor should you choose? Let’s look at the pros and cons of each one.

.NET Framework¶

The .NET Framework is the most well known and mature of the three options. The .NET Framework is a mature and fully featured framework that ships with Windows. The .NET Framework ecosystem is well established and has been around for well over a decade. The .NET Framework is production ready today and provides the highest level of compatibility for your existing applications and libraries.

The .NET Framework runs on Windows only. It is also a monolithic component with a large API surface area and a slower release cycle. While the code for the .NET Framework is available for reference it is not an active open source project.

.NET Core¶

.NET Core is a modular runtime and library implementation that includes a subset of the .NET Framework. .NET Core is supported on Windows, Mac and Linux. .NET Core consists of a set of libraries, called “CoreFX”, and a small, optimized runtime, called “CoreCLR”. .NET Core is open-source, so you can follow progress on the project and contribute to it on GitHub.

The CoreCLR runtime (Microsoft.CoreCLR) and CoreFX libraries are distributed via NuGet. Because .NET Core has been built as a componentized set of libraries you can limit the API surface area your application uses to just the pieces you need. You can also run .NET Core based applications on much more constrained environments (ex. ASP.NET Core on Nano Server).

The API factoring in .NET Core was updated to enable better componentization. This means that existing libraries built for the .NET Framework generally need to be recompiled to run on .NET Core. The .NET Core ecosystem is relatively new, but it is rapidly growing with the support of popular .NET packages like JSON.NET, AutoFac, xUnit.net and many others.

Developing on .NET Core allows you to target a single consistent platform that can run on multiple platforms.

MVC¶

🔧 Overview of ASP.NET MVC¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Models¶

Model Binding¶

By Rachel Appel

Sections:

Introduction to model binding
How model binding works
Customize model binding behavior with attributes
Binding formatted data from the request body

Introduction to model binding ¶

Model binding in ASP.NET Core MVC maps data from HTTP requests to action method parameters. The parameters may be simple types such as strings, integers, or floats, or they may be complex types. This is a great feature of MVC because mapping incoming data to a counterpart is an often repeated scenario, regardless of size or complexity of the data. MVC solves this problem by abstracting binding away so developers don’t have to keep rewriting a slightly different version of that same code in every app. Writing your own text to type converter code is tedious, and error prone.

How model binding works ¶

When MVC receives an HTTP request, it routes it to a specific action method of a controller. It determines which action method to run based on what is in the route data, then it binds values from the HTTP request to that action method’s parameters. For example, consider the following URL:

http://contoso.com/movies/edit/2

Since the route template looks like this, {controller=Home}/{action=Index}/{id?}, movies/edit/2 routes to the Movies controller, and its Edit action method. It also accepts an optional parameter called id. The code for the action method should look something like this:

public IActionResult Edit(int? id)

Note

The strings in the URL route are not case sensitive.

MVC will try to bind request data to the action parameters by name. MVC will look for values for each parameter using the parameter name and the names of its public settable properties. In the above example, the only action parameter is named id, which MVC binds to the value with the same name in the route values. In addition to route values MVC will bind data from various parts of the request and it does so in a set order. Below is a list of the data sources in the order that model binding looks through them:

Form values: These are form values that go in the HTTP request using the POST method. (including jQuery POST requests).
Route values: The set of route values provided by routing.
Query strings: The query string part of the URI.

Note

Form values, route data, and query strings are all stored as name-value pairs.

Since model binding asked for a key named id and there is nothing named id in the form values, it moved on to the route values looking for that key. In our example, it’s a match. Binding happens, and the value is converted to the integer 2. The same request using Edit(string id) would convert to the string “2”.

So far the example uses simple types. In MVC simple types are any .NET primitive type or type with a string type converter. If the action method’s parameter were a class such as the Movie type, which contains both simple and complex types as properties, MVC’s model binding will still handle it nicely. It uses reflection and recursion to traverse the properties of complex types looking for matches. Model binding looks for the pattern parameter_name.property_name to bind values to properties. If it doesn’t find matching values of this form, it will attempt to bind using just the property name. For those types such as Collection types, model binding looks for matches to parameter_name[index] or just [index]. Model binding treats Dictionary types similarly, asking for parameter_name[key] or just [key], as long as the keys are simple types. Keys that are supported match the field names HTML and tag helpers generated for the same model type. This enables round-tripping values so that the form fields remain filled with the user’s input for their convenience, for example, when bound data from a create or edit did not pass validation.

In order for binding to happen the class must have a public default constructor and member to be bound must be public writable properties. When model binding happens the class will only be instantiated using the public default constructor, then the properties can be set.

When a parameter is bound, model binding stops looking for values with that name and it moves on to bind the next parameter. If binding fails, MVC does not throw an error. You can query for model state errors by checking the ModelState.IsValid property.

Note

Each entry in the controller’s ModelState property is a ModelStateEntry containing an Errors property. It’s rarely necessary to query this collection yourself. Use ModelState.IsValid instead.

Additionally, there are some special data types that MVC must consider when performing model binding:

IFormFile, IEnumerable<IFormFile>: One or more uploaded files that are part of the HTTP request.
CancelationToken: Used to cancel activity in asynchronous controllers.

These types can be bound to action parameters or to properties on a class type.

Once model binding is complete, validation occurs. Default model binding works great for the vast majority of development scenarios. It is also extensible so if you have unique needs you can customize the built-in behavior.

Customize model binding behavior with attributes ¶

MVC contains several attributes that you can use to direct its default model binding behavior to a different source. For example, you can specify whether binding is required for a property, or if it should never happen at all by using the [BindRequired] or [BindNever] attributes. Alternatively, you can override the default data source, and specify the model binder’s data source. Below is a list of model binding attributes:

[BindRequired]: This attribute adds a model state error if binding cannot occur.
[BindNever]: Tells the model binder to never bind to this parameter.
[FromHeader], [FromQuery], [FromRoute], [FromForm]: Use these to specify the exact binding source you want to apply.
[FromServices]: This attribute uses dependency injection to bind parameters from services.
[FromBody]: Use the configured formatters to bind data from the request body. The formatter is selected based on content type of the request.
[ModelBinder]: Used to override the default model binder, binding source and name.

Attributes are very helpful tools when you need to override the default behavior of model binding.

Binding formatted data from the request body ¶

Request data can come in a variety of formats including JSON, XML and many others. When you use the [FromBody] attribute to indicate that you want to bind a parameter to data in the request body, MVC uses a configured set of formatters to handle the request data based on its content type. By default MVC includes a JsonInputFormatter class for handling JSON data, but you can add additional formatters for handling XML and other custom formats.

Note

There can be at most one parameter per action decorated with [FromBody]. The ASP.NET Core MVC run-time delegates the responsibility of reading the request stream to the formatter. Once the request stream is read for a parameter, it’s generally not possible to read the request stream again for binding other [FromBody] parameters.

Note

The JsonInputFormatter is the default formatter and it is based off of Json.NET.

ASP.NET selects input formatters based on the Content-Type header and the type of the parameter, unless there is an attribute applied to it specifying otherwise. If you’d like to use XML or another format you must configure it in the Startup.cs file, but you may first have to obtain a reference to Microsoft.AspNetCore.Mvc.Formatters.Xml using NuGet. Your startup code should look something like this:

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc()
       .AddXmlSerializerFormatters();
}

Code in the Startup.cs file contains a ConfigureServices method with a services argument you can use to build up services for your ASP.NET app. In the sample, we are adding an XML formatter as a service that MVC will provide for this app. The options argument passed into the AddMvc method allows you to add and manage filters, formatters, and other system options from MVC upon app startup. Then apply the Consumes attribute to controller classes or action methods to work with the format you want.

Model Validation¶

By Rachel Appel

In this article:

Sections

Introduction to model validation
Validation Attributes
Model State
Handling Model State Errors
Manual validation
Custom validation
Client side validation
IClientModelValidator
Remote validation

Introduction to model validation ¶

Before an app stores data in a database, the app must validate the data. Data must be checked for potential security threats, verified that it is appropriately formatted by type and size, and it must conform to your rules. Validation is necessary although it can be redundant and tedious to implement. In MVC, validation happens on both the client and server.

Fortunately, .NET has abstracted validation into validation attributes. These attributes contain validation code, thereby reducing the amount of code you must write.

Validation Attributes ¶

Validation attributes are a way to configure model validation so it’s similar conceptually to validation on fields in database tables. This includes constraints such as assigning data types or required fields. Other types of validation include applying patterns to data to enforce business rules, such as a credit card, phone number, or email address. Validation attributes make enforcing these requirements much simpler and easier to use.

Below is an annotated Movie model from an app that stores information about movies and TV shows. Most of the properties are required and several string properties have length requirements. Additionally, there is a numeric range restriction in place for the Price property from 0 to $999.99, along with a custom validation attribute.

public class Movie
{
    public int Id { get; set; }

    [Required]
    [StringLength(100)]
    public string Title { get; set; }

    [Required]
    [ClassicMovie(1960)]
    [DataType(DataType.Date)]
    public DateTime ReleaseDate { get; set; }

    [Required]
    [StringLength(1000)]
    public string Description { get; set; }

    [Required]
    [Range(0, 999.99)]
    public decimal Price { get; set; }

    [Required]
    public Genre Genre { get; set; }

    public bool Preorder { get; set; }
}

Simply reading through the model reveals the rules about data for this app, making it easier to maintain the code. Below are several popular built-in validation attributes:

[CreditCard]: Validates the property has a credit card format.
[Compare]: Validates two properties in a model match.
[EmailAddress]: Validates the property has an email format.
[Phone]: Validates the property has a telephone format.
[Range]: Validates the property value falls within the given range.
[RegularExpression]: Validates that the data matches the specified regular expression.
[Required]: Makes a property required.
[StringLength]: Validates that a string property has at most the given maximum length.
[Url]: Validates the property has a URL format.

MVC supports any attribute that derives from ValidationAttribute for validation purposes. Many useful validation attributes can be found in the System.ComponentModel.DataAnnotations namespace.

There may be instances where you need more features than built-in attributes provide. For those times, you can create custom validation attributes by deriving from ValidationAttribute or changing your model to implement IValidatableObject.

Model State ¶

Model state represents validation errors in submitted HTML form values.

MVC will continue validating fields until reaches the maximum number of errors (200 by default). You can configure this number by inserting the following code into the ConfigureServices method in the Startup.cs file:

services.AddMvc(options => options.MaxModelValidationErrors = 50);

Handling Model State Errors ¶

Model validation occurs prior to each controller action being invoked, and it is the action method’s responsibility to inspect ModelState.IsValid and react appropriately. In many cases, the appropriate reaction is to return some kind of error response, ideally detailing the reason why model validation failed.

Some apps will choose to follow a standard convention for dealing with model validation errors, in which case a filter may be an appropriate place to implement such a policy. You should test how your actions behave with valid and invalid model states.

Manual validation ¶

After model binding and validation are complete, you may want to repeat parts of it. For example, a user may have entered text in a field expecting an integer, or you may need to compute a value for a model’s property.

You may need to run validation manually. To do so, call the TryValidateModel method, as shown here:

TryValidateModel(movie);

Custom validation ¶

Validation attributes work for most validation needs. However, some validation rules are specific to your business, as they’re not just generic data validation such as ensuring a field is required or that it conforms to a range of values. For these scenarios, custom validation attributes are a great solution. Creating your own custom validation attributes in MVC is easy. Just inherit from the ValidationAttribute, and override the IsValid method. The IsValid method accepts two parameters, the first is an object named value and the second is a ValidationContext object named validationContext. Value refers to the actual value from the field that your custom validator is validating.

In the following sample, a business rule states that users may not set the genre to Classic for a movie released after 1960. The [ClassicMovie] attribute checks the genre first, and if it is a classic, then it checks the release date to see that it is later than 1960. If it is released after 1960, validation fails. The attribute accepts an integer parameter representing the year that you can use to validate data. You can capture the value of the parameter in the attribute’s constructor, as shown here:

public class ClassicMovieAttribute : ValidationAttribute, IClientModelValidator
{
    private int _year;

    public ClassicMovieAttribute(int Year)
    {
        _year = Year;
    }

    protected override ValidationResult IsValid(object value, ValidationContext validationContext)
    {
        Movie movie = (Movie)validationContext.ObjectInstance;

        if (movie.Genre == Genre.Classic && movie.ReleaseDate.Year > _year)
        {
            return new ValidationResult(GetErrorMessage());
        }

        return ValidationResult.Success;
    }

The movie variable above represents a Movie object that contains the data from the form submission to validate. In this case, the validation code checks the date and genre in the IsValid method of the ClassicMovieAttribute class as per the rules. Upon successful validation IsValid returns a ValidationResult.Success code, and when validation fails, a ValidationResult with an error message. When a user modifies the Genre field and submits the form, the IsValid method of the ClassicMovieAttribute will verify whether the movie is a classic. Like any built-in attribute, apply the ClassicMovieAttribute to a property such as ReleaseDate to ensure validation happens, as shown in the previous code sample. Since the example works only with Movie types, a better option is to use IValidatableObject as shown in the following paragraph.

Alternatively, this same code could be placed in the model by implementing the Validate method on the IValidatableObject interface. While custom validation attributes work well for validating individual properties, implementing IValidatableObject can be used to implement class-level validation as seen here.

public IEnumerable<ValidationResult> Validate(ValidationContext validationContext)
{
    if (Genre == Genre.Classic && ReleaseDate.Year > _classicYear)
    {
        yield return new ValidationResult(
            "Classic movies must have a release year earlier than " + _classicYear,
            new[] { "ReleaseDate" });
    }
}

Client side validation ¶

Client side validation is a great convenience for users. It saves time they would otherwise spend waiting for a round trip to the server. In business terms, even a few fractions of seconds multiplied hundreds of times each day adds up to be a lot of time, expense, and frustration. Straightforward and immediate validation enables users to work more efficiently and produce better quality input and output.

You must have a view with the proper JavaScript script references in place for client side validation to work as you see here.

<script src="https://ajax.aspnetcdn.com/ajax/jQuery/jquery-1.11.3.min.js"></script>

<script src="https://ajax.aspnetcdn.com/ajax/jquery.validate/1.14.0/jquery.validate.min.js"></script>
<script src="https://ajax.aspnetcdn.com/ajax/jquery.validation.unobtrusive/3.2.6/jquery.validate.unobtrusive.min.js"></script>

MVC uses validation attributes in addition to type metadata from model properties to validate data and display any error messages using JavaScript. When you use MVC to render form elements from a model using Tag Helpers or HTML helpers it will add HTML 5 data- attributes in the form elements that need validation, as shown below. MVC generates the data- attributes for both built-in and custom attributes. You can display validation errors on the client using the relevant tag helpers as shown here:

<div class="form-group">
    <label asp-for="ReleaseDate" class="col-md-2 control-label"></label>
    <div class="col-md-10">
        <input asp-for="ReleaseDate" class="form-control" />
        <span asp-validation-for="ReleaseDate" class="text-danger"></span>
    </div>
</div>

The tag helpers above render the HTML below. Notice that the data- attributes in the HTML output correspond to the validation attributes for the ReleaseDate property. The data-val-required attribute below contains an error message to display if the user doesn’t fill in the release date field, and that message displays in the accompanying <span> element.

<form action="/movies/Create" method="post">
  <div class="form-horizontal">
    <h4>Movie</h4>
    <div class="text-danger"></div>
    <div class="form-group">
      <label class="col-md-2 control-label" for="ReleaseDate">ReleaseDate</label>
      <div class="col-md-10">
        <input class="form-control" type="datetime"
        data-val="true" data-val-required="The ReleaseDate field is required."
        id="ReleaseDate" name="ReleaseDate" value="" />
        <span class="text-danger field-validation-valid"
        data-valmsg-for="ReleaseDate" data-valmsg-replace="true"></span>
      </div>
    </div>
    </div>
</form>

Client-side validation prevents submission until the form is valid. The Submit button runs JavaScript that either submits the form or displays error messages.

MVC determines type attribute values based on the .NET data type of a property, possibly overridden using [DataType] attributes. The base [DataType] attribute does no real server-side validation. Browsers choose their own error messages and display those errors however they wish, however the jQuery Validation Unobtrusive package can override the messages and display them consistently with others. This happens most obviously when users apply [DataType] subclasses such as [EmailAddress].

IClientModelValidator ¶

You may create client side logic for your custom attribute, and unobtrusive validation will execute it on the client for you automatically as part of validation. The first step is to control what data- attributes are added by implementing the IClientModelValidator interface as shown here:

public void AddValidation(ClientModelValidationContext context)
{
    if (context == null)
    {
        throw new ArgumentNullException(nameof(context));
    }

    MergeAttribute(context.Attributes, "data-val", "true");
    MergeAttribute(context.Attributes, "data-val-classicmovie", GetErrorMessage());

    var year = _year.ToString(CultureInfo.InvariantCulture);
    MergeAttribute(context.Attributes, "data-val-classicmovie-year", year);
}

Attributes that implement this interface can add HTML attributes to generated fields. Examining the output for the ReleaseDate element reveals HTML that is similar to the previous example, except now there is a data-val-classicmovie attribute that was defined in the AddValidation method of IClientModelValidator.

<input class="form-control" type="datetime"
data-val="true"
data-val-classicmovie="Classic movies must have a release year earlier than 1960"
data-val-classicmovie-year="1960"
data-val-required="The ReleaseDate field is required."
id="ReleaseDate" name="ReleaseDate" value="" />

Unobtrusive validation uses the data in the data- attributes to display error messages. However, jQuery doesn’t know about rules or messages until you add them to jQuery’s validator object. This is shown in the example below that adds a method named classicmovie containing custom client validation code to the jQuery validator object.

$(function () {
    jQuery.validator.addMethod('classicmovie',
        function (value, element, params) {
            // Get element value. Classic genre has value '0'.
            var genre = $(params[0]).val(),
                year = params[1],
                date = new Date(value);
            if (genre && genre.length > 0 && genre[0] === '0') {
                // Since this is a classic movie, invalid if release date is after given year.
                return date.getFullYear() <= year;
            }

            return true;
        });

    jQuery.validator.unobtrusive.adapters.add('classicmovie',
        [ 'element', 'year' ],
        function (options) {
            var element = $(options.form).find('select#Genre')[0];
            options.rules['classicmovie'] = [element, parseInt(options.params['year'])];
            options.messages['classicmovie'] = options.message;
        });
}(jQuery));

Now jQuery has the information to execute the custom JavaScript validation as well as the error message to display if that validation code returns false.

Remote validation ¶

Remote validation is a great feature to use when you need to validate data on the client against data on the server. For example, your app may need to verify whether an email or user name is already in use, and it must query a large amount of data to do so. Downloading large sets of data for validating one or a few fields consumes too many resources. It may also expose sensitive information. An alternative is to make a round-trip request to validate a field.

You can implement remote validation in a two step process. First, you must annotate your model with the [Remote] attribute. The [Remote] attribute accepts multiple overloads you can use to direct client side JavaScript to the appropriate code to call. The example points to the VerifyEmail action method of the Users controller.

public class User
{
    [Remote(action: "VerifyEmail", controller: "Users")]
    public string Email { get; set; }
}

The second step is putting the validation code in the corresponding action method as defined in the [Remote] attribute. It returns a JsonResult that the client side can use to proceed or pause and display an error if needed.

[AcceptVerbs("Get", "Post")]
public IActionResult VerifyEmail(string email)
{
    if (!_userRepository.VerifyEmail(email))
    {
        return Json(data: $"Email {email} is already in use.");
    }

    return Json(data: true);
}

Now when users enter an email, JavaScript in the view makes a remote call to see if that email has been taken, and if so, then displays the error message. Otherwise, the user can submit the form as usual.

Formatting Response Data¶

View or download sample from GitHub.

ASP.NET Core MVC has built-in support for formatting response data, using fixed formats or in response to client specifications.

Sections

Format-Specific Action Results
Content Negotiation
Configuring Formatters
Response Format URL Mappings

Format-Specific Action Results ¶

Some action result types are specific to a particular format, such as JsonResult and ContentResult. Actions can return specific results that are always formatted in a particular manner. For example, returning a JsonResult will return JSON-formatted data, regardless of client preferences. Likewise, returning a ContentResult will return plain-text-formatted string data (as will simply returning a string).

Note

An action isn’t required to return any particular type; MVC supports any object return value. If an action returns an IActionResult implementation and the controller inherits from Controller, developers have many helper methods corresponding to many of the choices. Results from actions that return objects that are not IActionResult types will be serialized using the appropriate IOutputFormatter implementation.

To return data in a specific format from a controller that inherits from the Controller base class, use the built-in helper method Json to return JSON and Content for plain text. Your action method should return either the specific result type (for instance, JsonResult) or IActionResult.

Returning JSON-formatted data:

// GET: api/authors
[HttpGet]
public JsonResult Get()
{
    return Json(_authorRepository.List());
}

Sample response from this action:

Note that the content type of the response is application/json, shown both in the list of network requests and in the Response Headers section. Also note the list of options presented by the browser (in this case, Microsoft Edge) in the Accept header in the Request Headers section. The current technique is ignoring this header; obeying it is discussed below.

To return plain text formatted data, use ContentResult and the Content helper:

// GET api/authors/about
[HttpGet("About")]
public ContentResult About()
{
    return Content("An API listing authors of docs.asp.net.");
}

A response from this action:

Note in this case the Content-Type returned is text/plain. You can also achieve this same behavior using just a string response type:

// GET api/authors/version
[HttpGet("version")]
public string Version()
{
    return "Version 1.0.0";
}

Tip

For non-trivial actions with multiple return types or options (for example, different HTTP status codes based on the result of operations performed), prefer IActionResult as the return type.

Content Negotiation ¶

Content negotiation (conneg for short) occurs when the client specifies an Accept header. The default format used by ASP.NET Core MVC is JSON. Content negotiation is implemented by ObjectResult. It is also built into the status code specific action results returned from the helper methods (which are all based on ObjectResult). You can also return a model type (a class you’ve defined as your data transfer type) and the framework will automatically wrap it in an ObjectResult for you.

The following action method uses the Ok and NotFound helper methods:

// GET: api/authors/search?namelike=th
[HttpGet("Search")]
public IActionResult Search(string namelike)
{
    var result = _authorRepository.GetByNameSubstring(namelike);
    if (!result.Any())
    {
        return NotFound(namelike);
    }
    return Ok(result);
}

A JSON-formatted response will be returned unless another format was requested and the server can return the requested format. You can use a tool like Fiddler to create a request that includes an Accept header and specify another format. In that case, if the server has a formatter that can produce a response in the requested format, the result will be returned in the client-preferred format.

In the above screenshot, the Fiddler Composer has been used to generate a request, specifying Accept: application/xml. By default, ASP.NET Core MVC only supports JSON, so even when another format is specified, the result returned is still JSON-formatted. You’ll see how to add additional formatters in the next section.

Controller actions can return POCOs (Plain Old CLR Objects), in which case ASP.NET MVC will automatically create an ObjectResult for you that wraps the object. The client will get the formatted serialized object (JSON format is the default; you can configure XML or other formats). If the object being returned is null, then the framework will return a 204 No Content response.

Returning an object type:

// GET api/authors/ardalis
[HttpGet("{alias}")]
public Author Get(string alias)
{
    return _authorRepository.GetByAlias(alias);
}

In the sample, a request for a valid author alias will receive a 200 OK response with the author’s data. A request for an invalid alias will receive a 204 No Content response. Screenshots showing the response in XML and JSON formats are shown below.

Content Negotiation Process¶

Content negotiation only takes place if an Accept header appears in the request. When a request contains an accept header, the framework will enumerate the media types in the accept header in preference order and will try to find a formatter that can produce a response in one of the formats specified by the accept header. In case no formatter is found that can satisfy the client’s request, the framework will try to find the first formatter that can produce a response (unless the developer has configured the option on MvcOptions to return 406 Not Acceptable instead). If the request specifies XML, but the XML formatter has not been configured, then the JSON formatter will be used. More generally, if no formatter is configured that can provide the requested format, then the first formatter than can format the object is used. If no header is given, the first formatter that can handle the object to be returned will be used to serialize the response. In this case, there isn’t any negotiation taking place - the server is determining what format it will use.

Note

If the Accept header contains /, the Header will be ignored unless RespectBrowserAcceptHeader is set to true on MvcOptions.

Browsers and Content Negotiation¶

Unlike typical API clients, web browsers tend to supply Accept headers that include a wide array of formats, including wildcards. By default, when the framework detects that the request is coming from a browser, it will ignore the Accept header and instead return the content in the application’s configured default format (JSON unless otherwise configured). This provides a more consistent experience when using different browsers to consume APIs.

If you would prefer your application honor browser accept headers, you can configure this as part of MVC’s configuration by setting RespectBrowserAcceptHeader to true in the ConfigureServices method in Startup.cs.

services.AddMvc(options =>
{
  options.RespectBrowserAcceptHeader = true; // false by default
}

Configuring Formatters ¶

If your application needs to support additional formats beyond the default of JSON, you can add these as additional dependencies in project.json and configure MVC to support them. There are separate formatters for input and output. Input formatters are used by Model Binding; output formatters are used to format responses. You can also configure 🔧 Custom Formatters.

Adding XML Format Support¶

To add support for XML formatting, add the “Microsoft.AspNetCore.Mvc.Formatters.Xml” package to your project.json‘s list of dependencies.

Add the XmlSerializerFormatters to MVC’s configuration in Startup.cs:

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc()
        .AddXmlSerializerFormatters();

    services.AddScoped<IAuthorRepository, AuthorRepository>();
}

Alternately, you can add just the output formatter:

services.AddMvc(options =>
{
  options.OutputFormatters.Add(new XmlSerializerOutputFormatter());
});

These two approaches will serialize results using System.Xml.Serialization.XmlSerializer. If you prefer, you can use the System.Runtime.Serialization.DataContractSerializer by adding its associated formatter:

services.AddMvc(options =>
{
  options.OutputFormatters.Add(new XmlDataContractSerializerOutputFormatter());
});

Once you’ve added support for XML formatting, your controller methods should return the appropriate format based on the request’s Accept header, as this Fiddler example demonstrates:

You can see in the Inspectors tab that the Raw GET request was made with an Accept: application/xml header set. The response pane shows the Content-Type: application/xml header, and the Author object has been serialized to XML.

Use the Composer tab to modify the request to specify application/json in the Accept header. Execute the request, and the response will be formatted as JSON:

In this screenshot, you can see the request sets a header of Accept: application/json and the response specifies the same as its Content-Type. The Author object is shown in the body of the response, in JSON format.

Forcing a Particular Format¶

If you would like to restrict the response formats for a specific action you can, you can apply the [Produces] filter. The [Produces] filter specifies the response formats for a specific action (or controller). Like most Filters, this can be applied at the action, controller, or global scope.

[Produces("application/json")]
public class AuthorsController

The [Produces] filter will force all actions within the AuthorsController to return JSON-formatted responses, even if other formatters were configured for the application and the client provided an Accept header requesting a different, available format. See Filters to learn more, including how to apply filters globally.

Special Case Formatters¶

Some special cases are implemented using built-in formatters. By default, string return types will be formatted as text/plain (text/html if requested via Accept header). This behavior can be removed by removing the TextOutputFormatter. You remove formatters in the Configure method in Startup.cs (shown below). Actions that have a model object return type will return a 204 No Content response when returning null. This behavior can be removed by removing the HttpNoContentOutputFormatter. The following code removes the TextOutputFormatter and HttpNoContentOutputFormatter`.

services.AddMvc(options =>
{
  options.OutputFormatters.RemoveType<TextOutputFormatter>();
  options.OutputFormatters.RemoveType<HttpNoContentOutputFormatter>();
});

Without the TextOutputFormatter, string return types return 406 Not Acceptable, for example. Note that if an XML formatter exists, it will format string return types if the TextOutputFormatter is removed.

Without the HttpNoContentOutputFormatter, null objects are formatted using the configured formatter. For example, the JSON formatter will simply return a response with a body of null, while the XML formatter will return an empty XML element with the attribute xsi:nil="true" set.

Response Format URL Mappings ¶

Clients can request a particular format as part of the URL, such as in the query string or part of the path, or by using a format-specific file extension such as .xml or .json. The mapping from request path should be specified in the route the API is using. For example:

[FormatFilter]
public class ProductsController
{
  [Route("[controller]/[action]/{id}.{format?}")]
  public Product GetById(int id)

This route would allow the requested format to be specified as an optional file extension. The [FormatFilter] attribute checks for the existence of the format value in the RouteData and will map the response format to the appropriate formatter when the response is created.

Examples¶
Route	Formatter
`/products/GetById/5`	The default output formatter
`/products/GetById/5.json`	The JSON formatter (if configured)
`/products/GetById/5.xml`	The XML formatter (if configured)

🔧 Custom Formatters¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Views¶

Views Overview¶

ASP.NET MVC Core controllers can return formatted results using views.

Sections

What are Views?
Benefits of Using Views
Creating a View
How do Controllers Specify Views?
Passing Data to Views
More View Features

What are Views?¶

In the Model-View-Controller (MVC) pattern, the view encapsulates the presentation details of the user’s interaction with the app. Views are HTML templates with embedded code that generate content to send to the client. Views use Razor syntax, which allows code to interact with HTML with minimal code or ceremony.

ASP.NET Core MVC views are .cshtml files stored by default in a Views folder within the application. Typically, each controller will have its own folder, in which are views for specific controller actions.

In addition to action-specific views, partial views, layouts, and other special view files can be used to help reduce repetition and allow for reuse within the app’s views.

Benefits of Using Views ¶

Views provide separation of concerns within an MVC app, encapsulating user interface level markup separately from business logic. ASP.NET MVC views use Razor syntax to make switching between HTML markup and server side logic painless. Common, repetitive aspects of the app’s user interface can easily be reused between views using layout and shared directives or partial views.

Creating a View ¶

Views that are specific to a controller are created in the Views/[ControllerName] folder. Views that are shared among controllers are placed in the /Views/Shared folder. Name the view file the same as its associated controller action, and add the .cshtml file extension. For example, to create a view for the About action on the Home controller, you would create the About.cshtml file in the /Views/Home folder.

A sample view file (About.cshtml):

@{
    ViewData["Title"] = "About";
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<p>Use this area to provide additional information.</p>

Razor code is denoted by the @ symbol. C# statements are run within Razor code blocks set off by curly braces ({ }), such as the assignment of “About” to the ViewData["Title"] element shown above. Razor can be used to display values within HTML by simply referencing the value with the @ symbol, as shown within the <h2> and <h3> elements above.

This view focuses on just the portion of the output for which it is responsible. The rest of the page’s layout, and other common aspects of the view, are specified elsewhere. Learn more about layout and shared view logic.

How do Controllers Specify Views?¶

Views are typically returned from actions as a ViewResult. Your action method can create and return a ViewResult directly, but more commonly if your controller inherits from Controller, you’ll simply use the View helper method, as this example demonstrates:

HomeController.cs

    public IActionResult About()
    {
        ViewData["Message"] = "Your application description page.";

        return View();
    }

The View helper method has several overloads to make returning views easier for app developers. You can optionally specify a view to return, as well as a model object to pass to the view.

When this action returns, the About.cshtml view shown above is rendered:

View Discovery¶

When an action returns a view, a process called view discovery takes place. This process determines which view file will be used. Unless a specific view file is specified, the runtime looks for a controller-specific view first, then looks for matching view name in the Shared folder.

When an action returns the View method, like so return View();, the action name is used as the view name. For example, if this were called from an action method named “Index”, it would be equivalent to passing in a view name of “Index”. A view name can be explicitly passed to the method (return View("SomeView");). In both of these cases, view discovery searches for a matching view file in:

Views/<ControllerName>/<ViewName>.cshtml

Views/Shared/<ViewName>.cshtml

Tip

We recommend following the convention of simply returning View() from actions when possible, as it results in more flexible, easier to refactor code.

A view file path can be provided, instead of a view name. In this case, the .cshtml extension must be specified as part of the file path. The path should be relative to the application root (and can optionally start with “/” or “~/”). For example: return View("Views/Home/About.cshtml");

Note

Partial views and view components use similar (but not identical) discovery mechanisms.

Note

You can customize the default convention regarding where views are located within the app by using a custom IViewLocationExpander.

Tip

View names may be case sensitive depending on the underlying file system. For compatibility across operating systems, always match case between controller and action names and associated view folders and filenames.

Passing Data to Views ¶

You can pass data to views using several mechanisms. The most robust approach is to specify a model type in the view (commonly referred to as a viewmodel, to distinguish it from business domain model types), and then pass an instance of this type to the view from the action. We recommend you use a model or view model to pass data to a view. This allows the view to take advantage of strong type checking. You can specify a model for a view using the @model directive:

@model WebApplication1.ViewModels.Address
<h2>Contact</h2>
<address>
    @Model.Street<br />
    @Model.City, @Model.State @Model.PostalCode<br />
    <abbr title="Phone">P:</abbr>
    425.555.0100
</address>

Once a model has been specified for a view, the instance sent to the view can be accessed in a strongly-typed manner using @Model as shown above. To provide an instance of the model type to the view, the controller passes it as a parameter:

public IActionResult Contact()
{
    ViewData["Message"] = "Your contact page.";

    var viewModel = new Address()
    {
        Name = "Microsoft",
        Street = "One Microsoft Way",
        City = "Redmond",
        State = "WA",
        PostalCode = "98052-6399"
    };
    return View(viewModel);
}

There are no restrictions on the types that can be provided to a view as a model. We recommend passing Plain Old CLR Object (POCO) view models with little or no behavior, so that business logic can be encapsulated elsewhere in the app. An example of this approach is the Address viewmodel used in the example above:

namespace WebApplication1.ViewModels
{
    public class Address
    {
        public string Name { get; set; }
        public string Street { get; set; }
        public string City { get; set; }
        public string State { get; set; }
        public string PostalCode { get; set; }
    }
}

Note

Nothing prevents you from using the same classes as your business model types and your display model types. However, keeping them separate allows your views to vary independently from your domain or persistence model, and can offer some security benefits as well (for models that users will send to the app using model binding).

Loosely Typed Data¶

In addition to strongly typed views, all views have access to a loosely typed collection of data. This same collection can be referenced through either the ViewData or ViewBag properties on controllers and views. The ViewBag property is a wrapper around ViewData that provides a dynamic view over that collection. It is not a separate collection.

ViewData is a dictionary object accessed through string keys. You can store and retrieve objects in it, and you’ll need to cast them to a specific type when you extract them. You can use ViewData to pass data from a controller to views, as well as within views (and partial views and layouts). String data can be stored and used directly, without the need for a cast.

Set some values for ViewData in an action:

public IActionResult SomeAction()
{
    ViewData["Greeting"] = "Hello";
    ViewData["Address"]  = new Address()
    {
        Name = "Steve",
        Street = "123 Main St",
        City = "Hudson",
        State = "OH",
        PostalCode = "44236"
    };

    return View();
}

Work with the data in a view:

@{
    // Requires cast
    var address = ViewData["Address"] as Address;
}

@ViewData["Greeting"] World!

<address>
    @address.Name<br />
    @address.Street<br />
    @address.City, @address.State @address.PostalCode
</address>

The ViewBag objects provides dynamic access to the objects stored in ViewData. This can be more convenient to work with, since it doesn’t require casting. The same example as above, using ViewBag instead of a strongly typed address instance in the view:

@ViewBag.Greeting World!

<address>
    @ViewBag.Address.Name<br />
    @ViewBag.Address.Street<br />
    @ViewBag.Address.City, @ViewBag.Address.State @ViewBag.Address.PostalCode
</address>

Note

Since both refer to the same underlying ViewData collection, you can mix and match between ViewData and ViewBag when reading and writing values, if convenient.

Dynamic Views¶

Views that do not declare a model type but have a model instance passed to them can reference this instance dynamically. For example, if an instance of Address is passed to a view that doesn’t declare an @model, the view would still be able to refer to the instance’s properties dynamically as shown:

<address>
    @Model.Street<br />
    @Model.City, @Model.State @Model.PostalCode<br />
    <abbr title="Phone">P:</abbr>
    425.555.0100
</address>

This feature can offer some flexibility, but does not offer any compilation protection or IntelliSense. If the property doesn’t exist, the page will fail at runtime.

More View Features ¶

Tag helpers make it easy to add server-side behavior to existing HTML tags, avoiding the need to use custom code or helpers within views. Tag helpers are applied as attributes to HTML elements, which are ignored by editors that aren’t familiar with them, allowing view markup to be edited and rendered in a variety of tools. Tag helpers have many uses, and in particular can make working with forms much easier.

Generating custom HTML markup can be achieved with many built-in HTML Helpers, and more complex UI logic (potentially with its own data requirements) can be encapsulated in View Components. View components provide the same separation of concerns that controllers and views offer, and can eliminate the need for actions and views to deal with data used by common UI elements.

Like many other aspects of ASP.NET Core, views support dependency injection, allowing services to be injected into views.

🔧 Razor Syntax¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Layout¶

Views frequently share visual and programmatic elements. In this article, you’ll learn how to use common layouts, share directives, and run common code before rendering views in your ASP.NET app.

Sections

What is a Layout
Specifying a Layout
Importing Shared Directives
Running Code Before Each View

What is a Layout ¶

Most web apps have a common layout that provides the user with a consistent experience as they navigate from page to page. The layout typically includes common user interface elements such as the app header, navigation or menu elements, and footer.

Common HTML structures such as scripts and stylesheets are also frequently used by many pages within an app. All of these shared elements may be defined in a layout file, which can then be referenced by any view used within the app. Layouts reduce duplicate code in views, helping them follow the Don’t Repeat Yourself (DRY) principle.

By convention, the default layout for an ASP.NET app is named _Layout.cshtml. The Visual Studio ASP.NET Core MVC project template includes this layout file in the Views/Shared folder:

This layout defines a top level template for views in the app. Apps do not require a layout, and apps can define more than one layout, with different views specifying different layouts.

An example _Layout.cshtml:

<!DOCTYPE html>
<html>
<head>
    <meta charset="utf-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>@ViewData["Title"] - WebApplication1</title>

    <environment names="Development">
        <link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />
        <link rel="stylesheet" href="~/css/site.css" />
    </environment>
    <environment names="Staging,Production">
        <link rel="stylesheet" href="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.6/css/bootstrap.min.css"
              asp-fallback-href="~/lib/bootstrap/dist/css/bootstrap.min.css"
              asp-fallback-test-class="sr-only" asp-fallback-test-property="position" asp-fallback-test-value="absolute" />
        <link rel="stylesheet" href="~/css/site.min.css" asp-append-version="true" />
    </environment>
</head>
<body>
    <div class="navbar navbar-inverse navbar-fixed-top">
        <div class="container">
            <div class="navbar-header">
                <button type="button" class="navbar-toggle" data-toggle="collapse" data-target=".navbar-collapse">
                    <span class="sr-only">Toggle navigation</span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                </button>
                <a asp-area="" asp-controller="Home" asp-action="Index" class="navbar-brand">WebApplication1</a>
            </div>
            <div class="navbar-collapse collapse">
                <ul class="nav navbar-nav">
                    <li><a asp-area="" asp-controller="Home" asp-action="Index">Home</a></li>
                    <li><a asp-area="" asp-controller="Home" asp-action="About">About</a></li>
                    <li><a asp-area="" asp-controller="Home" asp-action="Contact">Contact</a></li>
                </ul>
                @await Html.PartialAsync("_LoginPartial")
            </div>
        </div>
    </div>
    <div class="container body-content">
        @RenderBody()
        <hr />
        <footer>
            <p>&copy; 2016 - WebApplication1</p>
        </footer>
    </div>

    <environment names="Development">
        <script src="~/lib/jquery/dist/jquery.js"></script>
        <script src="~/lib/bootstrap/dist/js/bootstrap.js"></script>
        <script src="~/js/site.js" asp-append-version="true"></script>
    </environment>
    <environment names="Staging,Production">
        <script src="https://ajax.aspnetcdn.com/ajax/jquery/jquery-2.2.0.min.js"
                asp-fallback-src="~/lib/jquery/dist/jquery.min.js"
                asp-fallback-test="window.jQuery">
        </script>
        <script src="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.6/bootstrap.min.js"
                asp-fallback-src="~/lib/bootstrap/dist/js/bootstrap.min.js"
                asp-fallback-test="window.jQuery && window.jQuery.fn && window.jQuery.fn.modal">
        </script>
        <script src="~/js/site.min.js" asp-append-version="true"></script>
    </environment>

    @RenderSection("scripts", required: false)
</body>
</html>

Specifying a Layout ¶

Razor views have a Layout property. Individual views specify a layout by setting this property:

@{
    Layout = "_Layout";
}

The layout specified can use a full path (example: /Views/Shared/_Layout.cshtml) or a partial name (example: _Layout). When a partial name is provided, the Razor view engine will search for the layout file using its standard discovery process. The controller-associated folder is searched first, followed by the Shared folder. This discovery process is identical to the one used to discover partial views.

By default, every layout must call RenderBody. Wherever the call to RenderBody is placed, the contents of the view will be rendered.

Sections¶

A layout can optionally reference one or more sections, by calling RenderSection. Sections provide a way to organize where certain page elements should be placed. Each call to RenderSection can specify whether that section is required or optional. If a required section is not found, an exception will be thrown. Individual views specify the content to be rendered within a section using the @section Razor syntax. If a view defines a section, it must be rendered (or an error will occur).

An example @section definition in a view:

@section Scripts {
  <script type="text/javascript" src="/scripts/main.js"></script>
}

In the code above, validation scripts are added to the scripts section on a view that includes a form. Other views in the same application might not require any additional scripts, and so wouldn’t need to define a scripts section.

Sections defined in a view are available only in its immediate layout page. They cannot be referenced from partials, view components, or other parts of the view system.

Ignoring sections¶

By default, the body and all sections in a content page must all be rendered by the layout page. The Razor view engine enforces this by tracking whether the body and each section have been rendered.

To instruct the view engine to ignore the body or sections, call the IgnoreBody and IgnoreSection methods.

The body and every section in a Razor page must be either rendered or ignored.

Importing Shared Directives ¶

Views can use Razor directives to do many things, such as importing namespaces or performing dependency injection. Directives shared by many views may be specified in a common _ViewImports.cshtml file. The _ViewImports file supports the following directives:

@addTagHelper
@removeTagHelper
@tagHelperPrefix
@using
@model
@inherits
@inject

The file does not support other Razor features, such as functions and section definitions.

A sample _ViewImports.cshtml file:

@using WebApplication1
@using WebApplication1.Models
@using WebApplication1.Models.AccountViewModels
@using WebApplication1.Models.ManageViewModels
@using Microsoft.AspNetCore.Identity
@addTagHelper *, Microsoft.AspNetCore.Mvc.TagHelpers

The _ViewImports.cshtml file for an ASP.NET Core MVC app is typically placed in the Views folder. A _ViewImports.cshtml file can be placed within any folder, in which case it will only be applied to views within that folder and its subfolders. _ViewImports files are processed starting at the root level, and then for each folder leading up to the location of the view itself, so settings specified at the root level may be overridden at the folder level.

For example, if a root level _ViewImports.cshtml file specifies @model and @addTagHelper, and another _ViewImports.cshtml file in the controller-associated folder of the view specifies a different @model and adds another @addTagHelper, the view will have access to both tag helpers and will use the latter @model.

If multiple _ViewImports.cshtml files are run for a view, combined behavior of the directives included in the ViewImports.cshtml files will be as follows:

@addTagHelper, @removeTagHelper: all run, in order
@tagHelperPrefix: the closest one to the view overrides any others
@model: the closest one to the view overrides any others
@inherits: the closest one to the view overrides any others
@using: all are included; duplicates are ignored
@inject: for each property, the closest one to the view overrides any others with the same property name

Running Code Before Each View ¶

If you have code you need to run before every view, this should be placed in the _ViewStart.cshtml file. By convention, the _ViewStart.cshtml file is located in the Views folder. The statements listed in _ViewStart.cshtml are run before every full view (not layouts, and not partial views). Like ViewImports.cshtml, _ViewStart.cshtml is hierarchical. If a _ViewStart.cshtml file is defined in the controller-associated view folder, it will be run after the one defined in the root of the Views folder (if any).

A sample _ViewStart.cshtml file:

@{
    Layout = "_Layout";
}

The file above specifies that all views will use the _Layout.cshtml layout.

Note

Neither _ViewStart.cshtml nor _ViewImports.cshtml are typically placed in the /Views/Shared folder. The app-level versions of these files should be placed directly in the /Views folder.

Working with Forms¶

By Rick Anderson, Dave Paquette and Jerrie Pelser

This document demonstrates working with Forms and the HTML elements commonly used on a Form. The HTML Form element provides the primary mechanism web apps use to post back data to the server. Most of this document describes Tag Helpers and how they can help you productively create robust HTML forms. We recommend you read Introduction to Tag Helpers before you read this document.

In many cases, HTML Helpers provide an alternative approach to a specific Tag Helper, but it’s important to recognize that Tag Helpers do not replace HTML Helpers and there is not a Tag Helper for each HTML Helper. When an HTML Helper alternative exists, it is mentioned.

Sections:

The Form Tag Helper
The Input Tag Helper
The Textarea Tag Helper
The Label Tag Helper
The Validation Tag Helpers
The Select Tag Helper
Additional Resources

The Form Tag Helper ¶

The Form Tag Helper:

Generates the HTML <FORM> action attribute value for a MVC controller action or named route
Generates a hidden Request Verification Token to prevent cross-site request forgery (when used with the [ValidateAntiForgeryToken] attribute in the HTTP Post action method)
Provides the asp-route-<Parameter Name> attribute, where <Parameter Name> is added to the route values. The routeValues parameters to Html.BeginForm and Html.BeginRouteForm provide similar functionality.
Has an HTML Helper alternative Html.BeginForm and Html.BeginRouteForm

Sample:

<form asp-controller="Demo" asp-action="Register" method="post">
    <!-- Input and Submit elements -->
</form>

The Form Tag Helper above generates the following HTML:

<form method="post" action="/Demo/Register">
  <!-- Input and Submit elements -->
  <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
 </form>

The MVC runtime generates the action attribute value from the Form Tag Helper attributes asp-controller and asp-action. The Form Tag Helper also generates a hidden Request Verification Token to prevent cross-site request forgery (when used with the [ValidateAntiForgeryToken] attribute in the HTTP Post action method). Protecting a pure HTML Form from cross-site request forgery is very difficult, the Form Tag Helper provides this service for you.

Using a named route¶

The asp-route Tag Helper attribute can also generate markup for the HTML action attribute. An app with a route named register could use the following markup for the registration page:

<form asp-route="register" method="post">
    <!-- Input and Submit elements -->
</form>

Many of the views in the Views/Account folder (generated when you create a new web app with Individual User Accounts) contain the asp-route-returnurl attribute:

<form asp-controller="Account" asp-action="Login"
  asp-route-returnurl="@ViewData["ReturnUrl"]"
  method="post" class="form-horizontal" role="form">

Note:	With the built in templates, `returnUrl` is only populated automatically when you try to access an authorized resource but are not authenticated or authorized. When you attempt an unauthorized access, the security middleware redirects you to the login page with the `returnUrl` set.

The Input Tag Helper ¶

The Input Tag Helper binds an HTML <input> element to a model expression in your razor view.

Syntax:

<input asp-for="<Expression Name>" />

The Input Tag Helper:

Generates the id and name HTML attributes for the expression name specified in the asp-for attribute. asp-for="Property1.Property2" is equivalent to m => m.Property1.Property2, that is the attribute value literally is part of an expression. The name of the expression is what’s used for the asp-for attribute value.
Sets the HTML type attribute value based on the model type and data annotation attributes applied to the model property
Will not overwrite the HTML type attribute value when one is specified
Generates HTML5 validation attributes from data annotation attributes applied to model properties
Has an HTML Helper feature overlap with Html.TextBoxFor and Html.EditorFor. See the HTML Helper alternatives to Input Tag Helper section for details.
Provides strong typing. If the name of the property changes and you don’t update the Tag Helper you’ll get an error similar to the following:

An error occurred during the compilation of a resource required to process
this request. Please review the following specific error details and modify
your source code appropriately.

Type expected
 'RegisterViewModel' does not contain a definition for 'Email' and no
 extension method 'Email' accepting a first argument of type 'RegisterViewModel'
 could be found (are you missing a using directive or an assembly reference?)

The Input Tag Helper sets the HTML type attribute based on the .NET type. The following table lists some common .NET types and generated HTML type (not every .NET type is listed).

.NET type	Input Type
Bool	type=”checkbox”
String	type=”text”
DateTime	type=”datetime”
Byte	type=”number”
Int	type=”number”
Single, Double	type=”number”

The following table shows some common data annotations attributes that the input tag helper will map to specific input types (not every validation attribute is listed):

Attribute	Input Type
[EmailAddress]	type=”email”
[Url]	type=”url”
[HiddenInput]	type=”hidden”
[Phone]	type=”tel”
[DataType(DataType.Password)]	type=”password”
[DataType(DataType.Date)]	type=”date”
[DataType(DataType.Time)]	type=”time”

Sample:

using System.ComponentModel.DataAnnotations;

namespace FormsTagHelper.ViewModels
{
    public class RegisterViewModel
    {
        [Required]
        [EmailAddress]
        [Display(Name = "Email Address")]
        public string Email { get; set; }

        [Required]
        [DataType(DataType.Password)]
        public string Password { get; set; }
    }
}

@model RegisterViewModel

<form asp-controller="Demo" asp-action="RegisterInput" method="post">
    Email:  <input asp-for="Email" /> <br />
    Password: <input asp-for="Password" /><br />
    <button type="submit">Register</button>
</form>

The code above generates the following HTML:

  <form method="post" action="/Demo/RegisterInput">
    Email:
    <input type="email" data-val="true"
           data-val-email="The Email Address field is not a valid e-mail address."
           data-val-required="The Email Address field is required."
           id="Email" name="Email" value="" /> <br>
    Password:
    <input type="password" data-val="true"
           data-val-required="The Password field is required."
           id="Password" name="Password" /><br>
    <button type="submit">Register</button>
  <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
</form>

The data annotations applied to the Email and Password properties generate metadata on the model. The Input Tag Helper consumes the model metadata and produces HTML5 data-val-* attributes (see Model Validation). These attributes describe the validators to attach to the input fields. This provides unobtrusive HTML5 and jQuery validation. The unobtrusive attributes have the format data-val-rule="Error Message", where rule is the name of the validation rule (such as data-val-required, data-val-email, data-val-maxlength, etc.) If an error message is provided in the attribute, it is displayed as the value for the data-val-rule attribute. There are also attributes of the form data-val-ruleName-argumentName="argumentValue" that provide additional details about the rule, for example, data-val-maxlength-max="1024" .

HTML Helper alternatives to Input Tag Helper¶

Html.TextBox, Html.TextBoxFor, Html.Editor and Html.EditorFor have overlapping features with the Input Tag Helper. The Input Tag Helper will automatically set the type attribute; Html.TextBox and Html.TextBoxFor will not. Html.Editor and Html.EditorFor handle collections, complex objects and templates; the Input Tag Helper does not. The Input Tag Helper, Html.EditorFor and Html.TextBoxFor are strongly typed (they use lambda expressions); Html.TextBox and Html.Editor are not (they use expression names).

Expression names¶

The asp-for attribute value is a ModelExpression and the right hand side of a lambda expression. Therefore, asp-for="Property1" becomes m => m.Property1 in the generated code which is why you don’t need to prefix with Model. You can use the “@” character to start an inline expression and move before the m.:

@{
    var joe = "Joe";
}
<input asp-for="@joe" />

Generates the following:

<input type="text" id="joe" name="joe" value="Joe" />

Navigating child properties¶

You can also navigate to child properties using the property path of the view model. Consider a more complex model class that contains a child Address property.

 public class AddressViewModel
 {
     public string AddressLine1 { get; set; }
 }

 public class RegisterAddressViewModel
 {
     public string Email { get; set; }

     [DataType(DataType.Password)]
     public string Password { get; set; }

     public AddressViewModel Address { get; set; }
 }

In the view, we bind to Address.AddressLine1:

@model RegisterAddressViewModel

<form asp-controller="Demo" asp-action="RegisterAddress" method="post">
    Email:  <input asp-for="Email" /> <br />
    Password: <input asp-for="Password" /><br />
    Address: <input asp-for="Address.AddressLine1" /><br />
    <button type="submit">Register</button>
</form>

The following HTML is generated for Address.AddressLine1:

<input type="text" id="Address_AddressLine1" name="Address.AddressLine1" value="" />

Expression names and Collections¶

Sample, a model containing an array of Colors:

 public class Person
 {
     public List<string> Colors { get; set; }

     public int Age { get; set; }
 }

The action method:

public IActionResult Edit(int id, int colorIndex)
{
    ViewData["Index"] = colorIndex;
    return View(GetPerson(id));
}

The following Razor shows how you access a specific Color element:

@model Person
@{
    var index = (int)ViewData["index"];
}

<form asp-controller="ToDo" asp-action="Edit" method="post">
    @Html.EditorFor(m => m.Colors[index])
    <label asp-for="Age"></label>
    <input asp-for="Age" /><br />
    <button type="submit">Post</button>
</form>

The Views/Shared/EditorTemplates/String.cshtml template:

@model string

<label asp-for="@Model"></label>
<input asp-for="@Model" /> <br />

Sample using List<T>:

 public class ToDoItem
 {
     public string Name { get; set; }

     public bool IsDone { get; set; }

The following Razor shows how to iterate over a collection:

@model List<ToDoItem>

<form asp-controller="ToDo" asp-action="Edit" method="post">
    <table>
        <tr> <th>Name</th> <th>Is Done</th> </tr>

        @for (int i = 0; i < Model.Count; i++)
        {
            <tr>
                @Html.EditorFor(model => model[i])
            </tr>
        }

    </table>
    <button type="submit">Save</button>
</form>

The Views/Shared/EditorTemplates/ToDoItem.cshtml template:

@model ToDoItem

<td>
    <label asp-for="@Model.Name"></label>
    @Html.DisplayFor(model => model.Name)
</td>
<td>
    <input asp-for="@Model.IsDone" />
</td>

@*
    This template replaces the following Razor which evaluates the indexer three times.
    <td>
         <label asp-for="@Model[i].Name"></label>
         @Html.DisplayFor(model => model[i].Name)
     </td>
     <td>
         <input asp-for="@Model[i].IsDone" />
     </td>
*@

Note:	Always use `for` (and not `foreach`) to iterate over a list. Evaluating an indexer in a LINQ expression can be expensive and should be minimized.
Note:	The commented sample code above shows how you would replace the lambda expression with the `@` operator to access each `ToDoItem` in the list.

The Textarea Tag Helper ¶

The Textarea Tag Helper tag helper is similar to the Input Tag Helper.

Generates the id and name attributes, and the data validation attributes from the model for a <textarea> element.
Provides strong typing.
HTML Helper alternative: Html.TextAreaFor

Sample:

using System.ComponentModel.DataAnnotations;

namespace FormsTagHelper.ViewModels
{
    public class DescriptionViewModel
    {
        [MinLength(5)]
        [MaxLength(1024)]
        public string Description { get; set; }
    }
}

@model DescriptionViewModel

<form asp-controller="Demo" asp-action="RegisterTextArea" method="post">
    <textarea asp-for="Description"></textarea>
    <button type="submit">Test</button>
</form>

The following HTML is generated:

<form method="post" action="/Demo/RegisterTextArea">
  <textarea data-val="true"
   data-val-maxlength="The field Description must be a string or array type with a maximum length of &#x27;1024&#x27;."
   data-val-maxlength-max="1024"
   data-val-minlength="The field Description must be a string or array type with a minimum length of &#x27;5&#x27;."
   data-val-minlength-min="5"
   id="Description" name="Description">
  </textarea>
  <button type="submit">Test</button>
  <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
</form>

The Label Tag Helper ¶

Generates the label caption and for attribute on a <label> element for an expression name
HTML Helper alternative: Html.LabelFor.

The Label Tag Helper provides the following benefits over a pure HTML label element:

You automatically get the descriptive label value from the Display attribute. The intended display name might change over time, and the combination of Display attribute and Label Tag Helper will apply the Display everywhere it’s used.
Less markup in source code
Strong typing with the model property.

Sample:

using System.ComponentModel.DataAnnotations;

namespace FormsTagHelper.ViewModels
{
    public class SimpleViewModel
    {
        [Required]
        [EmailAddress]
        [Display(Name = "Email Address")]
        public string Email { get; set; }
    }
}

@model SimpleViewModel

<form asp-controller="Demo" asp-action="RegisterLabel" method="post">
    <label asp-for="Email"></label>
    <input asp-for="Email" /> <br />
</form>

The following HTML is generated for the <label> element:

<label for="Email">Email Address</label>

The Label Tag Helper generated the for attribute value of “Email”, which is the ID associated with the <input> element. The Tag Helpers generate consistent id and for elements so they can be correctly associated. The caption in this sample comes from the Display attribute. If the model didn’t contain a Display attribute, the caption would be the property name of the expression.

The Validation Tag Helpers ¶

There are two Validation Tag Helpers. The Validation Message Tag Helper (which displays a validation message for a single property on your model), and the Validation Summary Tag Helper (which displays a summary of validation errors). The Input Tag Helper adds HTML5 client side validation attributes to input elements based on data annotation attributes on your model classes. Validation is also performed on the server. The Validation Tag Helper displays these error messages when a validation error occurs.

The Validation Message Tag Helper¶

Adds the HTML5 data-valmsg-for="property" attribute to the span element, which attaches the validation error messages on the input field of the specified model property. When a client side validation error occurs, jQuery displays the error message in the <span> element.
Validation also takes place on the server. Clients may have JavaScript disabled and some validation can only be done on the server side.
HTML Helper alternative: Html.ValidationMessageFor

The Validation Message Tag Helper is used with the asp-validation-for attribute on a HTML span element.

<span asp-validation-for="Email"></span>

The Validation Message Tag Helper will generate the following HTML:

<span class="field-validation-valid"
  data-valmsg-for="Email"
  data-valmsg-replace="true"></span>

You generally use the Validation Message Tag Helper after an Input Tag Helper for the same property. Doing so displays any validation error messages near the input that caused the error.

Note:	You must have a view with the correct JavaScript and jQuery script references in place for client side validation. See Model Validation for more information.

When a server side validation error occurs (for example when you have custom server side validation or client-side validation is disabled), MVC places that error message as the body of the <span> element.

<span class="field-validation-error" data-valmsg-for="Email"
            data-valmsg-replace="true">
   The Email Address field is required.
</span>

The Validation Summary Tag Helper¶

Targets <div> elements with the asp-validation-summary attribute
HTML Helper alternative: @Html.ValidationSummary

The Validation Summary Tag Helper is used to display a summary of validation messages. The asp-validation-summary attribute value can be any of the following:

asp-validation-summary	Validation messages displayed
ValidationSummary.All	Property and model level
ValidationSummary.ModelOnly	Model
ValidationSummary.None	None

Sample¶

In the following example, the data model is decorated with DataAnnotation attributes, which generates validation error messages on the <input> element. When a validation error occurs, the Validation Tag Helper displays the error message:

using System.ComponentModel.DataAnnotations;

namespace FormsTagHelper.ViewModels
{
    public class RegisterViewModel
    {
        [Required]
        [EmailAddress]
        [Display(Name = "Email Address")]
        public string Email { get; set; }

        [Required]
        [DataType(DataType.Password)]
        public string Password { get; set; }
    }
}

@model RegisterViewModel

<form asp-controller="Demo" asp-action="RegisterValidation" method="post">
    <div asp-validation-summary="ValidationSummary.ModelOnly"></div>
    Email:  <input asp-for="Email" /> <br />
    <span asp-validation-for="Email"></span><br />
    Password: <input asp-for="Password" /><br />
    <span asp-validation-for="Password"></span><br />
    <button type="submit">Register</button>
</form>

The generated HTML (when the model is valid):

<form action="/DemoReg/Register" method="post">
  <div class="validation-summary-valid" data-valmsg-summary="true">
  <ul><li style="display:none"></li></ul></div>
  Email:  <input name="Email" id="Email" type="email" value=""
   data-val-required="The Email field is required."
   data-val-email="The Email field is not a valid e-mail address."
   data-val="true"> <br>
  <span class="field-validation-valid" data-valmsg-replace="true"
   data-valmsg-for="Email"></span><br>
  Password: <input name="Password" id="Password" type="password"
   data-val-required="The Password field is required." data-val="true"><br>
  <span class="field-validation-valid" data-valmsg-replace="true"
   data-valmsg-for="Password"></span><br>
  <button type="submit">Register</button>
  <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
</form>

The Select Tag Helper ¶

Generates select and associated option elements for properties of your model.
Has an HTML Helper alternative Html.DropDownListFor and Html.ListBoxFor

The Select Tag Helper asp-for specifies the model property name for the select element and asp-items specifies the option elements. For example:

 <select asp-for="Country" asp-items="Model.Countries"></select> 

Sample:

using Microsoft.AspNetCore.Mvc.Rendering;
using System.Collections.Generic;

namespace FormsTagHelper.ViewModels
{
    public class CountryViewModel
    {
        public string Country { get; set; }

        public List<SelectListItem> Countries { get; } = new List<SelectListItem>
        {
            new SelectListItem { Value = "MX", Text = "Mexico" },
            new SelectListItem { Value = "CA", Text = "Canada" },
            new SelectListItem { Value = "US", Text = "USA"  },
        };
    }
}

The Index method initializes the CountryViewModel, sets the selected country and passes it to the Index view.

  public IActionResult Index()
  {
      var model = new CountryViewModel();
      model.Country = "CA";
      return View(model);
  }

The HTTP POST Index method displays the selection:

  [HttpPost]
  [ValidateAntiForgeryToken]
  public IActionResult Index(CountryViewModel model)
  {
      if (ModelState.IsValid)
      {
          var msg = model.Country +  " selected";
          return RedirectToAction("IndexSuccess", new { message = msg});
      }

      // If we got this far, something failed; redisplay form.
      return View(model);
  }

The Index view:

@model CountryViewModel

<form asp-controller="Home" asp-action="Index" method="post">
    <select asp-for="Country" asp-items="Model.Countries"></select> 
    <br /><button type="submit">Register</button>
</form>

Which generates the following HTML (with “CA” selected):

<form method="post" action="/">
  <select id="Country" name="Country">
    <option value="MX">Mexico</option>
    <option selected="selected" value="CA">Canada</option>
    <option value="US">USA</option>
  </select>
    <br /><button type="submit">Register</button>
  <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
</form>

Note:	We do not recommend using `ViewBag` or `ViewData` with the Select Tag Helper. A view model is more robust at providing MVC metadata and generally less problematic.

The asp-for attribute value is a special case and doesn’t require a Model prefix, the other Tag Helper attributes do (such as asp-items)

 <select asp-for="Country" asp-items="Model.Countries"></select> 

Enum binding¶

It’s often convenient to use <select> with an enum property and generate the SelectListItem elements from the enum values.

Sample:

 public class CountryEnumViewModel
 {
     public CountryEnum EnumCountry { get; set; }
 }

using System.ComponentModel.DataAnnotations;

namespace FormsTagHelper.ViewModels
{
    {
        Mexico,
        [Display(Name = "United States of America")]
        USA,
        Canada,
        France,
        Germany,
        Spain
    }
}

The GetEnumSelectList method generates a SelectList object for an enum.

@model CountryEnumViewModel

<form asp-controller="Home" asp-action="IndexEnum" method="post">
    <select asp-for="EnumCountry" 
            asp-items="Html.GetEnumSelectList<CountryEnum>()"> >
    </select> 
    <br /><button type="submit">Register</button>
</form>

You can decorate your enumerator list with the Display attribute to get a richer UI:

using System.ComponentModel.DataAnnotations;

namespace FormsTagHelper.ViewModels
{
    public enum CountryEnum
    {
        [Display(Name = "United Mexican States")]
        Mexico,
        [Display(Name = "United States of America")]
        USA,
        Canada,
        France,
        Germany,
        Spain
    }
}

The following HTML is generated:

  <form method="post" action="/Home/IndexEnum">
      <select data-val="true" data-val-required="The EnumCountry field is required."
              id="EnumCountry" name="EnumCountry">
          <option value="0">United Mexican States</option>
          <option value="1">United States of America</option>
          <option value="2">Canada</option>
          <option value="3">France</option>
          <option value="4">Germany</option>
          <option selected="selected" value="5">Spain</option>
      </select>
      <br /><button type="submit">Register</button>
      <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
 </form>

Option Group¶

The HTML <optgroup> element is generated when the view model contains one or more SelectListGroup objects.

The CountryViewModelGroup groups the SelectListItem elements into the “North America” and “Europe” groups:

 public class CountryViewModelGroup
 {
     public CountryViewModelGroup()
     {
         var NorthAmericaGroup = new SelectListGroup { Name = "North America" };
         var EuropeGroup = new SelectListGroup { Name = "Europe" };

         Countries = new List<SelectListItem>
         {
             new SelectListItem
             {
                 Value = "MEX",
                 Text = "Mexico",
                 Group = NorthAmericaGroup
             },
             new SelectListItem
             {
                 Value = "CAN",
                 Text = "Canada",
                 Group = NorthAmericaGroup
             },
             new SelectListItem
             {
                 Value = "US",
                 Text = "USA",
                 Group = NorthAmericaGroup
             },
             new SelectListItem
             {
                 Value = "FR",
                 Text = "France",
                 Group = EuropeGroup
             },
             new SelectListItem
             {
                 Value = "ES",
                 Text = "Spain",
                 Group = EuropeGroup
             },
             new SelectListItem
             {
                 Value = "DE",
                 Text = "Germany",
                 Group = EuropeGroup
             }
       };
     }

     public string Country { get; set; }

     public List<SelectListItem> Countries { get; }
 }

The two groups are shown below:

The generated HTML:

  <form method="post" action="/Home/IndexGroup">
      <select id="Country" name="Country">
          <optgroup label="North America">
              <option value="MEX">Mexico</option>
              <option value="CAN">Canada</option>
              <option value="US">USA</option>
          </optgroup>
          <optgroup label="Europe">
              <option value="FR">France</option>
              <option value="ES">Spain</option>
              <option value="DE">Germany</option>
          </optgroup>
      </select>
      <br /><button type="submit">Register</button>
      <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
 </form>

Multiple select¶

The Select Tag Helper will automatically generate the multiple = “multiple” attribute if the property specified in the asp-for attribute is an IEnumerable. For example, given the following model:

using Microsoft.AspNetCore.Mvc.Rendering;
using System.Collections.Generic;

namespace FormsTagHelper.ViewModels
{
    public class CountryViewModelIEnumerable
    {
        public IEnumerable<string> CountryCodes { get; set; }

        public List<SelectListItem> Countries { get; } = new List<SelectListItem>
        {
            new SelectListItem { Value = "MX", Text = "Mexico" },
            new SelectListItem { Value = "CA", Text = "Canada" },
            new SelectListItem { Value = "US", Text = "USA"    },
            new SelectListItem { Value = "FR", Text = "France" },
            new SelectListItem { Value = "ES", Text = "Spain"  },
            new SelectListItem { Value = "DE", Text = "Germany"}
         };
    }
}

With the following view:

@model CountryViewModelIEnumerable

<form asp-controller="Home" asp-action="IndexMultiSelect" method="post">
    <select asp-for="CountryCodes" asp-items="Model.Countries"></select> 
    <br /><button type="submit">Register</button>
</form>

Generates the following HTML:

<form method="post" action="/Home/IndexMultiSelect">
    <select id="CountryCodes"
    multiple="multiple"
    name="CountryCodes"><option value="MX">Mexico</option>
<option value="CA">Canada</option>
<option value="US">USA</option>
<option value="FR">France</option>
<option value="ES">Spain</option>
<option value="DE">Germany</option>
</select>
    <br /><button type="submit">Register</button>
  <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
</form>

No selection¶

To allow for no selection, add a “not specified” option to the select list. If the property is a value type, you’ll have to make it nullable.

@model CountryViewModel

<form asp-controller="Home" asp-action="IndexEmpty" method="post">
    <select asp-for="Country" asp-items="Model.Countries">
        <option value="">&lt;none&gt;</option>
    </select> 
    <br /><button type="submit">Register</button>
</form>

If you find yourself using the “not specified” option in multiple pages, you can create a template to eliminate repeating the HTML:

@model CountryViewModel

<form asp-controller="Home" asp-action="IndexEmpty" method="post">
    @Html.EditorForModel()
    <br /><button type="submit">Register</button>
</form>

The Views/Shared/EditorTemplates/CountryViewModel.cshtml template:

@model CountryViewModel

<select asp-for="Country" asp-items="Model.Countries">
    <option value="">--none--</option>
</select>

Adding HTML <option> elements is not limited to the No selection case. For example, the following view and action method will generate HTML similar to the code above:

  public IActionResult IndexOption(int id)
  {
      var model = new CountryViewModel();
      model.Country = "CA";
      return View(model);
  }

@model CountryViewModel

<form asp-controller="Home" asp-action="IndexEmpty" method="post">
    <select asp-for="Country">
        <option value="">&lt;none&gt;</option>
        <option value="MX">Mexico</option>
        <option value="CA">Canada</option>
        <option value="US">USA</option>
    </select> 
    <br /><button type="submit">Register</button>
</form>

The correct <option> element will be selected ( contain the selected="selected" attribute) depending on the current Country value.

 <form method="post" action="/Home/IndexEmpty">
     <select id="Country" name="Country">
         <option value="">&lt;none&gt;</option>
         <option value="MX">Mexico</option>
         <option value="CA" selected="selected">Canada</option>
         <option value="US">USA</option>
     </select>
     <br /><button type="submit">Register</button>
  <input name="__RequestVerificationToken" type="hidden" value="<removed for brevity>" />
</form>

Additional Resources ¶

🔧 HTML Helpers¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Tag Helpers¶

Introduction to Tag Helpers¶

What are Tag Helpers?

What Tag Helpers provide

Managing Tag Helper scope

IntelliSense support for Tag Helpers

Tag Helpers compared to HTML Helpers

Tag Helpers compared to Web Server Controls

Customizing the Tag Helper element font

Additional Resources

What are Tag Helpers?¶

Tag Helpers enable server-side code to participate in creating and rendering HTML elements in Razor files. For example, the built-in ImageTagHelper can append a version number to the image name. Whenever the image changes, the server generates a new unique version for the image, so clients are guaranteed to get the current image (instead of a stale cached image). There are many built-in Tag Helpers for common tasks - such as creating forms, links, loading assets and more - and even more available in public GitHub repositories and as NuGet packages. Tag Helpers are authored in C#, and they target HTML elements based on element name, attribute name, or parent tag. For example, the built-in LabelTagHelper can target the HTML <label> element when the LabelTagHelper attributes are applied. If you’re familiar with HTML Helpers, Tag Helpers reduce the explicit transitions between HTML and C# in Razor views. Tag Helpers compared to HTML Helpers explains the differences in more detail.

What Tag Helpers provide¶

An HTML-friendly development experience: For the most part, Razor markup using Tag Helpers looks like standard HTML. Front-end designers conversant with HTML/CSS/JavaScript can edit Razor without learning C# Razor syntax.
A rich IntelliSense environment for creating HTML and Razor markup: This is in sharp contrast to HTML Helpers, the previous approach to server-side creation of markup in Razor views. Tag Helpers compared to HTML Helpers explains the differences in more detail. IntelliSense support for Tag Helpers explains the IntelliSense environment. Even developers experienced with Razor C# syntax are more productive using Tag Helpers than writing C# Razor markup.
A way to make you more productive and able to produce more robust, reliable, and maintainable code using information only available on the server: For example, historically the mantra on updating images was to change the name of the image when you change the image. Images should be aggressively cached for performance reasons, and unless you change the name of an image, you risk clients getting a stale copy. Historically, after an image was edited, the name had to be changed and each reference to the image in the web app needed to be updated. Not only is this very labor intensive, it’s also error prone (you could miss a reference, accidentally enter the wrong string, etc.) The built-in ImageTagHelper can do this for you automatically. The ImageTagHelper can append a version number to the image name, so whenever the image changes, the server automatically generates a new unique version for the image. Clients are guaranteed to get the current image. This robustness and labor savings comes essentially free by using the ImageTagHelper.

Most of the built-in Tag Helpers target existing HTML elements and provide server-side attributes for the element. For example, the <input> element used in many of the views in the Views/Account folder contains the asp-for attribute, which extracts the name of the specified model property into the rendered HTML. The following Razor markup:

<label asp-for="Email"></label>

Generates the following HTML:

<label for="Email">Email</label>

The asp-for attribute is made available by the For property in the LabelTagHelper. See Authoring Tag Helpers for more information.

Managing Tag Helper scope¶

Tag Helpers scope is controlled by a combination of @addTagHelper, @removeTagHelper, and the ”!” opt-out character.

@addTagHelper makes Tag Helpers available¶

If you create a new ASP.NET Core web app named AuthoringTagHelpers (with no authentication), the following Views/_ViewImports.cshtml file will be added to your project:

@using AuthoringTagHelpers
@addTagHelper *, Microsoft.AspNetCore.Mvc.TagHelpers

The @addTagHelper directive makes Tag Helpers available to the view. In this case, the view file is Views/_ViewImports.cshtml, which by default is inherited by all view files in the Views folder and sub-directories; making Tag Helpers available. The code above uses the wildcard syntax (“*”) to specify that all Tag Helpers in the specified assembly (Microsoft.AspNetCore.Mvc.TagHelpers) will be available to every view file in the Views directory or sub-directory. The first parameter after @addTagHelper specifies the Tag Helpers to load (we are using “*” for all Tag Helpers), and the second parameter “Microsoft.AspNetCore.Mvc.TagHelpers” specifies the assembly containing the Tag Helpers. Microsoft.AspNetCore.Mvc.TagHelpers is the assembly for the built-in ASP.NET Core Tag Helpers.

To expose all of the Tag Helpers in this project (which creates an assembly named AuthoringTagHelpers), you would use the following:

@using AuthoringTagHelpers
@addTagHelper *, Microsoft.AspNetCore.Mvc.TagHelpers
@addTagHelper "*, AuthoringTagHelpers"

If your project contains an EmailTagHelper with the default namespace (AuthoringTagHelpers.TagHelpers.EmailTagHelper), you can provide the fully qualified name (FQN) of the Tag Helper:

 @using AuthoringTagHelpers
 @addTagHelper *, Microsoft.AspNetCore.Mvc.TagHelpers
 @addTagHelper "AuthoringTagHelpers.TagHelpers.EmailTagHelper, AuthoringTagHelpers"

To add a Tag Helper to a view using an FQN, you first add the FQN (AuthoringTagHelpers.TagHelpers.EmailTagHelper), and then the assembly name (AuthoringTagHelpers). Most developers prefer to use the “*” wildcard syntax. The wildcard syntax allows you to insert the wildcard character “*” as the suffix in an FQN. For example, any of the following directives will bring in the EmailTagHelper:

@addTagHelper "AuthoringTagHelpers.TagHelpers.E*, AuthoringTagHelpers"
@addTagHelper "AuthoringTagHelpers.TagHelpers.Email*, AuthoringTagHelpers"

As mentioned previously, adding the @addTagHelper directive to the Views/_ViewImports.cshtml file makes the Tag Helper available to all view files in the Views directory and sub-directories. You can use the @addTagHelper directive in specific view files if you want to opt-in to exposing the Tag Helper to only those views.

@removeTagHelper removes Tag Helpers¶

The @removeTagHelper has the same two parameters as @addTagHelper, and it removes a Tag Helper that was previously added. For example, @removeTagHelper applied to a specific view removes the specified Tag Helper from the view. Using @removeTagHelper in a Views/Folder/_ViewImports.cshtml file removes the specified Tag Helper from all of the views in Folder.

Controlling Tag Helper scope with the _ViewImports.cshtml file¶

You can add a _ViewImports.cshtml to any view folder, and the view engine adds the directives from that _ViewImports.cshtml file to those contained in the Views/_ViewImports.cshtml file. If you added an empty Views/Home/_ViewImports.cshtml file for the Home views, there would be no change because the _ViewImports.cshtml file is additive. Any @addTagHelper directives you add to the Views/Home/_ViewImports.cshtml file (that are not in the default Views/_ViewImports.cshtml file) would expose those Tag Helpers to views only in the Home folder.

Opting out of individual elements¶

You can disable a Tag Helper at the element level with the Tag Helper opt-out character (”!”). For example, Email validation is disabled in the <span> with the Tag Helper opt-out character:

<!span asp-validation-for="Email" class="text-danger"></!span>

You must apply the Tag Helper opt-out character to the opening and closing tag. (The Visual Studio editor automatically adds the opt-out character to the closing tag when you add one to the opening tag). After you add the opt-out character, the element and Tag Helper attributes are no longer displayed in a distinctive font.

Using @tagHelperPrefix to make Tag Helper usage explicit¶

The @tagHelperPrefix directive allows you to specify a tag prefix string to enable Tag Helper support and to make Tag Helper usage explicit. In the code image below, the Tag Helper prefix is set to th:, so only those elements using the prefix th: support Tag Helpers (Tag Helper-enabled elements have a distinctive font). The <label> and <input> elements have the Tag Helper prefix and are Tag Helper-enabled, while the <span> element does not.

The same hierarchy rules that apply to @addTagHelper also apply to @tagHelperPrefix.

IntelliSense support for Tag Helpers¶

When you create a new ASP.NET web app in Visual Studio, it adds “Microsoft.AspNetCore.Razor.Tools” to the project.json file. This is the package that adds Tag Helper tooling.

Consider writing an HTML <label> element. As soon as you enter <l in the Visual Studio editor, IntelliSense displays matching elements:

Not only do you get HTML help, but the icon (the “@” symbol with “<>” under it).

identifies the element as targeted by Tag Helpers. Pure HTML elements (such as the fieldset) display the “<>”icon.

A pure HTML <label> tag displays the HTML tag (with the default Visual Studio color theme) in a brown font, the attributes in red, and the attribute values in blue.

After you enter <label, IntelliSense lists the available HTML/CSS attributes and the Tag Helper-targeted attributes:

IntelliSense statement completion allows you to enter the tab key to complete the statement with the selected value:

As soon as a Tag Helper attribute is entered, the tag and attribute fonts change. Using the default Visual Studio “Blue” or “Light” color theme, the font is bold purple. If you’re using the “Dark” theme the font is bold teal. The images in this document were taken using the default theme.

You can enter the Visual Studio CompleteWord shortcut (Ctrl +spacebar is the default) inside the double quotes (“”), and you are now in C#, just like you would be in a C# class. IntelliSense displays all the methods and properties on the page model. The methods and properties are available because the property type is ModelExpression. In the image below, I’m editing the Register view, so the RegisterViewModel is available.

IntelliSense lists the properties and methods available to the model on the page. The rich IntelliSense environment helps you select the CSS class:

Tag Helpers compared to HTML Helpers¶

Tag Helpers attach to HTML elements in Razor views, while HTML Helpers are invoked as methods interspersed with HTML in Razor views. Consider the following Razor markup, which creates an HTML label with the CSS class “caption”:

@Html.Label("FirstName", "First Name:", new {@class="caption"})

The at (@) symbol tells Razor this is the start of code. The next two parameters (“FirstName” and “First Name:”) are strings, so IntelliSense can’t help. The last argument:

new {@class="caption"}

Is an anonymous object used to represent attributes. Because class is a reserved keyword in C#, you use the @ symbol to force C# to interpret “@class=” as a symbol (property name). To a front-end designer (someone familiar with HTML/CSS/JavaScript and other client technologies but not familiar with C# and Razor), most of the line is foreign. The entire line must be authored with no help from IntelliSense.

Using the LabelTagHelper, the same markup can be written as:

With the Tag Helper version, as soon as you enter <l in the Visual Studio editor, IntelliSense displays matching elements:

IntelliSense helps you write the entire line. The LabelTagHelper also defaults to setting the content of the asp-for attribute value (“FirstName”) to “First Name”; It converts camel-cased properties to a sentence composed of the property name with a space where each new upper-case letter occurs. In the following markup:

generates:

<label class="caption" for="FirstName">First Name</label>

The camel-cased to sentence-cased content is not used if you add content to the <label>. For example:

generates:

<label class="caption" for="FirstName">Name First</label>

The following code image shows the Form portion of the Views/Account/Register.cshtml Razor view generated from the legacy ASP.NET 4.5.x MVC template included with Visual Studio 2015.

The Visual Studio editor displays C# code with a grey background. For example, the AntiForgeryToken HTML Helper:

@Html.AntiForgeryToken()

is displayed with a grey background. Most of the markup in the Register view is C#. Compare that to the equivalent approach using Tag Helpers:

The markup is much cleaner and easier to read, edit, and maintain than the HTML Helpers approach. The C# code is reduced to the minimum that the server needs to know about. The Visual Studio editor displays markup targeted by a Tag Helper in a distinctive font.

Consider the Email group:

<div class="form-group">
    <label asp-for="Email" class="col-md-2 control-label"></label>
    <div class="col-md-10">
        <input asp-for="Email" class="form-control" />
        <span asp-validation-for="Email" class="text-danger"></span>
    </div>
</div>

Each of the “asp-” attributes has a value of “Email”, but “Email” is not a string. In this context, “Email” is the C# model expression property for the RegisterViewModel.

The Visual Studio editor helps you write all of the markup in the Tag Helper approach of the register form, while Visual Studio provides no help for most of the code in the HTML Helpers approach. IntelliSense support for Tag Helpers goes into detail on working with Tag Helpers in the Visual Studio editor.

Tag Helpers compared to Web Server Controls¶

Tag Helpers don’t own the element they’re associated with; they simply participate in the rendering of the element and content. ASP.NET Web Server controls are declared and invoked on a page.
Web Server controls have a non-trivial lifecycle that can make developing and debugging difficult.
Web Server controls allow you to add functionality to the client Document Object Model (DOM) elements by using a client control. Tag Helpers have no DOM.
Web Server controls include automatic browser detection. Tag Helpers have no knowledge of the browser.
Multiple Tag Helpers can act on the same element (see Avoiding Tag Helper conflicts ) while you typically can’t compose Web Server controls.
Tag Helpers can modify the tag and content of HTML elements that they’re scoped to, but don’t directly modify anything else on a page. Web Server controls have a less specific scope and can perform actions that affect other parts of your page; enabling unintended side effects.
Web Server controls use type converters to convert strings into objects. With Tag Helpers, you work natively in C#, so you don’t need to do type conversion.
Web Server controls use System.ComponentModel to implement the run-time and design-time behavior of components and controls. System.ComponentModel includes the base classes and interfaces for implementing attributes and type converters, binding to data sources, and licensing components. Contrast that to Tag Helpers, which typically derive from TagHelper, and the TagHelper base class exposes only two methods, Process and ProcessAsync.

Customizing the Tag Helper element font¶

You can customize the font and colorization from Tools > Options > Environment > Fonts and Colors:

Additional Resources¶

Authoring Tag Helpers
Working with Forms (Tag Helpers)
TagHelperSamples on GitHub contains Tag Helper samples for working with Bootstrap.

Authoring Tag Helpers¶

Sections:

Getting started with Tag Helpers
Starting the email Tag Helper
A working email Tag Helper
The bold Tag Helper
Web site information Tag Helper
Condition Tag Helper
Inspecting and retrieving child content

Getting started with Tag Helpers ¶

This tutorial provides an introduction to programming Tag Helpers. Introduction to Tag Helpers describes the benefits that Tag Helpers provide.

A tag helper is any class that implements the ITagHelper interface. However, when you author a tag helper, you generally derive from TagHelper, doing so gives you access to the Process method. We will introduce the TagHelper methods and properties as we use them in this tutorial.

Create a new ASP.NET Core project called AuthoringTagHelpers. You won’t need authentication for this project.
Create a folder to hold the Tag Helpers called TagHelpers. The TagHelpers folder is not required, but it is a reasonable convention. Now let’s get started writing some simple tag helpers.

Starting the email Tag Helper ¶

In this section we will write a tag helper that updates an email tag. For example:

<email>Support</email>

The server will use our email tag helper to convert that markup into the following:

<a href="mailto:Support@contoso.com">Support@contoso.com</a>

That is, an anchor tag that makes this an email link. You might want to do this if you are writing a blog engine and need it to send email for marketing, support, and other contacts, all to the same domain.

Add the following EmailTagHelper class to the TagHelpers folder.

using Microsoft.AspNetCore.Razor.TagHelpers;
using System.Threading.Tasks;

namespace AuthoringTagHelpers.TagHelpers
{
    public class EmailTagHelper : TagHelper
    {
        public override void Process(TagHelperContext context, TagHelperOutput output)
        {
            output.TagName = "a";    // Replaces <email> with <a> tag
        }
    }
}

Notes:

Tag helpers use a naming convention that targets elements of the root class name (minus the TagHelper portion of the class name). In this example, the root name of EmailTagHelper is email, so the <email> tag will be targeted. This naming convention should work for most tag helpers, later on I’ll show how to override it.
The EmailTagHelper class derives from TagHelper. The TagHelper class provides methods and properties for writing Tag Helpers.
The overridden Process method controls what the tag helper does when executed. The TagHelper class also provides an asynchronous version (ProcessAsync) with the same parameters.
The context parameter to Process (and ProcessAsync) contains information associated with the execution of the current HTML tag.
The output parameter to Process (and ProcessAsync) contains a stateful HTML element representative of the original source used to generate an HTML tag and content.
Our class name has a suffix of TagHelper, which is not required, but it’s considered a best practice convention. You could declare the class as:

public class Email : TagHelper

To make the EmailTagHelper class available to all our Razor views, we will add the addTagHelper directive to the Views/_ViewImports.cshtml file:

@using AuthoringTagHelpers
@addTagHelper *, Microsoft.AspNetCore.Mvc.TagHelpers
@addTagHelper "AuthoringTagHelpers.TagHelpers.EmailTagHelper, AuthoringTagHelpers"

The code above uses the wildcard syntax to specify all the tag helpers in our assembly will be available. The first string after @addTagHelper specifies the tag helper to load (we are using “*” for all tag helpers), and the second string “AuthoringTagHelpers” specifies the assembly the tag helper is in. Also, note that the second line brings in the ASP.NET Core MVC tag helpers using the wildcard syntax (those helpers are discussed in Introduction to Tag Helpers.) It’s the @addTagHelper directive that makes the tag helper available to the Razor view. Alternatively, you can provide the fully qualified name (FQN) of a tag helper as shown below:

@using AuthoringTagHelpers
@addTagHelper *, Microsoft.AspNetCore.Mvc.TagHelpers
@addTagHelper "AuthoringTagHelpers.TagHelpers3.EmailTagHelper, AuthoringTagHelpers"

To add a tag helper to a view using a FQN, you first add the FQN (AuthoringTagHelpers.TagHelpers.EmailTagHelper), and then the assembly name (AuthoringTagHelpers). Most developers will prefer to use the wildcard syntax. Introduction to Tag Helpers goes into detail on tag helper adding, removing, hierarchy, and wildcard syntax.

Update the markup in the Views/Home/Contact.cshtml file with these changes:

@{
    ViewData["Title"] = "Contact";
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<address>
    One Microsoft Way<br />
    Redmond, WA 98052<br />
    <abbr title="Phone">P:</abbr>
    425.555.0100
</address>

<address>
    <strong>Support:</strong><email>Support</email><br />
    <strong>Marketing:</strong><email>Marketing</email>
</address>

Run the app and use your favorite browser to view the HTML source so you can verify that the email tags are replaced with anchor markup (For example, <a>Support</a>). Support and Marketing are rendered as a links, but they don’t have an href attribute to make them functional. We’ll fix that in the next section.

Note

Like HTML tags and attributes, tags, class names and attributes in Razor, and C# are not case-sensitive.

A working email Tag Helper ¶

In this section, we’ll update the EmailTagHelper so that it will create a valid anchor tag for email. We’ll update it to take information from a Razor view (in the form of a mail-to attribute) and use that in generating the anchor.

Update the EmailTagHelper class with the following:

public class EmailTagHelper : TagHelper
{
    private const string EmailDomain = "contoso.com";

    // Can be passed via <email mail-to="..." />. 
    // Pascal case gets translated into lower-kebab-case.
    public string MailTo { get; set; }

    public override void Process(TagHelperContext context, TagHelperOutput output)
    {
        output.TagName = "a";    // Replaces <email> with <a> tag

        var address = MailTo + "@" + EmailDomain;
        output.Attributes.SetAttribute("href", "mailto:" + address);
        output.Content.SetContent(address);
    }
}

Notes:

Pascal-cased class and property names for tag helpers are translated into their lower kebab case. Therefore, to use the MailTo attribute, you’ll use <email mail-to="value"/> equivalent.
The last line sets the completed content for our minimally functional tag helper.
The highlighted line shows the syntax for adding attributes:

public override void Process(TagHelperContext context, TagHelperOutput output)
{
    output.TagName = "a";    // Replaces <email> with <a> tag

    var address = MailTo + "@" + EmailDomain;
    output.Attributes.SetAttribute("href", "mailto:" + address);
    output.Content.SetContent(address);
}

That approach works for the attribute “href” as long as it doesn’t currently exist in the attributes collection. You can also use the output.Attributes.Add method to add a tag helper attribute to the end of the collection of tag attributes.

Update the markup in the Views/Home/Contact.cshtml file with these changes:

@{
    ViewData["Title"] = "Contact Copy";
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<address>
    One Microsoft Way Copy Version <br />
    Redmond, WA 98052-6399<br />
    <abbr title="Phone">P:</abbr>
    425.555.0100
</address>

<address>
    <strong>Support:</strong><email mail-to="Support"></email><br />
    <strong>Marketing:</strong><email mail-to="Marketing"></email>
</address>

Run the app and verify that it generates the correct links.

Note: If you were to write the email tag self-closing (<email mail-to="Rick" />), the final output would also be self-closing. To enable the ability to write the tag with only a start tag (<email mail-to="Rick">) you must decorate the class with the following:

[HtmlTargetElement("email", TagStructure = TagStructure.WithoutEndTag)] 
public class EmailVoidTagHelper : TagHelper
{
    private const string EmailDomain = "contoso.com";
    // Code removed for brevity

With a self-closing email tag helper, the output would be <a href="mailto:Rick@contoso.com" />. Self-closing anchor tags are not valid HTML, so you wouldn’t want to create one, but you might want to create a tag helper that is self-closing. Tag helpers set the type of the TagMode property after reading a tag.

An asynchronous email helper¶

In this section we’ll write an asynchronous email helper.

Replace the EmailTagHelper class with the following code:

public class EmailTagHelper : TagHelper
{
    private const string EmailDomain = "contoso.com";
    public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
    {
        output.TagName = "a";                                 // Replaces <email> with <a> tag
        var content = await output.GetChildContentAsync();
        var target = content.GetContent() + "@" + EmailDomain;
        output.Attributes.SetAttribute("href", "mailto:" + target);
        output.Content.SetContent(target);
    }
}

Notes:

This version uses the asynchronous ProcessAsync method. The asynchronous GetChildContentAsync returns a Task containing the TagHelperContent.
We use the output parameter to get contents of the HTML element.

Make the following change to the Views/Home/Contact.cshtml file so the tag helper can get the target email.

@{
    ViewData["Title"] = "Contact";
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<address>
    One Microsoft Way<br />
    Redmond, WA 98052<br />
    <abbr title="Phone">P:</abbr>
    425.555.0100
</address>

<address>
    <strong>Support:</strong><email>Support</email><br />
    <strong>Marketing:</strong><email>Marketing</email>
</address>

Run the app and verify that it generates valid email links.

The bold Tag Helper ¶

Add the following BoldTagHelper class to the TagHelpers folder.

using Microsoft.AspNetCore.Razor.TagHelpers;

namespace AuthoringTagHelpers.TagHelpers
{
    [HtmlTargetElement(Attributes = "bold")]
    public class BoldTagHelper : TagHelper
    {
        public override void Process(TagHelperContext context, TagHelperOutput output)
        {
            output.Attributes.RemoveAll("bold");
            output.PreContent.SetHtmlContent("<strong>");
            output.PostContent.SetHtmlContent("</strong>");
        }
    }
}

/*
 * public IActionResult About()
{
    ViewData["Message"] = "Your application description page.";
    return View("AboutBoldOnly");
    // return View();
}
*/

Notes:

The [HtmlTargetElement] attribute passes an attribute parameter that specifies that any HTML element that contains an HTML attribute named “bold” will match, and the Process override method in the class will run. In our sample, the Process method removes the “bold” attribute and surrounds the containing markup with <strong></strong>.
Because we don’t want to replace the existing tag content, we must write the opening <strong> tag with the PreContent.SetHtmlContent method and the closing </strong> tag with the PostContent.SetHtmlContent method.

Modify the About.cshtml view to contain a bold attribute value. The completed code is shown below.

@{
    ViewData["Title"] = "About";
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<p bold>Use this area to provide additional information.</p>

<bold> Is this bold?</bold>

Run the app. You can use your favorite browser to inspect the source and verify the markup.

The [HtmlTargetElement] attribute above only targets HTML markup that provides an attribute name of “bold”. The <bold> element was not modified by the tag helper.

Comment out the [HtmlTargetElement] attribute line and it will default to targeting <bold> tags, that is, HTML markup of the form <bold>. Remember, the default naming convention will match the class name BoldTagHelper to <bold> tags.
Run the app and verify that the <bold> tag is processed by the tag helper.

Decorating a class with multiple [HtmlTargetElement] attributes results in a logical-OR of the targets. For example, using the code below, a bold tag or a bold attribute will match.

[HtmlTargetElement("bold")]
[HtmlTargetElement(Attributes = "bold")]
public class BoldTagHelper : TagHelper
{
    public override void Process(TagHelperContext context, TagHelperOutput output)
    {
        output.Attributes.RemoveAll("bold");
        output.PreContent.SetHtmlContent("<strong>");
        output.PostContent.SetHtmlContent("</strong>");
    }
}

When multiple attributes are added to the same statement, the runtime treats them as a logical-AND. For example, in the code below, an HTML element must be named “bold” with an attribute named “bold” ( <bold bold /> ) to match.

[HtmlTargetElement("bold", Attributes = "bold")]

You can also use the [HtmlTargetElement] to change the name of the targeted element. For example if you wanted the BoldTagHelper to target <MyBold> tags, you would use the following attribute:

[HtmlTargetElement("MyBold")]

Web site information Tag Helper ¶

Add a Models folder.
Add the following WebsiteContext class to the Models folder:

using System;

namespace AuthoringTagHelpers.Models
{
    public class WebsiteContext
    {
        public Version Version { get; set; }
        public int CopyrightYear { get; set; }
        public bool Approved { get; set; }
        public int TagsToShow { get; set; }
    }
}

Add the following WebsiteInformationTagHelper class to the TagHelpers folder.

using System;
using AuthoringTagHelpers.Models;
using Microsoft.AspNetCore.Razor.TagHelpers;

namespace AuthoringTagHelpers.TagHelpers
{
    public class WebsiteInformationTagHelper : TagHelper
    {
        public WebsiteContext Info { get; set; }

      public override void Process(TagHelperContext context, TagHelperOutput output)
      {
         output.TagName = "section";
         output.Content.SetHtmlContent(
$@"<ul><li><strong>Version:</strong> {Info.Version}</li>
<li><strong>Copyright Year:</strong> {Info.CopyrightYear}</li>
<li><strong>Approved:</strong> {Info.Approved}</li>
<li><strong>Number of tags to show:</strong> {Info.TagsToShow}</li></ul>");
         output.TagMode = TagMode.StartTagAndEndTag;
      }
   }
}

Notes:

As mentioned previously, tag helpers translates Pascal-cased C# class names and properties for tag helpers into lower kebab case. Therefore, to use the WebsiteInformationTagHelper in Razor, you’ll write <website-information />.
We are not explicitly identifying the target element with the [HtmlTargetElement] attribute, so the default of website-information will be targeted. If you applied the following attribute (note it’s not kebab case but matches the class name):

[HtmlTargetElement("WebsiteInformation")]

The lower kebab case tag <website-information /> would not match. If you want use the [HtmlTargetElement] attribute, you would use kebab case as shown below:

[HtmlTargetElement("Website-Information")]

Elements that are self-closing have no content. For this example, the Razor markup will use a self-closing tag, but the tag helper will be creating a section element (which is not self-closing and we are writing content inside the section element). Therefore, we need to set TagMode to StartTagAndEndTag to write output. Alternatively, you can comment out the line setting TagMode and write markup with a closing tag. (Example markup is provided later in this tutorial.)
The $ (dollar sign) in the following line uses an interpolated string:

$@"<ul><li><strong>Version:</strong> {Info.Version}</li>

Add the following markup to the About.cshtml view. The highlighted markup displays the web site information.

@using AuthoringTagHelpers.Models
@{
    ViewData["Title"] = "About";
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<p bold>Use this area to provide additional information.</p>

<bold> Is this bold?</bold>

<h3> web site info </h3>
<website-information info="new WebsiteContext {
                                    Version = new Version(1, 3),
                                    CopyrightYear = 1638,
                                    Approved = true,
                                    TagsToShow = 131 }" />

Note: In the Razor markup shown below:

<website-information info="new WebsiteContext {
                                    Version = new Version(1, 3),
                                    CopyrightYear = 1638,
                                    Approved = true,
                                    TagsToShow = 131 }" />

Razor knows the info attribute is a class, not a string, and you want to write C# code. Any non-string tag helper attribute should be written without the @ character.

Run the app, and navigate to the About view to see the web site information.

Note:

You can use the following markup with a closing tag and remove the line with TagMode.StartTagAndEndTag in the tag helper:

<website-information info="new WebsiteContext {
                                    Version = new Version(1, 3),
                                    CopyrightYear = 1638,
                                    Approved = true,
                                    TagsToShow = 131 }" >
</website-information>

Condition Tag Helper ¶

The condition tag helper renders output when passed a true value.

Add the following ConditionTagHelper class to the TagHelpers folder.

using Microsoft.AspNetCore.Razor.TagHelpers;

namespace AuthoringTagHelpers.TagHelpers
{
    [HtmlTargetElement(Attributes = nameof(Condition))]
    public class ConditionTagHelper : TagHelper
    {
        public bool Condition { get; set; }

        public override void Process(TagHelperContext context, TagHelperOutput output)
        {
            if (!Condition)
            {
                output.SuppressOutput();
            }
        }
    }
}

Replace the contents of the Views/Home/Index.cshtml file with the following markup:

@using AuthoringTagHelpers.Models
@model WebsiteContext

@{
    ViewData["Title"] = "Home Page";
}

<div>
    <h3>Information about our website (outdated):</h3>
    <Website-InforMation info=Model />
    <div condition="Model.Approved">
        <p>
            This website has <strong surround="em"> @Model.Approved </strong> been approved yet.
            Visit www.contoso.com for more information.
        </p>
    </div>
</div>

Replace the Index method in the Home controller with the following code:

  public IActionResult Index(bool approved = false)
  {
      return View(new WebsiteContext
      {
          Approved = approved,
          CopyrightYear = 2015,
          Version = new Version(1, 3, 3, 7),
          TagsToShow = 20
      });
  }

Run the app and browse to the home page. The markup in the conditional div will not be rendered. Append the query string ?approved=true to the URL (for example, http://localhost:1235/Home/Index?approved=true). approved is set to true and the conditional markup will be displayed.

Note: We use the nameof operator to specify the attribute to target rather than specifying a string as we did with the bold tag helper:

[HtmlTargetElement(Attributes = nameof(Condition))]
 //   [HtmlTargetElement(Attributes = "condition")]
 public class ConditionTagHelper : TagHelper
{
   public bool Condition { get; set; }

   public override void Process(TagHelperContext context, TagHelperOutput output)
   {
      if (!Condition)
      {
         output.SuppressOutput();
      }
   }
}

The nameof operator will protect the code should it ever be refactored (we might want to change the name to RedCondition).

Avoiding Tag Helper conflicts¶

In this section, we will write a pair of auto-linking tag helpers. The first will replace markup containing a URL starting with HTTP to an HTML anchor tag containing the same URL (and thus yielding a link to the URL). The second will do the same for a URL starting with WWW.

Because these two helpers are closely related and we may refactor them in the future, we’ll keep them in the same file.

Add the following AutoLinkerHttpTagHelper class to the TagHelpers folder.

[HtmlTargetElement("p")]
public class AutoLinkerHttpTagHelper : TagHelper
{
    public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
    {
        var childContent = await output.GetChildContentAsync();
        // Find Urls in the content and replace them with their anchor tag equivalent.
        output.Content.SetHtmlContent(Regex.Replace(
             childContent.GetContent(),
             @"\b(?:https?://)(\S+)\b",
              "<a target=\"_blank\" href=\"$0\">$0</a>"));  // http link version}
    }
}

Notes: The AutoLinkerHttpTagHelper class targets p elements and uses Regex to create the anchor.

Add the following markup to the end of the Views/Home/Contact.cshtml file:

@{
    ViewData["Title"] = "Contact";
}
<h2>@ViewData["Title"].</h2>
<h3>@ViewData["Message"]</h3>

<address>
    One Microsoft Way<br />
    Redmond, WA 98052<br />
    <abbr title="Phone">P:</abbr>
    425.555.0100
</address>

<address>
    <strong>Support:</strong><email>Support</email><br />
    <strong>Marketing:</strong><email>Marketing</email>
</address>

<p>Visit us at http://docs.asp.net or at www.microsoft.com</p>

Run the app and verify that the tag helper renders the anchor correctly.
Update the AutoLinker class to include the AutoLinkerWwwTagHelper which will convert www text to an anchor tag that also contains the original www text. The updated code is highlighted below:

[HtmlTargetElement("p")]
public class AutoLinkerHttpTagHelper : TagHelper
{
    public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
    {
        var childContent = await output.GetChildContentAsync();
        // Find Urls in the content and replace them with their anchor tag equivalent.
        output.Content.SetHtmlContent(Regex.Replace(
             childContent.GetContent(),
             @"\b(?:https?://)(\S+)\b",
              "<a target=\"_blank\" href=\"$0\">$0</a>"));  // http link version}
    }
}

[HtmlTargetElement("p")]
public class AutoLinkerWwwTagHelper : TagHelper
{
    public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
    {
        var childContent = await output.GetChildContentAsync();
        // Find Urls in the content and replace them with their anchor tag equivalent.
        output.Content.SetHtmlContent(Regex.Replace(
            childContent.GetContent(),
             @"\b(www\.)(\S+)\b",
             "<a target=\"_blank\" href=\"http://$0\">$0</a>"));  // www version
    }
}

Run the app. Notice the www text is rendered as a link but the HTTP text is not. If you put a break point in both classes, you can see that the HTTP tag helper class runs first. The problem is that the tag helper output is cached, and when the WWW tag helper is run, it overwrites the cached output from the HTTP tag helper. Later in the tutorial we’ll see how to control the order that tag helpers run in. We’ll fix the code with the following:

    public class AutoLinkerHttpTagHelper : TagHelper
    {
        public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
        {
            var childContent = output.Content.IsModified ? output.Content.GetContent() :
                (await output.GetChildContentAsync()).GetContent();

            // Find Urls in the content and replace them with their anchor tag equivalent.
            output.Content.SetHtmlContent(Regex.Replace(
                 childContent,
                 @"\b(?:https?://)(\S+)\b",
                  "<a target=\"_blank\" href=\"$0\">$0</a>"));  // http link version}
        }
    }

    [HtmlTargetElement("p")]
    public class AutoLinkerWwwTagHelper : TagHelper
    {
        public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
        {
            var childContent = output.Content.IsModified ? output.Content.GetContent() : 
                (await output.GetChildContentAsync()).GetContent();
  
            // Find Urls in the content and replace them with their anchor tag equivalent.
            output.Content.SetHtmlContent(Regex.Replace(
                 childContent,
                 @"\b(www\.)(\S+)\b",
                 "<a target=\"_blank\" href=\"http://$0\">$0</a>"));  // www version
        }
    }
}

Note: In the first edition of the auto-linking tag helpers, we got the content of the target with the following code:

var childContent = await output.GetChildContentAsync();

That is, we call GetChildContentAsync using the TagHelperOutput passed into the ProcessAsync method. As mentioned previously, because the output is cached, the last tag helper to run wins. We fixed that problem with the following code:

var childContent = output.Content.IsModified ? output.Content.GetContent() :
    (await output.GetChildContentAsync()).GetContent();

The code above checks to see if the content has been modified, and if it has, it gets the content from the output buffer.

Run the app and verify that the two links work as expected. While it might appear our auto linker tag helper is correct and complete, it has a subtle problem. If the WWW tag helper runs first, the www links will not be correct. Update the code by adding the Order overload to control the order that the tag runs in. The Order property determines the execution order relative to other tag helpers targeting the same element. The default order value is zero and instances with lower values are executed first.

public class AutoLinkerHttpTagHelper : TagHelper
{
    // This filter must run before the AutoLinkerWwwTagHelper as it searches and replaces http and 
    // the AutoLinkerWwwTagHelper adds http to the markup.
    public override int Order
    {
        get  {  return int.MinValue;   }
    }

The above code will guarantee that the HTTP tag helper runs before the WWW tag helper. Change Order to MaxValue and verify that the markup generated for the WWW tag is incorrect.

Inspecting and retrieving child content ¶

The tag-helpers provide several properties to retrieve content.

The result of GetChildContentAsync can be appended to output.Content.
You can inspect the result of GetChildContentAsync with GetContent.
If you modify output.Content, the TagHelper body will not be executed or rendered unless you call GetChildContentAsync as in our auto-linker sample:

public class AutoLinkerHttpTagHelper : TagHelper
{
    public override async Task ProcessAsync(TagHelperContext context, TagHelperOutput output)
    {
        var childContent = output.Content.IsModified ? output.Content.GetContent() :
            (await output.GetChildContentAsync()).GetContent();

        // Find Urls in the content and replace them with their anchor tag equivalent.
        output.Content.SetHtmlContent(Regex.Replace(
             childContent,
             @"\b(?:https?://)(\S+)\b",
              "<a target=\"_blank\" href=\"$0\">$0</a>"));  // http link version}
    }
}

Multiple calls to GetChildContentAsync will return the same value and will not re-execute the TagHelper body unless you pass in a false parameter indicating not use the cached result.

Partial Views¶

ASP.NET Core MVC supports partial views, which are useful when you have reusable parts of web pages you want to share between different views.

Sections:

What are Partial Views?
When Should I Use Partial Views?
Declaring Partial Views
Referencing a Partial View
Accessing Data From Partial Views

What are Partial Views?¶

A partial view is a view that is rendered within another view. The HTML output generated by executing the partial view is rendered into the calling (or parent) view. Like views, partial views use the .cshtml file extension.

Note

If you’re coming from an ASP.NET Web Forms background, partial views are similar to user controls.

When Should I Use Partial Views?¶

Partial views are an effective way of breaking up large views into smaller components. They can reduce duplication of view content and allow view elements to be reused. Common layout elements should be specified in _Layout.cshtml. Non-layout reusable content can be encapsulated into partial views.

If you have a complex page made up of several logical pieces, it can be helpful to work with each piece as its own partial view. Each piece of the page can be viewed in isolation from the rest of the page, and the view for the page itself becomes much simpler since it only contains the overall page structure and calls to render the partial views.

Tip

Follow the Don’t Repeat Yourself Principle in your views.

Declaring Partial Views ¶

Partial views are created like any other view: you create a .cshtml file within the Views folder. There is no semantic difference between a partial view and a regular view - they are just rendered differently. You can have a view that is returned directly from a controller’s ViewResult, and the same view can be used as a partial view. The main difference between how a view and a partial view are rendered is that partial views do not run _ViewStart.cshtml (while views do - learn more about _ViewStart.cshtml in Layout).

Referencing a Partial View ¶

From within a view page, there are several ways in which you can render a partial view. The simplest is to use Html.Partial, which returns an IHtmlString and can be referenced by prefixing the call with @:

@Html.Partial("AuthorPartial")

The PartialAsync method is available for partial views containing asynchronous code (although code in views is generally discouraged):

@await Html.PartialAsync("AuthorPartial")

You can render a partial view with RenderPartial. This method doesn’t return a result; it streams the rendered output directly to the response. Because it doesn’t return a result, it must be called within a Razor code block (you can also call RenderPartialAsync if necessary):

@{
    Html.RenderPartial("AuthorPartial");
}

Because it streams the result directly, RenderPartial and RenderPartialAsync may perform better in some scenarios. However, in most cases it’s recommended you use Partial and PartialAsync.

Note

If your views need to execute code, the recommended pattern is to use a view component instead of a partial view.

Partial View Discovery¶

When referencing a partial view, you can refer to its location in several ways:

// Uses a view in current folder with this name
// If none is found, searches the Shared folder
@Html.Partial("ViewName")

// A view with this name must be in the same folder
@Html.Partial("ViewName.cshtml")

// Locate the view based on the application root
// Paths that start with "/" or "~/" refer to the application root
@Html.Partial("~/Views/Folder/ViewName.cshtml")
@Html.Partial("/Views/Folder/ViewName.cshtml")

// Locate the view using relative paths
@Html.Partial("../Account/LoginPartial.cshtml")

If desired, you can have different partial views with the same name in different view folders. When referencing the views by name (without file extension), views in each folder will use the partial view in the same folder with them. You can also specify a default partial view to use, placing it in the Shared folder. This view will be used by any views that don’t have their own copy of the partial view in their folder. In this way, you can have a default partial view (in Shared), which can be overridden by a partial view with the same name in the same folder as the parent view.

Partial views can be chained. That is, a partial view can call another partial view (as long as you don’t create a loop). Within each view or partial view, relative paths are always relative to that view, not the root or parent view.

Note

If you declare a Razor section in a partial view, it will not be visible to its parent(s); it will be limited to the partial view.

Accessing Data From Partial Views ¶

When a partial view is instantiated, it gets a copy of the parent view’s ViewData dictionary. Updates made to the data within the partial view are not persisted to the parent view. ViewData changed in a partial view is lost when the partial view returns.

You can pass an instance of ViewDataDictionary to the partial view:

@Html.Partial("PartialName", customViewData)

You can also pass a model into a partial view. This can be the page’s view model, or some portion of it, or a custom object. Simply pass in the model as the second parameter when calling Partial/PartialAsync or RenderPartial/RenderPartialAsync:

@Html.Partial("PartialName", viewModel)

You can pass an instance of ViewDataDictionary and a view model to a partial view:

@Html.Partial("PartialName", viewModel, customViewData)

An Example¶

The following view specifies a view model of type Article. Article has an AuthorName property that is passed to a partial view named AuthorPartial, and a property of type List<ArticleSection>, which is passed (in a loop) to a partial devoted to rendering that type:

@using PartialViewsSample.ViewModels
@model Article

<h2>@Model.Title</h2>
@Html.Partial("AuthorPartial", Model.AuthorName)
@Model.PublicationDate

@foreach (var section in @Model.Sections)
{
    @Html.Partial("ArticleSection", section)
}

The AuthorPartial (which in this case is in the /Views/Shared folder):

@model string
<div>
    <h3>@Model</h3>
    This partial view came from /Views/Shared/AuthorPartial.cshtml.<br/>
</div>

The ArticleSection partial:

@using PartialViewsSample.ViewModels
@model ArticleSection

<h3>@Model.Title</h3>
<div>
    @Model.Content
</div>

At runtime, the partials are rendered into the parent view, which itself is rendered within the shared _Layout.cshtml, resulting in output like this:

Injecting Services Into Views¶

ASP.NET Core supports dependency injection into views. This can be useful for view-specific services, such as localization or data required only for populating view elements. You should try to maintain separation of concerns between your controllers and views. Most of the data your views display should be passed in from the controller.

Sections:

A Simple Example
Populating Lookup Data
Overriding Services
See Also

A Simple Example ¶

You can inject a service into a view using the @inject directive. You can think of @inject as adding a property to your view, and populating the property using DI.

The syntax for @inject:: @inject <type> <name>

An example of @inject in action:

@using System.Threading.Tasks
@using ViewInjectSample.Model
@using ViewInjectSample.Model.Services
@model IEnumerable<ToDoItem>
@inject StatisticsService StatsService
<!DOCTYPE html>
<html>
<head>
    <title>To Do Items</title>
</head>
<body>
    <div>
        <h1>To Do Items</h1>
        <ul>
            <li>Total Items: @StatsService.GetCount()</li>
            <li>Completed: @StatsService.GetCompletedCount()</li>
            <li>Avg. Priority: @StatsService.GetAveragePriority()</li>
        </ul>
        <table>
            <tr>
                <th>Name</th>
                <th>Priority</th>
                <th>Is Done?</th>
            </tr>
            @foreach (var item in Model)
            {
                <tr>
                    <td>@item.Name</td>
                    <td>@item.Priority</td>
                    <td>@item.IsDone</td>
                </tr>
            }
        </table>
    </div>
</body>
</html>

This view displays a list of ToDoItem instances, along with a summary showing overall statistics. The summary is populated from the injected StatisticsService. This service is registered for dependency injection in ConfigureServices in Startup.cs:

// For more information on how to configure your application, visit http://go.microsoft.com/fwlink/?LinkID=398940
public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc();

    services.AddTransient<IToDoItemRepository, ToDoItemRepository>();
    services.AddTransient<StatisticsService>();
    services.AddTransient<ProfileOptionsService>();

The StatisticsService performs some calculations on the set of ToDoItem instances, which it accesses via a repository:

using System.Linq;
using ViewInjectSample.Interfaces;

namespace ViewInjectSample.Model.Services
{
    public class StatisticsService
    {
        private readonly IToDoItemRepository _toDoItemRepository;

        public StatisticsService(IToDoItemRepository toDoItemRepository)
        {
            _toDoItemRepository = toDoItemRepository;
        }

        public int GetCount()
        {
            return _toDoItemRepository.List().Count();
        }

        public int GetCompletedCount()
        {
            return _toDoItemRepository.List().Count(x => x.IsDone);
        }

        public double GetAveragePriority()
        {
            if (_toDoItemRepository.List().Count() == 0)
            {
                return 0.0;
            }

            return _toDoItemRepository.List().Average(x => x.Priority);
        }
    }
}

The sample repository uses an in-memory collection. The implementation shown above (which operates on all of the data in memory) is not recommended for large, remotely accessed data sets.

The sample displays data from the model bound to the view and the service injected into the view:

Populating Lookup Data ¶

View injection can be useful to populate options in UI elements, such as dropdown lists. Consider a user profile form that includes options for specifying gender, state, and other preferences. Rendering such a form using a standard MVC approach would require the controller to request data access services for each of these sets of options, and then populate a model or ViewBag with each set of options to be bound.

An alternative approach injects services directly into the view to obtain the options. This minimizes the amount of code required by the controller, moving this view element construction logic into the view itself. The controller action to display a profile editing form only needs to pass the form the profile instance:

using Microsoft.AspNetCore.Mvc;
using ViewInjectSample.Model;

namespace ViewInjectSample.Controllers
{
    public class ProfileController : Controller
    {
        [Route("Profile")]
        public IActionResult Index()
        {
            // TODO: look up profile based on logged-in user
            var profile = new Profile()
            {
                Name = "Steve",
                FavColor = "Blue",
                Gender = "Male",
                State = new State("Ohio","OH")
            };
            return View(profile);
        }
    }
}

The HTML form used to update these preferences includes dropdown lists for three of the properties:

These lists are populated by a service that has been injected into the view:

@using System.Threading.Tasks
@using ViewInjectSample.Model.Services
@model ViewInjectSample.Model.Profile
@inject ProfileOptionsService Options
<!DOCTYPE html>
<html>
<head>
    <title>Update Profile</title>
</head>
<body>
<div>
    <h1>Update Profile</h1>
    Name: @Html.TextBoxFor(m => m.Name)
    <br/>
    Gender: @Html.DropDownList("Gender",
           Options.ListGenders().Select(g => 
                new SelectListItem() { Text = g, Value = g }))
    <br/>

    State: @Html.DropDownListFor(m => m.State.Code,
           Options.ListStates().Select(s => 
                new SelectListItem() { Text = s.Name, Value = s.Code}))
    <br />

    Fav. Color: @Html.DropDownList("FavColor",
           Options.ListColors().Select(c => 
                new SelectListItem() { Text = c, Value = c }))
    </div>
</body>
</html>

The ProfileOptionsService is a UI-level service designed to provide just the data needed for this form:

using System.Collections.Generic;

namespace ViewInjectSample.Model.Services
{
    public class ProfileOptionsService
    {
        public List<string> ListGenders()
        {
            // keeping this simple
            return new List<string>() {"Female", "Male"};
        }

        public List<State> ListStates()
        {
            // a few states from USA
            return new List<State>()
            {
                new State("Alabama", "AL"),
                new State("Alaska", "AK"),
                new State("Ohio", "OH")
            };
        }

        public List<string> ListColors()
        {
            return new List<string>() { "Blue","Green","Red","Yellow" };
        }
    }
}

Tip

Don’t forget to register types you will request through dependency injection in the ConfigureServices method in Startup.cs.

Overriding Services ¶

In addition to injecting new services, this technique can also be used to override previously injected services on a page. The figure below shows all of the fields available on the page used in the first example:

As you can see, the default fields include Html, Component, and Url (as well as the StatsService that we injected). If for instance you wanted to replace the default HTML Helpers with your own, you could easily do so using @inject:

@using System.Threading.Tasks
@using ViewInjectSample.Helpers
@inject MyHtmlHelper Html
<!DOCTYPE html>
<html>
<head>
    <title>My Helper</title>
</head>
<body>
    <div>
        Test: @Html.Value
    </div>
</body>
</html>

If you want to extend existing services, you can simply use this technique while inheriting from or wrapping the existing implementation with your own.

View Components¶

Sections:

Introducing view components
Creating a view component
Invoking a view component
Walkthrough: Creating a simple view component
Additional Resources

Introducing view components ¶

New to ASP.NET Core MVC, view components are similar to partial views, but they are much more powerful. View components don’t use model binding, and only depend on the data you provide when calling into it. A view component:

Renders a chunk rather than a whole response
Includes the same separation-of-concerns and testability benefits found between a controller and view
Can have parameters and business logic
Is typically invoked from a layout page

View Components are intended anywhere you have reusable rendering logic that is too complex for a partial view, such as:

Dynamic navigation menus
Tag cloud (where it queries the database)
Login panel
Shopping cart
Recently published articles
Sidebar content on a typical blog
A login panel that would be rendered on every page and show either the links to log out or log in, depending on the log in state of the user

A view component consists of two parts, the class (typically derived from ViewComponent) and the result it returns (typically a view). Like controllers, a view component can be a POCO, but most developers will want to take advantage of the methods and properties available by deriving from ViewComponent.

Creating a view component ¶

This section contains the high level requirements to create a view component. Later in the article we’ll examine each step in detail and create a view component.

The view component class¶

A view component class can be created by any of the following:

Deriving from ViewComponent
Decorating a class with the [ViewComponent] attribute, or deriving from a class with the [ViewComponent] attribute
Creating a class where the name ends with the suffix ViewComponent

Like controllers, view components must be public, non-nested, and non-abstract classes. The view component name is the class name with the “ViewComponent” suffix removed. It can also be explicitly specified using the ViewComponentAttribute.Name property.

A view component class:

Fully supports constructor dependency injection
Does not take part in the controller lifecycle, which means you can’t use filters in a view component

View component methods¶

A view component defines its logic in an InvokeAsync method that returns an IViewComponentResult. Parameters come directly from invocation of the view component, not from model binding. A view component never directly handles a request. Typically a view component initializes a model and passes it to a view by calling the View method. In summary, view component methods:

Define an InvokeAsync` method that returns an IViewComponentResult
Typically initializes a model and passes it to a view by calling the ViewComponent View method
Parameters come from the calling method, not HTTP, there is no model binding
Are not reachable directly as an HTTP endpoint, they are invoked from your code (usually in a view). A view component never handles a request
Are overloaded on the signature rather than any details from the current HTTP request

View search path¶

The runtime searches for the view in the following paths:

Views/<controller_name>/Components/<view_component_name>/<view_name>

Views/Shared/Components/<view_component_name>/<view_name>

The default view name for a view component is Default, which means your view file will typically be named Default.cshtml. You can specify a different view name when creating the view component result or when calling the View method.

We recommend you name the view file Default.cshtml and use the Views/Shared/Components/<view_component_name>/<view_name> path. The PriorityList view component used in this sample uses Views/Shared/Components/PriorityList/Default.cshtml for the view component view.

Invoking a view component ¶

To use the view component, call @Component.InvokeAsync("Name of view component", <anonymous type containing parameters>) from a view. The parameters will be passed to the InvokeAsync method. The PriorityList view component developed in the article is invoked from the Views/Todo/Index.cshtml view file. In the following, the InvokeAsync method is called with two parameters:

@await Component.InvokeAsync("PriorityList", new { maxPriority = 2, isDone = false })

Invoking a view component directly from a controller¶

View components are typically invoked from a view, but you can invoke them directly from a controller method. While view components do not define endpoints like controllers, you can easily implement a controller action that returns the content of a ViewComponentResult.

In this example, the view component is called directly from the controller:

  public IActionResult IndexVC()
  {
      return ViewComponent("PriorityList", new { maxPriority = 3, isDone = false });
  }

Walkthrough: Creating a simple view component ¶

Download, build and test the starter code. It’s a simple project with a Todo controller that displays a list of Todo items.

Add a ViewComponent class¶

Create a ViewComponents folder and add the following PriorityListViewComponent class.

using Microsoft.AspNet.Mvc;
using Microsoft.Data.Entity;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using ViewComponentSample.Models;

namespace ViewComponentSample.ViewComponents
{
    public class PriorityListViewComponent : ViewComponent
    {
        private readonly ToDoContext db;

        public PriorityListViewComponent(ToDoContext context)
        {
            db = context;
        }

        public async Task<IViewComponentResult> InvokeAsync(
        int maxPriority, bool isDone)
        {
            var items = await GetItemsAsync(maxPriority, isDone);
            return View(items);
        }
        private Task<List<TodoItem>> GetItemsAsync(int maxPriority, bool isDone)
        {
            return db.ToDo.Where(x => x.IsDone == isDone &&
                                 x.Priority <= maxPriority).ToListAsync();
        }       
    }
}

Notes on the code:

View component classes can be contained in any folder in the project.
Because the class name PriorityListViewComponent ends with the suffix ViewComponent, the runtime will use the string “PriorityList” when referencing the class component from a view. I’ll explain that in more detail later.
The [ViewComponent] attribute can change the name used to reference a view component. For example, we could have named the class XYZ, and applied the ViewComponent attribute:
```
[ViewComponent(Name = "PriorityList")]
public class XYZ : ViewComponent
```
The [ViewComponent] attribute above tells the view component selector to use the name PriorityList when looking for the views associated with the component, and to use the string “PriorityList” when referencing the class component from a view. I’ll explain that in more detail later.
The component uses dependency injection to make the data context available.
InvokeAsync exposes a method which can be called from a view, and it can take an arbitrary number of arguments.
The InvokeAsync method returns the set of ToDo items that are not completed and have priority lower than or equal to maxPriority.

Create the view component Razor view¶

Create the Views/Shared/Components folder. This folder must be named Components.
Create the Views/Shared/Components/PriorityList folder. This folder name must match the name of the view component class, or the name of the class minus the suffix (if we followed convention and used the ViewComponent suffix in the class name). If you used the ViewComponent attribute, the class name would need to match the attribute designation.
Create a Views/Shared/Components/PriorityList/Default.cshtml Razor view.

@model IEnumerable<ViewComponentSample.Models.TodoItem>

<h3>Priority Items</h3>
<ul>
    @foreach (var todo in Model)
    {
        <li>@todo.Name</li>
    }
</ul>

The Razor view takes a list of TodoItem and displays them. If the view component InvokeAsync method doesn’t pass the name of the view (as in our sample), Default is used for the view name by convention. Later in the tutorial, I’ll show you how to pass the name of the view. To override the default styling for a specific controller, add a view to the controller specific view folder (for example Views/Todo/Components/PriorityList/Default.cshtml).

If the view component was controller specific, you could add it to the controller specific folder (Views/Todo/Components/PriorityList/Default.cshtml)

Add a div containing a call to the priority list component to the bottom of the Views/Todo/index.cshtml file:

    }
</table>
<div >
    @await Component.InvokeAsync("PriorityList", new { maxPriority = 2, isDone = false })
</div>

The markup @Component.InvokeAsync shows the syntax for calling view components. The first argument is the name of the component we want to invoke or call. Subsequent parameters are passed to the component. InvokeAsync can take an arbitrary number of arguments.

The following image shows the priority items:

You can also call the view component directly from the controller:

  public IActionResult IndexVC()
  {
      return ViewComponent("PriorityList", new { maxPriority = 3, isDone = false });
  }

Specifying a view name¶

A complex view component might need to specify a non-default view under some conditions. The following code shows how to specify the “PVC” view from the InvokeAsync method. Update the InvokeAsync method in the PriorityListViewComponent class.

public async Task<IViewComponentResult> InvokeAsync(
    int maxPriority, bool isDone)
{
    string MyView = "Default";
    // If asking for all completed tasks, render with the "PVC" view.
    if (maxPriority > 3 && isDone == true)
    {
        MyView = "PVC";
    }
    var items = await GetItemsAsync(maxPriority, isDone);
    return View(MyView, items);
}

Copy the Views/Shared/Components/PriorityList/Default.cshtml file to a view named Views/Shared/Components/PriorityList/PVC.cshtml. Add a heading to indicate the PVC view is being used.

@model IEnumerable<ViewComponentSample.Models.TodoItem>

<h2> PVC Named Priority Component View</h2>
<h4>@ViewBag.PriorityMessage</h4>
<ul>
    @foreach (var todo in Model)
    {
        <li>@todo.Name</li>
    }
</ul>

Update Views/TodoList/Index.cshtml

</table>

<div>
    @await Component.InvokeAsync("PriorityList", new { maxPriority = 4, isDone = true })
</div>

Run the app and verify PVC view.

If the PVC view is not rendered, verify you are calling the view component with a priority of 4 or higher.

Examine the view path¶

Change the priority parameter to three or less so the priority view is not returned.
Temporarily rename the Views/Todo/Components/PriorityList/Default.cshtml to Temp.cshtml.

Test the app, you’ll get the following error:

An unhandled exception occurred while processing the request.

InvalidOperationException: The view 'Components/PriorityList/Default'
   was not found. The following locations were searched:
   /Views/ToDo/Components/PriorityList/Default.cshtml
   /Views/Shared/Components/PriorityList/Default.cshtml.
Microsoft.AspNetCore.Mvc.ViewEngines.ViewEngineResult.EnsureSuccessful()

Copy Views/Shared/Components/PriorityList/Default.cshtml to *Views/Todo/Components/PriorityList/Default.cshtml.
Add some markup to the Todo view component view to indicate the view is from the Todo folder.
Test the non-shared component view.

Avoiding magic strings¶

If you want compile time safety you can replace the hard coded view component name with the class name. Create the view component without the “ViewComponent” suffix:

using Microsoft.AspNet.Mvc;
using Microsoft.Data.Entity;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using ViewComponentSample.Models;

namespace ViewComponentSample.ViewComponents
{
    public class PriorityList : ViewComponent
    {
        private readonly ToDoContext db;

        public PriorityList(ToDoContext context)
        {
            db = context;
        }

        public async Task<IViewComponentResult> InvokeAsync(
        int maxPriority, bool isDone)
        {
            var items = await GetItemsAsync(maxPriority, isDone);
            return View(items);
        }
        private Task<List<TodoItem>> GetItemsAsync(int maxPriority, bool isDone)
        {
            return db.ToDo.Where(x => x.IsDone == isDone &&
                                 x.Priority <= maxPriority).ToListAsync();
        }
    }
}

Add a using statement to your Razor view file and use the nameof operator:

@using ViewComponentSample.Models
@using ViewComponentSample.ViewComponents
@model IEnumerable<TodoItem>

<h2>ToDo nameof</h2>
<!-- Markup removed for brevity.  -->
    }
</table>

<div>

    @await Component.InvokeAsync(nameof(PriorityList), new { maxPriority = 4, isDone = true })
</div>

Additional Resources ¶

🔧 Creating a Custom View Engine¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

🔧 Building Mobile Specific Views¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Controllers¶

Controllers, Actions, and Action Results¶

Actions and action results are a fundamental part of how developers build apps using ASP.NET MVC.

Sections:

What is a Controller
Defining Actions

What is a Controller ¶

In ASP.NET MVC, a Controller is used to define and group a set of actions. An action (or action method) is a method on a controller that handles incoming requests. Controllers provide a logical means of grouping similar actions together, allowing common sets of rules (e.g. routing, caching, authorization) to be applied collectively. Incoming requests are mapped to actions through routing.

In ASP.NET Core MVC, a controller can be any instantiable class that ends in “Controller” or inherits from a class that ends with “Controller”. Controllers should follow the Explicit Dependencies Principle and request any dependencies their actions require through their constructor using dependency injection.

By convention, controller classes:

Are located in the root-level “Controllers” folder
Inherit from Microsoft.AspNetCore.Mvc.Controller

These two conventions are not required.

Within the Model-View-Controller pattern, a Controller is responsible for the initial processing of the request and instantiation of the Model. Generally, business decisions should be performed within the Model.

Note

The Model should be a Plain Old CLR Object (POCO), not a DbContext or database-related type.

The controller takes the result of the model’s processing (if any), returns the proper view along with the associated view data. Learn more: 🔧 Overview of ASP.NET MVC and Getting started with ASP.NET Core MVC and Visual Studio.

Tip

The Controller is a UI level abstraction. Its responsibility is to ensure incoming request data is valid and to choose which view (or result for an API) should be returned. In well-factored apps it will not directly include data access or business logic, but instead will delegate to services handling these responsibilities.

Defining Actions ¶

Any public method on a controller type is an action. Parameters on actions are bound to request data and validated using model binding.

Warning

Action methods that accept parameters should verify the ModelState.IsValid property is true.

Action methods should contain logic for mapping an incoming request to a business concern. Business concerns should typically be represented as services that your controller accesses through dependency injection. Actions then map the result of the business action to an application state.

Actions can return anything, but frequently will return an instance of IActionResult (or Task<IActionResult> for async methods) that produces a response. The action method is responsible for choosing what kind of response; the action result does the responding.

Controller Helper Methods¶

Although not required, most developers will want to have their controllers inherit from the base Controller class. Doing so provides controllers with access to many properties and helpful methods, including the following helper methods designed to assist in returning various responses:

View: Returns a view that uses a model to render HTML. Example: return View(customer);
HTTP Status Code: Return an HTTP status code. Example: return BadRequest();
Formatted Response: Return Json or similar to format an object in a specific manner. Example: return Json(customer);
Content negotiated response: Instead of returning an object directly, an action can return a content negotiated response (using Ok, Created, CreatedAtRoute or CreatedAtAction). Examples: return Ok(); or return CreatedAtRoute("routename",values,newobject");
Redirect: Returns a redirect to another action or destination (using Redirect, LocalRedirect, RedirectToAction or RedirectToRoute). Example: return RedirectToAction("Complete", new {id = 123});

In addition to the methods above, an action can also simply return an object. In this case, the object will be formatted based on the client’s request. Learn more about Formatting Response Data

Cross-Cutting Concerns¶

In most apps, many actions will share parts of their workflow. For instance, most of an app might be available only to authenticated users, or might benefit from caching. When you want to perform some logic before or after an action method runs, you can use a filter. You can help keep your actions from growing too large by using Filters to handle these cross-cutting concerns. This can help eliminate duplication within your actions, allowing them to follow the Don’t Repeat Yourself (DRY) principle.

In the case of authorization and authentication, you can apply the Authorize attribute to any actions that require it. Adding it to a controller will apply it to all actions within that controller. Adding this attribute will ensure the appropriate filter is applied to any request for this action. Some attributes can be applied at both controller and action levels to provide granular control over filter behavior. Learn more: Filters and 🔧 Authorization Filters.

Other examples of cross-cutting concerns in MVC apps may include:

Note

Many cross-cutting concerns can be handled using filters in MVC apps. Another option to keep in mind that is available to any ASP.NET Core app is custom middleware.

🔧 Routing to Controller Actions¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Filters¶

View or download sample from GitHub.

Filters in ASP.NET MVC allow you to run code before or after a particular stage in the execution pipeline. Filters can be configured globally, per-controller, or per-action.

Sections

How do filters work?
Configuring Filters
Authorization Filters
Resource Filters
Action Filters
Exception Filters
Result Filters
Filters vs. Middleware

How do filters work?¶

Each filter type is executed at a different stage in the pipeline, and thus has its own set of intended scenarios. Choose what type of filter to create based on the task you need it to perform, and where in the request pipeline it executes. Filters run within the MVC Action Invocation Pipeline, sometimes referred to as the Filter Pipeline, which runs after MVC selects the action to execute.

Different filter types run at different points within the pipeline. Some filters, like authorization filters, only run before the next stage in the pipeline, and take no action afterward. Other filters, like action filters, can execute both before and after other parts of the pipeline execute, as shown below.

Selecting a Filter¶

Authorization filters are used to determine whether the current user is authorized for the request being made.

Resource filters are the first filter to handle a request after authorization, and the last one to touch the request as it is leaving the filter pipeline. They’re especially useful to implement caching or otherwise short-circuit the filter pipeline for performance reasons.

Action filters wrap calls to individual action method calls, and can manipulate the arguments passed into an action as well as the action result returned from it.

Exception filters are used to apply global policies to unhandled exceptions in the MVC app.

Result filters wrap the execution of individual action results, and only run when the action method has executed successfully. They are ideal for logic that must surround view execution or formatter execution.

Implementation¶

All filters support both synchronous and asynchronous implementations through different interface definitions. Choose the sync or async variant depending on the kind of task you need to perform. They are interchangeable from the framework’s perspective.

Synchronous filters define both an OnStageExecuting and OnStageExecuted method (with noted exceptions). The OnStageExecuting method will be called before the event pipeline stage by the Stage name, and the OnStageExecuted method will be called after the pipeline stage named by the Stage name.

using FiltersSample.Helper;
using Microsoft.AspNetCore.Mvc.Filters;

namespace FiltersSample.Filters
{
    public class SampleActionFilter : IActionFilter
    {
        public void OnActionExecuting(ActionExecutingContext context)
        {
            // do something before the action executes
        }

        public void OnActionExecuted(ActionExecutedContext context)
        {
            // do something after the action executes
        }
    }
}

Asynchronous filters define a single OnStageExecutionAsync method that will surround execution of the pipeline stage named by Stage. The OnStageExecutionAsync method is provided a StageExecutionDelegate delegate which will execute the pipeline stage named by Stage when invoked and awaited.

using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc.Filters;

namespace FiltersSample.Filters
{
    public class SampleAsyncActionFilter : IAsyncActionFilter
    {
        public async Task OnActionExecutionAsync(
            ActionExecutingContext context,
            ActionExecutionDelegate next)
        {
            // do something before the action executes
            await next();
            // do something after the action executes
        }
    }
}

Note

You should only implement either the synchronous or the async version of a filter interface, not both. If you need to perform async work in the filter, implement the async interface. Otherwise, implement the synchronous interface. The framework will check to see if the filter implements the async interface first, and if so, it will call it. If not, it will call the synchronous interface’s method(s). If you were to implement both interfaces on one class, only the async method would be called by the framework. Also, it doesn’t matter whether your action is async or not, your filters can be synchronous or async independent of the action.

Filter Scopes¶

Filters can be scoped at three different levels. You can add a particular filter to a particular action as an attribute. You can add a filter to all actions within a controller by applying an attribute at the controller level. Or you can register a filter globally, to be run with every MVC action.

Global filters are added in the ConfigureServices method in Startup, when configuring MVC:

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc(options =>
    {
        options.Filters.Add(typeof(SampleActionFilter)); // by type
        options.Filters.Add(new SampleGlobalActionFilter()); // an instance
    });

    services.AddScoped<AddHeaderFilterWithDi>();
}

Filters can be added by type, or an instance can be added. If you add an instance, that instance will be used for every request. If you add a type, it will be type-activated, meaning an instance will be created for each request and any constructor dependencies will be populated by DI. Adding a filter by type is equivalent to filters.Add(new TypeFilterAttribute(typeof(MyFilter))).

It’s often convenient to implement filter interfaces as Attributes. Filter attributes are applied to controllers and action methods. The framework includes built-in attribute-based filters that you can subclass and customize. For example, the following filter inherits from ResultFilterAttribute, and overrides its OnResultExecuting method to add a header to the response.

using Microsoft.AspNetCore.Mvc.Filters;

namespace FiltersSample.Filters
{
    public class AddHeaderAttribute : ResultFilterAttribute
    {
        private readonly string _name;
        private readonly string _value;

        public AddHeaderAttribute(string name, string value)
        {
            _name = name;
            _value = value;
        }

        public override void OnResultExecuting(ResultExecutingContext context)
        {
            context.HttpContext.Response.Headers.Add(
                _name, new string[] { _value });
            base.OnResultExecuting(context);
        }
    }
}

Attributes allow filters to accept arguments, as shown in the example above. You would add this attribute to a controller or action method and specify the name and value of the HTTP header you wished to add to the response:

[AddHeader("Author", "Steve Smith @ardalis")]
public class SampleController : Controller
{
    public IActionResult Index()
    {
        return Content("Examine the headers using developer tools.");
    }
}

The result of the Index action is shown below - the response headers are displayed on the bottom right.

Several of the filter interfaces have corresponding attributes that can be used as base classes for custom implementations.

Filter attributes:

Cancellation and Short Circuiting¶

You can short-circuit the filter pipeline at any point by setting the Result property on the context parameter provided to the filter method. For instance, the following ShortCircuitingResourceFilter will prevent any other filters from running later in the pipeline, including any action filters.

using System;
using Microsoft.AspNetCore.Mvc;
using Microsoft.AspNetCore.Mvc.Filters;

namespace FiltersSample.Filters
{
    public class ShortCircuitingResourceFilterAttribute : Attribute,
            IResourceFilter
    {
        public void OnResourceExecuting(ResourceExecutingContext context)
        {
            context.Result = new ContentResult()
            {
                Content = "Resource unavailable - header should not be set"
            };
        }

        public void OnResourceExecuted(ResourceExecutedContext context)
        {
        }
    }
}

In the following code, both the ShortCircuitingResourceFilter and the AddHeader filter target the SomeResource action method. However, because the ShortCircuitingResourceFilter runs first and short-circuits the rest of the pipeline, the AddHeader filter never runs for the SomeResource action. This behavior would be the same if both filters were applied at the action method level, provided the ShortCircuitingResourceFilter ran first (see Ordering).

[AddHeader("Author", "Steve Smith @ardalis")]
public class SampleController : Controller
{
    [ShortCircuitingResourceFilter]
    public IActionResult SomeResource()
    {
        return Content("Successful access to resource - header should be set.");
    }

Configuring Filters ¶

Global filters are configured within Startup.cs. Attribute-based filters that do not require any dependencies can simply inherit from an existing attribute of the appropriate type for the filter in question. To create a filter without global scope that requires dependencies from DI, apply the ServiceFilterAttribute or TypeFilterAttribute attribute to the controller or action.

Dependency Injection¶

Filters that are implemented as attributes and added directly to controller classes or action methods cannot have constructor dependencies provided by dependency injection (DI). This is because attributes must have their constructor parameters supplied where they are applied. This is a limitation of how attributes work.

However, if your filters have dependencies you need to access from DI, there are several supported approaches. You can apply your filter to a class or action method using

ServiceFilterAttribute
TypeFilterAttribute
IFilterFactory implemented on your attribute

A TypeFilter will instantiate an instance, using services from DI for its dependencies. A ServiceFilter retrieves an instance of the filter from DI. The following example demonstrates using a ServiceFilter:

[ServiceFilter(typeof(AddHeaderFilterWithDi))]
public IActionResult Index()
{
    return View();
}

Using ServiceFilter without registering the filter type in ConfigureServices, throws the following exception:

System.InvalidOperationException: No service for type
'FiltersSample.Filters.AddHeaderFilterWithDI' has been registered.

To avoid this exception, you must register the AddHeaderFilterWithDI type in ConfigureServices:

services.AddScoped<AddHeaderFilterWithDi>();

ServiceFilterAttribute implements IFilterFactory, which exposes a single method for creating an IFilter instance. In the case of ServiceFilterAttribute, the IFilterFactory interface’s CreateInstance method is implemented to load the specified type from the services container (DI).

TypeFilterAttribute is very similar to ServiceFilterAttribute (and also implements IFilterFactory), but its type is not resolved directly from the DI container. Instead, it instantiates the type using a Microsoft.Extensions.DependencyInjection.ObjectFactory.

Because of this difference, types that are referenced using the TypeFilterAttribute do not need to be registered with the container first (but they will still have their dependencies fulfilled by the container). Also, TypeFilterAttribute can optionally accept constructor arguments for the type in question. The following example demonstrates how to pass arguments to a type using TypeFilterAttribute:

[TypeFilter(typeof(AddHeaderAttribute),
    Arguments = new object[] { "Author", "Steve Smith (@ardalis)" })]
public IActionResult Hi(string name)
{
    return Content($"Hi {name}");
}

If you have a simple filter that doesn’t require any arguments, but which has constructor dependencies that need to be filled by DI, you can inherit from TypeFilterAttribute, allowing you to use your own named attribute on classes and methods (instead of [TypeFilter(typeof(FilterType))]). The following filter shows how this can be implemented:

public class SampleActionFilterAttribute : TypeFilterAttribute
{
    public SampleActionFilterAttribute():base(typeof(SampleActionFilterImpl))
    {
    }

    private class SampleActionFilterImpl : IActionFilter
    {
        private readonly ILogger _logger;
        public SampleActionFilterImpl(ILoggerFactory loggerFactory)
        {
            _logger = loggerFactory.CreateLogger<SampleActionFilterAttribute>();
        }

        public void OnActionExecuting(ActionExecutingContext context)
        {
            _logger.LogInformation("Business action starting...");
            // perform some business logic work

        }

        public void OnActionExecuted(ActionExecutedContext context)
        {
            // perform some business logic work
            _logger.LogInformation("Business action completed.");
        }
    }
}

This filter can be applied to classes or methods using the [SampleActionFilter] syntax, instead of having to use [TypeFilter] or [ServiceFilter].

Note

Avoid creating and using filters purely for logging purposes, since the built-in framework logging features should already provide what you need for logging. If you’re going to add logging to your filters, it should focus on business domain concerns or behavior specific to your filter, rather than MVC actions or other framework events.

IFilterFactory implements IFilter. Therefore, an IFilterFactory instance can be used as an IFilter instance anywhere in the filter pipeline. When the framework prepares to invoke the filter, attempts to cast it to an IFilterFactory. If that cast succeeds, the CreateInstance method is called to create the IFilter instance that will be invoked. This provides a very flexible design, since the precise filter pipeline does not need to be set explicitly when the application starts.

You can implement IFilterFactory on your own attribute implementations as another approach to creating filters:

public class AddHeaderWithFactoryAttribute : Attribute, IFilterFactory
{
    // Implement IFilterFactory
    public IFilterMetadata CreateInstance(IServiceProvider serviceProvider)
    {
        return new InternalAddHeaderFilter();
    }

    private class InternalAddHeaderFilter : IResultFilter
    {
        public void OnResultExecuting(ResultExecutingContext context)
        {
            context.HttpContext.Response.Headers.Add(
                "Internal", new string[] { "Header Added" });
        }

        public void OnResultExecuted(ResultExecutedContext context)
        {
        }
    }

Ordering¶

Filters can be applied to action methods or controllers (via attribute) or added to the global filters collection. Scope also generally determines ordering. The filter closest to the action runs first; generally you get overriding behavior without having to explicitly set ordering. This is sometimes referred to as “Russian doll” nesting, as each increase in scope is wrapped around the previous scope, like a nesting doll.

In addition to scope, filters can override their sequence of execution by implementing IOrderedFilter. This interface simply exposes an int Order property, and filters execute in ascending numeric order based on this property. All of the built-in filters, including TypeFilterAttribute and ServiceFilterAttribute, implement IOrderedFilter, so you can specify the order of filters when you apply the attribute to a class or method. By default, the Order property is 0 for all of the built-in filters, so scope is used as a tie-breaker and (unless Order is set to a non-zero value) is the determining factor.

Every controller that inherits from the Controller base class includes OnActionExecuting and OnActionExecuted methods. These methods wrap the filters that run for a given action, running first and last. The scope-based order, assuming no Order has been set for any filter, is:

The Controller OnActionExecuting
The Global filter OnActionExecuting
The Class filter OnActionExecuting
The Method filter OnActionExecuting
The Method filter OnActionExecuted
The Class filter OnActionExecuted
The Global filter OnActionExecuted
The Controller OnActionExecuted

Note

IOrderedFilter trumps scope when determining the order in which filters will run. Filters are sorted first by order, then scope is used to break ties. Order defaults to 0 if not set.

To modify the default, scope-based order, you could explicitly set the Order property of a class-level or method-level filter. For example, adding Order=-1 to a method level attribute:

[MyFilter(Name = "Method Level Attribute", Order=-1)]

In this case, a value of less than zero would ensure this filter ran before both the Global and Class level filters (assuming their Order property was not set).

The new order would be:

The Controller OnActionExecuting
The Method filter OnActionExecuting
The Global filter OnActionExecuting
The Class filter OnActionExecuting
The Class filter OnActionExecuted
The Global filter OnActionExecuted
The Method filter OnActionExecuted
The Controller OnActionExecuted

Note

The Controller class’s methods always run before and after all filters. These methods are not implemented as IFilter instances and do not participate in the IFilter ordering algorithm.

Authorization Filters ¶

Authorization Filters control access to action methods, and are the first filters to be executed within the filter pipeline. They have only a before stage, unlike most filters that support before and after methods. You should only write a custom authorization filter if you are writing your own authorization framework. Note that you should not throw exceptions within authorization filters, since nothing will handle the exception (exception filters won’t handle them). Instead, issue a challenge or find another way.

Learn more about Authorization.

Resource Filters ¶

Resource Filters implement either the IResourceFilter or IAsyncResourceFilter interface, and their execution wraps most of the filter pipeline (only Authorization Filters run before them - all other filters and action processing happens between their OnResourceExecuting and OnResourceExecuted methods). Resource filters are especially useful if you need to short-circuit most of the work a request is doing. Caching would be one example use case for a resource filter, since if the response is already in the cache, the filter can immediately set a result and avoid the rest of the processing for the action.

The short circuiting resource filter shown above is one example of a resource filter. A very naive cache implementation (do not use this in production) that only works with ContentResult action results is shown below:

public class NaiveCacheResourceFilterAttribute : Attribute,
    IResourceFilter
{
    private static readonly Dictionary<string, object> _cache
                = new Dictionary<string, object>();
    private string _cacheKey;

    public void OnResourceExecuting(ResourceExecutingContext context)
    {
        _cacheKey = context.HttpContext.Request.Path.ToString();
        if (_cache.ContainsKey(_cacheKey))
        {
            var cachedValue = _cache[_cacheKey] as string;
            if (cachedValue != null)
            {
                context.Result = new ContentResult()
                { Content = cachedValue };
            }
        }
    }

    public void OnResourceExecuted(ResourceExecutedContext context)
    {
        if (!String.IsNullOrEmpty(_cacheKey) &&
            !_cache.ContainsKey(_cacheKey))
        {
            var result = context.Result as ContentResult;
            if (result != null)
            {
                _cache.Add(_cacheKey, result.Content);
            }
        }
    }
}

In OnResourceExecuting, if the result is already in the static dictionary cache, the Result property is set on context, and the action short-circuits and returns with the cached result. In the OnResourceExecuted method, if the current request’s key isn’t already in use, the current Result is stored in the cache, to be used by future requests.

Adding this filter to a class or method is shown here:

[TypeFilter(typeof(NaiveCacheResourceFilterAttribute))]
public class CachedController : Controller
{
    public IActionResult Index()
    {
        return Content("This content was generated at " + DateTime.Now);
    }
}

Action Filters ¶

Action Filters implement either the IActionFilter or IAsyncActionFilter interface and their execution surrounds the execution of action methods. Action filters are ideal for any logic that needs to see the results of model binding, or modify the controller or inputs to an action method. Additionally, action filters can view and directly modify the result of an action method.

The OnActionExecuting method runs before the action method, so it can manipulate the inputs to the action by changing ActionExecutingContext.ActionArguments or manipulate the controller through ActionExecutingContext.Controller. An OnActionExecuting method can short-circuit execution of the action method and subsequent action filters by setting ActionExecutingContext.Result. Throwing an exception in an OnActionExecuting method will also prevent execution of the action method and subsequent filters, but will be treated as a failure instead of successful result.

The OnActionExecuted method runs after the action method and can see and manipulate the results of the action through the ActionExecutedContext.Result property. ActionExecutedContext.Canceled will be set to true if the action execution was short-circuited by another filter. ActionExecutedContext.Exception will be set to a non-null value if the action or a subsequent action filter threw an exception. Setting ActionExecutedContext.Exception to null effectively ‘handles’ an exception, and ActionExectedContext.Result will then be executed as if it were returned from the action method normally.

For an IAsyncActionFilter the OnActionExecutionAsync combines all the possibilities of OnActionExecuting and OnActionExecuted. A call to await next() on the ActionExecutionDelegate will execute any subsequent action filters and the action method, returning an ActionExecutedContext. To short-circuit inside of an OnActionExecutionAsync, assign ActionExecutingContext.Result to some result instance and do not call the ActionExectionDelegate.

Exception Filters ¶

Exception Filters implement either the IExceptionFilter or IAsyncExceptionFilter interface.

Exception filters handle unhandled exceptions, including those that occur during controller creation and model binding. They are only called when an exception occurs in the pipeline. They can provide a single location to implement common error handling policies within an app. The framework provides an abstract ExceptionFilterAttribute that you should be able to subclass for your needs. Exception filters are good for trapping exceptions that occur within MVC actions, but they’re not as flexible as error handling middleware. Prefer middleware for the general case, and use filters only where you need to do error handling differently based on which MVC action was chosen.

Tip

One example where you might need a different form of error handling for different actions would be in an app that exposes both API endpoints and actions that return views/HTML. The API endpoints could return error information as JSON, while the view-based actions could return an error page as HTML.

Exception filters do not have two events (for before and after) - they only implement OnException (or OnExceptionAsync). The ExceptionContext provided in the OnException parameter includes the Exception that occurred. If you set context.ExceptionHandled to true, the effect is that you’ve handled the exception, so the request will proceed as if it hadn’t occurred (generally returning a 200 OK status). The following filter uses a custom developer error view to display details about exceptions that occur when the application is in development:

using Microsoft.AspNetCore.Hosting;
using Microsoft.AspNetCore.Mvc;
using Microsoft.AspNetCore.Mvc.Filters;
using Microsoft.AspNetCore.Mvc.ModelBinding;
using Microsoft.AspNetCore.Mvc.ViewFeatures;

namespace FiltersSample.Filters
{
    public class CustomExceptionFilterAttribute : ExceptionFilterAttribute
    {
        private readonly IHostingEnvironment _hostingEnvironment;
        private readonly IModelMetadataProvider _modelMetadataProvider;

        public CustomExceptionFilterAttribute(
            IHostingEnvironment hostingEnvironment,
            IModelMetadataProvider modelMetadataProvider)
        {
            _hostingEnvironment = hostingEnvironment;
            _modelMetadataProvider = modelMetadataProvider;
        }

        public override void OnException(ExceptionContext context)
        {
            if (!_hostingEnvironment.IsDevelopment())
            {
                // do nothing
                return;
            }
            var result = new ViewResult {ViewName = "CustomError"};
            result.ViewData = new ViewDataDictionary(_modelMetadataProvider,context.ModelState);
            result.ViewData.Add("Exception", context.Exception);
            // TODO: Pass additional detailed data via ViewData
            context.ExceptionHandled = true; // mark exception as handled
            context.Result = result;
        }
    }
}

Result Filters ¶

Result Filters implement either the IResultFilter or IAsyncResultFilter interface and their execution surrounds the execution of action results. Result filters are only executed for successful results - when the action or action filters produce an action result. Result filters are not executed when exception filters handle an exception, unless the exception filter sets Exception = null.

Note

The kind of result being executed depends on the action in question. An MVC action returning a view would include all razor processing as part of the ViewResult being executed. An API method might perform some serialization as part of the execution of the result. Learn more about action results

Result filters are ideal for any logic that needs to directly surround view execution or formatter execution. Result filters can replace or modify the action result that’s responsible for producing the response.

The OnResultExecuting method runs before the action result is executed, so it can manipulate the action result through ResultExecutingContext.Result. An OnResultExecuting method can short-circuit execution of the action result and subsequent result filters by setting ResultExecutingContext.Cancel to true. If short-circuited, MVC will not modify the response; you should generally write to the response object directly when short-circuiting to avoid generating an empty response. Throwing an exception in an OnResultExecuting method will also prevent execution of the action result and subsequent filters, but will be treated as a failure instead of a successful result.

The OnResultExecuted method runs after the action result has executed. At this point if no exception was thrown, the response has likely been sent to the client and cannot be changed further. ResultExecutedContext.Canceled will be set to true if the action result execution was short-circuited by another filter. ResultExecutedContext.Exception will be set to a non-null value if the action result or a subsequent result filter threw an exception. Setting ResultExecutedContext.Exception to null effectively ‘handles’ an exception and will prevent the exeception from being rethrown by MVC later in the pipeline. If handling an exception in a result filter, consider whether or not it’s appropriate to write any data to the response. If the action result throws partway through its execution, and the headers have already been flushed to the client, there’s no reliable mechanism to send a failure code.

For an IAsyncResultFilter the OnResultExecutionAsync combines all the possibilities of OnResultExecuting and OnResultExecuted. A call to await next() on the ResultExecutionDelegate will execute any subsequent result filters and the action result, returning a ResultExecutedContext. To short-circuit inside of an OnResultExecutionAsync, set ResultExecutingContext.Cancel to true and do not call the ResultExectionDelegate.

You can override the built-in ResultFilterAttribute to create result filters. The AddHeaderAttribute class shown above is an example of a result filter.

Tip

If you need to add headers to the response, do so before the action result executes. Otherwise, the response may have been sent to the client, and it will be too late to modify it. For a result filter, this means adding the header in OnResultExecuting rather than OnResultExecuted.

Filters vs. Middleware ¶

In general, filters are meant to handle cross-cutting business and application concerns. This is often the same use case for middleware. Filters are very similar to middleware in capability, but let you scope that behavior and insert it into a location in your app where it makes sense, such as before a view, or after model binding. Filters are a part of MVC, and have access to its context and constructs. For instance, middleware can’t easily detect whether model validation on a request has generated errors, and respond accordingly, but a filter can easily do so.

To experiment with filters, download, test and modify the sample.

Dependency Injection and Controllers¶

ASP.NET Core MVC controllers should request their dependencies explicitly via their constructors. In some instances, individual controller actions may require a service, and it may not make sense to request at the controller level. In this case, you can also choose to inject a service as a parameter on the action method.

Sections:

Dependency Injection
Constructor Injection
Action Injection with FromServices
Accessing Settings from a Controller

Dependency Injection ¶

Dependency injection is a technique that follows the Dependency Inversion Principle, allowing for applications to be composed of loosely coupled modules. ASP.NET Core has built-in support for dependency injection, which makes applications easier to test and maintain.

Constructor Injection ¶

ASP.NET Core’s built-in support for constructor-based dependency injection extends to MVC controllers. By simply adding a service type to your controller as a constructor parameter, ASP.NET Core will attempt to resolve that type using its built in service container. Services are typically, but not always, defined using interfaces. For example, if your application has business logic that depends on the current time, you can inject a service that retrieves the time (rather than hard-coding it), which would allow your tests to pass in implementations that use a set time.

using System;

namespace ControllerDI.Interfaces
{
    public interface IDateTime
    {
        DateTime Now { get; }
    }
}

Implementing an interface like this one so that it uses the system clock at runtime is trivial:

using System;
using ControllerDI.Interfaces;

namespace ControllerDI.Services
{
    public class SystemDateTime : IDateTime
    {
        public DateTime Now
        {
            get { return DateTime.Now; }
        }
    }
}

With this in place, we can use the service in our controller. In this case, we have added some logic to the HomeController Index method to display a greeting to the user based on the time of day.

using ControllerDI.Interfaces;
using Microsoft.AspNetCore.Mvc;

namespace ControllerDI.Controllers
{
    public class HomeController : Controller
    {
        private readonly IDateTime _dateTime;

        public HomeController(IDateTime dateTime)
        {
            _dateTime = dateTime;
        }

        public IActionResult Index()
        {
            var serverTime = _dateTime.Now;
            if (serverTime.Hour < 12)
            {
                ViewData["Message"] = "It's morning here - Good Morning!";
            }
            else if (serverTime.Hour < 17)
            {
                ViewData["Message"] = "It's afternoon here - Good Afternoon!";
            }
            else
            {
                ViewData["Message"] = "It's evening here - Good Evening!";
            }
            return View();
        }
    }
}

If we run the application now, we will most likely encounter an error:

An unhandled exception occurred while processing the request.

InvalidOperationException: Unable to resolve service for type 'ControllerDI.Interfaces.IDateTime' while attempting to activate 'ControllerDI.Controllers.HomeController'.
Microsoft.Extensions.DependencyInjection.ActivatorUtilities.GetService(IServiceProvider sp, Type type, Type requiredBy, Boolean isDefaultParameterRequired)

This error occurs when we have not configured a service in the ConfigureServices method in our Startup class. To specify that requests for IDateTime should be resolved using an instance of SystemDateTime, add the highlighted line in the listing below to your ConfigureServices method:

public void ConfigureServices(IServiceCollection services)
{
    // Add application services.
    services.AddTransient<IDateTime, SystemDateTime>();
}

Note

This particular service could be implemented using any of several different lifetime options (Transient, Scoped, or Singleton). See Dependency Injection to understand how each of these scope options will affect the behavior of your service.

Once the service has been configured, running the application and navigating to the home page should display the time-based message as expected:

Tip

See Testing Controller Logic to learn how to explicitly request dependencies http://deviq.com/explicit-dependencies-principle in controllers makes code easier to test.

ASP.NET Core’s built-in dependency injection supports having only a single constructor for classes requesting services. If you have more than one constructor, you may get an exception stating:

An unhandled exception occurred while processing the request.

InvalidOperationException: Multiple constructors accepting all given argument types have been found in type 'ControllerDI.Controllers.HomeController'. There should only be one applicable constructor.
Microsoft.Extensions.DependencyInjection.ActivatorUtilities.FindApplicableConstructor(Type instanceType, Type[] argumentTypes, ConstructorInfo& matchingConstructor, Nullable`1[]& parameterMap)

As the error message states, you can correct this problem having just a single constructor. You can also replace the default dependency injection support with a third party implementation, many of which support multiple constructors.

Action Injection with FromServices ¶

Sometimes you don’t need a service for more than one action within your controller. In this case, it may make sense to inject the service as a parameter to the action method. This is done by marking the parameter with the attribute [FromServices] as shown here:

public IActionResult About([FromServices] IDateTime dateTime)
{
    ViewData["Message"] = "Currently on the server the time is " + dateTime.Now;

    return View();
}

Accessing Settings from a Controller ¶

Accessing application or configuration settings from within a controller is a common pattern. This access should use the Options pattern described in configuration. You generally should not request settings directly from your controller using dependency injection. A better approach is to request an IOptions<T> instance, where T is the configuration class you need.

To work with the options pattern, you need to create a class that represents the options, such as this one:

namespace ControllerDI.Model
{
    public class SampleWebSettings
    {
        public string Title { get; set; }
        public int Updates { get; set; }
    }
}

Then you need to configure the application to use the options model and add your configuration class to the services collection in ConfigureServices:

public Startup(IHostingEnvironment env)
{
    var builder = new ConfigurationBuilder()
        .SetBasePath(env.ContentRootPath)
        .AddJsonFile("samplewebsettings.json");
    Configuration = builder.Build();
}

public IConfigurationRoot Configuration { get; set; }

// This method gets called by the runtime. Use this method to add services to the container.
// For more information on how to configure your application, visit http://go.microsoft.com/fwlink/?LinkID=398940
public void ConfigureServices(IServiceCollection services)
{
    // Required to use the Options<T> pattern
    services.AddOptions();

    // Add settings from configuration
    services.Configure<SampleWebSettings>(Configuration);

    // Uncomment to add settings from code
    //services.Configure<SampleWebSettings>(settings =>
    //{
    //    settings.Updates = 17;
    //});

    services.AddMvc();

    // Add application services.
    services.AddTransient<IDateTime, SystemDateTime>();
}

Note

In the above listing, we are configuring the application to read the settings from a JSON-formatted file. You can also configure the settings entirely in code, as is shown in the commented code above. See Configuration for further configuration options.

Once you’ve specified a strongly-typed configuration object (in this case, SampleWebSettings) and added it to the services collection, you can request it from any Controller or Action method by requesting an instance of IOptions<T> (in this case, IOptions<SampleWebSettings>). The following code shows how one would request the settings from a controller:

    public class SettingsController : Controller
    {
        private readonly SampleWebSettings _settings;

        public SettingsController(IOptions<SampleWebSettings> settingsOptions)
        {
            _settings = settingsOptions.Value;
        }

        public IActionResult Index()
        {
            ViewData["Title"] = _settings.Title;
            ViewData["Updates"] = _settings.Updates;
            return View();
        }
    }

Following the Options pattern allows settings and configuration to be decoupled from one another, and ensures the controller is following separation of concerns, since it doesn’t need to know how or where to find the settings information. It also makes the controller easier to unit test Testing Controller Logic, since there is no static cling or direct instantiation of settings classes within the controller class.

Testing Controller Logic¶

View or download sample from GitHub

Controllers in ASP.NET MVC apps should be small and focused on user-interface concerns. Large controllers that deal with non-UI concerns are more difficult to test and maintain.

Sections

Why Test Controllers
Unit Testing
Integration Testing

Why Test Controllers ¶

Controllers are a central part of any ASP.NET Core MVC application. As such, you should have confidence they behave as intended for your app. Automated tests can provide you with this confidence and can detect errors before they reach production. It’s important to avoid placing unnecessary responsibilities within your controllers and ensure your tests focus only on controller responsibilities.

Controller logic should be minimal and not be focused on business logic or infrastructure concerns (for example, data access). Test controller logic, not the framework. Test how the controller behaves based on valid or invalid inputs. Test controller responses based on the result of the business operation it performs.

Typical controller responsibilities:

Verify ModelState.IsValid
Return an error response if ModelState is invalid
Retrieve a business entity from persistence
Perform an action on the business entity
Save the business entity to persistence
Return an appropriate IActionResult

Unit Testing ¶

Unit testing involves testing a part of an app in isolation from its infrastructure and dependencies. When unit testing controller logic, only the contents of a single action is tested, not the behavior of its dependencies or of the framework itself. As you unit test your controller actions, make sure you focus only on its behavior. A controller unit test avoids things like filters, routing, or model binding. By focusing on testing just one thing, unit tests are generally simple to write and quick to run. A well-written set of unit tests can be run frequently without much overhead. However, unit tests do not detect issues in the interaction between components, which is the purpose of integration testing.

If you’ve writing custom filters, routes, etc, you should unit test them, but not as part of your tests on a particular controller action. They should be tested in isolation.

Tip

Create and run unit tests with Visual Studio.

To demonstrate unit testing, review the following controller. It displays a list of brainstorming sessions and allows new brainstorming sessions to be created with a POST:

using System;
using System.ComponentModel.DataAnnotations;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc;
using TestingControllersSample.Core.Interfaces;
using TestingControllersSample.Core.Model;
using TestingControllersSample.ViewModels;

namespace TestingControllersSample.Controllers
{
    public class HomeController : Controller
    {
        private readonly IBrainstormSessionRepository _sessionRepository;

        public HomeController(IBrainstormSessionRepository sessionRepository)
        {
            _sessionRepository = sessionRepository;
        }

        public async Task<IActionResult> Index()
        {
            var sessionList = await _sessionRepository.ListAsync();

            var model = sessionList.Select(session => new StormSessionViewModel()
            {
                Id = session.Id,
                DateCreated = session.DateCreated,
                Name = session.Name,
                IdeaCount = session.Ideas.Count
            });

            return View(model);
        }

        public class NewSessionModel
        {
            [Required]
            public string SessionName { get; set; }
        }

        [HttpPost]
        public async Task<IActionResult> Index(NewSessionModel model)
        {
            if (!ModelState.IsValid)
            {
                return BadRequest(ModelState);
            }

            await _sessionRepository.AddAsync(new BrainstormSession()
            {
                DateCreated = DateTimeOffset.Now,
                Name = model.SessionName
            });

            return RedirectToAction("Index");
        }
    }
}

The controller is following the explicit dependencies principle, expecting dependency injection to provide it with an instance of IBrainstormSessionRepository. This makes it fairly easy to test using a mock object framework, like Moq. The HTTP GET Index method has no looping or branching and only calls one method. To test this Index method, we need to verify that a ViewResult is returned, with a ViewModel from the repository’s List method.

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc;
using Moq;
using TestingControllersSample.Controllers;
using TestingControllersSample.Core.Interfaces;
using TestingControllersSample.Core.Model;
using TestingControllersSample.ViewModels;
using Xunit;

namespace TestingControllersSample.Tests.UnitTests
{
    public class HomeControllerTests
    {
        [Fact]
        public async Task Index_ReturnsAViewResult_WithAListOfBrainstormSessions()
        {
            // Arrange
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.ListAsync()).Returns(Task.FromResult(GetTestSessions()));
            var controller = new HomeController(mockRepo.Object);

            // Act
            var result = await controller.Index();

            // Assert
            var viewResult = Assert.IsType<ViewResult>(result);
            var model = Assert.IsAssignableFrom<IEnumerable<StormSessionViewModel>>(
                viewResult.ViewData.Model);
            Assert.Equal(2, model.Count());
        }

        [Fact]
        public async Task IndexPost_ReturnsBadRequestResult_WhenModelStateIsInvalid()
        {
            // Arrange
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.ListAsync()).Returns(Task.FromResult(GetTestSessions()));
            var controller = new HomeController(mockRepo.Object);
            controller.ModelState.AddModelError("SessionName", "Required");
            var newSession = new HomeController.NewSessionModel();

            // Act
            var result = await controller.Index(newSession);

            // Assert
            var badRequestResult = Assert.IsType<BadRequestObjectResult>(result);
            Assert.IsType<SerializableError>(badRequestResult.Value);
        }

        [Fact]
        public async Task IndexPost_ReturnsARedirectAndAddsSession_WhenModelStateIsValid()
        {
            // Arrange
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.AddAsync(It.IsAny<BrainstormSession>()))
                .Returns(Task.CompletedTask)
                .Verifiable();
            var controller = new HomeController(mockRepo.Object);
            var newSession = new HomeController.NewSessionModel()
            {
                SessionName = "Test Name"
            };

            // Act
            var result = await controller.Index(newSession);

            // Assert
            var redirectToActionResult = Assert.IsType<RedirectToActionResult>(result);
            Assert.Null(redirectToActionResult.ControllerName);
            Assert.Equal("Index", redirectToActionResult.ActionName);
            mockRepo.Verify();
        }

        private List<BrainstormSession> GetTestSessions()
        {
            var sessions = new List<BrainstormSession>();
            sessions.Add(new BrainstormSession()
            {
                DateCreated = new DateTime(2016, 7, 2),
                Id = 1,
                Name = "Test One"
            });
            sessions.Add(new BrainstormSession()
            {
                DateCreated = new DateTime(2016, 7, 1),
                Id = 2,
                Name = "Test Two"
            });
            return sessions;
        }
    }
}

The HTTP POST Index method (shown below) should verify:

The action method returns a ViewResult with the appropriate data when ModelState.IsValid is false
The Add method on the repository is called and a RedirectToActionResult is returned with the correct arguments when ModelState.IsValid is true.

        var viewResult = Assert.IsType<ViewResult>(result);
        var model = Assert.IsAssignableFrom<IEnumerable<StormSessionViewModel>>(
            viewResult.ViewData.Model);
        Assert.Equal(2, model.Count());
    }

    [Fact]
    public async Task IndexPost_ReturnsBadRequestResult_WhenModelStateIsInvalid()
    {
        // Arrange
        var mockRepo = new Mock<IBrainstormSessionRepository>();
        mockRepo.Setup(repo => repo.ListAsync()).Returns(Task.FromResult(GetTestSessions()));
        var controller = new HomeController(mockRepo.Object);
        controller.ModelState.AddModelError("SessionName", "Required");
        var newSession = new HomeController.NewSessionModel();

        // Act
        var result = await controller.Index(newSession);

        // Assert
        var badRequestResult = Assert.IsType<BadRequestObjectResult>(result);
        Assert.IsType<SerializableError>(badRequestResult.Value);
    }

    [Fact]
    public async Task IndexPost_ReturnsARedirectAndAddsSession_WhenModelStateIsValid()
    {
        // Arrange
        var mockRepo = new Mock<IBrainstormSessionRepository>();

The first test confirms when ModelState is not valid, the same ViewResult is returned as for a GET request. Note that the test doesn’t attempt to pass in an invalid model. That wouldn’t work anyway since model binding isn’t running - we’re just calling the method directly. However, we’re not trying to test model binding - we’re only testing what our code in the action method does. The simplest approach is to add an error to ModelState.

The second test verifies that when ModelState is valid, a new BrainstormSession is added (via the repository), and the method returns a RedirectToActionResult with the expected properties. Mocked calls that aren’t called are normally ignored, but calling Verifiable at the end of the setup call allows it to be verified in the test. This is done with the call to mockRepo.Verify.

Note

The Moq library used in this sample makes it easy to mix verifiable, or “strict”, mocks with non-verifiable mocks (also called “loose” mocks or stubs). Learn more about customizing Mock behavior with Moq.

Another controller in the app displays information related to a particular brainstorming session. It includes some logic to deal with invalid id values:

using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc;
using TestingControllersSample.Core.Interfaces;
using TestingControllersSample.ViewModels;

namespace TestingControllersSample.Controllers
{
    public class SessionController : Controller
    {
        private readonly IBrainstormSessionRepository _sessionRepository;

        public SessionController(IBrainstormSessionRepository sessionRepository)
        {
            _sessionRepository = sessionRepository;
        }

        public async Task<IActionResult> Index(int? id)
        {
            if (!id.HasValue)
            {
                return RedirectToAction("Index", "Home");
            }

            var session = await _sessionRepository.GetByIdAsync(id.Value);
            if (session == null)
            {
                return Content("Session not found.");
            }

            var viewModel = new StormSessionViewModel()
            {
                DateCreated = session.DateCreated,
                Name = session.Name,
                Id = session.Id
            };

            return View(viewModel);
        }
    }
}

The controller action has three cases to test, one for each return statement:

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc;
using Moq;
using TestingControllersSample.Controllers;
using TestingControllersSample.Core.Interfaces;
using TestingControllersSample.Core.Model;
using TestingControllersSample.ViewModels;
using Xunit;

namespace TestingControllersSample.Tests.UnitTests
{
    public class SessionControllerTests
    {
        [Fact]
        public async Task IndexReturnsARedirectToIndexHomeWhenIdIsNull()
        {
            // Arrange
            var controller = new SessionController(sessionRepository: null);

            // Act
            var result = await controller.Index(id: null);

            // Arrange
            var redirectToActionResult = Assert.IsType<RedirectToActionResult>(result);
            Assert.Equal("Home", redirectToActionResult.ControllerName);
            Assert.Equal("Index", redirectToActionResult.ActionName);
        }

        [Fact]
        public async Task IndexReturnsContentWithSessionNotFoundWhenSessionNotFound()
        {
            // Arrange
            int testSessionId = 1;
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.GetByIdAsync(testSessionId))
                .Returns(Task.FromResult((BrainstormSession)null));
            var controller = new SessionController(mockRepo.Object);

            // Act
            var result = await controller.Index(testSessionId);

            // Assert
            var contentResult = Assert.IsType<ContentResult>(result);
            Assert.Equal("Session not found.", contentResult.Content);
        }

        [Fact]
        public async Task IndexReturnsViewResultWithStormSessionViewModel()
        {
            // Arrange
            int testSessionId = 1;
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.GetByIdAsync(testSessionId))
                .Returns(Task.FromResult(GetTestSessions().FirstOrDefault(s => s.Id == testSessionId)));
            var controller = new SessionController(mockRepo.Object);

            // Act
            var result = await controller.Index(testSessionId);

            // Assert
            var viewResult = Assert.IsType<ViewResult>(result);
            var model = Assert.IsType<StormSessionViewModel>(viewResult.ViewData.Model);
            Assert.Equal("Test One", model.Name);
            Assert.Equal(2, model.DateCreated.Day);
            Assert.Equal(testSessionId, model.Id);
        }

        private List<BrainstormSession> GetTestSessions()
        {
            var sessions = new List<BrainstormSession>();
            sessions.Add(new BrainstormSession()
            {
                DateCreated = new DateTime(2016, 7, 2),
                Id = 1,
                Name = "Test One"
            });
            sessions.Add(new BrainstormSession()
            {
                DateCreated = new DateTime(2016, 7, 1),
                Id = 2,
                Name = "Test Two"
            });
            return sessions;
        }
    }
}

The app exposes functionality as a web API (a list of ideas associated with a brainstorming session and a method for adding new ideas to a session):

using System;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc;
using TestingControllersSample.ClientModels;
using TestingControllersSample.Core.Interfaces;
using TestingControllersSample.Core.Model;

namespace TestingControllersSample.Api
{
    [Route("api/ideas")]
    public class IdeasController : Controller
    {
        private readonly IBrainstormSessionRepository _sessionRepository;

        public IdeasController(IBrainstormSessionRepository sessionRepository)
        {
            _sessionRepository = sessionRepository;
        }

        [Route("forsession/{sessionId}")]
        [HttpGet]
        public async Task<IActionResult> ForSession(int sessionId)
        {
            var session = await _sessionRepository.GetByIdAsync(sessionId);
            if (session == null)
            {
                return NotFound(sessionId);
            }

            var result = session.Ideas.Select(idea => new IdeaDTO()
            {
                Id = idea.Id,
                Name = idea.Name,
                Description = idea.Description,
                DateCreated = idea.DateCreated
            }).ToList();

            return Ok(result);
        }

        [Route("create")]
        [HttpPost]
        public async Task<IActionResult> Create([FromBody]NewIdeaModel model)
        {
            if (!ModelState.IsValid)
            {
                return BadRequest(ModelState);
            }

            var session = await _sessionRepository.GetByIdAsync(model.SessionId);
            if (session == null)
            {
                return NotFound(model.SessionId);
            }

            var idea = new Idea()
            {
                DateCreated = DateTimeOffset.Now,
                Description = model.Description,
                Name = model.Name
            };
            session.AddIdea(idea);

            await _sessionRepository.UpdateAsync(session);

            return Ok(session);
        }
    }
}

The ForSession method returns a list of IdeaDTO types, with property names camel cased to match JavaScript conventions. Avoid returning your business domain entities directly via API calls, since frequently they include more data than the API client requires, and they unnecessarily couple your app’s internal domain model with the API you expose externally. Mapping between domain entities and the types you will return over the wire can be done manually (using a LINQ Select as shown here) or using a library like AutoMapper

The unit tests for the Create and ForSession API methods:

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Mvc;
using Moq;
using TestingControllersSample.Api;
using TestingControllersSample.ClientModels;
using TestingControllersSample.Core.Interfaces;
using TestingControllersSample.Core.Model;
using Xunit;

namespace TestingControllersSample.Tests.UnitTests
{
    public class ApiIdeasControllerTests
    {
        [Fact]
        public async Task Create_ReturnsBadRequest_GivenInvalidModel()
        {
            // Arrange & Act
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            var controller = new IdeasController(mockRepo.Object);
            controller.ModelState.AddModelError("error","some error");

            // Act
            var result = await controller.Create(model: null);

            // Assert
            Assert.IsType<BadRequestObjectResult>(result);
        }

        [Fact]
        public async Task Create_ReturnsHttpNotFound_ForInvalidSession()
        {
            // Arrange
            int testSessionId = 123;
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.GetByIdAsync(testSessionId))
                .Returns(Task.FromResult((BrainstormSession)null));
            var controller = new IdeasController(mockRepo.Object);

            // Act
            var result = await controller.Create(new NewIdeaModel());

            // Assert
            Assert.IsType<NotFoundObjectResult>(result);
        }

        [Fact]
        public async Task Create_ReturnsNewlyCreatedIdeaForSession()
        {
            // Arrange
            int testSessionId = 123;
            string testName = "test name";
            string testDescription = "test description";
            var testSession = GetTestSession();
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.GetByIdAsync(testSessionId))
                .Returns(Task.FromResult(testSession));
            var controller = new IdeasController(mockRepo.Object);

            var newIdea = new NewIdeaModel()
            {
                Description = testDescription,
                Name = testName,
                SessionId = testSessionId
            };
            mockRepo.Setup(repo => repo.UpdateAsync(testSession))
                .Returns(Task.CompletedTask)
                .Verifiable();

            // Act
            var result = await controller.Create(newIdea);

            // Assert
            var okResult = Assert.IsType<OkObjectResult>(result);
            var returnSession = Assert.IsType<BrainstormSession>(okResult.Value);
            mockRepo.Verify();
            Assert.Equal(2, returnSession.Ideas.Count());
            Assert.Equal(testName, returnSession.Ideas.LastOrDefault().Name);
            Assert.Equal(testDescription, returnSession.Ideas.LastOrDefault().Description);
        }

        [Fact]
        public async Task ForSession_ReturnsHttpNotFound_ForInvalidSession()
        {
            // Arrange
            int testSessionId = 123;
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.GetByIdAsync(testSessionId))
                .Returns(Task.FromResult((BrainstormSession)null));
            var controller = new IdeasController(mockRepo.Object);

            // Act
            var result = await controller.ForSession(testSessionId);

            // Assert
            var notFoundObjectResult = Assert.IsType<NotFoundObjectResult>(result);
            Assert.Equal(testSessionId, notFoundObjectResult.Value);
        }

        [Fact]
        public async Task ForSession_ReturnsIdeasForSession()
        {
            // Arrange
            int testSessionId = 123;
            var mockRepo = new Mock<IBrainstormSessionRepository>();
            mockRepo.Setup(repo => repo.GetByIdAsync(testSessionId)).Returns(Task.FromResult(GetTestSession()));
            var controller = new IdeasController(mockRepo.Object);

            // Act
            var result = await controller.ForSession(testSessionId);

            // Assert
            var okResult = Assert.IsType<OkObjectResult>(result);
            var returnValue = Assert.IsType<List<IdeaDTO>>(okResult.Value);
            var idea = returnValue.FirstOrDefault();
            Assert.Equal("One", idea.Name);
        }

        private BrainstormSession GetTestSession()
        {
            var session = new BrainstormSession()
            {
                DateCreated = new DateTime(2016, 7, 2),
                Id = 1,
                Name = "Test One"
            };

            var idea = new Idea() { Name = "One" };
            session.AddIdea(idea);
            return session;
        }
    }
}

As stated previously, to test the behavior of the method when ModelState is invalid, add a model error to the controller as part of the test. Don’t try to test model validation or model binding in your unit tests - just test your action method’s behavior when confronted with a particular ModelState value.

The second test depends on the repository returning null, so the mock repository is configured to return null. There’s no need to create a test database (in memory or otherwise) and construct a query that will return this result - it can be done in a single line as shown.

The last test verifies that the repository’s Update method is called. As we did previously, the mock is called with Verifiable and then the mocked repository’s Verify method is called to confirm the verifiable method was executed. It’s not a unit test responsibility to ensure that the Update method saved the data; that can be done with an integration test.

Integration Testing ¶

Integration testing is done to ensure separate modules within your app work correctly together. Generally, anything you can test with a unit test, you can also test with an integration test, but the reverse isn’t true. However, integration tests tend to be much slower than unit tests. Thus, it’s best to test whatever you can with unit tests, and use integration tests for scenarios that involve multiple collaborators.

Although they may still be useful, mock objects are rarely used in integration tests. In unit testing, mock objects are an effective way to control how collaborators outside of the unit being tested should behave for the purposes of the test. In an integration test, real collaborators are used to confirm the whole subsystem works together correctly.

Application State¶

One important consideration when performing integration testing is how to set your app’s state. Tests need to run independent of one another, and so each test should start with the app in a known state. If your app doesn’t use a database or have any persistence, this may not be an issue. However, most real-world apps persist their state to some kind of data store, so any modifications made by one test could impact another test unless the data store is reset. Using the built-in TestServer, it’s very straightforward to host ASP.NET Core apps within our integration tests, but that doesn’t necessarily grant access to the data it will use. If you’re using an actual database, one approach is to have the app connect to a test database, which your tests can access and ensure is reset to a known state before each test executes.

In this sample application, I’m using Entity Framework Core’s InMemoryDatabase support, so I can’t just connect to it from my test project. Instead, I expose an InitializeDatabase method from the app’s Startup class, which I call when the app starts up if it’s in the Development environment. My integration tests automatically benefit from this as long as they set the environment to Development. I don’t have to worry about resetting the database, since the InMemoryDatabase is reset each time the app restarts.

The Startup class:

using System;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNetCore.Builder;
using Microsoft.AspNetCore.Hosting;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Logging;
using TestingControllersSample.Core.Interfaces;
using TestingControllersSample.Core.Model;
using TestingControllersSample.Infrastructure;

namespace TestingControllersSample
{
    public class Startup
    {
        public void ConfigureServices(IServiceCollection services)
        {
            services.AddDbContext<AppDbContext>(
                optionsBuilder => optionsBuilder.UseInMemoryDatabase());

            services.AddMvc();

            services.AddScoped<IBrainstormSessionRepository,
                EFStormSessionRepository>();
        }

        public void Configure(IApplicationBuilder app,
            IHostingEnvironment env,
            ILoggerFactory loggerFactory)
        {
            loggerFactory.AddConsole(LogLevel.Warning);

            if (env.IsDevelopment())
            {
                app.UseDeveloperExceptionPage();

                var repository = app.ApplicationServices.GetService<IBrainstormSessionRepository>();
                InitializeDatabaseAsync(repository).Wait();
            }

            app.UseStaticFiles();

            app.UseMvcWithDefaultRoute();
        }

        public async Task InitializeDatabaseAsync(IBrainstormSessionRepository repo)
        {
            var sessionList = await repo.ListAsync();
            if (!sessionList.Any())
            {
                await repo.AddAsync(GetTestSession());
            }
        }

        public static BrainstormSession GetTestSession()
        {
            var session = new BrainstormSession()
            {
                Name = "Test Session 1",
                DateCreated = new DateTime(2016, 8, 1)
            };
            var idea = new Idea()
            {
                DateCreated = new DateTime(2016, 8, 1),
                Description = "Totally awesome idea",
                Name = "Awesome idea"
            };
            session.AddIdea(idea);
            return session;
        }
    }
}

You’ll see the GetTestSession method used frequently in the integration tests below.

Accessing Views¶

Each integration test class configures the TestServer that will run the ASP.NET Core app. By default, TestServer hosts the web app in the folder where it’s running - in this case, the test project folder. Thus, when you attempt to test controller actions that return ViewResult, you may see this error:

The view 'Index' was not found. The following locations were searched:
(list of locations)

To correct this issue, you need to configure the server to use the ApplicationBasePath and ApplicationName of the web project. This is done in the call to UseServices in the integration test class shown:

using System;
using System.Collections.Generic;
using System.Net;
using System.Net.Http;
using System.Threading.Tasks;
using Xunit;

namespace TestingControllersSample.Tests.IntegrationTests
{
    public class HomeControllerTests : IClassFixture<TestFixture<TestingControllersSample.Startup>>
    {
        private readonly HttpClient _client;

        public HomeControllerTests(TestFixture<TestingControllersSample.Startup> fixture)
        {
            _client = fixture.Client;
        }

        [Fact]
        public async Task ReturnsInitialListOfBrainstormSessions()
        {
            // Arrange
            var testSession = Startup.GetTestSession();

            // Act
            var response = await _client.GetAsync("/");

            // Assert
            response.EnsureSuccessStatusCode();
            var responseString = await response.Content.ReadAsStringAsync();
            Assert.True(responseString.Contains(testSession.Name));
        }

        [Fact]
        public async Task PostAddsNewBrainstormSession()
        {
            // Arrange
            string testSessionName = Guid.NewGuid().ToString();
            var data = new Dictionary<string, string>();
            data.Add("SessionName", testSessionName);
            var content = new FormUrlEncodedContent(data);

            // Act
            var response = await _client.PostAsync("/", content);

            // Assert
            Assert.Equal(HttpStatusCode.Redirect, response.StatusCode);
            Assert.Equal("/", response.Headers.Location.ToString());
        }
    }
}

In the test above, the responseString gets the actual rendered HTML from the View, which can be inspected to confirm it contains expected results.

API Methods¶

If your app exposes web APIs, it’s a good idea to have automated tests confirm they execute as expected. The built-in TestServer makes it easy to test web APIs. If your API methods are using model binding, you should always check ModelState.IsValid, and integration tests are the right place to confirm that your model validation is working properly.

The following set of tests target the Create method in the IdeasController class shown above:

        [Fact]
        public async Task CreatePostReturnsBadRequestForMissingNameValue()
        {
            // Arrange
            var newIdea = new NewIdeaDto("", "Description", 1);

            // Act
            var response = await _client.PostAsJsonAsync("/api/ideas/create", newIdea);

            // Assert
            Assert.Equal(HttpStatusCode.BadRequest, response.StatusCode);
        }

        [Fact]
        public async Task CreatePostReturnsBadRequestForMissingDescriptionValue()
        {
            // Arrange
            var newIdea = new NewIdeaDto("Name", "", 1);

            // Act
            var response = await _client.PostAsJsonAsync("/api/ideas/create", newIdea);

            // Assert
            Assert.Equal(HttpStatusCode.BadRequest, response.StatusCode);
        }

        [Fact]
        public async Task CreatePostReturnsBadRequestForSessionIdValueTooSmall()
        {
            // Arrange
            var newIdea = new NewIdeaDto("Name", "Description", 0);

            // Act
            var response = await _client.PostAsJsonAsync("/api/ideas/create", newIdea);

            // Assert
            Assert.Equal(HttpStatusCode.BadRequest, response.StatusCode);
        }

        [Fact]
        public async Task CreatePostReturnsBadRequestForSessionIdValueTooLarge()
        {
            // Arrange
            var newIdea = new NewIdeaDto("Name", "Description", 1000001);

            // Act
            var response = await _client.PostAsJsonAsync("/api/ideas/create", newIdea);

            // Assert
            Assert.Equal(HttpStatusCode.BadRequest, response.StatusCode);
        }

        [Fact]
        public async Task CreatePostReturnsNotFoundForInvalidSession()
        {
            // Arrange
            var newIdea = new NewIdeaDto("Name", "Description", 123);

            // Act
            var response = await _client.PostAsJsonAsync("/api/ideas/create", newIdea);

            // Assert
            Assert.Equal(HttpStatusCode.NotFound, response.StatusCode);
        }

        [Fact]
        public async Task CreatePostReturnsCreatedIdeaWithCorrectInputs()
        {
            // Arrange
            var testIdeaName = Guid.NewGuid().ToString();
            var newIdea = new NewIdeaDto(testIdeaName, "Description", 1);

            // Act
            var response = await _client.PostAsJsonAsync("/api/ideas/create", newIdea);

            // Assert
            response.EnsureSuccessStatusCode();
            var returnedSession = await response.Content.ReadAsJsonAsync<BrainstormSession>();
            Assert.Equal(2, returnedSession.Ideas.Count);
            Assert.True(returnedSession.Ideas.Any(i => i.Name == testIdeaName));
        }

        [Fact]
        public async Task ForSessionReturnsNotFoundForBadSessionId()
        {
            // Arrange & Act
            var response = await _client.GetAsync("/api/ideas/forsession/500");

            // Assert
            Assert.Equal(HttpStatusCode.NotFound, response.StatusCode);
        }

        [Fact]
        public async Task ForSessionReturnsIdeasForValidSessionId()
        {
            // Arrange
            var testSession = Startup.GetTestSession();

            // Act
            var response = await _client.GetAsync("/api/ideas/forsession/1");

            // Assert
            response.EnsureSuccessStatusCode();
            var ideaList = JsonConvert.DeserializeObject<List<IdeaDTO>>(
                await response.Content.ReadAsStringAsync());
            var firstIdea = ideaList.First();

Unlike integration tests of actions that returns HTML views, web API methods that return results can usually be deserialized as strongly typed objects, as the last test above shows. In this case, the test deserializes the result to a BrainstormSession instance, and confirms that the idea was correctly added to its collection of ideas.

You’ll find additional examples of integration tests in this article’s sample project.

Areas¶

By Tom Archer

Areas provide a way to separate a large MVC application into semantically-related groups of models, views, and controllers. Let’s take a look at an example to illustrate how Areas are created and used. Let’s say you have a store app that has two distinct groupings of controllers and views: Products and Services.

Instead of having all of the controllers located under the Controllers parent directory, and all the views located under the Views parent directory, you could use Areas to group your views and controllers according to the area (or logical grouping) with which they’re associated.

Project name
- Areas
  - Products
    - Controllers
      - HomeController.cs
    - Views
      - Home
        
        Index.cshtml
  - Services
    - Controllers
      - HomeController.cs
    - Views
      - Home
        
        Index.cshtml

Looking at the preceding directory hierarchy example, there are a few guidelines to keep in mind when defining areas:

A directory called Areas must exist as a child directory of the project.
The Areas directory contains a subdirectory for each of your project’s areas (Products and Services, in this example).
Your controllers should be located as follows: /Areas/[area]/Controllers/[controller].cs
Your views should be located as follows: /Areas/[area]/Views/[controller]/[action].cshtml

Note that if you have a view that is shared across controllers, it can be located in either of the following locations:

/Areas/[area]/Views/Shared/[action].cshtml
/Views/Shared/[action].cshtml

Once you’ve defined the folder hierarchy, you need to tell MVC that each controller is associated with an area. You do that by decorating the controller name with the [Area] attribute.

...
namespace MyStore.Areas.Products.Controllers
{
    [Area("Products")]
    public class HomeController : Controller
    {
        // GET: /<controller>/
        public IActionResult Index()
        {
            return View();
        }
    }
}

The final step is to set up a route definition that works with your newly created areas. The 🔧 Routing to Controller Actions article goes into detail about how to create route definitions, including using conventional routes versus attribute routes. In this example, we’ll use a conventional route. To do so, simply open the Startup.cs file and modify it by adding the highlighted route definition below.

...
app.UseMvc(routes =>
{
  routes.MapRoute(name: "areaRoute",
    template: "{area:exists}/{controller=Home}/{action=Index}");

  routes.MapRoute(
      name: "default",
      template: "{controller=Home}/{action=Index}");
});

Now, when the user browses to http://<yourApp>/products, the Index action method of the HomeController in the Products area will be invoked.

Linking between areas¶

To link between areas, you simply specify the area in which the controller is defined by using Tag Helpers.

The following snippet shows how to link to a controller action that is defined within an area named Products.

<a asp-area="Products" asp-controller="Home" asp-action="Index">See Products Home Page</a>

To link to a controller action that is not part of an area, supply an empty value to the asp-area attribute.

<a asp-area="" asp-controller="Home" asp-action="Index">Go to Home Page</a>

Summary¶

Areas are a very useful tool for grouping semantically-related controllers and actions under a common parent folder. In this article, you learned how to set up your folder hierarchy to support Areas, how to specify the [Area] attribute to denote a controller as belonging to a specified area, and how to define your routes with areas.

🔧 Working with the Application Model¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Testing¶

Integration Testing¶

Integration testing ensures that an application’s components function correctly when assembled together. ASP.NET Core supports integration testing using unit test frameworks and a built-in test web host that can be used to handle requests without network overhead.

Sections:

Introduction to Integration Testing
Integration Testing ASP.NET Core
Refactoring to use Middleware
Summary
Additional Resources

Introduction to Integration Testing ¶

Integration tests verify that different parts of an application work correctly together. Unlike Unit testing, integration tests frequently involve application infrastructure concerns, such as a database, file system, network resources, or web requests and responses. Unit tests use fakes or mock objects in place of these concerns, but the purpose of integration tests is to confirm that the system works as expected with these systems.

Integration tests, because they exercise larger segments of code and because they rely on infrastructure elements, tend to be orders of magnitude slower than unit tests. Thus, it’s a good idea to limit how many integration tests you write, especially if you can test the same behavior with a unit test.

Tip

If some behavior can be tested using either a unit test or an integration test, prefer the unit test, since it will be almost always be faster. You might have dozens or hundreds of unit tests with many different inputs, but just a handful of integration tests covering the most important scenarios.

Testing the logic within your own methods is usually the domain of unit tests. Testing how your application works within its framework (e.g. ASP.NET Core) or with a database is where integration tests come into play. It doesn’t take too many integration tests to confirm that you’re able to write a row to and then read a row from the database. You don’t need to test every possible permutation of your data access code - you only need to test enough to give you confidence that your application is working properly.

Integration Testing ASP.NET Core ¶

To get set up to run integration tests, you’ll need to create a test project, refer to your ASP.NET Core web project, and install a test runner. This process is described in the Unit testing documentation, along with more detailed instructions on running tests and recommendations for naming your tests and test classes.

Tip

Separate your unit tests and your integration tests using different projects. This helps ensure you don’t accidentally introduce infrastructure concerns into your unit tests, and lets you easily choose to run all tests, or just one set or the other.

The Test Host¶

ASP.NET Core includes a test host that can be added to integration test projects and used to host ASP.NET Core applications, serving test requests without the need for a real web host. The provided sample includes an integration test project which has been configured to use xUnit and the Test Host, as you can see from this excerpt from its project.json file:

"dependencies": {
  "PrimeWeb": "1.0.0",
  "xunit": "2.1.0",
  "dotnet-test-xunit": "1.0.0-rc2-build10025",
  "Microsoft.AspNetCore.TestHost": "1.0.0"
},

Once the Microsoft.AspNetCore.TestHost package is included in the project, you will be able to create and configure a TestServer in your tests. The following test shows how to verify that a request made to the root of a site returns “Hello World!” and should run successfully against the default ASP.NET Core Empty Web template created by Visual Studio.

private readonly TestServer _server;
private readonly HttpClient _client;
public PrimeWebDefaultRequestShould()
{
    // Arrange
    _server = new TestServer(new WebHostBuilder()
        .UseStartup<Startup>());
    _client = _server.CreateClient();
}

[Fact]
public async Task ReturnHelloWorld()
{
    // Act
    var response = await _client.GetAsync("/");
    response.EnsureSuccessStatusCode();

    var responseString = await response.Content.ReadAsStringAsync();

    // Assert
    Assert.Equal("Hello World!",
        responseString);
}

These tests are using the Arrange-Act-Assert pattern, but in this case all of the Arrange step is done in the constructor, which creates an instance of TestServer. A configured WebHostBuilder will be used to create a TestHost; in this example we are passing in the Configure method from our system under test (SUT)’s Startup class. This method will be used to configure the request pipeline of the TestServer identically to how the SUT server would be configured.

In the Act portion of the test, a request is made to the TestServer instance for the “/” path, and the response is read back into a string. This string is then compared with the expected string of “Hello World!”. If they match, the test passes, otherwise it fails.

Now we can add a few additional integration tests to confirm that the prime checking functionality works via the web application:

public class PrimeWebCheckPrimeShould
{
    private readonly TestServer _server;
    private readonly HttpClient _client;
    public PrimeWebCheckPrimeShould()
    {
        // Arrange
        _server = new TestServer(new WebHostBuilder()
            .UseStartup<Startup>());
        _client = _server.CreateClient();
    }

    private async Task<string> GetCheckPrimeResponseString(
        string querystring = "")
    {
        var request = "/checkprime";
        if(!string.IsNullOrEmpty(querystring))
        {
            request += "?" + querystring;
        }
        var response = await _client.GetAsync(request);
        response.EnsureSuccessStatusCode();

        return await response.Content.ReadAsStringAsync();
    }

    [Fact]
    public async Task ReturnInstructionsGivenEmptyQueryString()
    {
        // Act
        var responseString = await GetCheckPrimeResponseString();

        // Assert
        Assert.Equal("Pass in a number to check in the form /checkprime?5",
            responseString);
    }
    [Fact]
    public async Task ReturnPrimeGiven5()
    {
        // Act
        var responseString = await GetCheckPrimeResponseString("5");

        // Assert
        Assert.Equal("5 is prime!",
            responseString);
    }

    [Fact]
    public async Task ReturnNotPrimeGiven6()
    {
        // Act
        var responseString = await GetCheckPrimeResponseString("6");

        // Assert
        Assert.Equal("6 is NOT prime!",
            responseString);
    }
}

Note that we’re not really trying to test the correctness of our prime number checker with these tests, but rather that the web application is doing what we expect. We already have unit test coverage that gives us confidence in PrimeService, as you can see here:

Note

You can learn more about the unit tests in the Unit testing article.

Now that we have a set of passing tests, it’s a good time to think about whether we’re happy with the current way in which we’ve designed our application. If we see any code smells, now may be a good time to refactor the application to improve its design.

Refactoring to use Middleware ¶

Refactoring is the process of changing an application’s code to improve its design without changing its behavior. It should ideally be done when there is a suite of passing tests, since these help ensure the system’s behavior remains the same before and after the changes. Looking at the way in which the prime checking logic is implemented in our web application, we see:

  public void Configure(IApplicationBuilder app,
      IHostingEnvironment env)
  {
      if (env.IsDevelopment())
      {
          app.UseDeveloperExceptionPage();
      }

      app.Run(async (context) =>
      {
          if (context.Request.Path.Value.Contains("checkprime"))
          {
              int numberToCheck;
              try
              {
                  numberToCheck = int.Parse(context.Request.QueryString.Value.Replace("?",""));
                  var primeService = new PrimeService();
                  if (primeService.IsPrime(numberToCheck))
                  {
                      await context.Response.WriteAsync(numberToCheck + " is prime!");
                  }
                  else
                  {
                      await context.Response.WriteAsync(numberToCheck + " is NOT prime!");
                  }
              }
              catch
              {
                  await context.Response.WriteAsync("Pass in a number to check in the form /checkprime?5");
              }
          }
          else
          {
              await context.Response.WriteAsync("Hello World!");
          }
      });
  }

This code works, but it’s far from how we would like to implement this kind of functionality in an ASP.NET Core application, even as simple a one as this is. Imagine what the Configure method would look like if we needed to add this much code to it every time we added another URL endpoint!

One option we can consider is adding MVC to the application, and creating a controller to handle the prime checking. However, assuming we don’t currently need any other MVC functionality, that’s a bit overkill.

We can, however, take advantage of ASP.NET Core middleware, which will help us encapsulate the prime checking logic in its own class and achieve better separation of concerns within the Configure method.

We want to allow the path the middleware uses to be specified as a parameter, so the middleware class expects a RequestDelegate and a PrimeCheckerOptions instance in its constructor. If the path of the request doesn’t match what this middleware is configured to expect, we simply call the next middleware in the chain and do nothing further. The rest of the implementation code that was in Configure is now in the Invoke method.

Note

Since our middleware depends on the PrimeService service, we are also requesting an instance of this service via the constructor. The framework will provide this service via Dependency Injection, assuming it has been configured (e.g. in ConfigureServices).

using Microsoft.AspNetCore.Builder;
using Microsoft.AspNetCore.Http;
using PrimeWeb.Services;
using System;
using System.Threading.Tasks;

namespace PrimeWeb.Middleware
{
    public class PrimeCheckerMiddleware
    {
        private readonly RequestDelegate _next;
        private readonly PrimeCheckerOptions _options;
        private readonly PrimeService _primeService;

        public PrimeCheckerMiddleware(RequestDelegate next,
            PrimeCheckerOptions options,
            PrimeService primeService)
        {
            if (next == null)
            {
                throw new ArgumentNullException(nameof(next));
            }
            if (options == null)
            {
                throw new ArgumentNullException(nameof(options));
            }
            if (primeService == null)
            {
                throw new ArgumentNullException(nameof(primeService));
            }

            _next = next;
            _options = options;
            _primeService = primeService;
        }

        public async Task Invoke(HttpContext context)
        {
            var request = context.Request;
            if (!request.Path.HasValue ||
                request.Path != _options.Path)
            {
                await _next.Invoke(context);
            }
            else
            {
                int numberToCheck;
                if (int.TryParse(request.QueryString.Value.Replace("?", ""), out numberToCheck))
                {
                    if (_primeService.IsPrime(numberToCheck))
                    {
                        await context.Response.WriteAsync($"{numberToCheck} is prime!");
                    }
                    else
                    {
                        await context.Response.WriteAsync($"{numberToCheck} is NOT prime!");
                    }
                }
                else
                {
                    await context.Response.WriteAsync($"Pass in a number to check in the form {_options.Path}?5");
                }
            }
        }
    }
}

Note

Since this middleware acts as an endpoint in the request delegate chain when its path matches, there is no call to _next.Invoke in the case where this middleware handles the request.

With this middleware in place and some helpful extension methods created to make configuring it easier, the refactored Configure method looks like this:

public void Configure(IApplicationBuilder app,
    IHostingEnvironment env)
{
    if (env.IsDevelopment())
    {
        app.UseDeveloperExceptionPage();
    }

    app.UsePrimeChecker();

    app.Run(async (context) =>
    {
        await context.Response.WriteAsync("Hello World!");
    });
}

Following this refactoring, we are confident that the web application still works as before, since our integration tests are all passing.

Tip

It’s a good idea to commit your changes to source control after you complete a refactoring and your tests all pass. If you’re practicing Test Driven Development, consider adding Commit to your Red-Green-Refactor cycle.

Summary ¶

Integration testing provides a higher level of verification than unit testing. It tests application infrastructure and how different parts of an application work together. ASP.NET Core is very testable, and ships with a TestServer that makes wiring up integration tests for web server endpoints very easy.

Additional Resources ¶

Working with Data¶

Getting Started with ASP.NET Core and Entity Framework 6¶

By Paweł Grudzień and Damien Pontifex

This article will show you how to use Entity Framework 6 inside an ASP.NET Core application.

Sections:

Prerequisites
Setup connection strings and dependency injection
Migrate configuration from config to code
Summary
Additional Resources

Prerequisites ¶

Before you start, make sure that you compile against full .NET Framework in your project.json as Entity Framework 6 does not support .NET Core. If you need cross platform features you will need to upgrade to Entity Framework Core.

In your project.json file specify a single target for the full .NET Framework:

"frameworks": {
    "net46": {}
}

Setup connection strings and dependency injection ¶

The simplest change is to explicitly get your connection string and setup dependency injection of your DbContext instance.

In your DbContext subclass, ensure you have a constructor which takes the connection string as so:

public class ApplicationDbContext : DbContext
{
    public ApplicationDbContext(string nameOrConnectionString) : base(nameOrConnectionString)
    {
    }
}

In the Startup class within ConfigureServices add factory method of your context with it’s connection string. Context should be resolved once per scope to ensure performance and ensure reliable operation of Entity Framework.

public void ConfigureServices(IServiceCollection services)
{
    services.AddScoped((_) => new ApplicationDbContext(Configuration["Data:DefaultConnection:ConnectionString"]));

    // Configure remaining services
}

Migrate configuration from config to code ¶

Entity Framework 6 allows configuration to be specified in xml (in web.config or app.config) or through code. As of ASP.NET Core, all configuration is code-based.

Code-based configuration is achieved by creating a subclass of System.Data.Entity.Config.DbConfiguration and applying System.Data.Entity.DbConfigurationTypeAttribute to your DbContext subclass.

Our config file typically looked like this:

<entityFramework>
    <defaultConnectionFactory type="System.Data.Entity.Infrastructure.LocalDbConnectionFactory, EntityFramework">
        <parameters>
            <parameter value="mssqllocaldb" />
        </parameters>
    </defaultConnectionFactory>
    <providers>
        <provider invariantName="System.Data.SqlClient" type="System.Data.Entity.SqlServer.SqlProviderServices, EntityFramework.SqlServer" />
    </providers>
</entityFramework>

The defaultConnectionFactory element sets the factory for connections. If this attribute is not set then the default value is SqlConnectionProvider. If, on the other hand, value is provided, the given class will be used to create DbConnection with its CreateConnection method. If the given factory has no default constructor then you must add parameters that are used to construct the object.

[DbConfigurationType(typeof(CodeConfig))] // point to the class that inherit from DbConfiguration
public class ApplicationDbContext : DbContext
{
    [...]
}

public class CodeConfig : DbConfiguration
{
    public CodeConfig()
    {
        SetProviderServices("System.Data.SqlClient",
            System.Data.Entity.SqlServer.SqlProviderServices.Instance);
    }
}

SQL Server, SQL Server Express and LocalDB¶

This is the default and so no explicit configuration is needed. The above CodeConfig class can be used to explicitly set the provider services and the appropriate connection string should be passed to the DbContext constructor as shown above.

Summary ¶

Entity Framework 6 is an object relational mapping (ORM) library, that is capable of mapping your classes to database entities with little effort. These features made it very popular so migrating large portions of code may be undesirable for many projects. This article shows how to avoid migration to focus on other new features of ASP.NET.

Additional Resources ¶

Entity Framework - Code-Based Configuration

Azure Storage¶

Client-Side Development¶

Using Gulp¶

By Erik Reitan, Scott Addie and Daniel Roth

In a typical modern web application, the build process might:

Bundle and minify JavaScript and CSS files.
Run tools to call the bundling and minification tasks before each build.
Compile LESS or SASS files to CSS.
Compile CoffeeScript or TypeScript files to JavaScript.

A task runner is a tool which automates these routine development tasks and more. Visual Studio provides built-in support for two popular JavaScript-based task runners: Gulp and Grunt.

Sections:

Introducing Gulp
Gulp Starter Tasks in Visual Studio
Running Default Tasks
Defining and Running a New Task
Defining and Running Tasks in a Series
IntelliSense
Development, Staging, and Production Environments
Switching Between Environments
Task and Module Details
Gulp Recipes
Summary
See Also

Introducing Gulp ¶

Gulp is a JavaScript-based streaming build toolkit for client-side code. It is commonly used to stream client-side files through a series of processes when a specific event is triggered in a build environment. Some advantages of using Gulp include the automation of common development tasks, the simplification of repetitive tasks, and a decrease in overall development time. For instance, Gulp can be used to automate bundling and minification or the cleansing of a development environment before a new build.

The ASP.NET Core Web Application project template is used to help you get started designing and coding a new Web application in Visual Studio. It contains default functionality to demonstrate many aspects of ASP.NET. The template also includes Node Package Manager (npm) and Gulp, making it easier to add bundling and minification to a project.

Note

You don’t need the ASP.NET Core Web Application project template or Visual Studio to implement bundling and minification. For example, create an ASP.NET project using Yeoman, push it to GitHub, clone it on a Mac, and then bundle and minify the project.

When you create a new web project using ASP.NET Core Web Application template, Visual Studio includes the Gulp.js npm package, the gulpfile.js file, and a set of Gulp dependencies. The npm package contains all the prerequisites for running Gulp tasks in your Visual Studio project. The provided gulpfile.js file defines a set of Gulp tasks which can be run from the Task Runner Explorer window in Visual Studio. The devDependencies section of the package.json file specifies the development-time dependencies to install. These dependencies are not deployed with the application. You can add new packages to devDependencies and save the file:

{
  "name": "ASP.NET",
  "version": "0.0.0",
  "devDependencies": {
  "gulp": "3.8.11",
  "gulp-concat": "2.5.2",
  "gulp-cssmin": "0.1.7",
  "gulp-uglify": "1.2.0",
  "rimraf": "2.2.8"
  }
}

After adding a new key-value pair in devDependencies and saving the file, Visual Studio will download and install the corresponding version of the package. In Solution Explorer, these packages are found in Dependencies > npm.

Gulp Starter Tasks in Visual Studio ¶

A starter set of Gulp tasks is defined in gulpfile.js. These tasks delete and minify the CSS and JavaScript files. The following JavaScript, from the first half of gulpfile.js, includes Gulp modules and specifies file paths to be referenced within the forthcoming tasks:

/// <binding Clean='clean' />
"use strict";

var gulp = require("gulp"),
  rimraf = require("rimraf"),
  concat = require("gulp-concat"),
  cssmin = require("gulp-cssmin"),
  uglify = require("gulp-uglify");

var paths = {
  webroot: "./wwwroot/"
};

paths.js = paths.webroot + "js/**/*.js";
paths.minJs = paths.webroot + "js/**/*.min.js";
paths.css = paths.webroot + "css/**/*.css";
paths.minCss = paths.webroot + "css/**/*.min.css";
paths.concatJsDest = paths.webroot + "js/site.min.js";
paths.concatCssDest = paths.webroot + "css/site.min.css";

The above code specifies which Node modules are required. The require function imports each module so that the dependent tasks can utilize their features. Each of the imported modules is assigned to a variable. The modules can be located either by name or path. In this example, the modules named gulp, rimraf, gulp-concat, gulp-cssmin, and gulp-uglify are retrieved by name. Additionally, a series of paths are created so that the locations of CSS and JavaScript files can be reused and referenced within the tasks. The following table provides descriptions of the modules included in gulpfile.js.

Module Name	Description
gulp	The Gulp streaming build system. For more information, see gulp.
rimraf	A Node deletion module. For more information, see rimraf.
gulp-concat	A module that will concatenate files based on the operating system’s newline character. For more information, see gulp-concat.
gulp-cssmin	A module that will minify CSS files. For more information, see gulp-cssmin.
gulp-uglify	A module that minifies .js files using the UglifyJS toolkit. For more information, see gulp-uglify.

Once the requisite modules are imported in gulpfile.js, the tasks can be specified. Visual Studio registers six tasks, represented by the following code in gulpfile.js:

gulp.task("clean:js", function (cb) {
  rimraf(paths.concatJsDest, cb);
});

gulp.task("clean:css", function (cb) {
  rimraf(paths.concatCssDest, cb);
});

gulp.task("clean", ["clean:js", "clean:css"]);

gulp.task("min:js", function () {
  return gulp.src([paths.js, "!" + paths.minJs], { base: "." })
    .pipe(concat(paths.concatJsDest))
    .pipe(uglify())
    .pipe(gulp.dest("."));
});

gulp.task("min:css", function () {
  return gulp.src([paths.css, "!" + paths.minCss])
    .pipe(concat(paths.concatCssDest))
    .pipe(cssmin())
    .pipe(gulp.dest("."));
});

gulp.task("min", ["min:js", "min:css"]);

The following table provides an explanation of the tasks specified in the code above:

Task Name	Description
clean:js	A task that uses the rimraf Node deletion module to remove the minified version of the site.js file.
clean:css	A task that uses the rimraf Node deletion module to remove the minified version of the site.css file.
clean	A task that calls the `clean:js` task, followed by the `clean:css` task.
min:js	A task that minifies and concatenates all .js files within the js folder. The .min.js files are excluded.
min:css	A task that minifies and concatenates all .css files within the css folder. The .min.css files are excluded.
min	A task that calls the `min:js` task, followed by the `min:css` task.

Running Default Tasks ¶

If you haven’t already created a new Web app, create a new ASP.NET Web Application project in Visual Studio.

Select File > New > Project from the menu bar. The New Project dialog box is displayed.

Select the ASP.NET Web Application template, choose a project name, and click OK.
In the New ASP.NET Project dialog box, select the ASP.NET Core Web Application template and click OK.
In Solution Explorer, right-click gulpfile.js, and select Task Runner Explorer.

Task Runner Explorer shows the list of Gulp tasks. In the default ASP.NET Core Web Application template in Visual Studio, there are six tasks included from gulpfile.js.

Underneath Tasks in Task Runner Explorer, right-click clean, and select Run from the pop-up menu.

Task Runner Explorer will create a new tab named clean and execute the related clean task as it is defined in gulpfile.js.

Next, right-click the clean task, then select Bindings > Before Build.

The Before Build binding option allows the clean task to run automatically before each build of the project.

It’s worth noting that the bindings you set up with Task Runner Explorer are not stored in the project.json. Rather they are stored in the form of a comment at the top of your gulpfile.js. It is possible (as demonstrated in the default project templates) to have gulp tasks kicked off by the scripts section of your project.json. Task Runner Explorer is a way you can configure tasks to run using Visual Studio. If you are using a different editor (for example, Visual Studio Code) then using the project.json will probably be the most straightforward way to bring together the various stages (prebuild, build, etc.) and the running of gulp tasks.

Note

project.json stages are not triggered when building in Visual Studio by default. If you want to ensure that they are set this option in the Visual Studio project properties: Build tab -> Produce outputs on build. This will add a ProduceOutputsOnBuild element to your .xproj file which will cause Visual studio to trigger the project.json stages when building.

Defining and Running a New Task ¶

To define a new Gulp task, modify gulpfile.js.

Add the following JavaScript to the end of gulpfile.js:

gulp.task("first", function () {
  console.log('first task! <-----');
});

This task is named first, and it simply displays a string.

Save gulpfile.js.
In Solution Explorer, right-click gulpfile.js, and select Task Runner Explorer.
In Task Runner Explorer, right-click first, and select Run.

You’ll see that the output text is displayed. If you are interested in examples based on a common scenario, see Gulp Recipes.

Defining and Running Tasks in a Series ¶

When you run multiple tasks, the tasks run concurrently by default. However, if you need to run tasks in a specific order, you must specify when each task is complete, as well as which tasks depend on the completion of another task.

To define a series of tasks to run in order, replace the first task that you added above in gulpfile.js with the following:

gulp.task("series:first", function () {
  console.log('first task! <-----');
});

gulp.task("series:second", ["series:first"], function () {
  console.log('second task! <-----');
});

gulp.task("series", ["series:first", "series:second"], function () {});

You now have three tasks: series:first, series:second, and series. The series:second task includes a second parameter which specifies an array of tasks to be run and completed before the series:second task will run. As specified in the code above, only the series:first task must be completed before the series:second task will run.

Save gulpfile.js.
In Solution Explorer, right-click gulpfile.js and select Task Runner Explorer if it isn’t already open.
In Task Runner Explorer, right-click series and select Run.

IntelliSense ¶

IntelliSense provides code completion, parameter descriptions, and other features to boost productivity and to decrease errors. Gulp tasks are written in JavaScript; therefore, IntelliSense can provide assistance while developing. As you work with JavaScript, IntelliSense lists the objects, functions, properties, and parameters that are available based on your current context. Select a coding option from the pop-up list provided by IntelliSense to complete the code.

For more information about IntelliSense, see JavaScript IntelliSense.

Development, Staging, and Production Environments ¶

When Gulp is used to optimize client-side files for staging and production, the processed files are saved to a local staging and production location. The _Layout.cshtml file uses the environment tag helper to provide two different versions of CSS files. One version of CSS files is for development and the other version is optimized for both staging and production. In Visual Studio 2015, when you change the Hosting:Environment environment variable to Production, Visual Studio will build the Web app and link to the minimized CSS files. The following markup shows the environment tag helpers containing link tags to the Development CSS files and the minified Staging, Production CSS files.

<environment names="Development">
  <link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />
  <link rel="stylesheet" href="~/css/site.css" />
</environment>
<environment names="Staging,Production">
  <link rel="stylesheet" href="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.5/css/bootstrap.min.css"
      asp-fallback-href="~/lib/bootstrap/dist/css/bootstrap.min.css"
      asp-fallback-test-class="sr-only" asp-fallback-test-property="position" asp-fallback-test-value="absolute" />
  <link rel="stylesheet" href="~/css/site.min.css" asp-append-version="true" />
</environment>

Switching Between Environments ¶

To switch between compiling for different environments, modify the Hosting:Environment environment variable’s value.

In Task Runner Explorer, verify that the min task has been set to run Before Build.
In Solution Explorer, right-click the project name and select Properties.

The property sheet for the Web app is displayed.

Click the Debug tab.
Set the value of the Hosting:Environment environment variable to Production.
Press F5 to run the application in a browser.
In the browser window, right-click the page and select View Source to view the HTML for the page.

Notice that the stylesheet links point to the minified CSS files.

Close the browser to stop the Web app.
In Visual Studio, return to the property sheet for the Web app and change the Hosting:Environment environment variable back to Development.
Press F5 to run the application in a browser again.
In the browser window, right-click the page and select View Source to see the HTML for the page.

Notice that the stylesheet links point to the unminified versions of the CSS files.

For more information related to environments in ASP.NET Core, see Working with Multiple Environments 다중 환경에서 작업하기.

Task and Module Details ¶

A Gulp task is registered with a function name. You can specify dependencies if other tasks must run before the current task. Additional functions allow you to run and watch the Gulp tasks, as well as set the source (src) and destination (dest) of the files being modified. The following are the primary Gulp API functions:

Gulp Function	Syntax	Description
task	`gulp.task(name[, deps], fn) { }`	The `task` function creates a task. The `name` parameter defines the name of the task. The `deps` parameter contains an array of tasks to be completed before this task runs. The `fn` parameter represents a callback function which performs the operations of the task.
watch	`gulp.watch(glob [, opts], tasks) { }`	The `watch` function monitors files and runs tasks when a file change occurs. The `glob` parameter is a `string` or `array` that determines which files to watch. The `opts` parameter provides additional file watching options.
src	`gulp.src(globs[, options]) { }`	The `src` function provides files that match the `glob` value(s). The `glob` parameter is a `string` or `array` that determines which files to read. The `options` parameter provides additional file options.
dest	`gulp.dest(path[, options]) { }`	The `dest` function defines a location to which files can be written. The `path` parameter is a string or function that determines the destination folder. The `options` parameter is an object that specifies output folder options.

For additional Gulp API reference information, see Gulp Docs API.

Gulp Recipes ¶

The Gulp community provides Gulp recipes. These recipes consist of Gulp tasks to address common scenarios.

Summary ¶

Gulp is a JavaScript-based streaming build toolkit that can be used for bundling and minification. Visual Studio automatically installs Gulp along with a set of Gulp plugins. Gulp is maintained on GitHub. For additional information about Gulp, see the Gulp Documentation on GitHub.

Using Grunt¶

Noel Rice

Grunt is a JavaScript task runner that automates script minification, TypeScript compilation, code quality “lint” tools, CSS pre-processors, and just about any repetitive chore that needs doing to support client development. Grunt is fully supported in Visual Studio, though the ASP.NET project templates use Gulp by default (see Using Gulp).

Sections:

Preparing the application
Configuring NPM
Configuring Grunt
Watching for changes
Binding to Visual Studio Events
Summary
See Also

This example uses an empty ASP.NET Core project as its starting point, to show how to automate the client build process from scratch.

The finished example cleans the target deployment directory, combines JavaScript files, checks code quality, condenses JavaScript file content and deploys to the root of your web application. We will use the following packages:

grunt: The Grunt task runner package.
grunt-contrib-clean: A plugin that removes files or directories.
grunt-contrib-jshint: A plugin that reviews JavaScript code quality.
grunt-contrib-concat: A plugin that joins files into a single file.
grunt-contrib-uglify: A plugin that minifies JavaScript to reduce size.
grunt-contrib-watch: A plugin that watches file activity.

Preparing the application ¶

To begin, set up a new empty web application and add TypeScript example files. TypeScript files are automatically compiled into JavaScript using default Visual Studio settings and will be our raw material to process using Grunt.

In Visual Studio, create a new ASP.NET Web Application.
In the New ASP.NET Project dialog, select the ASP.NET Core Empty template and click the OK button.
In the Solution Explorer, review the project structure. The \src folder includes empty wwwroot and Dependencies nodes.

Add a new folder named TypeScript to your project directory.
Before adding any files, let’s make sure that Visual Studio has the option ‘compile on save’ for TypeScript files checked. Tools > Options > Text Editor > Typescript > Project

Right-click the TypeScript directory and select Add > New Item from the context menu. Select the JavaScript file item and name the file Tastes.ts (note the *.ts extension). Copy the line of TypeScript code below into the file (when you save, a new Tastes.js file will appear with the JavaScript source).

enum Tastes { Sweet, Sour, Salty, Bitter }

Add a second file to the TypeScript directory and name it Food.ts. Copy the code below into the file.

class Food {
  constructor(name: string, calories: number) {
    this._name = name;
    this._calories = calories;
  }

  private _name: string;
  get Name() {
    return this._name;
  }

  private _calories: number;
  get Calories() {
    return this._calories;
  }

  private _taste: Tastes;
  get Taste(): Tastes { return this._taste }
  set Taste(value: Tastes) {
    this._taste = value;
  }
}

Configuring NPM ¶

Next, configure NPM to download grunt and grunt-tasks.

In the Solution Explorer, right-click the project and select Add > New Item from the context menu. Select the NPM configuration file item, leave the default name, package.json, and click the Add button.
In the package.json file, inside the devDependencies object braces, enter “grunt”. Select grunt from the Intellisense list and press the Enter key. Visual Studio will quote the grunt package name, and add a colon. To the right of the colon, select the latest stable version of the package from the top of the Intellisense list (press Ctrl-Space if Intellisense does not appear).

Note

NPM uses semantic versioning to organize dependencies. Semantic versioning, also known as SemVer, identifies packages with the numbering scheme <major>.<minor>.<patch>. Intellisense simplifies semantic versioning by showing only a few common choices. The top item in the Intellisense list (0.4.5 in the example above) is considered the latest stable version of the package. The caret (^) symbol matches the most recent major version and the tilde (~) matches the most recent minor version. See the NPM semver version parser reference as a guide to the full expressivity that SemVer provides.

Add more dependencies to load grunt-contrib* packages for clean, jshint, concat, uglify and watch as shown in the example below. The versions do not need to match the example.

"devDependencies": {
    "grunt": "0.4.5",
    "grunt-contrib-clean": "0.6.0",
    "grunt-contrib-jshint": "0.11.0",
    "grunt-contrib-concat": "0.5.1",
    "grunt-contrib-uglify": "0.8.0",
    "grunt-contrib-watch": "0.6.1"
}

Save the package.json file.

The packages for each devDependencies item will download, along with any files that each package requires. You can find the package files in the node_modules directory by enabling the Show All Files button in the Solution Explorer.

Note

If you need to, you can manually restore dependencies in Solution Explorer by right-clicking on Dependencies\NPM and selecting the Restore Packages menu option.

Configuring Grunt ¶

Grunt is configured using a manifest named Gruntfile.js that defines, loads and registers tasks that can be run manually or configured to run automatically based on events in Visual Studio.

Right-click the project and select Add > New Item. Select the Grunt Configuration file option, leave the default name, Gruntfile.js, and click the Add button.

The initial code includes a module definition and the grunt.initConfig() method. The initConfig() is used to set options for each package, and the remainder of the module will load and register tasks.

module.exports = function (grunt) {
  grunt.initConfig({
  });
};

Inside the initConfig() method, add options for the clean task as shown in the example Gruntfile.js below. The clean task accepts an array of directory strings. This task removes files from wwwroot/lib and removes the entire /temp directory.

module.exports = function (grunt) {
  grunt.initConfig({
    clean: ["wwwroot/lib/*", "temp/"],
  });
};

Below the initConfig() method, add a call to grunt.loadNpmTasks(). This will make the task runnable from Visual Studio.

grunt.loadNpmTasks("grunt-contrib-clean");

Save Gruntfile.js. The file should look something like the screenshot below.

Right-click Gruntfile.js and select Task Runner Explorer from the context menu. The Task Runner Explorer window will open.

Verify that clean shows under Tasks in the Task Runner Explorer.

Right-click the clean task and select Run from the context menu. A command window displays progress of the task.

_images/task-runner-explorer-run-clean.png

Note

There are no files or directories to clean yet. If you like, you can manually create them in the Solution Explorer and then run the clean task as a test.

In the initConfig() method, add an entry for concat using the code below.

The src property array lists files to combine, in the order that they should be combined. The dest property assigns the path to the combined file that is produced.

concat: {
  all: {
    src: ['TypeScript/Tastes.js', 'TypeScript/Food.js'],
    dest: 'temp/combined.js'
  }
},

Note

The all property in the code above is the name of a target. Targets are used in some Grunt tasks to allow multiple build environments. You can view the built-in targets using Intellisense or assign your own.

Add the jshint task using the code below.

The jshint code-quality utility is run against every JavaScript file found in the temp directory.

jshint: {
  files: ['temp/*.js'],
  options: {
    '-W069': false,
  }
},

Note

The option “-W069” is an error produced by jshint when JavaScript uses bracket syntax to assign a property instead of dot notation, i.e. Tastes["Sweet"] instead of Tastes.Sweet. The option turns off the warning to allow the rest of the process to continue.

Add the uglify task using the code below.

The task minifies the combined.js file found in the temp directory and creates the result file in wwwroot/lib following the standard naming convention <file name>.min.js.

uglify: {
  all: {
    src: ['temp/combined.js'],
    dest: 'wwwroot/lib/combined.min.js'
  }
},

Under the call grunt.loadNpmTasks() that loads grunt-contrib-clean, include the same call for jshint, concat and uglify using the code below.

grunt.loadNpmTasks('grunt-contrib-jshint');
grunt.loadNpmTasks('grunt-contrib-concat');
grunt.loadNpmTasks('grunt-contrib-uglify');

Save Gruntfile.js. The file should look something like the example below.

Notice that the Task Runner Explorer Tasks list includes clean, concat, jshint and uglify tasks. Run each task in order and observe the results in Solution Explorer. Each task should run without errors.

_images/task-runner-explorer-run-each-task.png

The concat task creates a new combined.js file and places it into the temp directory. The jshint task simply runs and doesn’t produce output. The uglify task creates a new combined.min.js file and places it into wwwroot/lib. On completion, the solution should look something like the screenshot below:

_images/solution-explorer-after-all-tasks.png

Note

For more information on the options for each package, visit https://www.npmjs.com/ and lookup the package name in the search box on the main page. For example, you can look up the grunt-contrib-clean package to get a documentation link that explains all of its parameters.

All Together Now¶

Use the Grunt registerTask() method to run a series of tasks in a particular sequence. For example, to run the example steps above in the order clean -> concat -> jshint -> uglify, add the code below to the module. The code should be added to the same level as the loadNpmTasks() calls, outside initConfig.

grunt.registerTask("all", ['clean', 'concat', 'jshint', 'uglify']);

The new task shows up in Task Runner Explorer under Alias Tasks. You can right-click and run it just as you would other tasks. The all task will run clean, concat, jshint and uglify, in order.

Watching for changes ¶

A watch task keeps an eye on files and directories. The watch triggers tasks automatically if it detects changes. Add the code below to initConfig to watch for changes to *.js files in the TypeScript directory. If a JavaScript file is changed, watch will run the all task.

watch: {
  files: ["TypeScript/*.js"],
  tasks: ["all"]
}

Add a call to loadNpmTasks() to show the watch task in Task Runner Explorer.

grunt.loadNpmTasks('grunt-contrib-watch');

Right-click the watch task in Task Runner Explorer and select Run from the context menu. The command window that shows the watch task running will display a “Waiting…” message. Open one of the TypeScript files, add a space, and then save the file. This will trigger the watch task and trigger the other tasks to run in order. The screenshot below shows a sample run.

Binding to Visual Studio Events ¶

Unless you want to manually start your tasks every time you work in Visual Studio, you can bind tasks to Before Build, After Build, Clean, and Project Open events.

Let’s bind watch so that it runs every time Visual Studio opens. In Task Runner Explorer, right-click the watch task and select Bindings > Project Open from the context menu.

Unload and reload the project. When the project loads again, the watch task will start running automatically.

Summary ¶

Grunt is a powerful task runner that can be used to automate most client-build tasks. Grunt leverages NPM to deliver its packages, and features tooling integration with Visual Studio. Visual Studio’s Task Runner Explorer detects changes to configuration files and provides a convenient interface to run tasks, view running tasks, and bind tasks to Visual Studio events.

Manage Client-Side Packages with Bower¶

By Noel Rice, Scott Addie

Bower is a “package manager for the web.” Bower lets you install and restore client-side packages, including JavaScript and CSS libraries. For example, with Bower you can install CSS files, fonts, client frameworks, and JavaScript libraries from external sources. Bower resolves dependencies and will automatically download and install all the packages you need. For example, if you configure Bower to load the Bootstrap package, the necessary jQuery package will automatically come along for the ride. For .NET libraries you still use NuGet package manager.

Note

Visual Studio developers are already familiar with NuGet, so why not use NuGet instead of Bower? Mainly because Bower already has a rich ecosystem with over 34,000 packages in play; and, it integrates well with the Gulp and Grunt task runners.

Getting Started with Bower¶

ASP.NET Core project templates pre-construct the client build process for you. The ubiquitous jQuery and Bootstrap packages are installed, and the plumbing for NPM, Gulp, and Bower is already in place. The screenshot below depicts the initial project in Solution Explorer. It’s important to enable the “Show All Files” option, as the bower.json file is hidden by default.

Client-side packages are listed in the bower.json file. The ASP.NET Core project template pre-configures bower.json with jQuery, jQuery validation, and Bootstrap.

Let’s add support for photo albums by installing the Fotorama jQuery plugin. Bower packages can be installed either via the Manage Bower Packages UI or manually in the bower.json file.

Installation via Manage Bower Packages UI¶

Right-click the project name in Solution Explorer, and select the “Manage Bower Packages” menu option.
In the window that appears, click the “Browse” tab, and filter the packages list by typing “fotorama” into the search box:

Confirm that the “Save changes to bower.json” checkbox is checked, select the desired version from the drop-down list, and click the Install button.
Across the bottom status bar of the IDE, an Installing “fotorama” complete message appears to indicate a successful installation.

Manual Installation in bower.json¶

At the end of the dependencies section in bower.json, add a comma and type “fotorama”. Notice as you type that you get IntelliSense with a list of available packages. Select “fotorama” from the list.

Add a colon and then select the latest stable version of the package from the drop-down list. The double quotes will be added automatically.

Save the bower.json file.

Note

Visual Studio watches the bower.json file for changes. Upon saving, the bower install command is executed. See the Output window’s “Bower/npm” view for the exact command which was executed.

Now that the installation step has been completed, expand the twisty to the left of bower.json, and locate the .bowerrc file. Open it, and notice that the directory property is set to “wwwroot/lib”. This setting indicates the location at which Bower will install the package assets.

{
  "directory": "wwwroot/lib"
}

In Solution Explorer, expand the wwwroot node. The lib directory should now contain all of the packages, including the fotorama package.

Next, let’s add an HTML page to the project. In Solution Explorer, right-click wwwroot node and select Add > New Item > HTML Page. Name the page Index.html. Replace the contents of the file with the following:

<!DOCTYPE html>
<html>
<head>
  <meta charset="utf-8" />
  <title>Bower and Fotorama</title>
  <link href="lib/fotorama/fotorama.css" rel="stylesheet" />
</head>
<body>
  <div class="fotorama" data-nav="thumbs">
    <img src="images/asp-net-banners-01.png" />
    <img src="images/asp-net-banners-02.png" />
    <img src="images/banner-01-azure.png" />
    <img src="images/banner-02-vs.png" />
  </div>
  <script src="lib/jquery/dist/jquery.js"></script>
  <script src="lib/fotorama/fotorama.js"></script>
</body>
</html>

This example uses images currently available inside wwwroot/images, but you can add any images on hand.

Press Ctrl+Shift+W to display the page in the browser. The control displays the images and allows navigation by clicking the thumbnail list below the main image. This quick test shows that Bower installed the correct packages and dependencies.

Exploring the Client Build Process¶

Most ASP.NET Core project templates are already configured to use Bower. This next walkthrough starts with an empty ASP.NET Core project and adds each piece manually, so you can get a feel for how Bower is used in a project. See what happens to the project structure and the runtime output as each configuration change is made to the project.

The general steps to use the client-side build process with Bower are:

Define and download packages used in your project.
Reference packages from your web pages.

Define Packages¶

The first step is to define the packages your application needs and to download them. This example uses Bower to load jQuery and Bootstrap in the desired location.

In Visual Studio, create a new ASP.NET Web Application.
In the New ASP.NET Project dialog, select the ASP.NET Core Empty project template and click OK.
In Solution Explorer, the src directory includes a project.json file, and wwwroot and Dependencies nodes. The project directory will look like the screenshot below.

In Solution Explorer, right-click the project, and add the following item:

Bower Configuration File – bower.json

Note

The Bower Configuration File item template also adds a .bowerrc file.

Open bower.json, and add jquery and bootstrap to the dependencies section. As an alternative to the manual file editing, the “Manage Bower Packages” UI may be used. The resulting bower.json file should look like the example here. The versions will change over time, so use the latest stable build version from the drop-down list.

{
  "name": "ASP.NET",
  "private": true,
  "dependencies": {
    "jquery": "2.1.4",
    "bootstrap": "3.3.5"
  }
}

Save the bower.json file.

The project should now include bootstrap and jQuery directories in two locations: Dependencies/Bower and wwwroot/lib. It’s the .bowerrc file which instructed Bower to install the assets within wwwroot/lib.

Reference Packages¶

Now that Bower has copied the client support packages needed by the application, you can test that an HTML page can use the deployed jQuery and Bootstrap functionality.

Right-click wwwroot and select Add > New Item > HTML Page. Name the page Index.html.
Add the CSS and JavaScript references.

In Solution Explorer, expand wwwroot/lib/bootstrap and locate bootstrap.css. Drag this file into the head element of the HTML page.

Drag jquery.js and bootstrap.js to the end of the body element.

Make sure bootstrap.js follows jquery.js, so that jQuery is loaded first.

<!DOCTYPE html>
<html>
<head>
  <meta charset="utf-8" />
  <title>Bower Example</title>
  <link href="lib/bootstrap/dist/css/bootstrap.css" rel="stylesheet" />
</head>
<body>

  <script src="lib/jquery/dist/jquery.js"></script>
  <script src="lib/bootstrap/dist/js/bootstrap.js"></script>
</body>
</html>

Use the Installed Packages¶

Add jQuery and Bootstrap components to the page to verify that the web application is configured correctly.

Inside the body tag, above the script references, add a div element with the Bootstrap jumbotron class and an anchor tag.

<div class="jumbotron">
  <h1>Using the jumbotron style</h1>
  <p><a class="btn btn-primary btn-lg" role="button">
     Stateful button</a></p>
</div>

Add the following code after the jQuery and Bootstrap script references.

<script>
  $(".btn").click(function() {
    $(this).text('loading')
      .delay(1000)
      .queue(function () {
        $(this).text('reset');
        $(this).dequeue();
      });
  });
</script>

Within the Configure method of the Startup.cs file, add a call to the UseStaticFiles extension method. This middleware adds files, found within the web root, to the request pipeline. This line of code will look as follows:

app.UseStaticFiles();
Note

Be sure to install the Microsoft.AspNetCore.StaticFiles NuGet package. Without it, the UseStaticFiles extension method will not resolve.

With the Index.html file opened, press Ctrl+Shift+W to view the page in the browser. Verify that the jumbotron styling is applied, the jQuery code responds when the button is clicked, and that the Bootstrap button changes state.

Building Beautiful, Responsive Sites with Bootstrap¶

http://getbootstrap.com/examples/theme/

Bootstrap is currently the most popular web framework for developing responsive web applications. It offers a number of features and benefits that can improve your users’ experience with your web site, whether you’re a novice at front-end design and development or an expert. Bootstrap is deployed as a set of CSS and JavaScript files, and is designed to help your website or application scale efficiently from phones to tablets to desktops.

Sections:

Getting Started
Basic Templates and Features
More Themes
Components
JavaScript Support
Summary

Getting Started ¶

There are several ways to get started with Bootstrap. If you’re starting a new web application in Visual Studio, you can choose the default starter template for ASP.NET Core, in which case Bootstrap will come pre-installed:

_images/bootstrap-in-starter-template.png

Adding Bootstrap to an ASP.NET Core project is simply a matter of adding it to bower.json as a dependency:

{
  "name": "asp.net",
  "private": true,
  "dependencies": {
    "bootstrap": "3.3.6",
    "jquery": "2.2.0",
    "jquery-validation": "1.14.0",
    "jquery-validation-unobtrusive": "3.2.6"
  }
}

This is the recommended way to add Bootstrap to an ASP.NET Core project.

You can also install bootstrap using one of several package managers, such as bower, npm, or NuGet. In each case, the process is essentially the same:

Bower¶

bower install bootstrap

npm¶

npm install bootstrap

NuGet¶

Install-Package bootstrap

Note

The recommended way to install client-side dependencies like Bootstrap in ASP.NET Core is via Bower (using bower.json, as shown above). The use of npm/NuGet are shown to demonstrate how easily Bootstrap can be added to other kinds of web applications, including earlier versions of ASP.NET.

If you’re referencing your own local versions of Bootstrap, you’ll need to reference them in any pages that will use it. In production you should reference bootstrap using a CDN. In the default ASP.NET site template, the _Layout.cshtml file does so like this:

<!DOCTYPE html>
<html>
<head>
    <meta charset="utf-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0" />
    <title>@ViewData["Title"] - WebApplication1</title>

    <environment names="Development">
        <link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />
        <link rel="stylesheet" href="~/css/site.css" />
    </environment>
    <environment names="Staging,Production">
        <link rel="stylesheet" href="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.6/css/bootstrap.min.css"
              asp-fallback-href="~/lib/bootstrap/dist/css/bootstrap.min.css"
              asp-fallback-test-class="sr-only" asp-fallback-test-property="position" asp-fallback-test-value="absolute" />
        <link rel="stylesheet" href="~/css/site.min.css" asp-append-version="true" />
    </environment>
</head>
<body>
    <div class="navbar navbar-inverse navbar-fixed-top">
        <div class="container">
            <div class="navbar-header">
                <button type="button" class="navbar-toggle" data-toggle="collapse" data-target=".navbar-collapse">
                    <span class="sr-only">Toggle navigation</span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                </button>
                <a asp-area="" asp-controller="Home" asp-action="Index" class="navbar-brand">WebApplication1</a>
            </div>
            <div class="navbar-collapse collapse">
                <ul class="nav navbar-nav">
                    <li><a asp-area="" asp-controller="Home" asp-action="Index">Home</a></li>
                    <li><a asp-area="" asp-controller="Home" asp-action="About">About</a></li>
                    <li><a asp-area="" asp-controller="Home" asp-action="Contact">Contact</a></li>
                </ul>
                @await Html.PartialAsync("_LoginPartial")
            </div>
        </div>
    </div>
    <div class="container body-content">
        @RenderBody()
        <hr />
        <footer>
            <p>&copy; 2016 - WebApplication1</p>
        </footer>
    </div>

    <environment names="Development">
        <script src="~/lib/jquery/dist/jquery.js"></script>
        <script src="~/lib/bootstrap/dist/js/bootstrap.js"></script>
        <script src="~/js/site.js" asp-append-version="true"></script>
    </environment>
    <environment names="Staging,Production">
        <script src="https://ajax.aspnetcdn.com/ajax/jquery/jquery-2.2.0.min.js"
                asp-fallback-src="~/lib/jquery/dist/jquery.min.js"
                asp-fallback-test="window.jQuery">
        </script>
        <script src="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.6/bootstrap.min.js"
                asp-fallback-src="~/lib/bootstrap/dist/js/bootstrap.min.js"
                asp-fallback-test="window.jQuery && window.jQuery.fn && window.jQuery.fn.modal">
        </script>
        <script src="~/js/site.min.js" asp-append-version="true"></script>
    </environment>

    @RenderSection("scripts", required: false)
</body>
</html>

Note

If you’re going to be using any of Bootstrap’s jQuery plugins, you will also need to reference jQuery.

Basic Templates and Features ¶

The most basic Bootstrap template looks very much like the _Layout.cshtml file shown above, and simply includes a basic menu for navigation and a place to render the rest of the page.

Basic Navigation¶

The default template uses a set of <div> elements to render a top navbar and the main body of the page. If you’re using HTML5, you can replace the first <div> tag with a <nav> tag to get the same effect, but with more precise semantics. Within this first <div> you can see there are several others. First, a <div> with a class of “container”, and then within that, two more <div> elements: “navbar-header” and “navbar-collapse”. The navbar-header div includes a button that will appear when the screen is below a certain minimum width, showing 3 horizontal lines (a so-called “hamburger icon”). The icon is rendered using pure HTML and CSS; no image is required. This is the code that displays the icon, with each of the <span> tags rendering one of the white bars:

<button type="button" class="navbar-toggle" data-toggle="collapse" data-target=".navbar-collapse">
  <span class="icon-bar"></span>
  <span class="icon-bar"></span>
  <span class="icon-bar"></span>
</button>

It also includes the application name, which appears in the top left. The main navigation menu is rendered by the <ul> element within the second div, and includes links to Home, About, and Contact. Additional links for Register and Login are added by the _LoginPartial line on line 29. Below the navigation, the main body of each page is rendered in another <div>, marked with the “container” and “body-content” classes. In the simple default _Layout file shown here, the contents of the page are rendered by the specific View associated with the page, and then a simple <footer> is added to the end of the <div> element. You can see how the built-in About page appears using this template:

The collapsed navbar, with “hamburger” button in the top right, appears when the window drops below a certain width:

Clicking the icon reveals the menu items in a vertical drawer that slides down from the top of the page:

Typography and Links¶

Bootstrap sets up the site’s basic typography, colors, and link formatting in its CSS file. This CSS file includes default styles for tables, buttons, form elements, images, and more (learn more). One particularly useful feature is the grid layout system, covered next.

Grids¶

One of the most popular features of Bootstrap is its grid layout system. Modern web applications should avoid using the <table> tag for layout, instead restricting the use of this element to actual tabular data. Instead, columns and rows can be laid out using a series of <div> elements and the appropriate CSS classes. There are several advantages to this approach, including the ability to adjust the layout of grids to display vertically on narrow screens, such as on phones.

Bootstrap’s grid layout system is based on twelve columns. This number was chosen because it can be divided evenly into 1, 2, 3, or 4 columns, and column widths can vary to within 1/12th of the vertical width of the screen. To start using the grid layout system, you should begin with a container <div> and then add a row <div>, as shown here:

<div class="container">
  <div class="row">

  </div>
</div>

Next, add additional <div> elements for each column, and specify the number of columns that <div> should occupy (out of 12) as part of a CSS class starting with “col-md-”. For instance, if you want to simply have two columns of equal size, you would use a class of “col-md-6” for each one. In this case “md” is short for “medium” and refers to standard-sized desktop computer display sizes. There are four different options you can choose from, and each will be used for higher widths unless overridden (so if you want the layout to be fixed regardless of screen width, you can just specify xs classes).

CSS Class Prefix	Device Tier	Width
col-xs-	Phones	< 768px
col-sm-	Tablets	>= 768px
col-md-	Desktops	>= 992px
col-lg-	Larger Desktop Displays	>= 1200px

When specifying two columns both with “col-md-6” the resulting layout will be two columns at desktop resolutions, but these two columns will stack vertically when rendered on smaller devices (or a narrower browser window on a desktop), allowing users to easily view content without the need to scroll horizontally.

Bootstrap will always default to a single-column layout, so you only need to specify columns when you want more than one column. The only time you would want to explicitly specify that a <div> take up all 12 columns would be to override the behavior of a larger device tier. When specifying multiple device tier classes, you may need to reset the column rendering at certain points. Adding a clearfix div that is only visible within a certain viewport can achieve this, as shown here:

_images/narrow-and-wide-viewport-grid.png

In the above example, One and Two share a row in the “md” layout, while Two and Three share a row in the “xs” layout. Without the clearfix <div>, Two and Three are not shown correctly in the “xs” view (note that only One, Four, and Five are shown):

In this example, only a single row <div> was used, and Bootstrap still mostly did the right thing with regard to the layout and stacking of the columns. Typically, you should specify a row <div> for each horizontal row your layout requires, and of course you can nest Bootstrap grids within one another. When you do, each nested grid will occupy 100% of the width of the element in which it is placed, which can then be subdivided using column classes.

Jumbotron¶

If you’ve used the default ASP.NET MVC templates in Visual Studio 2012 or 2013, you’ve probably seen the Jumbotron in action. It refers to a large full-width section of a page that can be used to display a large background image, a call to action, a rotator, or similar elements. To add a jumbotron to a page, simply add a <div> and give it a class of “jumbotron”, then place a container <div> inside and add your content. We can easily adjust the standard About page to use a jumbotron for the main headings it displays:

Buttons¶

The default button classes and their colors are shown in the figure below.

Badges¶

Badges refer to small, usually numeric callouts next to a navigation item. They can indicate a number of messages or notifications waiting, or the presence of updates. Specifying such badges is as simple as adding a <span> containing the text, with a class of “badge”:

Alerts¶

You may need to display some kind of notification, alert, or error message to your application’s users. That’s where the standard alert classes come in. There are four different severity levels, with associated color schemes:

Navbars and Menus¶

Our layout already includes a standard navbar, but the Bootstrap theme supports additional styling options. We can also easily opt to display the navbar vertically rather than horizontally if that’s preferred, as well as adding sub-navigation items in flyout menus. Simple navigation menus, like tab strips, are built on top of <ul> elements. These can be created very simply by just providing them with the CSS classes “nav” and “nav-tabs”:

Navbars are built similarly, but are a bit more complex. They start with a <nav> or <div> with a class of “navbar”, within which a container div holds the rest of the elements. Our page includes a navbar in its header already – the one shown below simply expands on this, adding support for a dropdown menu:

Additional Elements¶

The default theme can also be used to present HTML tables in a nicely formatted style, including support for striped views. There are labels with styles that are similar to those of the buttons. You can create custom Dropdown menus that support additional styling options beyond the standard HTML <select> element, along with Navbars like the one our default starter site is already using. If you need a progress bar, there are several styles to choose from, as well as List Groups and panels that include a title and content. Explore additional options within the standard Bootstrap Theme here:

More Themes ¶

You can extend the standard Bootstrap Theme by overriding some or all of its CSS, adjusting the colors and styles to suit your own application’s needs. If you’d like to start from a ready-made theme, there are several theme galleries available online that specialize in Bootstrap Themes, such as WrapBootstrap.com (which has a variety of commercial themes) and Bootswatch.com (which offers free themes). Some of the paid templates available provide a great deal of functionality on top of the basic Bootstrap theme, such as rich support for administrative menus, and dashboards with rich charts and gauges. An example of a popular paid template is Inspinia, currently for sale for $18, which includes an ASP.NET MVC5 template in addition to AngularJS and static HTML versions. A sample screenshot is shown below.

If you’re interested in building your own dashboard, you may wish to start from the free example available here: http://getbootstrap.com/examples/dashboard/.

Components ¶

In addition to those elements already discussed, Bootstrap includes support for a variety of built-in UI components.

Glyphicons¶

Bootstrap includes icon sets from Glyphicons (http://glyphicons.com), with over 200 icons freely available for use within your Bootstrap-enabled web application. Here’s just a small sample:

Input Groups¶

Input groups allow bundling of additional text or buttons with an input element, providing the user with a more intuitive experience:

Breadcrumbs¶

Breadcrumbs are a common UI component used to show a user their recent history or depth within a site’s navigation hierarchy. Add them easily by applying the “breadcrumb” class to any <ol> list element. Include built-in support for pagination by using the “pagination” class on a <ul> element within a <nav>. Add responsive embedded slideshows and video by using <iframe>, <embed>, <video>, or <object> elements, which Bootstrap will style automatically. Specify a particular aspect ratio by using specific classes like “embed-responsive-16by9”.

JavaScript Support ¶

Bootstrap’s JavaScript library includes API support for the included components, allowing you to control their behavior programmatically within your application. In addition, bootstrap.js includes over a dozen custom jQuery plugins, providing additional features like transitions, modal dialogs, scroll detection (updating styles based on where the user has scrolled in the document), collapse behavior, carousels, and affixing menus to the window so they do not scroll off the screen. There’s not sufficient room to cover all of the JavaScript add-ons built into Bootstrap – to learn more please visit http://getbootstrap.com/javascript/.

Summary ¶

Bootstrap provides a web framework that can be used to quickly and productively lay out and style a wide variety of websites and applications. Its basic typography and styles provide a pleasant look and feel that can easily be manipulated through custom theme support, which can be hand-crafted or purchased commercially. It supports a host of web components that in the past would have required expensive third-party controls to accomplish, while supporting modern and open web standards.

Knockout.js MVVM Framework¶

http://ajax.aspnetcdn.com/ajax/knockout/knockout-3.3.0.js

Knockout is a popular JavaScript library that simplifies the creation of complex data-based user interfaces. It can be used alone or with other libraries, such as jQuery. Its primary purpose is to bind UI elements to an underlying data model defined as a JavaScript object, such that when changes are made to the UI, the model is updated, and vice versa. Knockout facilitates the use of a Model-View-ViewModel (MVVM) pattern in a web application’s client-side behavior. The two main concepts one must learn when working with Knockout’s MVVM implementation are Observables and Bindings.

Sections:

Getting Started with Knockout in ASP.NET Core
Observables, ViewModels, and Simple Binding
Control Flow
Templates
Components
Communicating with APIs
Summary

Getting Started with Knockout in ASP.NET Core ¶

Knockout is deployed as a single JavaScript file, so installing and using it is very straightforward using bower. Assuming you already have bower and gulp configured, open bower.json in your ASP.NET Core project and add the knockout dependency as shown here:

{
  "name": "KnockoutDemo",
  "private": true,
  "dependencies": {
    "knockout" : "^3.3.0"
  },
  "exportsOverride": {
  }
}

With this in place, you can then manually run bower by opening the Task Runner Explorer (under View ‣ Other Windows ‣ Task Runner Explorer) and then under Tasks, right-click on bower and select Run. The result should appear similar to this:

Now if you look in your project’s wwwroot folder, you should see knockout installed under the lib folder.

It’s recommended that in your production environment you reference knockout via a Content Delivery Network, or CDN, as this increases the likelihood that your users will already have a cached copy of the file and thus will not need to download it at all. Knockout is available on several CDNs, including the Microsoft Ajax CDN, here:

To include Knockout on a page that will use it, simply add a <script> element referencing the file from wherever you will be hosting it (with your application, or via a CDN):

<script type="text/javascript" src="knockout-3.3.0.js"></script>

Observables, ViewModels, and Simple Binding ¶

You may already be familiar with using JavaScript to manipulate elements on a web page, either via direct access to the DOM or using a library like jQuery. Typically this kind of behavior is achieved by writing code to directly set element values in response to certain user actions. With Knockout, a declarative approach is taken instead, through which elements on the page are bound to properties on an object. Instead of writing code to manipulate DOM elements, user actions simply interact with the ViewModel object, and Knockout takes care of ensuring the page elements are synchronized.

As a simple example, consider the page list below. It includes a <span> element with a data-bind attribute indicating that the text content should be bound to authorName. Next, in a JavaScript block a variable viewModel is defined with a single property, authorName, set to some value. Finally, a call to ko.applyBindings is made, passing in this viewModel variable.

<html>
  <head>
    <script type="text/javascript" src="lib/knockout/knockout.js"></script>
  </head>
  <body>
    <h1>Some Article</h1>
    <p>
      By <span data-bind="text: authorName"></span>
    </p>
    <script type="text/javascript">
      var viewModel = {
        authorName: 'Steve Smith'
      };
      ko.applyBindings(viewModel);
    </script>
  </body>
</html>

When viewed in the browser, the content of the <span> element is replaced with the value in the viewModel variable:

We now have simple one-way binding working. Notice that nowhere in the code did we write JavaScript to assign a value to the span’s contents. If we want to manipulate the ViewModel, we can take this a step further and add an HTML input textbox, and bind to its value, like so:

<p>
  Author Name: <input type="text" data-bind="value: authorName" />
</p>

Reloading the page, we see that this value is indeed bound to the input box:

However, if we change the value in the textbox, the corresponding value in the <span> element doesn’t change. Why not?

The issue is that nothing notified the <span> that it needed to be updated. Simply updating the ViewModel isn’t by itself sufficient, unless the ViewModel’s properties are wrapped in a special type. We need to use observables in the ViewModel for any properties that need to have changes automatically updated as they occur. By changing the ViewModel to use ko.observable("value") instead of just “value”, the ViewModel will update any HTML elements that are bound to its value whenever a change occurs. Note that input boxes don’t update their value until they lose focus, so you won’t see changes to bound elements as you type.

Note

Adding support for live updating after each keypress is simply a matter of adding valueUpdate: "afterkeydown" to the data-bind attribute’s contents.

Our viewModel, after updating it to use ko.observable:

var viewModel = {
  authorName: ko.observable('Steve Smith')
};
ko.applyBindings(viewModel);

Knockout supports a number of different kinds of bindings. So far we’ve seen how to bind to text and to value. You can also bind to any given attribute. For instance, to create a hyperlink with an anchor tag, the src attribute can be bound to the viewModel. Knockout also supports binding to functions. To demonstrate this, let’s update the viewModel to include the author’s twitter handle, and display the twitter handle as a link to the author’s twitter page. We’ll do this in three stages.

First, add the HTML to display the hyperlink, which we’ll show in parentheses after the author’s name:

<h1>Some Article</h1>
<p>
  By <span data-bind="text: authorName"></span>
  (<a data-bind="attr: { href: twitterUrl}, text: twitterAlias" ></a>)
</p>

Next, update the viewModel to include the twitterUrl and twitterAlias properties:

var viewModel = {
  authorName: ko.observable('Steve Smith'),
  twitterAlias: ko.observable('@ardalis'),
  twitterUrl: ko.computed(function() {
    return "https://twitter.com/";
  }, this)
};
ko.applyBindings(viewModel);

Notice that at this point we haven’t yet updated the twitterUrl to go to the correct URL for this twitter alias – it’s just pointing at twitter.com. Also notice that we’re using a new Knockout function, computed, for twitterUrl. This is an observable function that will notify any UI elements if it changes. However, for it to have access to other properties in the viewModel, we need to change how we are creating the viewModel, so that each property is its own statement.

The revised viewModel declaration is shown below. It is now declared as a function. Notice that each property is its own statement now, ending with a semicolon. Also notice that to access the twitterAlias property value, we need to execute it, so its reference includes ().

function viewModel() {
  this.authorName = ko.observable('Steve Smith');
  this.twitterAlias = ko.observable('@ardalis');

  this.twitterUrl = ko.computed(function() {
    return "https://twitter.com/" + this.twitterAlias().replace('@','');
  }, this)
};
ko.applyBindings(viewModel);

The result works as expected in the browser:

Knockout also supports binding to certain UI element events, such as the click event. This allows you to easily and declaratively bind UI elements to functions within the application’s viewModel. As a simple example, we can add a button that, when clicked, modifies the author’s twitterAlias to be all caps.

First, we add the button, binding to the button’s click event, and referencing the function name we’re going to add to the viewModel:

<p>
  <button data-bind="click: capitalizeTwitterAlias">Capitalize</button>
</p>

Then, add the function to the viewModel, and wire it up to modify the viewModel’s state. Notice that to set a new value to the twitterAlias property, we call it as a method and pass in the new value.

function viewModel() {
  this.authorName = ko.observable('Steve Smith');
  this.twitterAlias = ko.observable('@ardalis');

  this.twitterUrl = ko.computed(function() {
    return "https://twitter.com/" + this.twitterAlias().replace('@','');
  }, this);

  this.capitalizeTwitterAlias = function() {
    var currentValue = this.twitterAlias();
    this.twitterAlias(currentValue.toUpperCase());
  }
};
ko.applyBindings(viewModel);

Running the code and clicking the button modifies the displayed link as expected:

Control Flow ¶

Knockout includes bindings that can perform conditional and looping operations. Looping operations are especially useful for binding lists of data to UI lists, menus, and grids or tables. The foreach binding will iterate over an array. When used with an observable array, it will automatically update the UI elements when items are added or removed from the array, without re-creating every element in the UI tree. The following example uses a new viewModel which includes an observable array of game results. It is bound to a simple table with two columns using a foreach binding on the <tbody> element. Each <tr> element within <tbody> will be bound to an element of the gameResults collection.

<h1>Record</h1>
<table>
  <thead>
    <tr>
      <th>Opponent</th>
      <th>Result</th>
    </tr>
  </thead>
  <tbody data-bind="foreach: gameResults">
    <tr>
      <td data-bind="text:opponent"></td>
      <td data-bind="text:result"></td>
    </tr>
  </tbody>
</table>
<script type="text/javascript">
  function GameResult(opponent, result) {
    var self = this;
    self.opponent = opponent;
    self.result = ko.observable(result);
  }

  function ViewModel() {
    var self = this;

    self.resultChoices = ["Win", "Loss", "Tie"];

    self.gameResults = ko.observableArray([
      new GameResult("Brendan", self.resultChoices[0]),
      new GameResult("Brendan", self.resultChoices[0]),
      new GameResult("Michelle", self.resultChoices[1])
    ]);
  };
  ko.applyBindings(new ViewModel);
</script>

Notice that this time we’re using ViewModel with a capital “V” because we expect to construct it using “new” (in the applyBindings call). When executed, the page results in the following output:

To demonstrate that the observable collection is working, let’s add a bit more functionality. We can include the ability to record the results of another game to the ViewModel, and then add a button and some UI to work with this new function. First, let’s create the addResult method:

// add this to ViewModel()
self.addResult = function() {
  self.gameResults.push(new GameResult("", self.resultChoices[0]));
}

Bind this method to a button using the click binding:

<button data-bind="click: addResult">Add New Result</button>

Open the page in the browser and click the button a couple of times, resulting in a new table row with each click:

There are a few ways to support adding new records in the UI, typically either inline or in a separate form. We can easily modify the table to use textboxes and dropdownlists so that the whole thing is editable. Just change the <tr> element as shown:

<tbody data-bind="foreach: gameResults">
  <tr>
    <td><input data-bind="value:opponent" /></td>
    <td><select data-bind="options: $root.resultChoices,
      value:result, optionsText: $data"></select></td>
  </tr>
</tbody>

Note that $root refers to the root ViewModel, which is where the possible choices are exposed. $data refers to whatever the current model is within a given context - in this case it refers to an individual element of the resultChoices array, each of which is a simple string.

With this change, the entire grid becomes editable:

If we weren’t using Knockout, we could achieve all of this using jQuery, but most likely it would not be nearly as efficient. Knockout tracks which bound data items in the ViewModel correspond to which UI elements, and only updates those elements that need to be added, removed, or updated. It would take significant effort to achieve this ourselves using jQuery or direct DOM manipulation, and even then if we then wanted to display aggregate results (such as a win-loss record) based on the table’s data, we would need to once more loop through it and parse the HTML elements. With Knockout, displaying the win-loss record is trivial. We can perform the calculations within the ViewModel itself, and then display it with a simple text binding and a <span>.

To build the win-loss record string, we can use a computed observable. Note that references to observable properties within the ViewModel must be function calls, otherwise they will not retrieve the value of the observable (i.e. gameResults() not gameResults in the code shown):

self.displayRecord = ko.computed(function () {
  var wins = self.gameResults().filter(function (value) { return value.result() == "Win"; }).length;
  var losses = self.gameResults().filter(function (value) { return value.result() == "Loss"; }).length;
  var ties = self.gameResults().filter(function (value) { return value.result() == "Tie"; }).length;
  return wins + " - " + losses + " - " + ties;
}, this);

Bind this function to a span within the <h1> element at the top of the page:

<h1>Record <span data-bind="text: displayRecord"></span></h1>

The result:

Adding rows or modifying the selected element in any row’s Result column will update the record shown at the top of the window.

In addition to binding to values, you can also use almost any legal JavaScript expression within a binding. For example, if a UI element should only appear under certain conditions, such as when a value exceeds a certain threshold, you can specify this logically within the binding expression:

<div data-bind="visible: customerValue > 100"></div>

This <div> will only be visible when the customerValue is over 100.

Templates ¶

Knockout has support for templates, so that you can easily separate your UI from your behavior, or incrementally load UI elements into a large application on demand. We can update our previous example to make each row its own template by simply pulling the HTML out into a template and specifying the template by name in the data-bind call on <tbody>.

<tbody data-bind="template: { name: 'rowTemplate', foreach: gameResults }">
</tbody>
<script type="text/html" id="rowTemplate">
  <tr>
    <td><input data-bind="value:opponent" /></td>
    <td><select data-bind="options: $root.resultChoices,
      value:result, optionsText: $data"></select></td>
  </tr>
</script>

Knockout also supports other templating engines, such as the jQuery.tmpl library and Underscore.js’s templating engine.

Components ¶

Components allow you to organize and reuse UI code, usually along with the ViewModel data on which the UI code depends. To create a component, you simply need to specify its template and its viewModel, and give it a name. This is done by calling ko.components.register(). In addition to defining the templates and viewmodel inline, they can be loaded from external files using a library like require.js, resulting in very clean and efficient code.

Communicating with APIs ¶

Knockout can work with any data in JSON format. A common way to retrieve and save data using Knockout is with jQuery, which supports the $.getJSON() function to retrieve data, and the $.post() method to send data from the browser to an API endpoint. Of course, if you prefer a different way to send and receive JSON data, Knockout will work with it as well.

Summary ¶

Knockout provides a simple, elegant way to bind UI elements to the current state of the client application, defined in a ViewModel. Knockout’s binding syntax uses the data-bind attribute, applied to HTML elements that are to be processed. Knockout is able to efficiently render and update large data sets by tracking UI elements and only processing changes to affected elements. Large applications can break up UI logic using templates and components, which can be loaded on demand from external files. Currently version 3, Knockout is a stable JavaScript library that can improve web applications that require rich client interactivity.

Using Angular for Single Page Applications (SPAs)¶

By Venkata Koppaka and Scott Addie

In this article, you will learn how to build a SPA-style ASP.NET application using AngularJS.

Sections:

What is AngularJS?
Getting Started
Key Components
Angular 2.0
Summary
Related Resources

What is AngularJS?¶

AngularJS is a modern JavaScript framework from Google commonly used to work with Single Page Applications (SPAs). AngularJS is open sourced under MIT license, and the development progress of AngularJS can be followed on its GitHub repository. The library is called Angular because HTML uses angular-shaped brackets.

AngularJS is not a DOM manipulation library like jQuery, but it uses a subset of jQuery called jQLite. AngularJS is primarily based on declarative HTML attributes that you can add to your HTML tags. You can try AngularJS in your browser using the Code School website.

Version 1.5.x is the current stable version and the Angular team is working towards a big rewrite of AngularJS for V2.0 which is currently still in development. This article focuses on Angular 1.X with some notes on where Angular is heading with 2.0.

Getting Started ¶

To start using AngularJS in your ASP.NET application, you must either install it as part of your project, or reference it from a content delivery network (CDN).

Installation¶

There are several ways to add AngularJS to your application. If you’re starting a new ASP.NET Core web application in Visual Studio, you can add AngularJS using the built-in Bower support. Simply open bower.json, and add an entry to the dependencies property:

{
  "name": "ASP.NET",
  "private": true,
  "dependencies": {
    "bootstrap": "3.3.5",
    "jquery": "2.1.4",
    "jquery-validation": "1.14.0",
    "jquery-validation-unobtrusive": "3.2.4",
    "angular": "1.4.8",
    "angular-route": "1.4.8"
  }
}

Upon saving the bower.json file, Angular will be installed in your project’s wwwroot/lib folder. Additionally, it will be listed within the Dependencies/Bower folder. See the screenshot below.

Next, add a <script> reference to the bottom of the <body> section of your HTML page or _Layout.cshtml file, as shown here:

        <environment names="Development">
            <script src="~/lib/jquery/dist/jquery.js"></script>
            <script src="~/lib/bootstrap/dist/js/bootstrap.js"></script>
            <script src="~/lib/angular/angular.js"></script>
        </environment>

It’s recommended that production applications utilize CDNs for common libraries like Angular. You can reference Angular from one of several CDNs, such as this one:

        <environment names="Staging,Production">
            <script src="//ajax.aspnetcdn.com/ajax/jquery/jquery-2.1.4.min.js"
                    asp-fallback-src="~/lib/jquery/dist/jquery.min.js"
                    asp-fallback-test="window.jQuery">
            </script>
            <script src="//ajax.aspnetcdn.com/ajax/bootstrap/3.3.5/bootstrap.min.js"
                    asp-fallback-src="~/lib/bootstrap/dist/js/bootstrap.min.js"
                    asp-fallback-test="window.jQuery && window.jQuery.fn && window.jQuery.fn.modal">
            </script>
            <script src="https://ajax.googleapis.com/ajax/libs/angularjs/1.4.8/angular.min.js"
                    asp-fallback-src="~/lib/angular/angular.min.js"
                    asp-fallback-test="window.angular">
            </script>
            <script src="~/js/site.min.js" asp-append-version="true"></script>
        </environment>

Once you have a reference to the angular.js script file, you’re ready to begin using Angular in your web pages.

Key Components ¶

AngularJS includes a number of major components, such as directives, templates, repeaters, modules, controllers, and more. Let’s examine how these components work together to add behavior to your web pages.

Directives¶

AngularJS uses directives to extend HTML with custom attributes and elements. AngularJS directives are defined via data-ng-* or ng-* prefixes (ng is short for angular). There are two types of AngularJS directives:

Primitive Directives: These are predefined by the Angular team and are part of the AngularJS framework.

Custom Directives: These are custom directives that you can define.

One of the primitive directives used in all AngularJS applications is the ng-app directive, which bootstraps the AngularJS application. This directive can be applied to the <body> tag or to a child element of the body. Let’s see an example in action. Assuming you’re in an ASP.NET project, you can either add an HTML file to the wwwroot folder, or add a new controller action and an associated view. In this case, I’ve added a new Directives action method to HomeController.cs. The associated view is shown here:

@{
    Layout = "";
}
<html>
<body ng-app>
    <h1>Directives</h1>
    {{2+2}}
    <script src="~/lib/angular/angular.js"></script>
</body>
</html>

To keep these samples independent of one another, I’m not using the shared layout file. You can see that we decorated the body tag with the ng-app directive to indicate this page is an AngularJS application. The {{2+2}} is an Angular data binding expression that you will learn more about in a moment. Here is the result if you run this application:

Other primitive directives in AngularJS include:

ng-controller: Determines which JavaScript controller is bound to which view.
ng-model: Determines the model to which the values of an HTML element’s properties are bound.
ng-init: Used to initialize the application data in the form of an expression for the current scope.
ng-if: Removes or recreates the given HTML element in the DOM based on the truthiness of the expression provided.
ng-repeat: Repeats a given block of HTML over a set of data.
ng-show: Shows or hides the given HTML element based on the expression provided.

For a full list of all primitive directives supported in AngularJS, please refer to the directive documentation section on the AngularJS documentation website.

Data Binding¶

AngularJS provides data binding support out-of-the-box using either the ng-bind directive or a data binding expression syntax such as {{expression}}. AngularJS supports two-way data binding where data from a model is kept in synchronization with a view template at all times. Any changes to the view are automatically reflected in the model. Likewise, any changes in the model are reflected in the view.

Create either an HTML file or a controller action with an accompanying view named Databinding. Include the following in the view:

@{
    Layout = "";
}
<html>
<body ng-app>
    <h1>Databinding</h1>

    <div ng-init="firstName='John'; lastName='Doe';">
        <strong>First name:</strong> {{firstName}} <br />
        <strong>Last name:</strong> <span ng-bind="lastName" />
    </div>

    <script src="~/lib/angular/angular.js"></script>
</body>
</html>

Notice that you can display model values using either directives or data binding (ng-bind). The resulting page should look like this:

Templates¶

Templates in AngularJS are just plain HTML pages decorated with AngularJS directives and artifacts. A template in AngularJS is a mixture of directives, expressions, filters, and controls that combine with HTML to form the view.

Add another view to demonstrate templates, and add the following to it:

@{
    Layout = "";
}
<html>
<body ng-app>
    <h1>Templates</h1>

    <div ng-init="personName='John Doe'">
        <input ng-model="personName" /> {{personName}}
    </div>

    <script src="~/lib/angular/angular.js"></script>
</body>
</html>

The template has AngularJS directives like ng-app, ng-init, ng-model and data binding expression syntax to bind the personName property. Running in the browser, the view looks like the screenshot below:

If you change the name by typing in the input field, you will see the text next to the input field dynamically update, showing Angular two-way data binding in action.

Expressions¶

Expressions in AngularJS are JavaScript-like code snippets that are written inside the {{ expression }} syntax. The data from these expressions is bound to HTML the same way as ng-bind directives. The main difference between AngularJS expressions and regular JavaScript expressions is that AngularJS expressions are evaluated against the $scope object in AngularJS.

The AngularJS expressions in the sample below bind personName and a simple JavaScript calculated expression:

@{
    Layout = "";
}
<html>
<body ng-app>
    <h1>Expressions</h1>

    <div ng-init="personName='John Doe'">
        Person's name is: {{personName}} <br />
        Simple JavaScript calculation of 1 + 2: {{1+2}}
    </div>

    <script src="~/lib/angular/angular.js"></script>
</body>
</html>

The example running in the browser displays the personName data and the results of the calculation:

Repeaters¶

Repeating in AngularJS is done via a primitive directive called ng-repeat. The ng-repeat directive repeats a given HTML element in a view over the length of a repeating data array. Repeaters in AngularJS can repeat over an array of strings or objects. Here is a sample usage of repeating over an array of strings:

@{
    Layout = "";
}
<html>
<body ng-app>
    <h1>Repeaters</h1>

    <div ng-init="names=['John Doe', 'Mary Jane', 'Bob Parker']">
        <ul>
            <li ng-repeat="name in names">
                {{name}}
            </li>
        </ul>
    </div>

    <script src="~/lib/angular/angular.js"></script>
</body>
</html>

The repeat directive outputs a series of list items in an unordered list, as you can see in the developer tools shown in this screenshot:

Here is an example that repeats over an array of objects. The ng-init directive establishes a names array, where each element is an object containing first and last names. The ng-repeat assignment, name in names, outputs a list item for every array element.

@{
    Layout = "";
}
<html>
<body ng-app>
    <h1>Repeaters2</h1>

    <div ng-init="names=[
		{firstName:'John', lastName:'Doe'},
		{firstName:'Mary', lastName:'Jane'},
		{firstName:'Bob', lastName:'Parker'}]">
        <ul>
            <li ng-repeat="name in names">
                {{name.firstName + ' ' + name.lastName}}
            </li>
        </ul>
    </div>

    <script src="~/lib/angular/angular.js"></script>
</body>
</html>

The output in this case is the same as in the previous example.

Angular provides some additional directives that can help provide behavior based on where the loop is in its execution.

$index: Use $index in the ng-repeat loop to determine which index position your loop currently is on.
$even and $odd: Use $even in the ng-repeat loop to determine whether the current index in your loop is an even indexed row. Similarly, use $odd to determine if the current index is an odd indexed row.
$first and $last: Use $first in the ng-repeat loop to determine whether the current index in your loop is the first row. Similarly, use $last to determine if the current index is the last row.

Below is a sample that shows $index, $even, $odd, $first, and $last in action:

@{
    Layout = "";
}
<html>
<body ng-app>
    <h1>Repeaters2</h1>

    <div ng-init="names=[
		{firstName:'John', lastName:'Doe'},
		{firstName:'Mary', lastName:'Jane'},
		{firstName:'Bob', lastName:'Parker'}]">
        <ul>
            <li ng-repeat="name in names">
                {{name.firstName + ' ' + name.lastName}} at index {{$index}}
                <span ng-show="{{$first}}">, the first position</span>
                <span ng-show="{{$last}}">, the last position</span>
                <span ng-show="{{$odd}}">,which is odd-numbered.</span>
                <span ng-show="{{$even}}">,which is even-numbered.</span>
            </li>
        </ul>
    </div>

    <script src="~/lib/angular/angular.js"></script>
</body>
</html>

Here is the resulting output:

$scope¶

$scope is a JavaScript object that acts as glue between the view (template) and the controller (explained below). A view template in AngularJS only knows about the values attached to the $scope object in the controller.

Note

In the MVVM world, the $scope object in AngularJS is often defined as the ViewModel. The AngularJS team refers to the $scope object as the Data-Model. Learn more about Scopes in AngularJS.

Below is a simple example showing how to set properties on $scope within a separate JavaScript file, scope.js:

var personApp = angular.module('personApp', []);
personApp.controller('personController', ['$scope', function ($scope) {
    $scope.name = 'Mary Jane';
}]);

Observe the $scope parameter passed to the controller on line 2. This object is what the view knows about. On line 3, we are setting a property called “name” to “Mary Jane”.

What happens when a particular property is not found by the view? The view defined below refers to “name” and “age” properties:

@{
    Layout = "";
}
<html>
<body ng-app="personApp">
    <h1>Scope</h1>

    <div ng-controller="personController">
        <strong>Name:</strong> {{name}} <br />
        <strong>Missing Property (age):</strong> {{age}}
    </div>

    <script src="~/lib/angular/angular.js"></script>
    <script src="~/app/scope.js"></script>
</body>
</html>

Notice on line 9 that we are asking Angular to show the “name” property using expression syntax. Line 10 then refers to “age”, a property that does not exist. The running example shows the name set to “Mary Jane” and nothing for age. Missing properties are ignored.

Modules¶

A module in AngularJS is a collection of controllers, services, directives, etc. The angular.module() function call is used to create, register, and retrieve modules in AngularJS. All modules, including those shipped by the AngularJS team and third party libraries, should be registered using the angular.module() function.

Below is a snippet of code that shows how to create a new module in AngularJS. The first parameter is the name of the module. The second parameter defines dependencies on other modules. Later in this article, we will be showing how to pass these dependencies to an angular.module() method call.

var personApp = angular.module('personApp', []);

Use the ng-app directive to represent an AngularJS module on the page. To use a module, assign the name of the module, personApp in this example, to the ng-app directive in our template.

<body ng-app="personApp">

Controllers¶

Controllers in AngularJS are the first point of entry for your code. The <module name>.controller() function call is used to create and register controllers in AngularJS. The ng-controller directive is used to represent an AngularJS controller on the HTML page. The role of the controller in Angular is to set state and behavior of the data model ($scope). Controllers should not be used to manipulate the DOM directly.

Below is a snippet of code that registers a new controller. The personApp variable in the snippet references an Angular module, which is defined on line 2.

// module
var personApp = angular.module('personApp', []);

// controller
personApp.controller('personController', function ($scope) {
    $scope.firstName = "Mary";
    $scope.lastName = "Jane"
});

The view using the ng-controller directive assigns the controller name:

@{
    Layout = "";
}
<html>
<body ng-app="personApp">
    <h1>Controllers</h1>

    <div ng-controller="personController">
        <strong>First Name:</strong> {{firstName}} <br />
        <strong>Last Name:</strong> {{lastName}}
    </div>

    <script src="~/lib/angular/angular.js"></script>
    <script src="~/app/controllers.js"></script>
</body>
</html>

The page shows “Mary” and “Jane” that correspond to the firstName and lastName properties attached to the $scope object:

Services¶

Services in AngularJS are commonly used for shared code that is abstracted away into a file which can be used throughout the lifetime of an Angular application. Services are lazily instantiated, meaning that there will not be an instance of a service unless a component that depends on the service gets used. Factories are an example of a service used in AngularJS applications. Factories are created using the myApp.factory() function call, where myApp is the module.

Below is an example that shows how to use factories in AngularJS:

personApp.factory('personFactory', function () {
    function getName() {
        return "Mary Jane";
    }

    var service = {
        getName: getName
    };

    return service;
});

To call this factory from the controller, pass personFactory as a parameter to the controller function:

personApp.controller('personController', function($scope,personFactory) {
  $scope.name = personFactory.getName();
});

Using services to talk to a REST endpoint¶

Below is an end-to-end example using services in AngularJS to interact with an ASP.NET Core Web API endpoint. The example gets data from the Web API and displays the data in a view template. Let’s start with the view first:

@{
    Layout = "";
}
<html>
<body ng-app="PersonsApp">
    <h1>People</h1>

    <div ng-controller="personController">
        <ul>
            <li ng-repeat="person in people">
                <h2>{{person.FirstName}} {{person.LastName}}</h2>
            </li>
        </ul>
    </div>

    <script src="~/lib/angular/angular.js"></script>
    <script src="~/app/personApp.js"></script>
    <script src="~/app/personFactory.js"></script>
    <script src="~/app/personController.js"></script>
</body>
</html>

In this view, we have an Angular module called PersonsApp and a controller called personController. We are using ng-repeat to iterate over the list of persons. We are referencing three custom JavaScript files on lines 17-19.

The personApp.js file is used to register the PersonsApp module; and, the syntax is similar to previous examples. We are using the angular.module function to create a new instance of the module that we will be working with.

(function () {
    'use strict';
    var app = angular.module('PersonsApp', []);
})();

Let’s take a look at personFactory.js, below. We are calling the module’s factory method to create a factory. Line 12 shows the built-in Angular $http service retrieving people information from a web service.

(function () {
    'use strict';

    var serviceId = 'personFactory';

    angular.module('PersonsApp').factory(serviceId,
        ['$http', personFactory]);

    function personFactory($http) {

        function getPeople() {
            return $http.get('/api/people');
        }

        var service = {
            getPeople: getPeople
        };

        return service;
    }
})();

In personController.js, we are calling the module’s controller method to create the controller. The $scope object’s people property is assigned the data returned from the personFactory (line 13).

(function () {
    'use strict';

    var controllerId = 'personController';

    angular.module('PersonsApp').controller(controllerId,
        ['$scope', 'personFactory', personController]);

    function personController($scope, personFactory) {
        $scope.people = [];

        personFactory.getPeople().success(function (data) {
            $scope.people = data;
        }).error(function (error) {
            // log errors
        });
    }
})();

Let’s take a quick look at the Web API and the model behind it. The Person model is a POCO (Plain Old CLR Object) with Id, FirstName, and LastName properties:

namespace AngularSample.Models
{
    public class Person
    {
        public int Id { get; set; }
        public string FirstName { get; set; }
        public string LastName { get; set; }
    }
}

The Person controller returns a JSON-formatted list of Person objects:

using AngularSample.Models;
using Microsoft.AspNet.Mvc;
using System.Collections.Generic;

namespace AngularSample.Controllers.Api
{
    public class PersonController : Controller
    {
        [Route("/api/people")]
        public JsonResult GetPeople()
        {
            var people = new List<Person>()
            {
                new Person { Id = 1, FirstName = "John", LastName = "Doe" },
                new Person { Id = 1, FirstName = "Mary", LastName = "Jane" },
                new Person { Id = 1, FirstName = "Bob", LastName = "Parker" }
            };

            return Json(people);
        }
    }
}

Let’s see the application in action:

You can view the application’s structure on GitHub.

Note

For more on structuring AngularJS applications, see John Papa’s Angular Style Guide

Note

To create AngularJS module, controller, factory, directive and view files easily, be sure to check out Sayed Hashimi’s SideWaffle template pack for Visual Studio. Sayed Hashimi is a Senior Program Manager on the Visual Studio Web Team at Microsoft and SideWaffle templates are considered the gold standard. At the time of this writing, SideWaffle is available for Visual Studio 2012, 2013, and 2015.

Routing and Multiple Views¶

AngularJS has a built-in route provider to handle SPA (Single Page Application) based navigation. To work with routing in AngularJS, you must add the angular-route library using Bower. You can see in the bower.json file referenced at the start of this article that we are already referencing it in our project.

After you install the package, add the script reference (angular-route.js) to your view.

Now let’s take the Person App we have been building and add navigation to it. First, we will make a copy of the app by creating a new PeopleController action called Spa and a corresponding Spa.cshtml view by copying the Index.cshtml view in the People folder. Add a script reference to angular-route (see line 11). Also add a div marked with the ng-view directive (see line 6) as a placeholder to place views in. We are going to be using several additional .js files which are referenced on lines 13-16.

@{
    Layout = "";
}
<html>
<body ng-app="personApp">
    <div ng-view>

    </div>

    <script src="~/lib/angular/angular.js"></script>
    <script src="~/lib/angular-route/angular-route.js"></script>

    <script src="~/app/personModule.js"></script>
    <script src="~/app/personRoutes.js"></script>
    <script src="~/app/personListController.js"></script>
    <script src="~/app/personDetailController.js"></script>
</body>
</html>

Let’s take a look at personModule.js file to see how we are instantiating the module with routing. We are passing ngRoute as a library into the module. This module handles routing in our application.

var personApp = angular.module('personApp', ['ngRoute']);

The personRoutes.js file, below, defines routes based on the route provider. Lines 4-7 define navigation by effectively saying, when a URL with /persons is requested, use a template called partials/personlist by working through personListController. Lines 8-11 indicate a detail page with a route parameter of personId. If the URL doesn’t match one of the patterns, Angular defaults to the /persons view.

personApp.config(['$routeProvider',
    function ($routeProvider) {
        $routeProvider.
            when('/persons', {
                templateUrl: '/app/partials/personlist.html',
                controller: 'personListController'
            }).
            when('/persons/:personId', {
                templateUrl: '/app/partials/persondetail.html',
                controller: 'personDetailController'
            }).
            otherwise({
                redirectTo: '/persons'
            })
    }
]);

The personlist.html file is a partial view containing only the HTML needed to display person list.

<div>
    <h1>PERSONS PAGE</h1>
    <span ng-bind="message"/>
</div>

The controller is defined by using the module’s controller function in personListController.js.

personApp.controller('personListController', function ($scope) {
    $scope.message = "You are on the Persons List Page.";
})

If we run this application and navigate to the people/spa#/persons URL, we will see:

If we navigate to a detail page, for example people/spa#/persons/2, we will see the detail partial view:

You can view the full source and any files not shown in this article on GitHub.

Event Handlers¶

There are a number of directives in AngularJS that add event-handling capabilities to the input elements in your HTML DOM. Below is a list of the events that are built into AngularJS.

ng-click

ng-dbl-click

ng-mousedown

ng-mouseup

ng-mouseenter

ng-mouseleave

ng-mousemove

ng-keydown

ng-keyup

ng-keypress

ng-change

Note

You can add your own event handlers using the custom directives feature in AngularJS.

Let’s look at how the ng-click event is wired up. Create a new JavaScript file named eventHandlerController.js, and add the following to it:

personApp.controller('eventHandlerController', function ($scope) {
    $scope.firstName = 'Mary';
    $scope.lastName = 'Jane';

    $scope.sayName = function () {
        alert('Welcome, ' + $scope.firstName + ' ' + $scope.lastName);
    }
});

Notice the new sayName function in eventHandlerController on line 5 above. All the method is doing for now is showing a JavaScript alert to the user with a welcome message.

The view below binds a controller function to an AngularJS event. Line 9 has a button on which the ng-click Angular directive has been applied. It calls our sayName function, which is attached to the $scope object passed to this view.

@{
    Layout = "";
}
<html>
<body ng-app="personApp">
    <div ng-controller="eventHandlerController">
        <strong>First Name:</strong> {{firstName}} <br />
        <strong>Last Name:</strong> {{lastName}} <br />
        <input ng-click="sayName()" type="button" value="Say Name" />
    </div>
    <script src="~/lib/angular/angular.js"></script>
    <script src="~/lib/angular-route/angular-route.js"></script>

    <script src="~/app/personModule.js"></script>
    <script src="~/app/eventHandlerController.js"></script>
</body>
</html>

The running example demonstrates that the controller’s sayName function is called automatically when the button is clicked.

For more detail on AngularJS built-in event handler directives, be sure to head to the documentation website of AngularJS.

Angular 2.0 ¶

Angular 2.0 is the next version of AngularJS, which is completely reimagined with ES6 and mobile in mind. It’s built using Microsoft’s TypeScript language. Angular 2.0 is currently a RC product and is expected to be released in early 2016. Several breaking changes will be introduced in the Angular 2.0 release, so the Angular team is working hard to provide guidance to developers. A migration path will become more clear as the release date approaches. If you wish to play with Angular 2.0 now, the Angular team has created Angular.io to show their progress, to provide early documentation, and to gather feedback.

Summary ¶

This article provides an overview of AngularJS for ASP.NET developers. It aims to help developers who are new to this SPA framework get up-to-speed quickly.

Styling Applications with Less, Sass, and Font Awesome¶

http://fortawesome.github.io/Font-Awesome/icons/

Users of web applications have increasingly high expectations when it comes to style and overall experience. Modern web applications frequently leverage rich tools and frameworks for defining and managing their look and feel in a consistent manner. Frameworks like Bootstrap can go a long way toward defining a common set of styles and layout options for the web sites. However, most non-trivial sites also benefit from being able to effectively define and maintain styles and cascading style sheet (CSS) files, as well as having easy access to non-image icons that help make the site’s interface more intuitive. That’s where languages and tools that support Less and Sass, and libraries like Font Awesome, come in.

Sections:

CSS Preprocessor Languages
Less
Sass
Less or Sass?
Font Awesome
Summary

CSS Preprocessor Languages ¶

Languages that are compiled into other languages, in order to improve the experience of working with the underlying language, are referred to as pre-processors. There are two popular pre-processors for CSS: Less and Sass. These pre-processors add features to CSS, such as support for variables and nested rules, which improve the maintainability of large, complex stylesheets. CSS as a language is very basic, lacking support even for something as simple as variables, and this tends to make CSS files repetitive and bloated. Adding real programming language features via preprocessors can help reduce duplication and provide better organization of styling rules. Visual Studio provides built-in support for both Less and Sass, as well as extensions that can further improve the development experience when working with these languages.

As a quick example of how preprocessors can improve readability and maintainability of style information, consider this CSS:

.header {
   color: black;
   font-weight: bold;
   font-size: 18px;
   font-family: Helvetica, Arial, sans-serif;
}

.small-header {
   color: black;
   font-weight: bold;
   font-size: 14px;
   font-family: Helvetica, Arial, sans-serif;
}

Using Less, this can be rewritten to eliminate all of the duplication, using a mixin (so named because it allows you to “mix in” properties from one class or rule-set into another):

.header {
   color: black;
   font-weight: bold;
   font-size: 18px;
   font-family: Helvetica, Arial, sans-serif;
}

.small-header {
   .header;
   font-size: 14px;
}

Visual Studio adds a great deal of built-in support for Less and Sass. You can also add support for earlier versions of Visual Studio by installing the Web Essentials extension.

Less ¶

The Less CSS pre-processor runs using Node.js. You can quickly install it using the Node Package Manager (NPM), with:

npm install -g less

If you’re using Visual Studio, you can get started with Less by adding one or more Less files to your project, and then configuring Gulp (or Grunt) to process them at compile-time. Add a Styles folder to your project, and then add a new Less file called main.less to this folder.

Once added, your folder structure should look something like this:

Now we can add some basic styling to the file, which will be compiled into CSS and deployed to the wwwroot folder by Gulp.

Modify main.less to include the following content, which creates a simple color palette from a single base color.

@base: #663333;
@background: spin(@base, 180);
@lighter: lighten(spin(@base, 5), 10%);
@lighter2: lighten(spin(@base, 10), 20%);
@darker: darken(spin(@base, -5), 10%);
@darker2: darken(spin(@base, -10), 20%);

body {
  background-color:@background;
}
.baseColor  {color:@base}
.bgLight    {color:@lighter}
.bgLight2   {color:@lighter2}
.bgDark     {color:@darker}
.bgDark2    {color:@darker2}

@base and the other @-prefixed items are variables. Each of them represents a color. Except for @base, they are set using color functions: lighten, darken, and spin. Lighten and darken do pretty much what you would expect; spin adjusts the hue of a color by a number of degrees (around the color wheel). The less processor is smart enough to ignore variables that aren’t used, so to demonstrate how these variables work, we need to use them somewhere. The classes .baseColor, etc. will demonstrate the calculated values of each of the variables in the CSS file that is produced.

Getting Started¶

If you don’t already have one in your project, add a new Gulp configuration file. Make sure package.json includes gulp in its devDependencies, and add “gulp-less”:

"devDependencies": {
    "gulp": "3.8.11",
    "gulp-less": "3.0.2",
    "rimraf": "2.3.2"
  }

Save your changes to the package.json file, and you should see that the all of the files referenced can be found in the Dependencies folder under NPM. If not, right-click on the NPM folder and select “Restore Packages.”

Now open gulpfile.js. Add a variable at the top to represent less:

var gulp = require("gulp"),
          rimraf = require("rimraf"),
          fs = require("fs"),
          less = require("gulp-less");

add another variable to allow you to access project properties:

var project = require('./project.json');

Next, add a task to run less, using the syntax shown here:

gulp.task("less", function () {
  return gulp.src('Styles/main.less')
    .pipe(less())
    .pipe(gulp.dest(project.webroot + '/css'));
});

Open the Task Runner Explorer (view>Other Windows > Task Runner Explorer). Among the tasks, you should see a new task named less. Run it, and you should have output similar to what is shown here:

Now refresh your Solution Explorer and inspect the contents of the wwwroot/css folder. You should find a new file, main.css, there:

Open main.css and you should see something like the following:

body {
  background-color: #336666;
}
.baseColor {
  color: #663333;
}
.bgLight {
  color: #884a44;
}
.bgLight2 {
  color: #aa6355;
}
.bgDark {
  color: #442225;
}
.bgDark2 {
  color: #221114;
}

Add a simple HTML page to the wwwroot folder and reference main.css to see the color palette in action.

<!DOCTYPE html>
<html>
<head>
  <meta charset="utf-8" />
  <link href="css/main.css" rel="stylesheet" />
  <title></title>
</head>
<body>
  <div>
    <div class="baseColor">BaseColor</div>
    <div class="bgLight">Light</div>
    <div class="bgLight2">Light2</div>
    <div class="bgDark">Dark</div>
    <div class="bgDark2">Dark2</div>
  </div>
</body>
</html>

You can see that the 180 degree spin on @base used to produce @background resulted in the color wheel opposing color of @base:

Less also provides support for nested rules, as well as nested media queries. For example, defining nested hierarchies like menus can result in verbose CSS rules like these:

nav {
  height: 40px;
  width: 100%;
}
nav li {
  height: 38px;
  width: 100px;
}
nav li a:link {
  color: #000;
  text-decoration: none;
}
nav li a:visited {
  text-decoration: none;
  color: #CC3333;
}
nav li a:hover {
  text-decoration: underline;
  font-weight: bold;
}
nav li a:active {
  text-decoration: underline;
}

Ideally all of the related style rules will be placed together within the CSS file, but in practice there is nothing enforcing this rule except convention and perhaps block comments.

Defining these same rules using Less looks like this:

nav {
  height: 40px;
  width: 100%;
  li {
    height: 38px;
    width: 100px;
    a {
      color: #000;
      &:link { text-decoration:none}
      &:visited { color: #CC3333; text-decoration:none}
      &:hover { text-decoration:underline; font-weight:bold}
      &:active {text-decoration:underline}
    }
  }
}

Note that in this case, all of the subordinate elements of nav are contained within its scope. There is no longer any repetition of parent elements (nav, li, a), and the total line count has dropped as well (though some of that is a result of putting values on the same lines in the second example). It can be very helpful, organizationally, to see all of the rules for a given UI element within an explicitly bounded scope, in this case set off from the rest of the file by curly braces.

The & syntax is a Less selector feature, with & representing the current selector parent. So, within the a {...} block, & represents an a tag, and thus &:link is equivalent to a:link.

Media queries, extremely useful in creating responsive designs, can also contribute heavily to repetition and complexity in CSS. Less allows media queries to be nested within classes, so that the entire class definition doesn’t need to be repeated within different top-level @media elements. For example, this CSS for a responsive menu:

.navigation {
  margin-top: 30%;
  width: 100%;
}
@media screen and (min-width: 40em) {
  .navigation {
    margin: 0;
  }
}
@media screen and (min-width: 62em) {
  .navigation {
    width: 960px;
    margin: 0;
  }
}

This can be better defined in Less as:

.navigation {
  margin-top: 30%;
  width: 100%;
  @media screen and (min-width: 40em) {
    margin: 0;
  }
  @media screen and (min-width: 62em) {
    width: 960px;
    margin: 0;
  }
}

Another feature of Less that we have already seen is its support for mathematical operations, allowing style attributes to be constructed from pre-defined variables. This makes updating related styles much easier, since the base variable can be modified and all dependent values change automatically.

CSS files, especially for large sites (and especially if media queries are being used), tend to get quite large over time, making working with them unwieldy. Less files can be defined separately, then pulled together using @import directives. Less can also be used to import individual CSS files, as well, if desired.

Mixins can accept parameters, and Less supports conditional logic in the form of mixin guards, which provide a declarative way to define when certain mixins take effect. A common use for mixin guards is to adjust colors based on how light or dark the source color is. Given a mixin that accepts a parameter for color, a mixin guard can be used to modify the mixin based on that color:

.box (@color) when (lightness(@color) >= 50%) {
  background-color: #000;
}
.box (@color) when (lightness(@color) < 50%) {
  background-color: #FFF;
}
.box (@color) {
  color: @color;
}

.feature {
  .box (@base);
}

Given our current @base value of #663333, this Less script will produce the following CSS:

.feature {
  background-color: #FFF;
  color: #663333;
}

Less provides a number of additional features, but this should give you some idea of the power of this preprocessing language.

Sass ¶

Sass is similar to Less, providing support for many of the same features, but with slightly different syntax. It is built using Ruby, rather than JavaScript, and so has different setup requirements. The original Sass language did not use curly braces or semicolons, but instead defined scope using white space and indentation. In version 3 of Sass, a new syntax was introduced, SCSS (“Sassy CSS”). SCSS is similar to CSS in that it ignores indentation levels and whitespace, and instead uses semicolons and curly braces.

To install Sass, typically you would first install Ruby (pre-installed on Mac), and then run:

gem install sass

However, assuming you’re running Visual Studio, you can get started with Sass in much the same way as you would with Less. Open package.json and add the “gulp-sass” package to devDependencies:

"devDependencies": {
  "gulp": "3.8.11",
  "gulp-less": "3.0.2",
  "gulp-sass": "1.3.3",
  "rimraf": "2.3.2"
}

Next, modify gulpfile.js to add a sass variable and a task to compile your Sass files and place the results in the wwwroot folder:

var gulp = require("gulp"),
  rimraf = require("rimraf"),
  fs = require("fs"),
  less = require("gulp-less"),
  sass = require("gulp-sass");

// other content removed

gulp.task("sass", function () {
  return gulp.src('Styles/main2.scss')
    .pipe(sass())
    .pipe(gulp.dest(project.webroot + '/css'));
});

Now you can add the Sass file main2.scss to the Styles folder in the root of the project:

Open main2.scss and add the following:

$base: #CC0000;
body {
  background-color: $base;
}

Save all of your files. Now in Task Runner Explorer, you should see a sass task. Run it, refresh solution explorer, and look in the /wwwroot/css folder. There should be a main2.css file, with these contents:

body {
  background-color: #CC0000; }

Sass supports nesting in much the same was that Less does, providing similar benefits. Files can be split up by function and included using the @import directive:

@import 'anotherfile';

Sass supports mixins as well, using the @mixin keyword to define them and @include to include them, as in this example from sass-lang.com:

@mixin border-radius($radius) {
  -webkit-border-radius: $radius;
   -moz-border-radius: $radius;
    -ms-border-radius: $radius;
      border-radius: $radius;
}

.box { @include border-radius(10px); }

In addition to mixins, Sass also supports the concept of inheritance, allowing one class to extend another. It’s conceptually similar to a mixin, but results in less CSS code. It’s accomplished using the @extend keyword. First, let’s see how we might use mixins, and the resulting CSS code. Add the following to your main2.scss file:

@mixin alert {
  border: 1px solid black;
  padding: 5px;
  color: #333333;
}

.success {
  @include alert;
  border-color: green;
}

.error {
  @include alert;
  color: red;
  border-color: red;
  font-weight:bold;
}

Examine the output in main2.css after running the sass task in Task Runner Explorer:

.success {
  border: 1px solid black;
  padding: 5px;
  color: #333333;
  border-color: green;
 }

.error {
  border: 1px solid black;
  padding: 5px;
  color: #333333;
  color: red;
  border-color: red;
  font-weight: bold;
}

Notice that all of the common properties of the alert mixin are repeated in each class. The mixin did a good job of helping use eliminate duplication at development time, but it’s still creating CSS with a lot of duplication in it, resulting in larger than necessary CSS files - a potential performance issue. It would be great if we could follow the Don’t Repeat Yourself (DRY) Principle at both development time and runtime.

Now replace the alert mixin with a .alert class, and change @include to @extend (remembering to extend .alert, not alert):

.alert {
  border: 1px solid black;
  padding: 5px;
  color: #333333;
}

.success {
  @extend .alert;
  border-color: green;
}

.error {
  @extend .alert;
  color: red;
  border-color: red;
  font-weight:bold;
}

Run Sass once more, and examine the resulting CSS:

.alert, .success, .error {
  border: 1px solid black;
  padding: 5px;
  color: #333333; }

.success {
  border-color: green; }

.error {
  color: red;
  border-color: red;
  font-weight: bold; }

Now the properties are defined only as many times as needed, and better CSS is generated.

Sass also includes functions and conditional logic operations, similar to Less. In fact, the two languages’ capabilities are very similar.

Less or Sass?¶

There is still no consensus as to whether it’s generally better to use Less or Sass (or even whether to prefer the original Sass or the newer SCSS syntax within Sass). A recent poll conducted on twitter of mostly ASP.NET developers found that the majority preferred to use Less, by about a 2-to-1 margin. Probably the most important decision is to use one of these tools, as opposed to just hand-coding your CSS files. Once you’ve made that decision, both Less and Sass are good choices.

Font Awesome ¶

In addition to CSS pre-compilers, another great resource for styling modern web applications is Font Awesome. Font Awesome is a toolkit that provides over 500 scalable vector icons that can be freely used in your web applications. It was originally designed to work with Bootstrap, but has no dependency on that framework, or on any JavaScript libraries.

The easiest way to get started with Font Awesome is to add a reference to it, using its public content delivery network (CDN) location:

<link rel="stylesheet"
href="//maxcdn.bootstrapcdn.com/font-awesome/4.3.0/css/font-awesome.min.css">

Of course, you can also quickly add it to your Visual Studio project by adding it to the “dependencies” in bower.json:

{
  "name": "ASP.NET",
  "private": true,
  "dependencies": {
    "bootstrap": "3.0.0",
    "jquery": "1.10.2",
    "jquery-validation": "1.11.1",
    "jquery-validation-unobtrusive": "3.2.2",
    "hammer.js": "2.0.4",
    "bootstrap-touch-carousel": "0.8.0",
    "Font-Awesome": "4.3.0"
  }
}

Then, to get the stylesheet added to the wwwroot folder, modify gulpfile.js as follows:

gulp.task("copy", ["clean"], function () {
  var bower = {
    "angular": "angular/angular*.{js,map}",
    "bootstrap": "bootstrap/dist/**/*.{js,map,css,ttf,svg,woff,eot}",
    "bootstrap-touch-carousel": "bootstrap-touch-carousel/dist/**/*.{js,css}",
    "hammer.js": "hammer.js/hammer*.{js,map}",
    "jquery": "jquery/jquery*.{js,map}",
    "jquery-validation": "jquery-validation/jquery.validate.js",
    "jquery-validation-unobtrusive": "jquery-validation-unobtrusive/jquery.validate.unobtrusive.js",
    "font-awesome": "Font-Awesome/**/*.{css,otf,eot,svg,ttf,woff,wof2}"
  };

  for (var destinationDir in bower) {
    gulp.src(paths.bower + bower[destinationDir])
      .pipe(gulp.dest(paths.lib + destinationDir));
  }
});

Once this is in place (and saved), running the ‘copy’ task in Task Runner Explorer should copy the font awesome fonts and css files to /lib/font-awesome.

Once you have a reference to it on a page, you can add icons to your application by simply applying Font Awesome classes, typically prefixed with “fa-”, to your inline HTML elements (such as <span> or <i>). As a very simple example, you can add icons to simple lists and menus using code like this:

<!DOCTYPE html>
<html>
<head>
  <meta charset="utf-8" />
  <title></title>
  <link href="lib/font-awesome/css/font-awesome.css" rel="stylesheet" />
</head>
<body>
  <ul class="fa-ul">
    <li><i class="fa fa-li fa-home"></i> Home</li>
    <li><i class="fa fa-li fa-cog"></i> Settings</li>
  </ul>
</body>
</html>

This produces the following in the browser - note the icon beside each item:

You can view a complete list of the available icons here:

Summary ¶

Modern web applications increasingly demand responsive, fluid designs that are clean, intuitive, and easy to use from a variety of devices. Managing the complexity of the CSS stylesheets required to achieve these goals is best done using a pre-processor like Less or Sass. In addition, toolkits like Font Awesome quickly provide well-known icons to textual navigation menus and buttons, improving the overall user experience of your application.

Bundling and Minification¶

By Rick Anderson, Erik Reitan and Daniel Roth

Bundling and minification are two techniques you can use in ASP.NET to improve page load performance for your web application. Bundling combines multiple files into a single file. Minification performs a variety of different code optimizations to scripts and CSS, which results in smaller payloads. Used together, bundling and minification improves load time performance by reducing the number of requests to the server and reducing the size of the requested assets (such as CSS and JavaScript files).

This article explains the benefits of using bundling and minification, including how these features can be used with ASP.NET Core applications.

Sections:

Overview
Bundling
Minification
Impact of Bundling and Minification
Controlling Bundling and Minification
See Also

Overview ¶

In ASP.NET Core apps, you bundle and minify the client-side resources during design-time using third party tools, such as Gulp and Grunt. By using design-time bundling and minification, the minified files are created prior to the application’s deployment. Bundling and minifying before deployment provides the advantage of reduced server load. However, it’s important to recognize that design-time bundling and minification increases build complexity and only works with static files.

Bundling and minification primarily improve the first page request load time. Once a web page has been requested, the browser caches the assets (JavaScript, CSS and images) so bundling and minification won’t provide any performance boost when requesting the same page, or pages on the same site requesting the same assets. If you don’t set the expires header correctly on your assets, and you don’t use bundling and minification, the browsers freshness heuristics will mark the assets stale after a few days and the browser will require a validation request for each asset. In this case, bundling and minification provide a performance increase even after the first page request.

Bundling ¶

Bundling is a feature that makes it easy to combine or bundle multiple files into a single file. Because bundling combines multiple files into a single file, it reduces the number of requests to the server that is required to retrieve and display a web asset, such as a web page. You can create CSS, JavaScript and other bundles. Fewer files, means fewer HTTP requests from your browser to the server or from the service providing your application. This results in improved first page load performance.

Bundling can be accomplished using the gulp-concat plugin, which is installed with the Node Package Manager (npm). Add the gulp-concat package to the devDependencies section of your package.json file. To edit your package.json file from Visual Studio right-click on the npm node under Dependencies in the solution explorer and select Open package.json:

{
  "name": "asp.net",
  "version": "0.0.0",
  "private": true,
  "devDependencies": {
    "gulp": "3.8.11",
    "gulp-concat": "2.5.2",
    "gulp-cssmin": "0.1.7",
    "gulp-uglify": "1.2.0",
    "rimraf": "2.2.8"
  }
}

Run npm install to install the specified packages. Visual Studio will automatically install npm packages whenever package.json is modified.

In your gulpfile.js import the gulp-concat module:

var gulp = require("gulp"),
    rimraf = require("rimraf"),
    concat = require("gulp-concat"),
    cssmin = require("gulp-cssmin"),
    uglify = require("gulp-uglify");

Use globbing patterns to specify the files that you want to bundle and minify:

var paths = {
    js: webroot + "js/**/*.js",
    minJs: webroot + "js/**/*.min.js",
    css: webroot + "css/**/*.css",
    minCss: webroot + "css/**/*.min.css",
    concatJsDest: webroot + "js/site.min.js",
    concatCssDest: webroot + "css/site.min.css"
};

You can then define gulp tasks that run concat on the desired files and output the result to your webroot:

gulp.task("min:js", function () {
    return gulp.src([paths.js, "!" + paths.minJs], { base: "." })
        .pipe(concat(paths.concatJsDest))
        .pipe(uglify())
        .pipe(gulp.dest("."));
});

gulp.task("min:css", function () {
    return gulp.src([paths.css, "!" + paths.minCss])
        .pipe(concat(paths.concatCssDest))
        .pipe(cssmin())
        .pipe(gulp.dest("."));
});

The gulp.src function emits a stream of files that can be piped to gulp plugins. An array of globs specifies the files to emit using node-glob syntax. The glob beginning with ! excludes matching files from the glob results up to that point.

Minification ¶

Minification performs a variety of different code optimizations to reduce the size of requested assets (such as CSS, image, JavaScript files). Common results of minification include removing unnecessary white space and comments, and shortening variable names to one character.

Consider the following JavaScript function:

AddAltToImg = function (imageTagAndImageID, imageContext) {
  ///<signature>
  ///<summary> Adds an alt tab to the image
  // </summary>
  //<param name="imgElement" type="String">The image selector.</param>
  //<param name="ContextForImage" type="String">The image context.</param>
  ///</signature>
  var imageElement = $(imageTagAndImageID, imageContext);
  imageElement.attr('alt', imageElement.attr('id').replace(/ID/, ''));
}

After minification, the function is reduced to the following:

AddAltToImg=function(t,a){var r=$(t,a);r.attr("alt",r.attr("id").replace(/ID/,""))};

In addition to removing the comments and unnecessary whitespace, the following parameters and variable names were renamed (shortened) as follows:

Original	Renamed
imageTagAndImageID	t
imageContext	a
imageElement	r

To minify your JavaScript files you can use the gulp-uglify plugin. For CSS you can use the gulp-cssmin plugin. Install these packages using npm as before:

{
  "name": "asp.net",
  "version": "0.0.0",
  "private": true,
  "devDependencies": {
    "gulp": "3.8.11",
    "gulp-concat": "2.5.2",
    "gulp-cssmin": "0.1.7",
    "gulp-uglify": "1.2.0",
    "rimraf": "2.2.8"
  }
}

Import the gulp-uglify and gulp-cssmin modules in your gulpfile.js file:

var gulp = require("gulp"),
    rimraf = require("rimraf"),
    concat = require("gulp-concat"),
    cssmin = require("gulp-cssmin"),
    uglify = require("gulp-uglify");

Add uglify to minify your bundled JavaScript files and cssmin to minify your bundled CSS files.

gulp.task("min:js", function () {
    return gulp.src([paths.js, "!" + paths.minJs], { base: "." })
        .pipe(concat(paths.concatJsDest))
        .pipe(uglify())
        .pipe(gulp.dest("."));
});

gulp.task("min:css", function () {
    return gulp.src([paths.css, "!" + paths.minCss])
        .pipe(concat(paths.concatCssDest))
        .pipe(cssmin())
        .pipe(gulp.dest("."));
});

To run bundling and minification tasks from the command-line using gulp (gulp min), or you can also execute any of your gulp tasks from within Visual Studio using the Task Runner Explorer. To use the Task Runner Explorer select gulpfile.js in the Solution Explorer and then select Tools > Task Runner Explorer:

Note

The gulp tasks for bundling and minification do not general run when your project is built and must be run manually.

Impact of Bundling and Minification ¶

The following table shows several important differences between listing all the assets individually and using bundling and minification on a simple web page:

Action	With B/M	Without B/M	Change
File Requests	7	18	157%
KB Transferred	156	264.68	70%
Load Time (MS)	885	2360	167%

The bytes sent had a significant reduction with bundling as browsers are fairly verbose with the HTTP headers that they apply on requests. The load time shows a big improvement, however this example was run locally. You will get greater gains in performance when using bundling and minification with assets transferred over a network.

Controlling Bundling and Minification ¶

In general, you want to use the bundled and minified files of your app only in a production environment. During development, you want to use your original files so your app is easier to debug.

You can specify which scripts and CSS files to include in your pages using the environment tag helper in your layout pages(See Tag Helpers). The environment tag helper will only render its contents when running in specific environments. See Working with Multiple Environments 다중 환경에서 작업하기 for details on specifying the current environment.

The following environment tag will render the unprocessed CSS files when running in the Development environment:

<environment names="Development">
    <link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />
    <link rel="stylesheet" href="~/css/site.css" />
</environment>

This environment tag will render the bundled and minified CSS files only when running in Production or Staging:

<environment names="Staging,Production">
    <link rel="stylesheet" href="https://ajax.aspnetcdn.com/ajax/bootstrap/3.3.6/css/bootstrap.min.css"
          asp-fallback-href="~/lib/bootstrap/dist/css/bootstrap.min.css"
          asp-fallback-test-class="sr-only" asp-fallback-test-property="position" asp-fallback-test-value="absolute" />
    <link rel="stylesheet" href="~/css/site.min.css" asp-append-version="true" />
</environment>

🔧 Working with a Content Delivery Network (CDN)¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

🔧 Responsive Design for the Mobile Web¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Building Projects with Yeoman¶

By Steve Smith, Scott Addie, Rick Anderson and Noel Rice

Yeoman generates complete projects for a given set of client tools. Yeoman is an open-source tool that works like a Visual Studio project template. The Yeoman command line tool yo works alongside a Yeoman generator. Generators define the technologies that go into a project.

Sections:

Install Node.js, npm, and Yeoman
Create an ASP.NET app
Restore, build and run
Specifying the client-side task runner
Building and Running from Visual Studio
Restoring, Building, and Hosting from the Command Line
Adding to Your Project with Sub Generators
Related Resources

Install Node.js, npm, and Yeoman ¶

To get started with Yeoman install Node.js. The installer includes Node.js and npm.

Follow the instructions on http://yeoman.io/learning/ to install yo, bower, grunt, and gulp.

npm install -g yo bower grunt-cli gulp

Note

If you get the error npm ERR! Please try running this command again as root/Administrator. on Mac OS, run the following command using sudo: sudo npm install -g yo bower grunt-cli gulp

From the command line, install the ASP.NET generator:

npm install -g generator-aspnet

Note

If you get a permission error, run the command under sudo as described above.

The –g flag installs the generator globally, so that it can be used from any path.

Create an ASP.NET app ¶

Create a directory for your projects

mkdir src
cd src

Run the ASP.NET generator for yo

yo aspnet

The generator displays a menu. Arrow down to the Empty Web Application project and tap Enter:

Use “EmptyWeb1” for the app name and then tap Enter

Yeoman will scaffold the project and its supporting files. Suggested next steps are also provided in the form of commands.

The ASP.NET generator creates ASP.NET Core projects that can be loaded into Visual Studio Code, Visual Studio, or run from the command line.

Restore, build and run ¶

Follow the suggested commands by changing directories to the EmptyWeb1 directory. Then run dotnet restore.

Build and run the app using dotnet build and dotnet run:

At this point you can navigate to the URL shown to test the newly created ASP.NET Core app.

Tip

If you were directed to this tutorial from Your First ASP.NET Core Application on a Mac Using Visual Studio Code, you can return now.

Specifying the client-side task runner ¶

The ASP.NET generator creates supporting files to configure client-side build tools. A Grunt or Gulp task runner file is added to your project to automate build tasks for Web projects. The default generator creates gulpfile.js to run tasks. Running the generator with the --grunt argument generates Gruntfile.js:

yo aspnet --grunt

The generator also configures package.json to load Grunt or Gulp dependencies. It also adds bower.json and .bowerrc files to restore client-side packages using the Bower client-side package manager.

Building and Running from Visual Studio ¶

You can load your generated ASP.NET Core web project directly into Visual Studio, then build and run your project from there. Follow the instructions above to scaffold a new ASP.NET Core app using yeoman. This time, choose Web Application from the menu and name the app MyWebApp.

Open Visual Studio. From the File menu, select Open ‣ Project/Solution.

In the Open Project dialog, navigate to the project.json file, select it, and click the Open button. In the Solution Explorer, the project should look something like the screenshot below.

Yeoman scaffolds a MVC web application, complete with both server- and client-side build support. Server-side dependencies are listed under the References node, and client-side dependencies in the Dependencies node of Solution Explorer. Dependencies are restored automatically when the project is loaded.

When all the dependencies are restored, press F5 to run the project. The default home page displays in the browser.

Restoring, Building, and Hosting from the Command Line ¶

You can prepare and host your web application using the .NET Core command-line interface.

From the command line, change the current directory to the folder containing the project (that is, the folder containing the project.json file):

cd src\MyWebApp

From the command line, restore the project’s NuGet package dependencies:

dotnet restore

Run the application:

dotnet run

The cross-platform Kestrel web server will begin listening on port 5000.

Open a web browser, and navigate to http://localhost:5000.

Adding to Your Project with Sub Generators ¶

You can add new generated files using Yeoman even after the project is created. Use sub generators to add any of the file types that make up your project. For example, to add a new class to your project, enter the yo aspnet:Class command followed by the name of the class. Execute the following command from the directory in which the file should be created:

yo aspnet:Class Person

The result is a file named Person.cs with a class named Person:

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;

namespace MyNamespace
{
    public class Person
    {
        public Person()
        {
        }
    }
}

Mobile¶

Publishing and Deployment¶

Publishing to IIS¶

By Luke Latham and Rick Anderson

Sections:

Supported operating systems
IIS configuration
Install the .NET Core Windows Server Hosting bundle
Application configuration
Deploy the application
Configure the website in IIS
Create a Data Protection Registry Hive
Common errors
Additional resources

Supported operating systems ¶

The following operating systems are supported:

Windows 7 and newer
Windows Server 2008 R2 and newer*

*Conceptually, the IIS configuration described in this document also applies to hosting ASP.NET Core applications on Nano Server IIS, but refer to ASP.NET Core on Nano Server for specific instructions.

IIS configuration ¶

Enable the Web Server (IIS) server role and establish role services.

Windows desktop operating systems¶

Navigate to Control Panel > Programs > Programs and Features > Turn Windows features on or off (left side of the screen). Open the group for Internet Information Services and Web Management Tools. Check the box for IIS Management Console. Check the box for World Wide Web Services. Accept the default features for World Wide Web Services or customize the IIS features to suit your needs.

Windows Server operating systems¶

For server operating systems, use the Add Roles and Features Wizard via the Manage menu or the link in Server Manager. On the Server Roles step, check the box for Web Server (IIS).

On the Role services step, select the IIS role services you desire or accept the default role services provided.

Proceed through the Confirmation step to enable the web server role and services.

Install the .NET Core Windows Server Hosting bundle ¶

Install the .NET Core Windows Server Hosting bundle on the server. The bundle will install the .NET Core Runtime, .NET Core Library, and the ASP.NET Core Module. The module creates the reverse-proxy between IIS and the Kestrel server.
Execute iisreset at the command line or restart the server to pickup changes to the system PATH.

For more information on the ASP.NET Core Module, including configuration of the module and setting environment variables with web.config, the use of app_offline.htm to suspend request processing, and activation of module logging, see ASP.NET Core Module Configuration Reference.

Application configuration ¶

Enabling the IISIntegration components¶

Include a dependency on the Microsoft.AspNetCore.Server.IISIntegration package in the application dependencies. Incorporate IIS Integration middleware into the application by adding the .UseIISIntegration() extension method to WebHostBuilder().

var host = new WebHostBuilder()
  .UseKestrel()
  .UseContentRoot(Directory.GetCurrentDirectory())
  .UseIISIntegration()
  .UseStartup<Startup>()
  .Build();

Note that code calling .UseIISIntegration() does not affect code portability.

Setting IISOptions for the IISIntegration service¶

To configure IISIntegration service options, include a service configuration for IISOptions in ConfigureServices.

services.Configure<IISOptions>(options => {
  ...
});

Option	Setting
AutomaticAuthentication	If true, the authentication middleware will alter the request user arriving and respond to generic challenges. If false, the authentication middleware will only provide identity and respond to challenges when explicitly indicated by the AuthenticationScheme.
ForwardClientCertificate	If true and the MS-ASPNETCORE-CLIENTCERT request header is present, the ITLSConnectionFeature will be populated.
ForwardWindowsAuthentication	If true, authentication middleware will attempt to authenticate using platform handler windows authentication. If false, authentication middleware won’t be added.

publish-iis tool¶

The publish-iis tool can be added to any .NET Core application and will configure the ASP.NET Core Module by creating or modifying the web.config file. The tool runs after publishing with the dotnet publish command or publishing with Visual Studio and will configure the processPath and arguments for you. If you’re publishing a web.config file by including the file in your project and listing the file in the publishOptions section of project.json, the tool will not modify other IIS settings you have included in the file.

To include the publish-iis tool in your application, add entries to the tools and scripts sections of project.json.

"tools": {
  "Microsoft.AspNetCore.Server.IISIntegration.Tools": "1.0.0-preview2-final"
},
"scripts": {
  "postpublish": "dotnet publish-iis --publish-folder %publish:OutputPath% --framework %publish:FullTargetFramework%"
}

Deploy the application ¶

On the target IIS server, create a folder to contain the application’s assets.
Within the folder you created, create a logs folder to hold application logs (if you plan to enable logging). If you plan to deploy your application with a logs folder in the payload, you may skip this step.
Deploy the application to the folder you created on the target IIS server. MSDeploy (Web Deploy) is the recommended mechanism for deployment, but you may use any of several methods to move the application to the server (for example, Xcopy, Robocopy, or PowerShell). Visual Studio users may use the default Visual Studio web publish script. For information on using Web Deploy, see Publishing to IIS with Web Deploy using Visual Studio.

Warning

.NET Core applications are hosted via a reverse-proxy between IIS and the Kestrel server. In order to create the reverse-proxy, the web.config file must be present at the content root path (typically the app base path) of the deployed application, which is the website physical path provided to IIS.

Sensitive files exist on the app’s physical path, including subfolders, such as my_application.runtimeconfig.json, my_application.xml (XML Documentation comments), and my_application.deps.json. The web.config file is required to create the reverse proxy to Kestrel, which prevents IIS from serving these and other sensitive files. Therefore, it is important that the web.config file is never accidently renamed or removed from the deployment.

Configure the website in IIS ¶

In IIS Manager, create a new website. Provide a Site name and set the Physical path to the application’s assets folder that you created. Provide the Binding configuration and create the website.
Set the application pool to No Managed Code. ASP.NET Core runs in a separate process and manages the runtime.

Note

If you change the default identity of the application pool from ApplicationPoolIdentity, verify the new identity has the required permissions to access the application’s assets and database.

Open the Add Website window.

Configure the website.

In the Application Pools panel, open the Edit Application Pool window by right-clicking on the website’s application pool and selecting Basic Settings... from the popup menu.

Set the .NET CLR version to No Managed Code.

Browse the website.

Create a Data Protection Registry Hive ¶

Data Protection keys used by ASP.NET applications are stored in registry hives external to the applications. To persist the keys for a given application, you must create a registry hive for the application’s application pool.

For standalone IIS installations, you may use the Data Protection Provision-AutoGenKeys.ps1 PowerShell script for each application pool used with an ASP.NET Core application. The keys will be persisted in the registry.

In web farm scenarios, an application can be configured to use a UNC path to store its data protection key ring. By default, the data protection keys are not encrypted. You can deploy an x509 certificate to each machine to encrypt the key ring. See Configuring Data Protection for details.

Warning

Data Protection is used by various ASP.NET middlewares, including those used in authentication. Even if you do not specifically call any Data Protection APIs from your own code you should configure Data Protection with the deployment script or in your own code. If you do not configure data protection when using IIS by default the keys will be held in memory and discarded when your application closes or restarts. This will then, for example, invalidate any cookies written by the cookie authentication and users will have to login again.

Common errors ¶

The following is not a complete list of errors. Should you encounter an error not listed here, please leave a detailed error message in the DISQUS section below (click Show comments to open the DISQUS panel).

To diagnose problems with IIS deployments, study browser output, examine the server’s Application log through Event Viewer, and enable module logging. The ASP.NET Core Module log will be found on the path provided in the stdoutLogFile attribute of the <aspNetCore> element in web.config. Any folders on the path provided in the attribute value must exist in the deployment. You must also set stdoutLogEnabled=”true” to enable module logging. Applications that use the publish-iis tooling to create the web.config file will default the stdoutLogEnabled setting to false, so you must manually provide the file or modify the file in order to enable module logging.

Several of the common errors do not appear in the browser, Application Log, and ASP.NET Core Module Log until the module startupTimeLimit (default: 120 seconds) and startupRetryCount (default: 2) have passed. Therefore, wait a full six minutes before deducing that the module has failed to start a process for the application.

A quick way to determine if the application is working properly is to run the application directly on Kestrel. If the application was published as a portable app, execute dotnet <my_app>.dll in the deployment folder. If the application was published as a self-contained app, run the application’s executable directly on the command line, <my_app>.exe, in the deployment folder. If Kestrel is listening on default port 5000, you should be able to browse the application at http://localhost:5000/. If the application responds normally at the Kestrel endpoint address, the problem is more likely related to the IIS-ASP.NET Core Module-Kestrel configuration and less likely within the application itself.

A way to determine if the IIS reverse proxy to the Kestrel server is working properly is to perform a simple static file request for a stylesheet, script, or image from the application’s static assets in wwwroot using Static File middleware. If the application can serve static files but MVC Views and other endpoints are failing, the problem is less likely related to the IIS-ASP.NET Core Module-Kestrel configuration and more likely within the application itself (for example, MVC routing or 500 Internal Server Error).

In most cases, enabling application logging will assist in troubleshooting problems with application or the reverse proxy. See Logging for more information.

Common errors and general troubleshooting instructions:

Installer unable to obtain VC++ Redistributable¶

Installer Exception: Installation of the .NET Core Windows Server Hosting Bundle fails with 0x80070002 - The system cannot find the file specified.

Troubleshooting:

If the server does not have Internet access while installing the server hosting bundle, this exception will ensue when the installer is prevented from obtaining the Microsoft Visual C++ 2015 Redistributable (x64) packages online. You may obtain an installer for the packages from the Microsoft Download Center.

Platform conflicts with RID¶

Browser: HTTP Error 502.5 - Process Failure
Application Log: - Application Error: Faulting module: KERNELBASE.dll Exception code: 0xe0434352 Faulting module path: C:\WINDOWS\system32\KERNELBASE.dll - IIS AspNetCore Module: Failed to start process with commandline ‘“dotnet” .\my_application.dll’ (portable app) or ‘“PATH\my_application.exe”’ (self-contained app), ErrorCode = ‘0x80004005’.
ASP.NET Core Module Log: Unhandled Exception: System.BadImageFormatException: Could not load file or assembly ‘teststandalone.dll’ or one of its dependencies. An attempt was made to load a program with an incorrect format.

Troubleshooting:

If you published a self-contained application, confirm that you didn’t set a platform in buildOptions of project.json that conflicts with the publishing RID. For example, do not specify a platform of x86 and publish with an RID of win81-x64 (dotnet publish -c Release -r win81-x64). The project will publish without warning or error but fail with the above logged exceptions on the server.

URI endpoint wrong or stopped website¶

Browser: ERR_CONNECTION_REFUSED
Application Log: No entry
ASP.NET Core Module Log: Log file not created

Troubleshooting:

Confirm you are using the correct URI endpoint for the application. Check your bindings.
Confirm that the IIS website is not in the Stopped state.

CoreWebEngine or W3SVC server features disabled¶

OS Exception: The IIS 7.0 CoreWebEngine and W3SVC features must be installed to use the Microsoft HTTP Platform Handler 1.x.

Troubleshooting:

Confirm that you have enabled the proper server role and features. See IIS Configuration.

Incorrect website physical path or application missing¶

Browser: 403 Forbidden: Access is denied –OR– 403.14 Forbidden: The Web server is configured to not list the contents of this directory.
Application Log: No entry
ASP.NET Core Module Log: Log file not created

Troubleshooting:

Check the IIS website Basic Settings and the physical application assets folder. Confirm that the application is in the folder at the IIS website Physical path.

Incorrect server role, module not installed, or incorrect permissions¶

Browser: 500.19 Internal Server Error: The requested page cannot be accessed because the related configuration data for the page is invalid.
Application Log: No entry
ASP.NET Core Module Log: Log file not created

Troubleshooting:

Confirm that you have enabled the proper server role. See IIS Configuration.
Check Programs & Features and confirm that the Microsoft ASP.NET Core Module has been installed. If the Microsoft ASP.NET Core Module is not present in the list of installed programs, install the module. See IIS Configuration.
Make sure that the Application Pool Process Model Identity is either set to ApplicationPoolIdentity; or if a custom identity is in use, confirm the identity has the correct permissions to access the application’s assets folder.

Hosting bundle not installed or server not restarted¶

Browser: 502.3 Bad Gateway: There was a connection error while trying to route the request.
Application Log: Process ‘0’ failed to start. Port = PORT, Error Code = ‘-2147024894’.
ASP.NET Core Module Log: Log file created but empty

Troubleshooting:

You may have deployed a portable application without installing .NET Core on the server. If you are attempting to deploy a portable application and have not installed .NET Core, run the .NET Core Windows Server Hosting Bundle Installer on the server. See Install the .NET Core Windows Server Hosting Bundle.
You may have deployed a portable application and installed .NET Core without restarting the server. Restart the server.

Incorrect proecessPath, missing PATH variable, or dotnet.exe access violation¶

Browser: HTTP Error 502.5 - Process Failure
Application Log: Failed to start process with commandline ‘“dotnet” .\my_application.dll’ (portable app) or ‘”.\my_application_Foo.exe”’ (self-contained app), ErrorCode = ‘0x80070002’.
ASP.NET Core Module Log: Log file created but empty

Troubleshooting:

Check the processPath attribute on the <aspNetCore> element in web.config to confirm that it is dotnet for a portable application or .\my_application.exe for a self-contained application.
For a portable application, dotnet.exe might not be accessible via the PATH settings. Confirm that C:\Program Files\dotnet\ exists in the System PATH settings.
For a portable application, dotnet.exe might not be accessible for the user identity of the Application Pool. Confirm that the AppPool user identity has access to the C:\Program Files\dotnet directory.

Incorrect arguments of <aspNetCore> element¶

Browser: HTTP Error 502.5 - Process Failure
Application Log: Failed to start process with commandline ‘“dotnet” .\my_application_Foo.dll’, ErrorCode = ‘0x80004005’.
ASP.NET Core Module Log: The application to execute does not exist: ‘PATH\my_application_Foo.dll’

Troubleshooting:

Examine the arguments attribute on the <aspNetCore> element in web.config to confirm that it is either (a) .\my_applciation.dll for a portable application; or (b) not present, an empty string (arguments=”“), or a list of your application’s arguments (arguments=”arg1, arg2, ...”) for a self-contained application.

Missing .NET Framework version¶

Browser: 502.3 Bad Gateway: There was a connection error while trying to route the request.
Application Log: Failed to start process with commandline ‘[IIS_WEBSITE_PHYSICAL_PATH] ‘, Error Code = ‘0x80004005’.
ASP.NET Core Module Log: Missing method, file, or assembly exception. The method, file, or assembly specified in the exception is a .NET Framework method, file, or assembly.

Troubleshooting:

Install the .NET Framework version missing from the server.

Stopped Application Pool¶

Browser: 503 Service Unavailable
Application Log: No entry
ASP.NET Core Module Log: Log file not created

Troubleshooting

Confirm that the Application Pool is not in the Stopped state.

IIS Integration middleware not implemented or .UseUrls() after .UseIISIntegration()¶

Browser: HTTP Error 502.5 - Process Failure
Application Log: Process was created with commandline ‘“dotnet” .\my_application.dll’ (portable app) or ‘”.\my_application.exe”’ (self-contained app) but either crashed or did not reponse within given time or did not listen on the given port ‘PORT’, ErrorCode = ‘0x800705b4’
ASP.NET Core Module Log: Log file created and shows normal operation.

Troubleshooting

Confirm that you have correctly referenced the IIS Integration middleware by calling the .UseIISIntegration() method of the application’s WebHostBuilder().
If you are using the .UseUrls() extension method when self-hosting with Kestrel, confirm that it is positioned before the .UseIISIntegration() extension method on WebHostBuilder(). .UseIISIntegration() must set the Url for the reverse-proxy when running Kestrel behind IIS and not have its value overridden by .UseUrls().

Additional resources ¶

Publishing to IIS with Web Deploy using Visual Studio¶

By Sayed Ibrahim Hashimi

Publishing an ASP.NET Core project to an IIS server with Web Deploy requires a few additional steps in comparison to an ASP.NET 4 project. We are working on simplifying the experience for the next version. Until then you can use these instructions to get started with publishing an ASP.NET Core web application using Web Deploy to any IIS host.

To publish an ASP.NET Core application to a remote IIS server the following steps are required.

Configure your remote IIS server to support ASP.NET Core
Create a publish profile
Customize the profile to support Web Deploy publish

In this document we will walk through each step.

Preparing your web server for ASP.NET Core¶

The first step is to ensure that your remote server is configured for ASP.NET Core. At a high level you’ll need.

An IIS server with IIS 7.5+
Install HttpPlatformHandler
Install Web Deploy v3.6

The HttpPlatformHandler is a new component that connects IIS with your ASP.NET Core application. You can get that from the following download links.

In addition to installing the HttpPlatformHandler, you’ll need to install the latest version of Web Deploy (version 3.6). To install Web Deploy 3.6 you can use the the Web Platform Installer. (WebPI) or directly from the download center. The preferred method is to use WebPI. WebPI offers a standalone setup as well as a configuration for hosting providers.

Configure Data Protection¶

To persist Data Protection keys you must create registry hives for each application pool to store the keys. You should use the Provisioning PowerShell script for each application pool you will be hosting ASP.NET Core applications under.

For web farm scenarios developers can configure their applications to use a UNC path to store the data protection key ring. By default this does not encrypt the key ring. You can deploy an x509 certificate to each machine and use that to encrypt the keyring. See the configuration APIs for more details.

Warning

Data Protection is used by various ASP.NET middlewares, including those used in authentication. Even if you do not specifically call any Data Protection APIs from your own code you should configure Data Protection with the deployment script or in your own code. If you do not configure data protection when using IIS by default the keys will be held in memory and discarded when your application closes or restarts. This will then, for example, invalidate any cookies written by the cookie authentication and users will have to login again.

You can find more info on configuring your IIS server for ASP.NET Core at Publishing to IIS. Now let’s move on to the Visual Studio experience.

Publishing with Visual Studio¶

After you have configured your web server, the next thing to do is to create a publish profile in Visual Studio. The easiest way to get started with publishing an ASP.NET Core application to a standard IIS host is to use a publish profile. If your hosting provider has support for creating a publish profile, download that and then import it into the Visual Studio publish dialog with the Import button. You can see that dialog shown below.

After importing the publish profile, there is one additional step that needs to be taken before being able to publish to a standard IIS host. In the publish PowerShell script generated (under PropertiesPublishProfiles) update the publish module version number from 1.0.1 to 1.0.2-beta2. After changing 1.0.1 to 1.0.2-beta2 you can use the Visual Studio publish dialog to publish and preview changes.

How Web Publishing In Visual Studio Works¶

By Sayed Ibrahim Hashimi

The web publish experience for ASP.NET Core projects has significantly changed from ASP.NET 4. This doc will provide an overview of the changes and instructions on how to customize the publish process. Unless stated otherwise, the instructions in this article are for publishing from Visual Studio. For an overview of how to publish a web app on ASP.NET Core see Publishing and Deployment.

In ASP.NET when you publish a Visual Studio web project MSBuild is used to drive the entire process. The project file (.csproj or .vbproj) is used to gather the files that need to be published as well as perform any updates during publish (for example updating web.config). The project file contains the full list of files as well as their type. That information is used to determine the files to publish. The logic was implemented in an MSBuild .targets file. During the publish process MSBuild will call Web Deploy (msdeploy.exe) to transfer the files to the final location. To customize the publish process you would need to update the publish profile, or project file, with custom MSBuild elements.

In ASP.NET Core the publish process has been simplified, we no longer store a reference to files that the project contains. All files are included in the project by default (files can be excluded from the project or from publish by updating project.json). When you publish an ASP.NET Core project from Visual Studio the following happens:

A publish profile is created at Properties\PublishProfiles\profilename.pubxml. The publish profile is an MSBuild file.

A PowerShell script is created at Properties\PublishProfiles\profilename.ps1.

dotnet publish is called to gather the files to publish to a temporary folder.

A PowerShell script is called passing in the properties from the publish profile and the location where dotnet publish has placed the files to publish.

To create a publish profile in Visual Studio, right click on the project in Solution Explorer and then select Publish.

The following image shows a visualization of this process.

In the image above each black circle ● indicates an extension point, we will cover each extension point later in this document.

When you start a publish operation, the publish dialog is closed and then MSBuild is called to start the process. Visual Studio calls MSBuild to do this so that you can have parity with publishing when using Visual Studio or the command line. The MSBuild layer is pretty thin, for the most part it just calls dotnet publish. Let’s take a closer look at dotnet publish.

The dotnet publish command will inspect project.json and the project folder to determine the files which need to be published. It will place the files needed to run the application in a single folder ready to be transferred to the final destination.

After dotnet publish has completed, the PowerShell script for the publish profile is called. Now that we have briefly discussed how publishing works at a high level let’s take a look at the structure of the PowerShell script created for publishing.

When you create a publish profile in Visual Studio for an ASP.NET Core project a PowerShell script is created that has the following structure.

[cmdletbinding(SupportsShouldProcess=$true)]
param($publishProperties=@{}, $packOutput,$pubProfilePath, $nugetUrl)

$publishModuleVersion = '1.0.2-beta2'

# functions to bootstrap the process when Visual Studio is not installed
# have been removed to simplify this doc

try{
    if (!(Enable-PublishModule)){
        Enable-PackageDownloader
        Enable-NuGetModule -name 'publish-module' -version $publishModuleVersion -nugetUrl $nugetUrl
    }

    'Calling Publish-AspNet' | Write-Verbose
    # call Publish-AspNet to perform the publish operation
    Publish-AspNet -publishProperties $publishProperties -packOutput $packOutput -pubProfilePath $pubProfilePath
}
catch{
    "An error occurred during publish.n{0}" -f $_.Exception.Message | Write-Error
}

In the above snippet some functions have been removed for readability. Those functions are used to bootstrap the script in the case that it’s executed from a machine which doesn’t have Visual Studio installed. The script contains the following important elements:

Script parameters

Publish module version

Call to Publish-AspNet

The parameters of the script define the contract between Visual Studio and the PowerShell script. You should not change the declared parameters because Visual Studio depends on those. You can add additional parameters, but they must be added at the end.

The publish module version, denoted by $publishModuleVersion, defines the version of the web publish module that will be used. Valid version numbers can be found from published versions of the publish-module NuGet package on nuget.org. Once you create a publish profile the script definition is locked to a particular version of the publish-module package. If you need to update the version of the script you can delete the .ps1 file and then publish again in Visual Studio to get a new script created.

The call to Publish-AspNet moves the files from your local machine to the final destination. Publish-AspNet will be passed all the properties defined in the .pubxml file, even custom properties. For Web Deploy publish, msdeploy.exe will be called to publish the files to the destination. Publish-AspNet is passed the same parameters as the original script. You can get more info on the parameters for Publish-AspNet use Get-Help Publish-AspNet. If you get an error that the publish-module is not loaded, you can load it with

Import-Module “${env:ProgramFiles(x86)}\Microsoft Visual Studio 14.0\Common7\IDE\Extensions\Microsoft\Web Tools\Publish\Scripts\1.0.1\publish-module.psm1"

from a machine which has Visual Studio installed. Now let’s move on to discuss how to customize the publish process.

How to customize publishing In the previous section we saw the visualization of the publish process. The image is shown again to make this easier to follow.

The image above shows the three main extension points, you’re most likely to use is #3.

Customize the call to dotnet publish

Most developers will not need to customize this extension point. Visual Studio starts the publish process by calling an MSBuild target. This target will take care of initializing the environment and calling dotnet publish to layout the files. If you need to customize that call in a way that is not enabled by the publish dialog then you can use MSBuild elements in either the project file (.xproj file) or the publish profile (.pubxml file). We won’t get into details of how to do that here as it’s an advanced scenario that few will need to extend.

Customize dotnet publish

As stated previously dotnet publish is a command line utility that can be used to help publish your ASP.NET Core application. This is a cross platform command line utility (that is, you can use it on Windows, Mac or Linux) and does not require Visual Studio. If you are working on a team in which some developers are not using Visual Studio, then you may want to script building and publishing. When dotnet publish is executed it can be configured to execute custom commands before or after execution. The commands will be listed in project.json in the scripts section.

The supported scripts for publish are prepublish and postpublish. The ASP.NET Core Web Application template uses the prepublish step by default. The relevant snippet from project.json is shown below.

"scripts": {
  "prepublish": [ "npm install", "bower install", "gulp clean", "gulp min" ]
}

Here multiple comma separated calls are declared.

When Visual Studio is used the prepublish and postpublish steps are executed as a part of the call to dotnet publish. The postpublish script from project.json is executed before the files are published to the remote destination because that takes place immediately after dotnet publish completes. In the next step we cover customizing the PowerShell script to control what happens to the files after they reach the target destination.

Customize the publish profile PowerShell Script

After creating a publish profile in Visual Studio the PowerShell script Properties\PublishProfiles\ProfileName.ps1 is created. The script does the following:

Runs dotnet publish, which will package the web project into a temporary folder to prepare it for the next phase of publishing.

The profile PowerShell script is directly invoked. The publish properties and the path to the temporary folder are passed in as parameters. Note, the temporary folder will be deleted on each publish.

As mentioned previously the most important line in the default publish script is the call to Publish-AspNet. The call to Publish-AspNet:

Takes the contents of the folder at $packOutput, which contains the results of dotnet publish, and publishes it to the destination.

The publish properties are passed in the script parameter $publishProperties.

$publishProperties is a PowerShell hashtable which contains all the properties declared in the profile .pubxml file. It also includes values for file text replacements or files to exclude. For more info on the values for $publishProperties use Get-Help publish-aspnet –Examples.

To customize this process, you can edit the PowerShell script directly. To perform an action before publish starts, add the action before the call to Publish-AspNet. To have an action performed after publish, add the appropriate calls after Publish-AspNet. When Publish-AspNet is called the contents of the $packOutput directory are published to the destination. For example, if you need add a file to the publish process, just copy it to the correct location in $packOutput before Publish-AspNet is called. The snippet below shows how to do that.

# copy files from image repo to the wwwroot\external-images folder
$externalImagesSourcePath = 'C:\resources\external-images'
$externalImagesDestPath = (Join-Path "$packOutput\wwwroot" 'external-images')
if(-not (Test-Path $externalImagesDestPath)){
  -Item -Path $externalImagesDestPath -ItemType Directory
}

Get-ChildItem $externalImagesSourcePath -File | Copy-Item -Destination $externalImagesDestPath

'Calling Publish-AspNet' | Write-Verbose
# call Publish-AspNet to perform the publish operation
Publish-AspNet -publishProperties $publishProperties -packOutput $packOutput -pubProfilePath $pubProfilePath

In this snippet external images are copied from c:\resources\external-images to $packOutput\wwwroot\external-images. Before starting the copy operation the script ensures that the destination folder exists. Since the copy operation takes place before the call to Publish-AspNet the new files will be included in the published content. To perform actions after the files have reached the destination then you can place those commands after the call to Publish-AspNet.

You are free to customize, or even completely replace, the Publish-AspNet script provided. As previously mentioned, you will need to preserve the parameter declaration, but the rest is up to you.

Publishing to an Azure Web App with Continuous Deployment¶

By Erik Reitan

This tutorial shows you how to create an ASP.NET Core web app using Visual Studio and deploy it from Visual Studio to Azure App Service using continuous deployment.

Note

To complete this tutorial, you need a Microsoft Azure account. If you don’t have an account, you can activate your MSDN subscriber benefits or sign up for a free trial.

Sections:

Prerequisites
Create an ASP.NET Core web app
Create a web app in the Azure Portal
Enable Git publishing for the new web app
Publish your web app to Azure App Service
Run the app in Azure
Update your web app and republish
View the updated web app in Azure
Additional Resources

Prerequisites ¶

This tutorial assumes you have already installed the following:

Visual Studio
ASP.NET Core (runtime and tooling)
Git for Windows

Create an ASP.NET Core web app ¶

Start Visual Studio.
From the File menu, select New > Project.
Select the ASP.NET Web Application project template. It appears under Installed > Templates > Visual C# > Web. Name the project SampleWebAppDemo. Select the Add to source control option and click OK.

In the New ASP.NET Project dialog, select the ASP.NET Core Empty template, then click OK.

From the Choose Source Control dialog box, select Git as the source control system for the new project.

Running the web app locally¶

Once Visual Studio finishes creating the app, run the app by selecting Debug -> Start Debugging. As an alternative, you can press F5.

It may take time to initialize Visual Studio and the new app. Once it is complete, the browser will show the running app.

After reviewing the running Web app, close the browser and click the “Stop Debugging” icon in the toolbar of Visual Studio to stop the app.

Create a web app in the Azure Portal ¶

The following steps will guide you through creating a web app in the Azure Portal.

Log in to the Azure Portal.
Click NEW at the top left of the Portal.
Click Web + Mobile > Web App.

In the Web App blade, enter a unique value for the App Service Name.

Note

The App Service Name name needs to be unique. The portal will enforce this rule when you attempt to enter the name. After you enter a different value, you’ll need to substitute that value for each occurrence of SampleWebAppDemo that you see in this tutorial.

Also in the Web App blade, select an existing App Service Plan/Location or create a new one. If you create a new plan, select the pricing tier, location, and other options. For more information on App Service plans, Azure App Service plans in-depth overview.

Click Create. Azure will provision and start your web app.

Enable Git publishing for the new web app ¶

Git is a distributed version control system that you can use to deploy your Azure App Service web app. You’ll store the code you write for your web app in a local Git repository, and you’ll deploy your code to Azure by pushing to a remote repository.

Log into the Azure Portal, if you’re not already logged in.
Click Browse, located at the bottom of the navigation pane.
Click Web Apps to view a list of the web apps associated with your Azure subscription.
Select the web app you created in the previous section of this tutorial.
If the Settings blade is not shown, select Settings in the Web App blade.
In the Settings blade, select Deployment source > Choose Source > Local Git Repository.

Click OK.
If you have not previously set up deployment credentials for publishing a web app or other App Service app, set them up now:

Click Settings > Deployment credentials. The Set deployment credentials blade will be displayed.

Create a user name and password. You’ll need this password later when setting up Git.

Click Save.

In the Web App blade, click Settings > Properties. The URL of the remote Git repository that you’ll deploy to is shown under GIT URL.
Copy the GIT URL value for later use in the tutorial.

Publish your web app to Azure App Service ¶

In this section, you will create a local Git repository using Visual Studio and push from that repository to Azure to deploy your web app. The steps involved include the following:

Add the remote repository setting using your GIT URL value, so you can deploy your local repository to Azure.

Commit your project changes.

Push your project changes from your local repository to your remote repository on Azure.

In Solution Explorer right-click Solution ‘SampleWebAppDemo’ and select Commit. The Team Explorer will be displayed.

In Team Explorer, select the Home (home icon) > Settings > Repository Settings.
In the Remotes section of the Repository Settings select Add. The Add Remote dialog box will be displayed.
Set the Name of the remote to Azure-SampleApp.
Set the value for Fetch to the Git URL that you copied from Azure earlier in this tutorial. Note that this is the URL that ends with .git.

Note

As an alternative, you can specify the remote repository from the Command Window by opening the Command Window, changing to your project directory, and entering the command. For example:

git remote add Azure-SampleApp https://me@sampleapp.scm.azurewebsites.net:443/SampleApp.git

Select the Home (home icon) > Settings > Global Settings. Make sure you have your name and your email address set. You may also need to select Update.
Select Home > Changes to return to the Changes view.
Enter a commit message, such as Initial Push #1 and click Commit. This action will create a commit locally. Next, you need to sync with Azure.

Note

As an alternative, you can commit your changes from the Command Window by opening the Command Window, changing to your project directory, and entering the git commands. For example:

git add .

git commit -am "Initial Push #1"

Select Home > Sync > Actions > Open Command Prompt. The command prompt will open to your project directory.
Enter the following command in the command window:

git push -u Azure-SampleApp master

Enter your Azure deployment credentials password that you created earlier in Azure.

Note

Your password will not be visible as you enter it.

This command will start the process of pushing your local project files to Azure. The output from the above command ends with a message that deployment was successful.

remote: Finished successfully.
remote: Running post deployment command(s)...
remote: Deployment successful.
To https://username@samplewebappdemo01.scm.azurewebsites.net:443/SampleWebAppDemo01.git
* [new branch]      master -> master
Branch master set up to track remote branch master from Azure-SampleApp.

Note

If you need to collaborate on a project, you should consider pushing to GitHub in between pushing to Azure.

Verify the Active Deployment¶

You can verify that you successfully transferred the web app from your local environment to Azure. You’ll see the listed successful deployment.

In the Azure Portal, select your web app. Then, select Settings > Continuous deployment.

Run the app in Azure ¶

Now that you have deployed your web app to Azure, you can run the app.

This can be done in two ways:

In the Azure Portal, locate the web app blade for your web app, and click Browse to view your app in your default browser.
Open a browser and enter the URL for your web app. For example:

http://SampleWebAppDemo.azurewebsites.net

Update your web app and republish ¶

After you make changes to your local code, you can republish.

In Solution Explorer of Visual Studio, open the Startup.cs file.
In the Configure method, modify the Response.WriteAsync method so that it appears as follows:

await context.Response.WriteAsync("Hello World! Deploy to Azure.");

Save changes to Startup.cs.
In Solution Explorer, right-click Solution ‘SampleWebAppDemo’ and select Commit. The Team Explorer will be displayed.
Enter a commit message, such as:

Update #2

Press the Commit button to commit the project changes.
Select Home > Sync > Actions > Push.

Note

As an alternative, you can push your changes from the Command Window by opening the Command Window, changing to your project directory, and entering a git command. For example:

git push -u Azure-SampleApp master

View the updated web app in Azure ¶

View your updated web app by selecting Browse from the web app blade in the Azure Portal or by opening a browser and entering the URL for your web app. For example:

http://SampleWebAppDemo.azurewebsites.net

Additional Resources ¶

🔧 Publishing to a Windows Virtual Machine on Azure¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Publish to a Linux Production Environment¶

In this guide, we will cover setting up a production-ready ASP.NET environment on an Ubuntu 14.04 Server.

We will take an existing ASP.NET Core application and place it behind a reverse-proxy server. We will then setup the reverse-proxy server to forward requests to our Kestrel web server.

Additionally we will ensure our web application runs on startup as a daemon and configure a process management tool to help restart our web application in the event of a crash to guarantee high availability.

Sections:

Prerequisites
Copy over your app
Configure a reverse proxy server
Monitoring our Web Application
Start our web application on startup
Viewing logs
Securing our application

Prerequisites ¶

Access to an Ubuntu 14.04 Server with a standard user account with sudo privilege.
An existing ASP.NET Core application.

Copy over your app ¶

Run dotnet publish from your dev environment to package your application into a self-contained directory that can run on your server.

Before we proceed, copy your ASP.NET Core application to your server using whatever tool (SCP, FTP, etc) integrates into your workflow. Try and run the app and navigate to http://<serveraddress>:<port> in your browser to see if the application runs fine on Linux. I recommend you have a working app before proceeding.

Note

You can use Yeoman to create a new ASP.NET Core application for a new project.

Configure a reverse proxy server ¶

A reverse proxy is a common setup for serving dynamic web applications. The reverse proxy terminates the HTTP request and forwards it to the ASP.NET application.

Why use a reverse-proxy server?¶

Kestrel is great for serving dynamic content from ASP.NET, however the web serving parts aren’t as feature rich as full-featured servers like IIS, Apache or Nginx. A reverse proxy-server can allow you to offload work like serving static content, caching requests, compressing requests, and SSL termination from the HTTP server. The reverse proxy server may reside on a dedicated machine or may be deployed alongside an HTTP server.

For the purposes of this guide, we are going to use a single instance of Nginx that runs on the same server alongside your HTTP server. However, based on your requirements you may choose a different setup.

Install Nginx¶

sudo apt-get install nginx

Note

If you plan to install optional Nginx modules you may be required to build Nginx from source.

We are going to apt-get to install Nginx. The installer also creates a System V init script that runs Nginx as daemon on system startup. Since we just installed Nginx for the first time, we can explicitly start it by running

sudo service nginx start

At this point you should be able to navigate to your browser and see the default landing page for Nginx.

Configure Nginx¶

We will now configure Nginx as a reverse proxy to forward requests to our ASP.NET application

We will be modifying the /etc/nginx/sites-available/default, so open it up in your favorite text editor and replace the contents with the following.

server {
    listen 80;
    location / {
        proxy_pass http://localhost:5000;
        proxy_http_version 1.1;
        proxy_set_header Upgrade $http_upgrade;
        proxy_set_header Connection keep-alive;
        proxy_set_header Host $host;
        proxy_cache_bypass $http_upgrade;
    }
}

This is one of the simplest configuration files for Nginx that forwards incoming public traffic on your port 80 to a port 5000 that your web application will listen on.

Once you have completed making changes to your nginx configuration you can run sudo nginx -t to verify the syntax of your configuration files. If the configuration file test is successful you can ask nginx to pick up the changes by running sudo nginx -s reload.

Monitoring our Web Application ¶

Nginx will forward requests to your Kestrel server, however unlike IIS on Windows, it does not mangage your Kestrel process. In this tutorial, we will use supervisor to start our application on system boot and restart our process in the event of a failure.

Installing supervisor¶

sudo apt-get install supervisor

Note

supervisor is a python based tool and you can acquire it through pip or easy_install instead.

Configuring supervisor¶

Supervisor works by creating child processes based on data in its configuration file. When a child process dies, supervisor is notified via the SIGCHILD signal and supervisor can react accordingly and restart your web application.

To have supervisor monitor our application, we will add a file to the /etc/supervisor/conf.d/ directory.

/etc/supervisor/conf.d/hellomvc.conf¶

[program:hellomvc]
command=/usr/bin/dotnet /var/aspnetcore/HelloMVC/HelloMVC.dll
directory=/var/aspnetcore/HelloMVC/
autostart=true
autorestart=true
stderr_logfile=/var/log/hellomvc.err.log
stdout_logfile=/var/log/hellomvc.out.log
environment=ASPNETCORE__ENVIRONMENT=Production
user=www-data
stopsignal=INT

Once you are done editing the configuration file, restart the supervisord process to change the set of programs controlled by supervisord.

sudo service supervisor stop
sudo service supervisor start

Start our web application on startup ¶

In our case, since we are using supervisor to manage our application, the application will be automatically started by supervisor. Supervisor uses a System V Init script to run as a daemon on system boot and will susbsequently launch your application. If you chose not to use supervisor or an equivalent tool, you will need to write a systemd or upstart or SysVinit script to start your application on startup.

Viewing logs ¶

Supervisord logs messages about its own health and its subprocess’ state changes to the activity log. The path to the activity log is configured via the logfile parameter in the configuration file.

sudo tail -f /var/log/supervisor/supervisord.log

You can redirect application logs (STDOUT and STERR) in the program section of your configuration file.

tail -f /var/log/hellomvc.out.log

Securing our application ¶

Enable `AppArmor`¶

Linux Security Modules (LSM) is a framework that is part of the Linux kernel since Linux 2.6 that supports different implementations of security modules. AppArmor is a LSM that implements a Mandatory Access Control system which allows you to confine the program to a limited set of resources. Ensure AppArmor is enabled and properly configured.

Configuring our firewall¶

Close off all external ports that are not in use. Uncomplicated firewall (ufw) provides a frontend for iptables by providing a command-line interface for configuring the firewall. Verify that ufw is configured to allow traffic on any ports you need.

sudo apt-get install ufw
sudo ufw enable

sudo ufw allow 80/tcp
sudo ufw allow 443/tcp

Securing Nginx¶

The default distribution of Nginx doesn’t enable SSL. To enable all the security features we require, we will build from source.

Download the source and install the build dependencies¶

# Install the build dependencies
sudo apt-get update
sudo apt-get install build-essential zlib1g-dev libpcre3-dev libssl-dev libxslt1-dev libxml2-dev libgd2-xpm-dev libgeoip-dev libgoogle-perftools-dev libperl-dev

# Download nginx 1.10.0 or latest
wget http://www.nginx.org/download/nginx-1.10.0.tar.gz
tar zxf nginx-1.10.0.tar.gz

Change the Nginx response name¶

Edit src/http/ngx_http_header_filter_module.c

static char ngx_http_server_string[] = "Server: Your Web Server" CRLF;
static char ngx_http_server_full_string[] = "Server: Your Web Server" CRLF;

Configure the options and build¶

The PCRE library is required for regular expressions. Regular expressions are used in the location directive for the ngx_http_rewrite_module. The http_ssl_module adds HTTPS protocol support.

Consider using a web application firewall like ModSecurity to harden your application.

./configure
--with-pcre=../pcre-8.38
--with-zlib=../zlib-1.2.8
--with-http_ssl_module
--with-stream
--with-mail=dynamic

Configure SSL¶

Configure your server to listen to HTTPS traffic on port 443 by specifying a valid certificate issued by a trusted Certificate Authority (CA).
Harden your security by employing some of the practices suggested below like choosing a stronger cipher and redirecting all traffic over HTTP to HTTPS.
Adding an HTTP Strict-Transport-Security (HSTS) header ensures all subsequent requests made by the client are over HTTPS only.
Do not add the Strict-Transport-Security header or chose an appropriate max-age if you plan to disable SSL in the future.

http {
    include    /etc/nginx/proxy.conf;
    limit_req_zone $binary_remote_addr zone=one:10m rate=5r/s;
    server_tokens off;

    sendfile on;
    keepalive_timeout 29; # Adjust to the lowest possible value that makes sense for your use case.
    client_body_timeout 10; client_header_timeout 10; send_timeout 10;

    upstream hellomvc{
        server localhost:5000;
    }

    server {
        listen *:80;
        add_header Strict-Transport-Security max-age=15768000;
        return 301 https://$host$request_uri;
    }

    server {
        listen *:443    ssl;
        server_name     example.com;
        ssl_certificate /etc/ssl/certs/testCert.crt;
        ssl_certificate_key /etc/ssl/certs/testCert.key;
        ssl_protocols TLSv1.1 TLSv1.2;
        ssl_prefer_server_ciphers on;
        ssl_ciphers "EECDH+AESGCM:EDH+AESGCM:AES256+EECDH:AES256+EDH";
        ssl_ecdh_curve secp384r1;
        ssl_session_cache shared:SSL:10m;
        ssl_session_tickets off;
        ssl_stapling on; #ensure your cert is capable
        ssl_stapling_verify on; #ensure your cert is capable

        add_header Strict-Transport-Security max-age=63072000; includeSubdomains; preload";
        add_header X-Frame-Options DENY;
        add_header X-Content-Type-Options nosniff;

        #Redirects all traffic
        location / {
            proxy_pass  http://hellomvc;
            limit_req   zone=one burst=10;
        }
    }
}

proxy_redirect 			off;
proxy_set_header 		Host 			$host;
proxy_set_header		X-Real-IP 		$remote_addr;
proxy_set_header		X-Forwarded-For	$proxy_add_x_forwarded_for;
client_max_body_size 	10m;
client_body_buffer_size 128k;
proxy_connect_timeout 	90;
proxy_send_timeout 		90;
proxy_read_timeout 		90;
proxy_buffers			32 4k;

Use VSTS to Build and Publish to an Azure Web App with Continuous Deployment¶

By Damien Pontifex

This tutorial shows you how to create an ASP.NET Core web app using Visual Studio and deploy it from Visual Studio to Azure App Service using continuous deployment.

Note

To complete this tutorial, you need a Microsoft Azure account. If you don’t have an account, you can activate your MSDN subscriber benefits or sign up for a free trial. You will also need a Visual Studio Team Services account. If you don’t have an account, you can sign up for free.

Sections:

Prerequisites
Setup VSTS Build
Use VSTS Release
Additional Resources

Prerequisites ¶

This tutorial assumes you already have the following:

ASP.NET Core (runtime and tooling). Hosted Build Pool servers in VSTS already have RC2 tooling installed.
Git
The Trackyon Advantage extension installed into your team services account. This adds an available zip task for later steps.

Setup VSTS Build ¶

Setup some build variables to make later steps clearer and easier to retain consistent paths across build steps.

Create a variable for PublishOutput and set it to your desired path. We have used $(Build.StagingDirectory)/WebApplication

Create a variable for DeployPackage and set it to the path you would like the zipped web package to be at. We have used $(Build.StagingDirectory)/WebApplication.zip to have it alongside our published output.

Use a Command Line build step to restore packages.

Click Add build step... and choose Utility > Command Line > Add

Set the arguments for the build step as:

Tool: dotnet

Arguments: restore

Use another Command Line build step to publish the project.

Click Add build step... and choose Utility > Command Line > Add

Set the arguments for the build step as:

Tool: dotnet

Arguments: publish src/WebApplication –configuration $(BuildConfiguration) –output $(PublishOutput)

Replace src/WebApplication to the path of your app to be deployed as appropriate

Compress the published output so it can be deployed to Azure App Service. We will use the Trackyon Advantage task we installed to zip the contents of our published output for deployment.

Click Add build step... and choose Utility > Trackyon Zip > Add

Set the arguments for the zip build step as:

Folder to Zip: $(PublishOutput)

Path to final Zip file: $(DeployPackage)

Use the Azure Web App Deployment build step to publish the compressed publish output to your Azure Web App. The Web Deploy Package will be the output of the contents compressed in step 4. In this case, we re-use the variable for it’s path we setup earlier.

Click Add build step... and choose Deploy > Azure Web App Deployment > Add

Set the arguments for the deployment step as:

Azure Subscription: <your configured azure connection>

Web App Location: <desired region>

Web App Name: <desired app service name>

Web Deploy Package: $(DeployPackage)

Use VSTS Release ¶

VSTS Release management can alternatively be used to manage the release pipeline from the VSTS build. We require a small change to the build pipeline and setup of the release process.

If configured, remove the Azure Web App Deployment step from the VSTS build setup in the previous section.
Add a Copy and Publish Build Artifacts step to the build pipeline

Click Add build step... and choose Utility > Copy and Publish Build Artifacts > Add

Set the arguments for the copy and publish step as:

Contents: $(DeployPackage)

Arifact Name: DeployPackage

Artifact Type: Server

You will be able to create a release definition and link to the Build definition and utilise the artifacts copied from step 2 here for publishing.

Additional Resources ¶

Guidance for Hosting Providers¶

ASP.NET Core Module Configuration Reference¶

By Luke Latham, Rick Anderson and Sourabh Shirhatti

In ASP.NET Core, the web application is hosted by an external process outside of IIS. The ASP.NET Core Module is an IIS 7.5+ module, which is responsible for process management of ASP.NET Core http listeners and to proxy requests to processes that it manages. This document provides an overview of how to configure the ASP.NET Core Module for shared hosting of ASP.NET Core.

Sections:

Installing the ASP.NET Core Module
Configuring the ASP.NET Core Module
- Configuration Attributes
- Child Elements
ASP.NET Core Module app_offline.htm
ASP.NET Core Module Start-up Error Page
ASP.NET Core Module configuration examples
- Log creation and redirection
- Setting environment variables

Installing the ASP.NET Core Module ¶

Install the .NET Core Windows Server Hosting bundle on the server. The bundle will install the .NET Core Runtime, .NET Core Library, and the ASP.NET Core Module.

Configuring the ASP.NET Core Module ¶

The ASP.NET Core Module is configured via a site or application web.config file and has its own configuration section within system.webServer - aspNetCore.

Configuration Attributes ¶

Attribute	Description
processPath	Required string attribute. Path to the executable or script that will launch a process listening for HTTP requests. Relative paths are supported. If the path begins with ‘.’, the path is considered to be relative to the site root. There is no default value.
arguments	Optional string attribute. Arguments to the executable or script specified in processPath. The default value is an empty string.
startupTimeLimit	Optional integer attribute. Duration in seconds for which the the handler will wait for the executable or script to start a process listening on the port. If this time limit is exceeded, the handler will kill the process and attempt to launch it again startupRetryCount times. The default value is 120.
shutdownTimeLimit	Optional integer attribute. Duration in seconds for which the the handler will wait for the executable or script to gracefully shutdown when the app_offline.htm file is detected The default value is 10.
rapidFailsPerMinute	Optional integer attribute. Specifies the number of times the process specified in processPath is allowed to crash per minute. If this limit is exceeded, the handler will stop launching the process for the remainder of the minute. The default value is 10.
requestTimeout	Optional timespan attribute. Specifies the duration for which the ASP.NET Core Module will wait for a response from the process listening on %ASPNETCORE_PORT%. The default value is “00:02:00”.
stdoutLogEnabled	Optional Boolean attribute. If true, stdout and stderr for the process specified in processPath will be redirected to the file specified in stdoutLogFile. The default value is false.
stdoutLogFile	Optional string attribute. Specifies the relative or absolute file path for which stdout and stderr from the process specified in processPath will be logged. Relative paths are relative to the root of the site. Any path starting with ‘.’ will be relative to the site root and all other paths will be treated as absolute paths. The default value is `aspnetcore-stdout`.
forwardWindowsAuthToken	True or False. If true, the token will be forwarded to the child process listening on %ASPNETCORE_PORT% as a header ‘MS-ASPNETCORE-WINAUTHTOKEN’ per request. It is the responsibility of that process to call CloseHandle on this token per request. The default value is false.
disableStartUpErrorPage	True or False. If true, the 502.5 - Process Failure page will be supressed and the 502 status code page configured in your web.config will take precedence. The default value is false.

Child Elements ¶

Attribute	Description
environmentVariables	Configures environmentVariables collection for the process specified in processPath.
recycleOnFileChange	Specify a list of files to monitor. If any of these files are updated/deleted, the Core Module will restart the backend process.

ASP.NET Core Module app_offline.htm¶

If you place a file with the name app_offline.htm at the root of a web application directory, the ASP.NET Core Module will attempt to gracefully shut-down the application and stop processing any new incoming requests. If the application is still running after shutdownTimeLimit number of seconds, the ASP.NET Core Module will kill the running process.

While the app_offline..htm file is present, the ASP.NET Core Module will repond to all requests by sending back the contents of the app_offline.htm file. Once the app_offline.htm file is removed, the next request loads the application, which then responds to requests.

ASP.NET Core Module Start-up Error Page ¶

If the ASP.NET Core Module fails to launch the backend process or the backend process starts but fails to listen on the configured port, you will see an HTTP 502.5 status code page. To supress this page and revert to the default IIS 502 status code page, use the disableStartUpErrorPage attribute. Look at the HTTP Errors attribute to override this page with a custom error page.

ASP.NET Core Module configuration examples ¶

Log creation and redirection ¶

To save logs, you must create the logs directory. The ASP.NET Core Module can redirect stdout and stderr logs to disk by setting the stdoutLogEnabled and stdoutLogFile attributes of the aspNetCore element. Logs are not rotated (unless process recycling/restart occurs). It is the responsibilty of the hoster to limit the disk space the logs consume.

    <aspNetCore processPath="dotnet"
            arguments=".\MyApp.dll"
            stdoutLogEnabled="true"
            stdoutLogFile=".\logs\stdout">
    </aspNetCore>

Setting environment variables ¶

The ASP.NET Core Module allows you specify environment variables for the process specified in the processPath setting by specifying them in environmentVariables child attribute to the aspNetCore attribute.

    <aspNetCore processPath="dotnet"
            arguments=".\MyApp.dll"
            stdoutLogEnabled="true"
            stdoutLogFile=".\logs\stdout">
      <environmentVariables>
        <environmentVariable name="DEMO" value="demo_value" />
      </environmentVariables>
    </aspNetCore>

Directory Structure¶

By Luke Latham

In ASP.NET Core, the application directory, publish, is comprised of application files, config files, static assets, packages, and the runtime (for self-contained apps). This is the same directory structure as previous versions of ASP.NET, where the entire application lives inside the web root directory.

App Type	Directory Structure
Portable	publish* logs* (if included in publishOptions) refs* runtimes* Views* (if included in publishOptions) wwwroot* (if included in publishOptions) .dll files myapp.deps.json myapp.dll myapp.pdb myapp.runtimeconfig.json web.config (if included in publishOptions)
Self-contained	publish* logs* (if included in publishOptions) refs* Views* (if included in publishOptions) wwwroot* (if included in publishOptions) .dll files myapp.deps.json myapp.exe myapp.pdb myapp.runtimeconfig.json web.config (if included in publishOptions)

* Indicates a directory

The contents of the publish directory represent the content root path, also called the application base path, of the deployment. Whatever name is given to the publish directory in the deployment, its location serves as the server’s physical path to the hosted application. The wwwroot directory, if present, only contains static assets. The logs directory may be included in the deployment by creating it in the project and adding it to publishOptions of project.json or by physically creating the directory on the server.

The deployment directory requires Read/Execute permissions, while the logs directory requires Read/Write permissions. Additional directories where assets will be written require Read/Write permissions.

Application Pools¶

When hosting multiple web sites on a single server, you should consider isolating the applications from each other by running each application in its own application pool. This document provides an overview of how to set up Application Pools to securely host multiple web sites on a single server.

Application Pool Identity Account¶

An application pool identity account allows you to run an application under a unique account without having to create and manage domains or local accounts. On IIS 8.0+ the IIS Admin Worker Process (WAS) will create a virtual account with the name of the new application pool and run the application pool’s worker processes under this account by default.

Configuring IIS Application Pool Identities¶

In the IIS Management Console, under Advanced Settings for your application pool ensure that Identity list item is set to use ApplicationPoolIdentity as shown in the image below.

Securing Resources¶

The IIS management process creates a secure identifier with the name of the application pool in the Windows Security System. Resources can be secured by using this identity, however this identity is not a real user account and will not show up in the Windows User Management Console.

To grant the IIS worker process access to your application, you will need to modify the Access Control List (ACL) for the the directory containing your application.

Open Windows Explorer and navigate to the directory.
Right click on the directory and click properties.
Under the Security tab, click the Edit button and then the Add button
Click the Locations and make sure you select your server.

Enter IIS AppPool\DefaultAppPool in Enter the object names to select textbox.
Click the Check Names button and then click OK.

You can also do this via the command-line by using ICACLS tool.

ICACLS C:\sites\MyWebApp /grant "IIS AppPool\DefaultAppPool" :F

Servicing¶

By Sourabh Shirhatti, Daniel Roth

.NET Core supports servicing of runtime components and packages to patch any vulnerabilities when they are discovered. For information on how to enable servicing for applications in a hosted environment please refer to the .NET Core Servicing documentation.

Data Protection¶

The ASP.NET Core data protection stack provides a simple and easy to use cryptographic API a developer can use to protect data, including key management and rotation. This document provides an overview of how to configure Data Protection on your server to enable developers to use data protection.

Configuring Data Protection¶

Warning

Data Protection is used by various ASP.NET middleware, including those used in authentication. The default behavior on IIS hosted web sites is to store keys in memory and discard them when the process restarts. This behavior will have side effects, for example, discarding keys invalidate any cookies written by the cookie authentication and users will have to login again.

To automatically persist keys for an application hosted in IIS, you must create registry hives for each application pool. Use the Provisioning PowerShell script for each application pool you will be hosting ASP.NET Core applications under. This script will create a special registry key in the HKLM registry that is ACLed only to the worker process account. Keys are encrypted at rest using DPAPI.

Note

A developer can configure their applications Data Protection APIs to store data on the file system. Data Protection can be configured by the developer to use a UNC share to store keys, to enable load balancing. A hoster should ensure that the file permissions for such a share are limited to the Windows account the application runs as. In addition a developer may choose to protect keys at rest using an X509 certificate. A hoster may wish to consider a mechanism to allow users to upload certificates, place them into the user’s trusted certificate store and ensure they are available on all machines the users application will run on.

Machine Wide Policy¶

The data protection system has limited support for setting default machine-wide policy for all applications that consume the data protection APIs. See the data protection documentation for more details.

Security¶

Authentication¶

Introduction to Identity¶

By Pranav Rastogi, Rick Anderson, Tom Dykstra, Jon Galloway and Erik Reitan

ASP.NET Core Identity is a membership system which allows you to add login functionality to your application. Users can create an account and login with a user name and password or they can use an external login providers such as Facebook, Google, Microsoft Account, Twitter and more.

You can configure ASP.NET Core Identity to use a SQL Server database to store user names, passwords, and profile data. Alternatively, you can use your own persistent store to store data in another other persistent storage, such as Azure Table Storage.

Overview of Identity¶

In this topic, you’ll learn how to use ASP.NET Core Identity to add functionality to register, log in, and log out a user. You can follow along step by step or just read the details. For more detailed instructions about creating apps using ASP.NET Core Identity, see the Next Steps section at the end of this article.

Create an ASP.NET Core Web Application project in Visual Studio with Individual User Accounts.

In Visual Studio, select File -> New -> Project. Then, select the ASP.NET Web Application from the New Project dialog box. Continue by selecting an ASP.NET Core Web Application with Individual User Accounts as the authentication method.

The created project contains the Microsoft.AspNetCore.Identity.EntityFrameworkCore package, which will persist the identity data and schema to SQL Server using Entity Framework Core.

Note

In Visual Studio, you can view NuGet packages details by selecting Tools -> NuGet Package Manager -> Manage NuGet Packages for Solution. You also see a list of packages in the dependencies section of the project.json file within your project.

The identity services are added to the application in the ConfigureServices method in the Startup class:
// This method gets called by the runtime. Use this method to add services to the container.
public void ConfigureServices(IServiceCollection services)
{
    // Add framework services.
    services.AddEntityFramework()
        .AddSqlServer()
        .AddDbContext<ApplicationDbContext>(options =>
            options.UseSqlServer(Configuration["Data:DefaultConnection:ConnectionString"]));

    services.AddIdentity<ApplicationUser, IdentityRole>()
        .AddEntityFrameworkStores<ApplicationDbContext>()
        .AddDefaultTokenProviders();

    services.AddMvc();

    // Add application services.
    services.AddTransient<IEmailSender, AuthMessageSender>();
    services.AddTransient<ISmsSender, AuthMessageSender>();
}
These services are then made available to the application through dependency injection.

Identity is enabled for the application by calling UseIdentity in the Configure method of the Startup class. This adds cookie-based authentication to the request pipeline.
// This method gets called by the runtime. Use this method to configure the HTTP request pipeline.
public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    loggerFactory.AddConsole(Configuration.GetSection("Logging"));
    loggerFactory.AddDebug();

    if (env.IsDevelopment())
    {
        app.UseBrowserLink();
        app.UseDeveloperExceptionPage();
        app.UseDatabaseErrorPage();
    }
    else
    {
        app.UseExceptionHandler("/Home/Error");
    }

    app.UseIISPlatformHandler(options => options.AuthenticationDescriptions.Clear());

    app.UseStaticFiles();

    app.UseIdentity();

    // To configure external authentication please see http://go.microsoft.com/fwlink/?LinkID=532715

    app.UseMvc(routes =>
    {
        routes.MapRoute(
            name: "default",
            template: "{controller=Home}/{action=Index}/{id?}");
    });
}
For more information about the application start up process, see Application Startup.

Creating a user.

Launch the application from Visual Studio (Debug -> Start Debugging) and then click on the Register link in the browser to create a user. The following image shows the Register page which collects the user name and password.

When the user clicks the Register link, the UserManager and SignInManager services are injected into the Controller:

    public class AccountController : Controller
    {
        private readonly UserManager<ApplicationUser> _userManager;
        private readonly SignInManager<ApplicationUser> _signInManager;
        private readonly IEmailSender _emailSender;
        private readonly ISmsSender _smsSender;
        private static bool _databaseChecked;
        private readonly ILogger _logger;

        public AccountController(
            UserManager<ApplicationUser> userManager,
            SignInManager<ApplicationUser> signInManager,
            IEmailSender emailSender,
            ISmsSender smsSender,
            ILoggerFactory loggerFactory)
        {
            _userManager = userManager;
            _signInManager = signInManager;
            _emailSender = emailSender;
            _smsSender = smsSender;
            _logger = loggerFactory.CreateLogger<AccountController>();
        }

        //
        // GET: /Account/Login

Then, the Register action creates the user by calling CreateAsync function of the UserManager object, as shown below:

[HttpPost]
[AllowAnonymous]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Register(RegisterViewModel model)
{
    if (ModelState.IsValid)
    {
        var user = new ApplicationUser { UserName = model.Email, Email = model.Email };
        var result = await _userManager.CreateAsync(user, model.Password);
        if (result.Succeeded)
        {
            // For more information on how to enable account confirmation and password reset please visit http://go.microsoft.com/fwlink/?LinkID=532713
            // Send an email with this link
            //var code = await _userManager.GenerateEmailConfirmationTokenAsync(user);
            //var callbackUrl = Url.Action("ConfirmEmail", "Account", new { userId = user.Id, code = code }, protocol: HttpContext.Request.Scheme);
            //await _emailSender.SendEmailAsync(model.Email, "Confirm your account",
            //    "Please confirm your account by clicking this link: <a href=\"" + callbackUrl + "\">link</a>");
            await _signInManager.SignInAsync(user, isPersistent: false);
            _logger.LogInformation(3, "User created a new account with password.");
            return RedirectToAction(nameof(HomeController.Index), "Home");
        }
        AddErrors(result);
    }

    // If we got this far, something failed, redisplay form
    return View(model);
}

Log in.

If the user was successfully created, the user is logged in by the SignInAsync method, also contained in the Register action. By signing in, the SignInAsync method stores a cookie with the user’s claims.

[HttpPost]
[AllowAnonymous]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Register(RegisterViewModel model)
{
    if (ModelState.IsValid)
    {
        var user = new ApplicationUser { UserName = model.Email, Email = model.Email };
        var result = await _userManager.CreateAsync(user, model.Password);
        if (result.Succeeded)
        {
            // For more information on how to enable account confirmation and password reset please visit http://go.microsoft.com/fwlink/?LinkID=532713
            // Send an email with this link
            //var code = await _userManager.GenerateEmailConfirmationTokenAsync(user);
            //var callbackUrl = Url.Action("ConfirmEmail", "Account", new { userId = user.Id, code = code }, protocol: HttpContext.Request.Scheme);
            //await _emailSender.SendEmailAsync(model.Email, "Confirm your account",
            //    "Please confirm your account by clicking this link: <a href=\"" + callbackUrl + "\">link</a>");
            await _signInManager.SignInAsync(user, isPersistent: false);
            _logger.LogInformation(3, "User created a new account with password.");
            return RedirectToAction(nameof(HomeController.Index), "Home");
        }
        AddErrors(result);
    }

    // If we got this far, something failed, redisplay form
    return View(model);
}

The above SignInAsync method calls the below SignInAsync task, which is contained in the SignInManager class.

If needed, you can access the user’s identity details inside a controller action. For instance, by setting a breakpoint inside the HomeController.Index action method, you can view the User.claims details. By having the user signed-in, you can make authorization decisions. For more information, see Authorization.

As a registered user, you can log in to the web app by clicking the Log in link. When a registered user logs in, the Login action of the AccountController is called. Then, the Login action signs in the user using the PasswordSignInAsync method contained in the Login action.

[HttpPost]
[AllowAnonymous]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Login(LoginViewModel model, string returnUrl = null)
{
    ViewData["ReturnUrl"] = returnUrl;
    if (ModelState.IsValid)
    {
        // This doesn't count login failures towards account lockout
        // To enable password failures to trigger account lockout, set lockoutOnFailure: true
        var result = await _signInManager.PasswordSignInAsync(model.Email, model.Password, model.RememberMe, lockoutOnFailure: false);
        if (result.Succeeded)
        {
            _logger.LogInformation(1, "User logged in.");
            return RedirectToLocal(returnUrl);
        }
        if (result.RequiresTwoFactor)
        {
            return RedirectToAction(nameof(SendCode), new { ReturnUrl = returnUrl, RememberMe = model.RememberMe });
        }
        if (result.IsLockedOut)
        {
            _logger.LogWarning(2, "User account locked out.");
            return View("Lockout");
        }
        else
        {
            ModelState.AddModelError(string.Empty, "Invalid login attempt.");
            return View(model);
        }
    }

    // If we got this far, something failed, redisplay form
    return View(model);
}

Log off.

Clicking the Log off link calls the LogOff action in the account controller.
[HttpPost]
[ValidateAntiForgeryToken]
public async Task<IActionResult> LogOff()
{
    await _signInManager.SignOutAsync();
    _logger.LogInformation(4, "User logged out.");
    return RedirectToAction(nameof(HomeController.Index), "Home");
}
The code above shows the SignInManager.SignOutAsync method. The SignOutAsync method clears the users claims stored in a cookie.

View the database.

After stopping the application, view the user database from Visual Studio by selecting View -> SQL Server Object Explorer. Then, expand the following within the SQL Server Object Explorer:

(localdb)MSSQLLocalDB

Databases

aspnet5-<the name of your application>

Tables

Next, right-click the dbo.AspNetUsers table and select View Data to see the properties of the user you created.

Identity Components¶

The primary reference assembly for the identity system is Microsoft.AspNetCore.Identity. This package contains the core set of interfaces for ASP.NET Core Identity.

These dependencies are needed to use the identity system in ASP.NET Core applications:

EntityFramework.SqlServer - Entity Framework is Microsoft’s recommended data access technology for relational databases.
Microsoft.AspNetCore.Authentication.Cookies - Middleware that enables an application to use cookie based authentication, similar to ASP.NET’s Forms Authentication.
Microsoft.AspNetCore.Cryptography.KeyDerivation - Utilities for key derivation.
Microsoft.AspNetCore.Hosting.Abstractions - Hosting abstractions.

Migrating to ASP.NET Core Identity¶

For additional information and guidance on migrating your existing identity store see Migrating Authentication and Identity

Next Steps¶

Enabling authentication using Facebook, Google and other external providers¶

By Rick Anderson and Pranav Rastogi

This tutorial shows you how to build an ASP.NET Core app that enables users to log in using OAuth 2.0 with credentials from an external authentication provider, such as Facebook, Twitter, LinkedIn, Microsoft, and Google. For simplicity, this tutorial focuses on working with credentials from Facebook and Google.

Enabling these credentials in your web sites provides a significant advantage because millions of users already have accounts with these external providers. These users may be more inclined to sign up for your site if they do not have to create and remember new credentials.

Sections:

Create a New ASP.NET Core Project
Creating the app in Facebook
Use SecretManager to store Facebook AppId and AppSecret
Enable Facebook middleware
Login with Facebook
Optionally set password
Next steps

Create a New ASP.NET Core Project ¶

Note

The tutorial requires the latest version of Visual Studio 2015 and ASP.NET Core.

In Visual Studio, create a New Project (from the Start Page, or via File > New > Project)

Tap Web Application and verify Authentication is set to Individual User Accounts

Enable SSL
- In solution explorer, right click the project and select Properties
- On the left pane, tap Debug
- Check Enable SSL
- Copy the SSL URL and paste it into the App URL

Require SSL. Add the following code to ConfigureServices in Startup:

services.Configure<MvcOptions>(options =>
{
    options.Filters.Add(new RequireHttpsAttribute ());
});

Test the app

Creating the app in Facebook ¶

Each of the OAuth2 providers require provider specific keys to enable OAuth2.

Navigate to https://developers.facebook.com/apps and log in.
If you aren’t already registered as a Facebook developer, click Register as a Developer and follow the directions to register.
Tap Add a New App
Select Website from the platform choices.
Tap Skip and Create App ID

Enter a display name, category, contact email and tap Create App ID.

Tap Settings from the left menu bar.

On the Basic settings section of the page select Add Platform to specify that you are adding a website app.

Select Website from the platform choices.

Add your Site URL (https://localhost:44320/)
Make a note of your App ID and your App Secret so that you can add both into your ASP.NET Core app later in this tutorial. Also, Add your Site URL (https://localhost:44300/) to test your application.

Use SecretManager to store Facebook AppId and AppSecret ¶

The project created has code in Startup which reads the configuration values from a secret store. As a best practice, it is not recommended to store the secrets in a configuration file in the application since they can be checked into source control which may be publicly accessible.

Follow these steps to add the Facebook AppId and AppSecret to the Secret Manager:

Install the Secret Manager tool.

Set the Facebook AppId:

dotnet user-secrets set Authentication:Facebook:AppId <app-Id>

Set the Facebook AppSecret:

dotnet user-secrets set Authentication:Facebook:AppSecret <app-secret>

The following code reads the configuration values stored by the Secret Manager.

public Startup(IHostingEnvironment env)
{
    var builder = new ConfigurationBuilder()
        .SetBasePath(env.ContentRootPath)
        .AddJsonFile("appsettings.json", optional: true, reloadOnChange: true)
        .AddJsonFile($"appsettings.{env.EnvironmentName}.json", optional: true);

    if (env.IsDevelopment())
    {
        // For more details on using the user secret store see http://go.microsoft.com/fwlink/?LinkID=532709
        builder.AddUserSecrets();
    }

    builder.AddEnvironmentVariables();
    Configuration = builder.Build();
}

Enable Facebook middleware ¶

Note: You will need to use NuGet to install the Microsoft.AspNetCore.Authentication.Facebook package if it hasn’t already been installed.

Add the Facebook middleware in the Configure method in Startup:

public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
{
    loggerFactory.AddConsole(Configuration.GetSection("Logging"));
    loggerFactory.AddDebug();

    if (env.IsDevelopment())
    {
        app.UseDeveloperExceptionPage();
        app.UseDatabaseErrorPage();
        app.UseBrowserLink();
    }
    else
    {
        app.UseExceptionHandler("/Home/Error");
    }

    app.UseStaticFiles();

    app.UseIdentity();

    app.UseFacebookAuthentication(new FacebookOptions()
    {
        AppId = Configuration["Authentication:Facebook:AppId"],
        AppSecret = Configuration["Authentication:Facebook:AppSecret"]
    });

    app.UseMvc(routes =>
    {
        routes.MapRoute(
            name: "default",
            template: "{controller=Home}/{action=Index}/{id?}");
    });
}

Login with Facebook ¶

Run your application and click Login. You will see an option for Facebook.

When you click on Facebook, you will be redirected to Facebook for authentication.

Once you enter your Facebook credentials, then you will be redirected back to the Web site where you can set your email.

You are now logged in using your Facebook credentials.

Optionally set password ¶

When you register with an external login provider, you do not have a password registered with the app. This alleviates you from creating and remembering a password for the site, but it also makes you dependent on the external login provider. If the external login provider is unavailable, you won’t be able to log in to the web site.

To create a password and login using your email that you set during the login process with external providers:

Tap the Hello <email alias> link at the top right corner to navigate to the Manage view.

Tap Create

Set a valid password and you can use this to login with your email

Next steps ¶

This article showed how you can authenticate with Facebook. You can follow a similar approach to authenticate with Microsoft Account, Twitter, Google and other providers.
Once you publish your Web site to Azure Web App, you should reset the AppSecret in the Facebook developer portal.
Set the Facebook AppId and AppSecret as application setting in the Azure Web App portal. The configuration system is setup to read keys from environment variables.

Account Confirmation and Password Recovery¶

This tutorial shows you how to build an ASP.NET Core app with email confirmation and password reset support.

Sections:

Create a New ASP.NET Core Project
Require SSL
Require email confirmation
Enable account confirmation and password recovery
Register, confirm email, and reset password
Require email confirmation before login
Combine social and local login accounts

Create a New ASP.NET Core Project ¶

Note

The tutorial requires Visual Studio 2015 updated 2 and ASP.NET Core RC2 or higher.

In Visual Studio, create a New Project (from the Start Page, or via File > New > Project)

Tap Web Application and verify Authentication is set to Individual User Accounts

Run the app and then click on the Register link and register a user. At this point, the only validation on the email is with the [EmailAddress] attribute. After you submit the registration, you are logged into the app. Later in the tutorial we’ll change this so new users cannot log in until their email has been validated.

In SQL Server Object Explorer (SSOX), navigate to (localdb)MSSQLLocalDB(SQL Server 12). Right click on dbo.AspNetUsers > View Data:

Note the EmailConfirmed field is False.

Right-click on the row and from the context menu, select Delete. You might want to use this email again in the next step, when the app sends a confirmation email. Deleting the email alias now will make it easier in the following steps.

Require SSL ¶

In this section we’ll set up our Visual Studio project to use SSL and our project to require SSL.

Enable SSL in Visual Studio¶

In solution explorer, right click the project and select Properties

On the left pane, tap Debug

Check Enable SSL

Copy the SSL URL and paste it into the App URL

Add the following code to ConfigureServices in Startup:

services.Configure<MvcOptions>(options =>
{
    options.Filters.Add(new RequireHttpsAttribute ());
});

Add the [RequireHttps] attribute to each controller. The [RequireHttps] attribute will redirect all HTTP GET requests to HTTPS GET and will reject all HTTP POSTs. A security best practice is to use HTTPS for all requests.

[RequireHttps]
public class HomeController : Controller

Require email confirmation ¶

It’s a best practice to confirm the email of a new user registration to verify they are not impersonating someone else (that is, they haven’t registered with someone else’s email). Suppose you had a discussion forum, you would want to prevent “bob@example.com” from registering as “joe@contoso.com”. Without email confirmation, “joe@contoso.com” could get unwanted email from your app. Suppose Bob accidentally registered as “bib@example.com” and hadn’t noticed it, he wouldn’t be able to use password recovery because the app doesn’t have his correct email. Email confirmation provides only limited protection from bots and doesn’t provide protection from determined spammers who have many working email aliases they can use to register.

You generally want to prevent new users from posting any data to your web site before they have been confirmed by email, an SMS text message, or another mechanism. In the sections below, we will enable email confirmation and modify the code to prevent newly registered users from logging in until their email has been confirmed.

Configure email provider¶

We’ll use the Options pattern to access the user account and key settings. For more information, see configuration.

Create a class to fetch the secure email key. For this sample, the AuthMessageSenderOptions class is created in the Services/AuthMessageSenderOptions.cs file.
public class AuthMessageSenderOptions
{
    public string SendGridUser { get; set; }
    public string SendGridKey { get; set; }
}

Set the SendGridUser and SendGridKey with the secret-manager tool. For example:

C:\WebApplication3\src\WebApplication3>dotnet user-secrets set SendGridUser RickAndMSFT
info: Successfully saved SendGridUser = RickAndMSFT to the secret store.

On Windows, Secret Manager stores your keys/value pairs in a secrets.json file in the %APPDATA%/Microsoft/UserSecrets/<userSecretsId> directory. The userSecretsId directory can be found in your project.json file. For this example, the first few lines of the project.json file are shown below:

{
  "webroot": "wwwroot",
  "userSecretsId": "aspnet-WebApplication3-f1645c1b-3962-4e7f-99b2-4fb292b6dade",
  "version": "1.0.0-*",

  "dependencies": {

At this time, the contents of the secrets.json file are not encrypted. The secrets.json file is shown below (the sensitive keys have been removed.)

{
  "SendGridUser": "RickAndMSFT",
  "SendGridKey": "",
  "Authentication:Facebook:AppId": "",
  "Authentication:Facebook:AppSecret": ""
}

Configure startup to use AuthMessageSenderOptions¶

Add the dependecy Microsoft.Extensions.Options.ConfigurationExtensions in the project.json file.

Add AuthMessageSenderOptions to the service container at the end of the ConfigureServices method in the Startup.cs file:

    // Add application services.
    services.AddTransient<IEmailSender, AuthMessageSender>();
    services.AddTransient<ISmsSender, AuthMessageSender>();
    services.Configure<AuthMessageSenderOptions>(Configuration);

Configure the AuthMessageSender class¶

This tutorial shows how to add email notification through SendGrid, but you can send email using SMTP and other mechanisms.

Install the SendGrid.NetCore NuGet package. From the Package Manager Console, enter the following the following command:

Install-Package SendGrid.NetCore -Pre

Note

SendGrid.NetCore package is a prerelease version , to install it is necessary to use -Pre option on Install-Package.

Follow the instructions Create a SendGrid account to register for a free SendGrid account.
Add code in Services/MessageServices.cs similar to the following to configure SendGrid

public class AuthMessageSender : IEmailSender, ISmsSender
{
    public AuthMessageSender(IOptions<AuthMessageSenderOptions> optionsAccessor)
    {
        Options = optionsAccessor.Value;
    }

    public AuthMessageSenderOptions Options { get; } //set only via Secret Manager

    public Task SendEmailAsync(string email, string subject, string message)
    {
        // Plug in your email service here to send an email.
        var myMessage = new SendGrid.SendGridMessage();
        myMessage.AddTo(email);
        myMessage.From = new System.Net.Mail.MailAddress("Joe@contoso.com", "Joe Smith");
        myMessage.Subject = subject;
        myMessage.Text = message;
        myMessage.Html = message;
        var credentials = new System.Net.NetworkCredential(
            Options.SendGridUser,
            Options.SendGridKey);
        // Create a Web transport for sending email.
        var transportWeb = new SendGrid.Web(credentials);
        // Send the email.
        if (transportWeb != null)
        {
            return transportWeb.DeliverAsync(myMessage);
        }
        else
        {
            return Task.FromResult(0);
        }
    }

    public Task SendSmsAsync(string number, string message)
    {
        // Plug in your SMS service here to send a text message.
        return Task.FromResult(0);
    }
}

Enable account confirmation and password recovery ¶

The template already has the code for account confirmation and password recovery. Follow these steps to enable it:

Find the [HttpPost] Register method in the AccountController.cs file.
Uncomment the code to enable account confirmation.

[HttpPost]
[AllowAnonymous]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Register(RegisterViewModel model, string returnUrl = null)
{
    ViewData["ReturnUrl"] = returnUrl;
    if (ModelState.IsValid)
    {
        var user = new ApplicationUser { UserName = model.Email, Email = model.Email };
        var result = await _userManager.CreateAsync(user, model.Password);
        if (result.Succeeded)
        {
            // For more information on how to enable account confirmation and password reset please visit http://go.microsoft.com/fwlink/?LinkID=532713
            // Send an email with this link
            var code = await _userManager.GenerateEmailConfirmationTokenAsync(user);
            var callbackUrl = Url.Action("ConfirmEmail", "Account", new { userId = user.Id, code = code }, protocol: HttpContext.Request.Scheme);
            await _emailSender.SendEmailAsync(model.Email, "Confirm your account",
                $"Please confirm your account by clicking this link: <a href='{callbackUrl}'>link</a>");
            //await _signInManager.SignInAsync(user, isPersistent: false);
            _logger.LogInformation(3, "User created a new account with password.");
            return RedirectToLocal(returnUrl);
            
        }
        AddErrors(result);
    }

    // If we got this far, something failed, redisplay form
    return View(model);
}

Note: We’re also preventing a newly registered user from being automatically logged on by commenting out the following line:

//await _signInManager.SignInAsync(user, isPersistent: false);

Enable password recovery by uncommenting the code in the ForgotPassword action in the Controllers/AccountController.cs file.

[HttpPost]
[AllowAnonymous]
[ValidateAntiForgeryToken]
public async Task<IActionResult> ForgotPassword(ForgotPasswordViewModel model)
{
    if (ModelState.IsValid)
    {
        var user = await _userManager.FindByNameAsync(model.Email);
        if (user == null || !(await _userManager.IsEmailConfirmedAsync(user)))
        {
            // Don't reveal that the user does not exist or is not confirmed
            return View("ForgotPasswordConfirmation");
        }

        // For more information on how to enable account confirmation and password reset please visit http://go.microsoft.com/fwlink/?LinkID=532713
        // Send an email with this link
        var code = await _userManager.GeneratePasswordResetTokenAsync(user);
        var callbackUrl = Url.Action("ResetPassword", "Account", new { userId = user.Id, code = code }, protocol: HttpContext.Request.Scheme);
        await _emailSender.SendEmailAsync(model.Email, "Reset Password",
           $"Please reset your password by clicking here: <a href='{callbackUrl}'>link</a>");
        return View("ForgotPasswordConfirmation");
    }

    // If we got this far, something failed, redisplay form
    return View(model);
}

Uncomment the highlighted ForgotPassword from in the Views/Account/ForgotPassword.cshtml view file.

@model ForgotPasswordViewModel
@{
    ViewData["Title"] = "Forgot your password?";
}

<h2>@ViewData["Title"]</h2>
<p>
    For more information on how to enable reset password please see this <a href="http://go.microsoft.com/fwlink/?LinkID=532713">article</a>.
</p>

<form asp-controller="Account" asp-action="ForgotPassword" method="post" class="form-horizontal">
    <h4>Enter your email.</h4>
    <hr />
    <div asp-validation-summary="All" class="text-danger"></div>
    <div class="form-group">
        <label asp-for="Email" class="col-md-2 control-label"></label>
        <div class="col-md-10">
            <input asp-for="Email" class="form-control" />
            <span asp-validation-for="Email" class="text-danger"></span>
        </div>
    </div>
    <div class="form-group">
        <div class="col-md-offset-2 col-md-10">
            <button type="submit" class="btn btn-default">Submit</button>
        </div>
    </div>
</form>

@section Scripts {
    @{ await Html.RenderPartialAsync("_ValidationScriptsPartial"); }
}

Register, confirm email, and reset password ¶

In this section, run the web app and show the account confirmation and password recovery flow.

Run the application and register a new user

Check your email for the account confirmation link. If you don’t get the email notification:
- Check the SendGrid web site to verify your sent mail messages.
- Check your spam folder.
- Try another email alias on a different email provider (Microsoft, Yahoo, Gmail, etc.)
- In SSOX, navigate to dbo.AspNetUsers and delete the email entry and try again.
Click the link to confirm your email.
Log in with your email and password.
Log off.

Test password reset¶

Login and select Forgot your password?
Enter the email you used to register the account.
An email with a link to reset your password will be sent. Check your email and click the link to reset your password. After your password has been successfully reset, you can login with your email and new password.

Require email confirmation before login ¶

With the current templates, once a user completes the registration form, they are logged in (authenticated). You generally want to confirm their email before logging them in. In the section below, we will modify the code to require new users have a confirmed email before they are logged in. Update the [HttpPost] Login action in the AccountController.cs file with the following highlighted changes.

//
// POST: /Account/Login
[HttpPost]
[AllowAnonymous]
[ValidateAntiForgeryToken]
public async Task<IActionResult> Login(LoginViewModel model, string returnUrl = null)
{
    ViewData["ReturnUrl"] = returnUrl;
    if (ModelState.IsValid)
    {
        // Require the user to have a confirmed email before they can log on.
        var user = await _userManager.FindByNameAsync(model.Email);
        if (user != null)
        {
            if (!await _userManager.IsEmailConfirmedAsync(user))
            {
                ModelState.AddModelError(string.Empty, "You must have a confirmed email to log in.");
                return View(model);
            }
        }

        // This doesn't count login failures towards account lockout
        // To enable password failures to trigger account lockout, set lockoutOnFailure: true
        var result = await _signInManager.PasswordSignInAsync(model.Email, model.Password, model.RememberMe, lockoutOnFailure: false);
        if (result.Succeeded)
        {
            _logger.LogInformation(1, "User logged in.");
            return RedirectToLocal(returnUrl);
        }
        if (result.RequiresTwoFactor)
        {
            return RedirectToAction(nameof(SendCode), new { ReturnUrl = returnUrl, RememberMe = model.RememberMe });
        }
        if (result.IsLockedOut)
        {
            _logger.LogWarning(2, "User account locked out.");
            return View("Lockout");
        }
        else
        {
            ModelState.AddModelError(string.Empty, "Invalid login attempt.");
            return View(model);
        }
    }

Note

A security best practice is to not use production secrets in test and development. If you publish the app to Azure, you can set the SendGrid secrets as application settings in the Azure Web App portal. The configuration system is setup to read keys from environment variables.

Two-factor authentication with SMS¶

This tutorial will show you how to set up two-factor authentication (2FA) using SMS. Twilio is used, but you can use any other SMS provider. We recommend you complete Account Confirmation and Password Recovery before starting this tutorial.

Sections:

Create a new ASP.NET Core project
Setup up SMS for two-factor authentication with Twilio
Enable two-factor authentication
Log in with two-factor authentication
Account lockout for protecting against brute force attacks
Debugging Twilio

Create a new ASP.NET Core project ¶

Create a new ASP.NET Core web app with individual user accounts.

After you create the project, follow the instructions in Account Confirmation and Password Recovery to set up and require SSL.

Setup up SMS for two-factor authentication with Twilio ¶

Create a Twilio account.
On the Dashboard tab of your Twilio account, note the Account SID and Authentication token. Note: Tap Show API Credentials to see the Authentication token.
On the Numbers tab, note the Twilio phone number.
Install the Twilio NuGet package. From the Package Manager Console (PMC), enter the following the following command:
```
Install-Package Twilio
```
Add code in the Services/MessageServices.cs file to enable SMS.

public class AuthMessageSender : IEmailSender, ISmsSender
{
    public AuthMessageSender(IOptions<AuthMessageSMSSenderOptions> optionsAccessor)
    {
        Options = optionsAccessor.Value;
    }

    public AuthMessageSMSSenderOptions Options { get; }  // set only via Secret Manager

    public Task SendEmailAsync(string email, string subject, string message)
    {
        // Plug in your email service here to send an email.
        return Task.FromResult(0);
    }

    public Task SendSmsAsync(string number, string message)
    {
        var twilio = new Twilio.TwilioRestClient(
            Options.SID,           // Account Sid from dashboard
            Options.AuthToken);    // Auth Token

        var result = twilio.SendMessage(Options.SendNumber, number, message);
        // Use the debug output for testing without receiving a SMS message.
        // Remove the Debug.WriteLine(message) line after debugging.
        // System.Diagnostics.Debug.WriteLine(message);
        return Task.FromResult(0);
    }
}

Note

Twilio does not yet support .NET Core. To use Twilio from your application you need to either target the full .NET Framework or you can call the Twilio REST API to send SMS messages.

Note

You can remove // line comment characters from the System.Diagnostics.Debug.WriteLine(message); line to test the application when you can’t get SMS messages. A better approach to logging is to use the built in logging.

Configure the SMS provider key/value¶

We’ll use the Options pattern to access the user account and key settings. For more information, see configuration.

Create a class to fetch the secure SMS key. For this sample, the AuthMessageSMSSenderOptions class is created in the Services/AuthMessageSMSSenderOptions.cs file.

public class AuthMessageSMSSenderOptions
{
    public string SID { get; set; }
    public string AuthToken { get; set; }
    public string SendNumber { get; set; }
}

Set SID, AuthToken, and SendNumber with the secret-manager tool. For example:

C:/WebSMS/src/WebApplication1>dotnet user-secrets set SID abcdefghi
info: Successfully saved SID = abcdefghi to the secret store.

Configure startup to use AuthMessageSMSSenderOptions¶

Add AuthMessageSMSSenderOptions to the service container at the end of the ConfigureServices method in the Startup.cs file:

    // Register application services.
    services.AddTransient<IEmailSender, AuthMessageSender>();
    services.AddTransient<ISmsSender, AuthMessageSender>();
    services.Configure<AuthMessageSMSSenderOptions>(Configuration);
}

Enable two-factor authentication ¶

Open the Views/Manage/Index.cshtml Razor view file.
Uncomment the phone number markup which starts at

@*@(Model.PhoneNumber ?? "None")
Uncomment the Model.TwoFactor markup which starts at

@*@if (Model.TwoFactor)
Comment out or remove the <p>There are no two-factor authentication providers configured. markup.

The completed code is shown below:

<dt>Phone Number:</dt>
<dd>
    <p>
        Phone Numbers can used as a second factor of verification in two-factor authentication.
        See <a href="http://go.microsoft.com/fwlink/?LinkID=532713">this article</a>
        for details on setting up this ASP.NET application to support two-factor authentication using SMS.
    </p>
    @(Model.PhoneNumber ?? "None") [
        @if (Model.PhoneNumber != null)
        {
            <a asp-controller="Manage" asp-action="AddPhoneNumber">Change</a>
                @: &nbsp;|&nbsp;
                <a asp-controller="Manage" asp-action="RemovePhoneNumber">Remove</a>
        }
        else
        {
            <a asp-controller="Manage" asp-action="AddPhoneNumber">Add</a>
        }
        ]
</dd>

<dt>Two-Factor Authentication:</dt>
<dd>
    @*<p>
        There are no two-factor authentication providers configured. See <a href="http://go.microsoft.com/fwlink/?LinkID=532713">this article</a>
        for setting up this application to support two-factor authentication.
    </p>*@
    @if (Model.TwoFactor)
        {
            <form asp-controller="Manage" asp-action="DisableTwoFactorAuthentication" method="post" class="form-horizontal" role="form">
                <text>
                    Enabled
                    <button type="submit" class="btn btn-link">Disable</button>
                </text>
            </form>
        }
        else
        {
            <form asp-controller="Manage" asp-action="EnableTwoFactorAuthentication" method="post" class="form-horizontal" role="form">
                <text>
                    Disabled
                    <button type="submit" class="btn btn-link">Enable</button>
                </text>
            </form>
        }
</dd>

Log in with two-factor authentication ¶

Run the app and register a new user

Tap on your user name, which activates the Index action method in Manage controller. Then tap the phone number Add link.

Add a phone number that will receive the verification code, and tap Send verification code.

You will get a text message with the verification code. Enter it and tap Submit

If you don’t get a text message, see Debugging Twilio.

The Manage view shows your phone number was added successfully.

Tap Enable to enable two-factor authentication.

Test two-factor authentication¶

Log off.
Log in.
The user account has enabled two-factor authentication, so you have to provide the second factor of authentication . In this tutorial you have enabled phone verification. The built in templates also allow you to set up email as the second factor. You can set up additional second factors for authentication such as QR codes. Tap Submit.

Enter the code you get in the SMS message.
Clicking on the Remember this browser check box will exempt you from needing to use 2FA to log on when using the same device and browser. Enabling 2FA and clicking on Remember this browser will provide you with strong 2FA protection from malicious users trying to access your account, as long as they don’t have access to your device. You can do this on any private device you regularly use. By setting Remember this browser, you get the added security of 2FA from devices you don’t regularly use, and you get the convenience on not having to go through 2FA on your own devices.

Account lockout for protecting against brute force attacks ¶

We recommend you use account lockout with 2FA. Once a user logs in (through a local account or social account), each failed attempt at 2FA is stored, and if the maximum attempts (default is 5) is reached, the user is locked out for five minutes (you can set the lock out time with DefaultAccountLockoutTimeSpan). The following configures Account to be locked out for 10 minutes after 10 failed attempts.

    services.Configure<IdentityOptions>(options =>
    {
        options.Lockout.DefaultLockoutTimeSpan = TimeSpan.FromMinutes(10);
        options.Lockout.MaxFailedAccessAttempts = 10;
    });

    // Register application services.
    services.AddTransient<IEmailSender, AuthMessageSender>();
    services.AddTransient<ISmsSender, AuthMessageSender>();
    services.Configure<AuthMessageSMSSenderOptions>(Configuration);
}

Debugging Twilio ¶

If you’re able to use the Twilio API, but you don’t get an SMS message, try the following:

Log in to the Twilio site and navigate to the Logs > SMS & MMS Logs page. You can verify that messages were sent and delivered.
Use the following code in a console application to test Twilio:

static void Main(string[] args)
{
  string AccountSid = "";
  string AuthToken = "";
  var twilio = new Twilio.TwilioRestClient(AccountSid, AuthToken);
  string FromPhone = "";
  string toPhone = "";
  var message = twilio.SendMessage(FromPhone, toPhone, "Twilio Test");
  Console.WriteLine(message.Sid);
}

🔧 Supporting Third Party Clients using OAuth 2.0¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Azure Active Directory¶

Authorization¶

Introduction¶

Authorization refers to the process that determines what a user is able to do. For example user Adam may be able to create a document library, add documents, edit documents and delete them. User Bob may only be authorized to read documents in a single library.

Authorization is orthogonal and independent from authentication, which is the process of ascertaining who a user is. Authentication may create one or more identities for the current user.

Authorization Types¶

In ASP.NET Core authorization now provides simple declarative role and a richer policy based model where authorization is expressed in requirements and handlers evaluate a users claims against requirements. Imperative checks can be based on simple policies or polices which evaluate both the user identity and properties of the resource that the user is attempting to access.

Namespaces¶

Authorization components, including the AuthorizeAttribute and AllowAnonymousAttribute attributes are found in the Microsoft.AspNetCore.Authorization namespace.

Simple Authorization¶

Authorization in MVC is controlled through the AuthorizeAttribute attribute and its various parameters. At its simplest applying the AuthorizeAttribute attribute to a controller or action limits access to the controller or action to any authenticated user.

For example, the following code limits access to the AccountController to any authenticated user.

[Authorize]
public class AccountController : Controller
{
    public ActionResult Login()
    {
    }

    public ActionResult Logout()
    {
    }
}

If you want to apply authorization to an action rather than the controller simply apply the AuthorizeAttribute attribute to the action itself;

public class AccountController : Controller
{
    public ActionResult Login()
    {
    }

    [Authorize]
    public ActionResult Logout()
    {
    }
}

Now only authenticated users can access the logout function.

You can also use the AllowAnonymousAttribute attribute to allow access by non-authenticated users to individual actions; for example

[Authorize]
public class AccountController : Controller
{
    [AllowAnonymous]
    public ActionResult Login()
    {
    }

    public ActionResult Logout()
    {
    }
}

This would allow only authenticated users to the AccountController, except for the Login action, which is accessible by everyone, regardless of their authenticated or unauthenticated / anonymous status.

Warning

[AllowAnonymous] bypasses all authorization statements. If you apply combine [AllowAnonymous] and any [Authorize] attribute then the Authorize attributes will always be ignored. For example if you apply [AllowAnonymous] at the controller level any [Authorize] attributes on the same controller, or on any action within it will be ignored.

Role based Authorization¶

When an identity is created it may belong to one or more roles, for example Tracy may belong to the Administrator and User roles whilst Scott may only belong to the user role. How these roles are created and managed depends on the backing store of the authorization process. Roles are exposed to the developer through the IsInRole property on the ClaimsPrincipal class.

Adding role checks¶

Role based authorization checks are declarative - the developer embeds them within their code, against a controller or an action within a controller, specifying roles which the current user must be a member of to access the requested resource.

For example the following code would limit access to any actions on the AdministrationController to users who are a member of the Administrator group.

[Authorize(Roles = "Administrator")]
public class AdministrationController : Controller
{
}

You can specify multiple roles as a comma separated list;

[Authorize(Roles = "HRManager,Finance")]
public class SalaryController : Controller
{
}

This controller would be only accessible by users who are members of the HRManager role or the Finance role.

If you apply multiple attributes then an accessing user must be a member of all the roles specified; the following sample requires that a user must be a member of both the PowerUser and ControlPanelUser role.

[Authorize(Roles = "PowerUser")]
[Authorize(Roles = "ControlPanelUser")]
public class ControlPanelController : Controller
{
}

You can further limit access by applying additional role authorization attributes at the action level;

[Authorize(Roles = "Administrator, PowerUser")]
public class ControlPanelController : Controller
{
    public ActionResult SetTime()
    {
    }

    [Authorize(Roles = "Administrator")]
    public ActionResult ShutDown()
    {
    }
}

In the previous code snippet members of the Administrator role or the PowerUser role can access the controller and the SetTime action, but only members of the Administrator role can access the ShutDown action.

You can also lock down a controller but allow anonymous, unauthenticated access to individual actions.

[Authorize]
public class ControlPanelController : Controller
{
    public ActionResult SetTime()
    {
    }

    [AllowAnonymous]
    public ActionResult Login()
    {
    }
}

Policy based role checks¶

Role requirements can also be expressed using the new Policy syntax, where a developer registers a policy at startup as part of the Authorization service configuration. This normally takes part in ConfigureServices() in your Startup.cs file.

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc();

    services.AddAuthorization(options =>
    {
        options.AddPolicy("RequireAdministratorRole", policy => policy.RequireRole("Administrator"));
    });
}

Policies are applied using the Policy property on the AuthorizeAttribute attribute;

[Authorize(Policy = "RequireAdministratorRole")]
public IActionResult Shutdown()
{
    return View();
}

If you want to specify multiple allowed roles in a requirement then you can specify them as parameters to the RequireRole method;

options.AddPolicy("ElevatedRights", policy =>
                  policy.RequireRole("Administrator", "PowerUser", "BackupAdministrator"));

This example authorizes users who belong to the Administrator, PowerUser or BackupAdministrator roles.

Claims-Based Authorization¶

When an identity is created it may be assigned one or more claims issued by a trusted party. A claim is name value pair that represents what the subject is, not what the subject can do. For example you may have a Drivers License, issued by a local driving license authority. Your driver’s license has your date of birth on it. In this case the claim name would be DateOfBirth, the claim value would be your date of birth, for example 8th June 1970 and the issuer would be the driving license authority. Claims based authorization, at its simplest, checks the value of a claim and allows access to a resource based upon that value. For example if you want access to a night club the authorization process might be:

The door security officer would evaluate the value of your date of birth claim and whether they trust the issuer (the driving license authority) before granting you access.

An identity can contain multiple claims with multiple values and can contain multiple claims of the same type.

Adding claims checks¶

Claim based authorization checks are declarative - the developer embeds them within their code, against a controller or an action within a controller, specifying claims which the current user must possess, and optionally the value the claim must hold to access the requested resource. Claims requirements are policy based, the developer must build and register a policy expressing the claims requirements.

The simplest type of claim policy looks for the presence of a claim and does not check the value.

First you need to build and register the policy. This takes place as part of the Authorization service configuration, which normally takes part in ConfigureServices() in your Startup.cs file.

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc();

    services.AddAuthorization(options =>
    {
        options.AddPolicy("EmployeeOnly", policy => policy.RequireClaim("EmployeeNumber"));
    });
}

In this case the EmployeeOnly policy checks for the presence of an EmployeeNumber claim on the current identity.

You then apply the policy using the Policy property on the AuthorizeAttribute attribute to specify the policy name;

[Authorize(Policy = "EmployeeOnly")]
public IActionResult VacationBalance()
{
    return View();
}

The AuthorizeAttribute attribute can be applied to an entire controller, in this instance only identities matching the policy will be allowed access to any Action on the controller.

[Authorize(Policy = "EmployeeOnly")]
public class VacationController : Controller
{
    public ActionResult VacationBalance()
    {
    }
}

If you have a controller that is protected by the AuthorizeAttribute attribute, but want to allow anonymous access to particular actions you apply the AllowAnonymousAttribute attribute;

[Authorize(Policy = "EmployeeOnly")]
public class VacationController : Controller
{
    public ActionResult VacationBalance()
    {
    }

    [AllowAnonymous]
    public ActionResult VacationPolicy()
    {
    }
}

Most claims come with a value. You can specify a list of allowed values when creating the policy. The following example would only succeed for employees whose employee number was 1, 2, 3, 4 or 5.

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc();

    services.AddAuthorization(options =>
    {
        options.AddPolicy("Founders", policy =>
                          policy.RequireClaim("EmployeeNumber", "1", "2", "3", "4", "5"));
    }
}

Multiple Policy Evaluation¶

If you apply multiple policies to a controller or action then all policies must pass before access is granted. For example;

[Authorize(Policy = "EmployeeOnly")]
public class SalaryController : Controller
{
    public ActionResult Payslip()
    {
    }

    [Authorize(Policy = "HumanResources")]
    public ActionResult UpdateSalary()
    {
    }
}

In the above example any identity which fulfills the EmployeeOnly policy can access the Payslip action as that policy is enforced on the controller. However in order to call the UpdateSalary action the identity must fulfill both the EmployeeOnly policy and the HumanResources policy.

If you want more complicated policies, such as taking a date of birth claim, calculating an age from it then checking the age is 21 or older then you need to write custom policy handlers.

Custom Policy-Based Authorization¶

Underneath the covers the role authorization and claims authorization make use of a requirement, a handler for the requirement and a pre-configured policy. These building blocks allow you to express authorization evaluations in code, allowing for a richer, reusable, and easily testable authorization structure.

An authorization policy is made up of one or more requirements and registered at application startup as part of the Authorization service configuration, in ConfigureServices in the Startup.cs file.

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc();

    services.AddAuthorization(options =>
    {
        options.AddPolicy("Over21",
                          policy => policy.Requirements.Add(new MinimumAgeRequirement(21)));
    }
});

Here you can see an “Over21” policy is created with a single requirement, that of a minimum age, which is passed as a parameter to the requirement.

Policies are applied using the Authorize attribute by specifying the policy name, for example;

[Authorize(Policy="Over21")]
public class AlcoholPurchaseRequirementsController : Controller
{
    public ActionResult Login()
    {
    }

    public ActionResult Logout()
    {
    }
}

Requirements¶

An authorization requirement is a collection of data parameters that a policy can use to evaluate the current user principal. In our Minimum Age policy the requirement we have a single parameter, the minimum age. A requirement must implement IAuthorizationRequirement. This is an empty, marker interface. A parameterized minimum age requirement might be implemented as follows;

public class MinimumAgeRequirement : IAuthorizationRequirement
{
    public MinimumAgeRequirement(int age)
    {
        MinimumAge = age;
    }

    protected int MinimumAge { get; set; }
}

A requirement doesn’t need to have data or properties.

Authorization Handlers¶

An authorization handler is responsible for the evaluation of any properties of a requirement. The authorization handler must evaluate them against a provided AuthorizationContext to decide if authorization is allowed. A requirement can have multiple handlers. Handlers must inherit AuthorizationHandler<T> where T is the requirement it handles.

The minimum age handler might look like this:

public class MinimumAgeHandler : AuthorizationHandler<MinimumAgeRequirement>
{
    protected override Task HandleRequirementAsync(AuthorizationContext context, MinimumAgeRequirement requirement)
    {
        if (!context.User.HasClaim(c => c.Type == ClaimTypes.DateOfBirth &&
                                   c.Issuer == "http://contoso.com"))
        {
            return Task.FromResult(0);
        }

        var dateOfBirth = Convert.ToDateTime(context.User.FindFirst(
            c => c.Type == ClaimTypes.DateOfBirth && c.Issuer == "http://contoso.com").Value);

        int calculatedAge = DateTime.Today.Year - dateOfBirth.Year;
        if (dateOfBirth > DateTime.Today.AddYears(-calculatedAge))
        {
            calculatedAge--;
        }

        if (calculatedAge >= requirement.MinimumAge)
        {
            context.Succeed(requirement);
        }
        return Task.FromResult(0);
    }
}

In the code above we first look to see if the current user principal has a date of birth claim which has been issued by an Issuer we know and trust. If the claim is missing we can’t authorize so we return. If we have a claim, we figure out how old the user is, and if they meet the minimum age passed in by the requirement then authorization has been successful. Once authorization is successful we call context.Succeed() passing in the requirement that has been successful as a parameter.

Handlers must be registered in the services collection during configuration, for example;

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc();

    services.AddAuthorization(options =>
    {
        options.AddPolicy("Over21",
                          policy => policy.Requirements.Add(new MinimumAgeRequirement(21)));
    });

    services.AddSingleton<IAuthorizationHandler, MinimumAgeHandler>();
}

Each handler is added to the services collection by using services.AddSingleton<IAuthorizationHandler, YourHandlerClass>(); passing in your handler class.

What should a handler return?¶

You can see in our handler example that the Handle() method has no return value, so how do we indicate success or failure?

A handler indicates success by calling context.Succeed(IAuthorizationRequirement requirement), passing the requirement that has been successfully validated.
A handler does not need to handle failures generally, as other handlers for the same requirement may succeed.
To guarantee failure even if other handlers for a requirement succeed, call context.Fail.

Regardless of what you call inside your handler all handlers for a requirement will be called when a policy requires the requirement. This allows requirements to have side effects, such as logging, which will always take place even if context.Fail() has been called in another handler.

Why would I want multiple handlers for a requirement?¶

In cases where you want evaluation to be on an OR basis you implement multiple handlers for a single requirement. For example, Microsoft has doors which only open with key cards. If you leave your key card at home the receptionist prints a temporary sticker and opens the door for you. In this scenario you’d have a single requirement, EnterBuilding, but multiple handlers, each one examining a single requirement.

public class EnterBuildingRequirement : IAuthorizationRequirement
{
}

public class BadgeEntryHandler : AuthorizationHandler<EnterBuildingRequirement>
{
    protected override Task HandleRequirementAsync(AuthorizationContext context, EnterBuildingRequirement requirement)
    {
        if (context.User.HasClaim(c => c.Type == ClaimTypes.BadgeId &&
                                       c.Issuer == "http://microsoftsecurity"))
        {
            context.Succeed(requirement);
            return Task.FromResult(0);
        }
    }
}

public class HasTemporaryStickerHandler : AuthorizationHandler<EnterBuildingRequirement>
{
    protected override Task HandleRequirementAsync(AuthorizationContext context, EnterBuildingRequirement requirement)
    {
        if (context.User.HasClaim(c => c.Type == ClaimTypes.TemporaryBadgeId &&
                                       c.Issuer == "https://microsoftsecurity"))
        {
            // We'd also check the expiration date on the sticker.
            context.Succeed(requirement);
            return Task.FromResult(0);
        }
    }
}

Now, assuming both handlers are registered when a policy evaluates the EnterBuildingRequirement if either handler succeeds the policy evaluation will succeed.

Using a func to fufill a policy¶

There may be occasions where fufilling a policy is simple to express in code. It is possible to simply supply a Func<AuthorizationHandlerContext, bool> when configuring your policy with the RequireAssertion policy builder.

For example the previous BadgeEntryHandler could be rewritten as follows;

services.AddAuthorization(options =>
    {
        options.AddPolicy("BadgeEntry",
                          policy => policy.RequireAssertion(context =>
                                  context.User.HasClaim(c =>
                                     (c.Type == ClaimTypes.BadgeId ||
                                      c.Type == ClaimTypes.TemporaryBadgeId)
                                      && c.Issuer == "https://microsoftsecurity"));
                          }));
    }
 }

Accessing MVC Request Context In Handlers¶

The Handle method you must implement in an authorization handler has two parameters, an AuthorizationContext and the Requirement you are handling. Frameworks such as MVC or Jabbr are free to add any object to the Resource property on the AuthorizationContext to pass through extra information.

For example MVC passes an instance of Microsoft.AspNetCore.Mvc.Filters.AuthorizationFilterContext in the resource property which is used to access HttpContext, RouteData and everything else MVC provides.

The use of the Resource property is framework specific. Using information in the Resource property will limit your authorization policies to particular frameworks. You should cast the Resource property using the as keyword, and then check the cast has succeed to ensure your code doesn’t crash with InvalidCastExceptions when run on other frameworks;

var mvcContext = context.Resource as Microsoft.AspNetCore.Mvc.Filters.AuthorizationFilterContext;

if (mvcContext != null)
{
    // Examine MVC specific things like routing data.
}

Dependency Injection in Requirement Handlers¶

As handlers must be registered in the service collection they support dependency injection. If, for example, you had a repository of rules you want to evaluate inside a handler and that repository is registered in the service collection authorization will resolve and inject that into your constructor.

For example, if you wanted to use ASP.NET’s logging infrastructure you would to inject ILoggerFactory into your handler. Such a handler might look like this;

public class LoggingAuthorizationHandler : AuthorizationHandler<MyRequirement>
{
    ILogger _logger;

    public LoggingAuthorizationHandler(ILoggerFactory loggerFactory)
    {
        _logger = loggerFactory.CreateLogger(this.GetType().FullName);
    }

    protected override Task HandleRequirementAsync(AuthorizationHandlerContext context, MyRequirement requirement)
    {
        _logger.LogInformation("Inside my handler");
        // Check if the requirement is fulfilled.
        return Task.CompletedTask;
    }
}

Then you register handlers with services.AddSingleton(), for example

services.AddSingleton<IAuthorizationHandler, LoggingAuthorizationHandler>();

An instance of the handler will be created when your application starts, and DI will inject the registered ILoggerFactory into your constructor.

Resource Based Authorization¶

Often authorization depends upon the resource being accessed. For example a document may have an author property. Only the document author would be allowed to update it, so the resource must be loaded from the document repository before an authorization evaluation can be made. This cannot be done with an Authorize attribute, as attribute evaluation takes place before data binding and before your own code to load a resource runs inside an action. Instead of declarative authorization, the attribute method, we must use imperative authorization, where a developer calls an authorize function within his own code.

Authorizing within your code¶

Authorization is implemented as a service, IAuthorizationService, registered in the service collection and available via dependency injection for Controllers to access.

public class DocumentController : Controller
{
    IAuthorizationService _authorizationService;

    public DocumentController(IAuthorizationService authorizationService)
    {
        _authorizationService = authorizationService;
    }
}

IAuthorizationService has two methods, one where you pass the resource and the policy name and the other where you pass the resource and a list of requirements to evaluate.

Task<bool> AuthorizeAsync(ClaimsPrincipal user,
                          object resource,
                          IEnumerable<IAuthorizationRequirement> requirements);
Task<bool> AuthorizeAsync(ClaimsPrincipal user,
                          object resource,
                          string policyName);

To call the service load your resource within your action then call the AuthorizeAsync overload you require. For example

public async Task<IActionResult> Edit(Guid documentId)
{
    Document document = documentRepository.Find(documentId);

    if (document == null)
    {
        return new HttpNotFoundResult();
    }

    if (await authorizationService.AuthorizeAsync(User, document, "EditPolicy"))
    {
        return View(document);
    }
    else
    {
        return new ChallengeResult();
    }
}

Writing a resource based handler¶

Writing a handler for resource based authorization is not that much different to writing a plain requirements handler. You create a requirement, and then implement a handler for the requirement, specifying the requirement as before and also the resource type. For example, a handler which might accept a Document resource would look as follows;

public class DocumentAuthorizationHandler : AuthorizationHandler<MyRequirement, Document>
{
    public override Task HandleRequirementAsync(AuthorizationHandlerContext context,
                                                MyRequirement requirement,
                                                Document resource)
    {
        // Validate the requirement against the resource and identity.

        return Task.CompletedTask;
    }
}

Don’t forget you also need to register your handler in the ConfigureServices method;

services.AddSingleton<IAuthorizationHandler, DocumentAuthorizationHandler>();

Operational Requirements¶

If you are making decisions based on operations such as read, write, update and delete, you can use the OperationAuthorizationRequirement class in the Microsoft.AspNetCore.Authorization.Infrastructure namespace. This prebuilt requirement class enables you to write a single handler which has a parameterized operation name, rather than create individual classes for each operation. To use it provide some operation names:

public static class Operations
{
    public static OperationAuthorizationRequirement Create =
        new OperationAuthorizationRequirement { Name = "Create" };
    public static OperationAuthorizationRequirement Read =
        new OperationAuthorizationRequirement   { Name = "Read" };
    public static OperationAuthorizationRequirement Update =
        new OperationAuthorizationRequirement { Name = "Update" };
    public static OperationAuthorizationRequirement Delete =
        new OperationAuthorizationRequirement { Name = "Delete" };
}

Your handler could then be implemented as follows, using a hypothetical Document class as the resource;

public class DocumentAuthorizationHandler :
    AuthorizationHandler<OperationAuthorizationRequirement, Document>
{
    public override Task HandleRequirementAsync(AuthorizationHandlerContext context,
                                                OperationAuthorizationRequirement requirement,
                                                Document resource)
    {
        // Validate the operation using the resource, the identity and
        // the Name property value from the requirement.

        return Task.CompletedTask;
    }
}

You can see the handler works on OperationAuthorizationRequirement. The code inside the handler must take the Name property of the supplied requirement into account when making its evaluations.

To call an operational resource handler you need to specify the operation when calling AuthorizeAsync in your action. For example

if (await authorizationService.AuthorizeAsync(User, document, Operations.Read))
{
    return View(document);
}
else
{
    return new ChallengeResult();
}

This example checks if the User is able to perform the Read operation for the current document instance. If authorization succeeds the view for the document will be returned. If authorization fails returning ChallengeResult will inform any authentication middleware authorization has failed and the middleware can take the appropriate response, for example returning a 401 or 403 status code, or redirecting the user to a login page for interactive browser clients.

View Based Authorization¶

Often a developer will want to show, hide or otherwise modify a UI based on the current user identity. You can access the authorization service within MVC views via dependency injection. To inject the authorization service into a Razor view use the @inject directive, for example @inject IAuthorizationService AuthorizationService. If you want the authorization service in every view then place the @inject directive into the _ViewImports.cshtml file in the Views directory.

Once you have injected the authorization service you use it by calling the AuthorizeAsync method in exactly the same way as you would check during resource based authorization.

@if (await AuthorizationService.AuthorizeAsync(User, "PolicyName"))
{
    <p>This paragraph is displayed because you fulfilled PolicyName.</p>
}

In some cases the resource will be your view model, and you can call AuthorizeAsync in exactly the same way as you would check during resource based authorization;

@if (await AuthorizationService.AuthorizeAsync(User, Model, Operations.Edit))
{
    <p><a class="btn btn-default" role="button"
        href="@Url.Action("Edit", "Document", new { id = Model.Id })">Edit</a></p>
}

Here you can see the model is passed as the resource authorization should take into consideration.

Warning

Do not rely on showing or hiding parts of your UI as your only authorization method. Hiding a UI element does not mean a user cannot access it. You must also authorize the user within your controller code.

Limiting identity by scheme¶

In some scenarios, such as Single Page Applications it is possible to end up with multiple authentication methods. For example, your application may use cookie-based authentication to log in and bearer authentication for JavaScript requests. In some cases you may have multiple instances of an authentication middleware. For example, two cookie middlewares where one contains a basic identity and one is created when a multi-factor authentication has triggered because the user requested an operation that requires extra security.

Authentication schemes are named when authentication middleware is configured during authentication, for example

app.UseCookieAuthentication(new CookieAuthenticationOptions()
{
    AuthenticationScheme = "Cookie",
    LoginPath = new PathString("/Account/Unauthorized/"),
    AccessDeniedPath = new PathString("/Account/Forbidden/"),
    AutomaticAuthenticate = false
});

app.UseBearerAuthentication(options =>
{
    options.AuthenticationScheme = "Bearer";
    options.AutomaticAuthenticate = false;
});

In this configuration two authentication middlewares have been added, one for cookies and one for bearer.

Note

When adding multiple authentication middleware you should ensure that no middleware is configured to run automatically. You do this by setting the AutomaticAuthenticate options property to false. If you fail to do this filtering by scheme will not work.

Selecting the scheme with the Authorize attribute¶

As no authentication middleware is configured to automatically run and create an identity you must, at the point of authorization choose which middleware will be used. The simplest way to select the middleware you wish to authorize with is to use the ActiveAuthenticationSchemes property. This property accepts a comma delimited list of Authentication Schemes to use. For example;

[Authorize(ActiveAuthenticationSchemes = "Cookie,Bearer")]
public class MixedController : Controller

In the example above both the cookie and bearer middlewares will run and have a chance to create and append an identity for the current user. By specifying a single scheme only the specified middleware will run;

[Authorize(ActiveAuthenticationSchemes = "Bearer")]

In this case only the middleware with the Bearer scheme would run, and any cookie based identities would be ignored.

Selecting the scheme with policies¶

If you prefer to specify the desired schemes in policy you can set the AuthenticationSchemes collection when adding your policy.

options.AddPolicy("Over18", policy =>
{
    policy.AuthenticationSchemes.Add("Bearer");
    policy.RequireAuthenticatedUser();
    policy.Requirements.Add(new Over18Requirement());
});

In this example the Over18 policy will only run against the identity created by the Bearer middleware.

🔧 Authorization Filters¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Data Protection¶

Introduction to Data Protection¶

Web applications often need to store security-sensitive data. Windows provides DPAPI for desktop applications but this is unsuitable for web applications. The ASP.NET Core data protection stack provide a simple, easy to use cryptographic API a developer can use to protect data, including key management and rotation.

The ASP.NET Core data protection stack is designed to serve as the long-term replacement for the <machineKey> element in ASP.NET 1.x - 4.x. It was designed to address many of the shortcomings of the old cryptographic stack while providing an out-of-the-box solution for the majority of use cases modern applications are likely to encounter.

Problem statement¶

The overall problem statement can be succinctly stated in a single sentence: I need to persist trusted information for later retrieval, but I do not trust the persistence mechanism. In web terms, this might be written as “I need to round-trip trusted state via an untrusted client.”

The canonical example of this is an authentication cookie or bearer token. The server generates an “I am Groot and have xyz permissions” token and hands it to the client. At some future date the client will present that token back to the server, but the server needs some kind of assurance that the client hasn’t forged the token. Thus the first requirement: authenticity (a.k.a. integrity, tamper-proofing).

Since the persisted state is trusted by the server, we anticipate that this state might contain information that is specific to the operating environment. This could be in the form of a file path, a permission, a handle or other indirect reference, or some other piece of server-specific data. Such information should generally not be disclosed to an untrusted client. Thus the second requirement: confidentiality.

Finally, since modern applications are componentized, what we’ve seen is that individual components will want to take advantage of this system without regard to other components in the system. For instance, if a bearer token component is using this stack, it should operate without interference from an anti-CSRF mechanism that might also be using the same stack. Thus the final requirement: isolation.

We can provide further constraints in order to narrow the scope of our requirements. We assume that all services operating within the cryptosystem are equally trusted and that the data does not need to be generated or consumed outside of the services under our direct control. Furthermore, we require that operations are as fast as possible since each request to the web service might go through the cryptosystem one or more times. This makes symmetric cryptography ideal for our scenario, and we can discount asymmetric cryptography until such a time that it is needed.

Design philosophy¶

We started by identifying problems with the existing stack. Once we had that, we surveyed the landscape of existing solutions and concluded that no existing solution quite had the capabilities we sought. We then engineered a solution based on several guiding principles.

The system should offer simplicity of configuration. Ideally the system would be zero-configuration and developers could hit the ground running. In situations where developers need to configure a specific aspect (such as the key repository), consideration should be given to making those specific configurations simple.
Offer a simple consumer-facing API. The APIs should be easy to use correctly and difficult to use incorrectly.
Developers should not learn key management principles. The system should handle algorithm selection and key lifetime on the developer’s behalf. Ideally the developer should never even have access to the raw key material.
Keys should be protected at rest when possible. The system should figure out an appropriate default protection mechanism and apply it automatically.

With these principles in mind we developed a simple, easy to use data protection stack.

The ASP.NET Core data protection APIs are not primarily intended for indefinite persistence of confidential payloads. Other technologies like Windows CNG DPAPI and Azure Rights Management are more suited to the scenario of indefinite storage, and they have correspondingly strong key management capabilities. That said, there is nothing prohibiting a developer from using the ASP.NET Core data protection APIs for long-term protection of confidential data.

Audience¶

The data protection system is divided into five main packages. Various aspects of these APIs target three main audiences;

The Consumer APIs Overview target application and framework developers.

“I don’t want to learn about how the stack operates or about how it is configured. I simply want to perform some operation in as simple a manner as possible with high probability of using the APIs successfully.”

The configuration APIs target application developers and system administrators.

“I need to tell the data protection system that my environment requires non-default paths or settings.”

The extensibility APIs target developers in charge of implementing custom policy. Usage of these APIs would be limited to rare situations and experienced, security aware developers.

“I need to replace an entire component within the system because I have truly unique behavioral requirements. I am willing to learn uncommonly-used parts of the API surface in order to build a plugin that fulfills my requirements.”

Package Layout¶

The data protection stack consists of five packages.

Microsoft.AspNetCore.DataProtection.Abstractions contains the basic IDataProtectionProvider and IDataProtector interfaces. It also contains useful extension methods that can assist working with these types (e.g., overloads of IDataProtector.Protect). See the consumer interfaces section for more information. If somebody else is responsible for instantiating the data protection system and you are simply consuming the APIs, you’ll want to reference Microsoft.AspNetCore.DataProtection.Abstractions.
Microsoft.AspNetCore.DataProtection contains the core implementation of the data protection system, including the core cryptographic operations, key management, configuration, and extensibility. If you’re responsible for instantiating the data protection system (e.g., adding it to an IServiceCollection) or modifying or extending its behavior, you’ll want to reference Microsoft.AspNetCore.DataProtection.
Microsoft.AspNetCore.DataProtection.Extensions contains additional APIs which developers might find useful but which don’t belong in the core package. For instance, this package contains a simple “instantiate the system pointing at a specific key storage directory with no dependency injection setup” API (more info). It also contains extension methods for limiting the lifetime of protected payloads (more info).
Microsoft.AspNetCore.DataProtection.SystemWeb can be installed into an existing ASP.NET 4.x application to redirect its <machineKey> operations to instead use the new data protection stack. See compatibility for more information.
Microsoft.AspNetCore.Cryptography.KeyDerivation provides an implementation of the PBKDF2 password hashing routine and can be used by systems which need to handle user passwords securely. See Password Hashing for more information.

Getting Started with the Data Protection APIs¶

At its simplest protecting data consists of the following steps:

Create a data protector from a data protection provider.
Call the Protect method with the data you want to protect.
Call the Unprotect method with the data you want to turn back into plain text.

Most frameworks such as ASP.NET or SignalR already configure the data protection system and add it to a service container you access via dependency injection. The following sample demonstrates configuring a service container for dependency injection and registering the data protection stack, receiving the data protection provider via DI, creating a protector and protecting then unprotecting data

using System;
using Microsoft.AspNetCore.DataProtection;
using Microsoft.Extensions.DependencyInjection;

public class Program
{
    public static void Main(string[] args)
    {
        // add data protection services
        var serviceCollection = new ServiceCollection();
        serviceCollection.AddDataProtection();
        var services = serviceCollection.BuildServiceProvider();

        // create an instance of MyClass using the service provider
        var instance = ActivatorUtilities.CreateInstance<MyClass>(services);
        instance.RunSample();
    }

    public class MyClass
    {
        IDataProtector _protector;

        // the 'provider' parameter is provided by DI
        public MyClass(IDataProtectionProvider provider)
        {
            _protector = provider.CreateProtector("Contoso.MyClass.v1");
        }

        public void RunSample()
        {
            Console.Write("Enter input: ");
            string input = Console.ReadLine();

            // protect the payload
            string protectedPayload = _protector.Protect(input);
            Console.WriteLine($"Protect returned: {protectedPayload}");

            // unprotect the payload
            string unprotectedPayload = _protector.Unprotect(protectedPayload);
            Console.WriteLine($"Unprotect returned: {unprotectedPayload}");
        }
    }
}

/*
 * SAMPLE OUTPUT
 *
 * Enter input: Hello world!
 * Protect returned: CfDJ8ICcgQwZZhlAlTZT...OdfH66i1PnGmpCR5e441xQ
 * Unprotect returned: Hello world!
 */

When you create a protector you must provide one or more Purpose Strings. A purpose string provides isolation between consumers, for example a protector created with a purpose string of “green” would not be able to unprotect data provided by a protector with a purpose of “purple”.

Tip

Instances of IDataProtectionProvider and IDataProtector are thread-safe for multiple callers. It is intended that once a component gets a reference to an IDataProtector via a call to CreateProtector, it will use that reference for multiple calls to Protect and Unprotect.

A call to Unprotect will throw CryptographicException if the protected payload cannot be verified or deciphered. Some components may wish to ignore errors during unprotect operations; a component which reads authentication cookies might handle this error and treat the request as if it had no cookie at all rather than fail the request outright. Components which want this behavior should specifically catch CryptographicException instead of swallowing all exceptions.

Consumer APIs¶

Consumer APIs Overview¶

The IDataProtectionProvider and IDataProtector interfaces are the basic interfaces through which consumers use the data protection system. They are located in the Microsoft.AspNetCore.DataProtection.Interfaces package.

IDataProtectionProvider¶

The provider interface represents the root of the data protection system. It cannot directly be used to protect or unprotect data. Instead, the consumer must get a reference to an IDataProtector by calling IDataProtectionProvider.CreateProtector(purpose), where purpose is a string that describes the intended consumer use case. See Purpose Strings for much more information on the intent of this parameter and how to choose an appropriate value.

IDataProtector¶

The protector interface is returned by a call to CreateProtector, and it is this interface which consumers can use to perform protect and unprotect operations.

To protect a piece of data, pass the data to the Protect method. The basic interface defines a method which converts byte[] -> byte[], but there is also an overload (provided as an extension method) which converts string -> string. The security offered by the two methods is identical; the developer should choose whichever overload is most convenient for his use case. Irrespective of the overload chosen, the value returned by the Protect method is now protected (enciphered and tamper-proofed), and the application can send it to an untrusted client.

To unprotect a previously-protected piece of data, pass the protected data to the Unprotect method. (There are byte[]-based and string-based overloads for developer convenience.) If the protected payload was generated by an earlier call to Protect on this same IDataProtector, the Unprotect method will return the original unprotected payload. If the protected payload has been tampered with or was produced by a different IDataProtector, the Unprotect method will throw CryptographicException.

The concept of same vs. different IDataProtector ties back to the concept of purpose. If two IDataProtector instances were generated from the same root IDataProtectionProvider but via different purpose strings in the call to IDataProtectionProvider.CreateProtector, then they are considered different protectors, and one will not be able to unprotect payloads generated by the other.

Consuming these interfaces¶

For a DI-aware component, the intended usage is that the component take an IDataProtectionProvider parameter in its constructor and that the DI system automatically provides this service when the component is instantiated.

Note

Some applications (such as console applications or ASP.NET 4.x applications) might not be DI-aware so cannot use the mechanism described here. For these scenarios consult the Non DI Aware Scenarios document for more information on getting an instance of an IDataProtection provider without going through DI.

The following sample demonstrates three concepts:

Adding the data protection system to the service container,
Using DI to receive an instance of an IDataProtectionProvider, and
Creating an IDataProtector from an IDataProtectionProvider and using it to protect and unprotect data.

using System;
using Microsoft.AspNetCore.DataProtection;
using Microsoft.Extensions.DependencyInjection;

public class Program
{
    public static void Main(string[] args)
    {
        // add data protection services
        var serviceCollection = new ServiceCollection();
        serviceCollection.AddDataProtection();
        var services = serviceCollection.BuildServiceProvider();

        // create an instance of MyClass using the service provider
        var instance = ActivatorUtilities.CreateInstance<MyClass>(services);
        instance.RunSample();
    }

    public class MyClass
    {
        IDataProtector _protector;

        // the 'provider' parameter is provided by DI
        public MyClass(IDataProtectionProvider provider)
        {
            _protector = provider.CreateProtector("Contoso.MyClass.v1");
        }

        public void RunSample()
        {
            Console.Write("Enter input: ");
            string input = Console.ReadLine();

            // protect the payload
            string protectedPayload = _protector.Protect(input);
            Console.WriteLine($"Protect returned: {protectedPayload}");

            // unprotect the payload
            string unprotectedPayload = _protector.Unprotect(protectedPayload);
            Console.WriteLine($"Unprotect returned: {unprotectedPayload}");
        }
    }
}

/*
 * SAMPLE OUTPUT
 *
 * Enter input: Hello world!
 * Protect returned: CfDJ8ICcgQwZZhlAlTZT...OdfH66i1PnGmpCR5e441xQ
 * Unprotect returned: Hello world!
 */

The package Microsoft.AspNetCore.DataProtection.Abstractions contains an extension method IServiceProvider.GetDataProtector as a developer convenience. It encapsulates as a single operation both retrieving an IDataProtectionProvider from the service provider and calling IDataProtectionProvider.CreateProtector. The following sample demonstrates its usage.

using System;
using Microsoft.AspNetCore.DataProtection;
using Microsoft.Extensions.DependencyInjection;
 
public class Program
{
    public static void Main(string[] args)
    {
        // add data protection services
        var serviceCollection = new ServiceCollection();
        serviceCollection.AddDataProtection();
        var services = serviceCollection.BuildServiceProvider();
 
        // get an IDataProtector from the IServiceProvider
        var protector = services.GetDataProtector("Contoso.Example.v2");
        Console.Write("Enter input: ");
        string input = Console.ReadLine();
 
        // protect the payload
        string protectedPayload = protector.Protect(input);
        Console.WriteLine($"Protect returned: {protectedPayload}");
 
        // unprotect the payload
        string unprotectedPayload = protector.Unprotect(protectedPayload);
        Console.WriteLine($"Unprotect returned: {unprotectedPayload}");
    }
}

Tip

Instances of IDataProtectionProvider and IDataProtector are thread-safe for multiple callers. It is intended that once a component gets a reference to an IDataProtector via a call to CreateProtector, it will use that reference for multiple calls to Protect and Unprotect.

A call to Unprotect will throw CryptographicException if the protected payload cannot be verified or deciphered. Some components may wish to ignore errors during unprotect operations; a component which reads authentication cookies might handle this error and treat the request as if it had no cookie at all rather than fail the request outright. Components which want this behavior should specifically catch CryptographicException instead of swallowing all exceptions.

Purpose Strings¶

Components which consume IDataProtectionProvider must pass a unique purposes parameter to the CreateProtector method. The purposes parameter is inherent to the security of the data protection system, as it provides isolation between cryptographic consumers, even if the root cryptographic keys are the same.

When a consumer specifies a purpose, the purpose string is used along with the root cryptographic keys to derive cryptographic subkeys unique to that consumer. This isolates the consumer from all other cryptographic consumers in the application: no other component can read its payloads, and it cannot read any other component’s payloads. This isolation also renders infeasible entire categories of attack against the component.

In the diagram above IDataProtector instances A and B cannot read each other’s payloads, only their own.

The purpose string doesn’t have to be secret. It should simply be unique in the sense that no other well-behaved component will ever provide the same purpose string.

Tip

Using the namespace and type name of the component consuming the data protection APIs is a good rule of thumb, as in practice this information will never conflict.

A Contoso-authored component which is responsible for minting bearer tokens might use Contoso.Security.BearerToken as its purpose string. Or - even better - it might use Contoso.Security.BearerToken.v1 as its purpose string. Appending the version number allows a future version to use Contoso.Security.BearerToken.v2 as its purpose, and the different versions would be completely isolated from one another as far as payloads go.

Since the purposes parameter to CreateProtector is a string array, the above could have been instead specified as [ “Contoso.Security.BearerToken”, “v1” ]. This allows establishing a hierarchy of purposes and opens up the possibility of multi-tenancy scenarios with the data protection system.

Warning

Components should not allow untrusted user input to be the sole source of input for the purposes chain.

For example, consider a component Contoso.Messaging.SecureMessage which is responsible for storing secure messages. If the secure messaging component were to call CreateProtector([ username ]), then a malicious user might create an account with username “Contoso.Security.BearerToken” in an attempt to get the component to call CreateProtector([ “Contoso.Security.BearerToken” ]), thus inadvertently causing the secure messaging system to mint payloads that could be perceived as authentication tokens.

A better purposes chain for the messaging component would be CreateProtector([ “Contoso.Messaging.SecureMessage”, “User: username” ]), which provides proper isolation.

The isolation provided by and behaviors of IDataProtectionProvider, IDataProtector, and purposes are as follows:

For a given IDataProtectionProvider object, the CreateProtector method will create an IDataProtector object uniquely tied to both the IDataProtectionProvider object which created it and the purposes parameter which was passed into the method.
The purpose parameter must not be null. (If purposes is specified as an array, this means that the array must not be of zero length and all elements of the array must be non-null.) An empty string purpose is technically allowed but is discouraged.
Two purposes arguments are equivalent if and only if they contain the same strings (using an ordinal comparer) in the same order. A single purpose argument is equivalent to the corresponding single-element purposes array.
Two IDataProtector objects are equivalent if and only if they are created from equivalent IDataProtectionProvider objects with equivalent purposes parameters.
For a given IDataProtector object, a call to Unprotect(protectedData) will return the original unprotectedData if and only if protectedData := Protect(unprotectedData) for an equivalent IDataProtector object.

Note

We’re not considering the case where some component intentionally chooses a purpose string which is known to conflict with another component. Such a component would essentially be considered malicious, and this system is not intended to provide security guarantees in the event that malicious code is already running inside of the worker process.

Purpose hierarchy and multi-tenancy¶

Since an IDataProtector is also implicitly an IDataProtectionProvider, purposes can be chained together. In this sense provider.CreateProtector([ “purpose1”, “purpose2” ]) is equivalent to provider.CreateProtector(“purpose1”).CreateProtector(“purpose2”).

This allows for some interesting hierarchical relationships through the data protection system. In the earlier example of Contoso.Messaging.SecureMessage, the SecureMessage component can call provider.CreateProtector(“Contoso.Messaging.SecureMessage”) once upfront and cache the result into a private _myProvider field. Future protectors can then be created via calls to _myProvider.CreateProtector(“User: username”), and these protectors would be used for securing the individual messages.

This can also be flipped. Consider a single logical application which hosts multiple tenants (a CMS seems reasonable), and each tenant can be configured with its own authentication and state management system. The umbrella application has a single master provider, and it calls provider.CreateProtector(“Tenant 1”) and provider.CreateProtector(“Tenant 2”) to give each tenant its own isolated slice of the data protection system. The tenants could then derive their own individual protectors based on their own needs, but no matter how hard they try they cannot create protectors which collide with any other tenant in the system. Graphically this is represented as below.

Warning

This assumes the umbrella application controls which APIs are available to individual tenants and that tenants cannot execute arbitrary code on the server. If a tenant can execute arbitrary code, he could perform private reflection to break the isolation guarantees, or he could just read the master keying material directly and derive whatever subkeys he desires.

The data protection system actually uses a sort of multi-tenancy in its default out-of-the-box configuration. By default master keying material is stored in the worker process account’s user profile folder (or the registry, for IIS application pool identities). But it is actually fairly common to use a single account to run multiple applications, and thus all these applications would end up sharing the master keying material. To solve this, the data protection system automatically inserts a unique-per-application identifier as the first element in the overall purpose chain. This implicit purpose serves to isolate individual applications from one another by effectively treating each application as a unique tenant within the system, and the protector creation process looks identical to the image above.

Password Hashing¶

The data protection code base includes a package Microsoft.AspNetCore.Cryptography.KeyDerivation which contains cryptographic key derivation functions. This package is a standalone component and has no dependencies on the rest of the data protection system. It can be used completely independently. The source exists alongside the data protection code base as a convenience.

The package currently offers a method KeyDerivation.Pbkdf2 which allows hashing a password using the PBKDF2 algorithm. This API is very similar to the .NET Framework’s existing Rfc2898DeriveBytes type, but there are three important distinctions:

The KeyDerivation.Pbkdf2 method supports consuming multiple PRFs (currently HMACSHA1, HMACSHA256, and HMACSHA512), whereas the Rfc2898DeriveBytes type only supports HMACSHA1.
The KeyDerivation.Pbkdf2 method detects the current operating system and attempts to choose the most optimized implementation of the routine, providing much better performance in certain cases. (On Windows 8, it offers around 10x the throughput of Rfc2898DeriveBytes.)
The KeyDerivation.Pbkdf2 method requires the caller to specify all parameters (salt, PRF, and iteration count). The Rfc2898DeriveBytes type provides default values for these.

using System;
using System.Security.Cryptography;
using Microsoft.AspNetCore.Cryptography.KeyDerivation;
 
public class Program
{
    public static void Main(string[] args)
    {
        Console.Write("Enter a password: ");
        string password = Console.ReadLine();
 
        // generate a 128-bit salt using a secure PRNG
        byte[] salt = new byte[128 / 8];
        using (var rng = RandomNumberGenerator.Create())
        {
            rng.GetBytes(salt);
        }
        Console.WriteLine($"Salt: {Convert.ToBase64String(salt)}");
 
        // derive a 256-bit subkey (use HMACSHA1 with 10,000 iterations)
        string hashed = Convert.ToBase64String(KeyDerivation.Pbkdf2(
            password: password,
            salt: salt,
            prf: KeyDerivationPrf.HMACSHA1,
            iterationCount: 10000,
            numBytesRequested: 256 / 8));
        Console.WriteLine($"Hashed: {hashed}");
    }
}
 
/*
 * SAMPLE OUTPUT
 *
 * Enter a password: Xtw9NMgx
 * Salt: NZsP6NnmfBuYeJrrAKNuVQ==
 * Hashed: /OOoOer10+tGwTRDTrQSoeCxVTFr6dtYly7d0cPxIak=
 */

See the source code for ASP.NET Core Identity’s PasswordHasher type for a real-world use case.

Limiting the lifetime of protected payloads¶

There are scenarios where the application developer wants to create a protected payload that expires after a set period of time. For instance, the protected payload might represent a password reset token that should only be valid for one hour. It is certainly possible for the developer to create his own payload format that contains an embedded expiration date, and advanced developers may wish to do this anyway, but for the majority of developers managing these expirations can grow tedious.

To make this easier for our developer audience, the package Microsoft.AspNetCore.DataProtection.Extensions contains utility APIs for creating payloads that automatically expire after a set period of time. These APIs hang off of the ITimeLimitedDataProtector type.

API usage¶

The ITimeLimitedDataProtector interface is the core interface for protecting and unprotecting time-limited / self-expiring payloads. To create an instance of an ITimeLimitedDataProtector, you’ll first need an instance of a regular IDataProtector constructed with a specific purpose. Once the IDataProtector instance is available, call the IDataProtector.ToTimeLimitedDataProtector extension method to get back a protector with built-in expiration capabilities.

ITimeLimitedDataProtector exposes the following API surface and extension methods:

CreateProtector(string purpose) : ITimeLimitedDataProtector This API is similar to the existing IDataProtectionProvider.CreateProtector in that it can be used to create purpose chains from a root time-limited protector.
Protect(byte[] plaintext, DateTimeOffset expiration) : byte[]
Protect(byte[] plaintext, TimeSpan lifetime) : byte[]
Protect(byte[] plaintext) : byte[]
Protect(string plaintext, DateTimeOffset expiration) : string
Protect(string plaintext, TimeSpan lifetime) : string
Protect(string plaintext) : string

In addition to the core Protect methods which take only the plaintext, there are new overloads which allow specifying the payload’s expiration date. The expiration date can be specified as an absolute date (via a DateTimeOffset) or as a relative time (from the current system time, via a TimeSpan). If an overload which doesn’t take an expiration is called, the payload is assumed never to expire.

Unprotect(byte[] protectedData, out DateTimeOffset expiration) : byte[]
Unprotect(byte[] protectedData) : byte[]
Unprotect(string protectedData, out DateTimeOffset expiration) : string
Unprotect(string protectedData) : string

The Unprotect methods return the original unprotected data. If the payload hasn’t yet expired, the absolute expiration is returned as an optional out parameter along with the original unprotected data. If the payload is expired, all overloads of the Unprotect method will throw CryptographicException.

Warning

It is not advised to use these APIs to protect payloads which require long-term or indefinite persistence. “Can I afford for the protected payloads to be permanently unrecoverable after a month?” can serve as a good rule of thumb; if the answer is no then developers should consider alternative APIs.

The sample below uses the non-DI code paths for instantiating the data protection system. To run this sample, ensure that you have first added a reference to the Microsoft.AspNetCore.DataProtection.Extensions package.

using System;
using System.IO;
using System.Threading;
using Microsoft.AspNetCore.DataProtection;
 
public class Program
{
    public static void Main(string[] args)
    {
        // create a protector for my application
 
        var provider = DataProtectionProvider.Create(new DirectoryInfo(@"c:\myapp-keys\"));
        var baseProtector = provider.CreateProtector("Contoso.TimeLimitedSample");
 
        // convert the normal protector into a time-limited protector
        var timeLimitedProtector = baseProtector.ToTimeLimitedDataProtector();
 
        // get some input and protect it for five seconds
        Console.Write("Enter input: ");
        string input = Console.ReadLine();
        string protectedData = timeLimitedProtector.Protect(input, lifetime: TimeSpan.FromSeconds(5));
        Console.WriteLine($"Protected data: {protectedData}");
 
        // unprotect it to demonstrate that round-tripping works properly
        string roundtripped = timeLimitedProtector.Unprotect(protectedData);
        Console.WriteLine($"Round-tripped data: {roundtripped}");
 
        // wait 6 seconds and perform another unprotect, demonstrating that the payload self-expires
        Console.WriteLine("Waiting 6 seconds...");
        Thread.Sleep(6000);
        timeLimitedProtector.Unprotect(protectedData);
    }
}
 
/*
 * SAMPLE OUTPUT
 *
 * Enter input: Hello!
 * Protected data: CfDJ8Hu5z0zwxn...nLk7Ok
 * Round-tripped data: Hello!
 * Waiting 6 seconds...
 * <<throws CryptographicException with message 'The payload expired at ...'>>

 */

Unprotecting payloads whose keys have been revoked¶

The ASP.NET Core data protection APIs are not primarily intended for indefinite persistence of confidential payloads. Other technologies like Windows CNG DPAPI and Azure Rights Management are more suited to the scenario of indefinite storage, and they have correspondingly strong key management capabilities. That said, there is nothing prohibiting a developer from using the ASP.NET Core data protection APIs for long-term protection of confidential data. Keys are never removed from the key ring, so IDataProtector.Unprotect can always recover existing payloads as long as the keys are available and valid.

However, an issue arises when the developer tries to unprotect data that has been protected with a revoked key, as IDataProtector.Unprotect will throw an exception in this case. This might be fine for short-lived or transient payloads (like authentication tokens), as these kinds of payloads can easily be recreated by the system, and at worst the site visitor might be required to log in again. But for persisted payloads, having Unprotect throw could lead to unacceptable data loss.

IPersistedDataProtector¶

To support the scenario of allowing payloads to be unprotected even in the face of revoked keys, the data protection system contains an IPersistedDataProtector type. To get an instance of IPersistedDataProtector, simply get an instance of IDataProtector in the normal fashion and try casting the IDataProtector to IPersistedDataProtector.

Note

Not all IDataProtector instances can be cast to IPersistedDataProtector. Developers should use the C# as operator or similar to avoid runtime exceptions caused by invalid casts, and they should be prepared to handle the failure case appropriately.

IPersistedDataProtector exposes the following API surface:

DangerousUnprotect(byte[] protectedData, bool ignoreRevocationErrors,
  out bool requiresMigration, out bool wasRevoked) : byte[]

This API takes the protected payload (as a byte array) and returns the unprotected payload. There is no string-based overload. The two out parameters are as follows.

requiresMigration: will be set to true if the key used to protect this payload is no longer the active default key, e.g., the key used to protect this payload is old and a key rolling operation has since taken place. The caller may wish to consider reprotecting the payload depending on his business needs.
wasRevoked: will be set to true if the key used to protect this payload was revoked.

Warning

Exercise extreme caution when passing ignoreRevocationErrors: true to the DangerousUnprotect method. If after calling this method the wasRevoked value is true, then the key used to protect this payload was revoked, and the payload’s authenticity should be treated as suspect. In this case only continue operating on the unprotected payload if you have some separate assurance that it is authentic, e.g. that it’s coming from a secure database rather than being sent by an untrusted web client.

using System;
using System.IO;
using System.Text;
using Microsoft.AspNetCore.DataProtection;
using Microsoft.AspNetCore.DataProtection.KeyManagement;
using Microsoft.Extensions.DependencyInjection;

public class Program
{
    public static void Main(string[] args)
    {
        var serviceCollection = new ServiceCollection();
        serviceCollection.AddDataProtection()
            // point at a specific folder and use DPAPI to encrypt keys
            .PersistKeysToFileSystem(new DirectoryInfo(@"c:\temp-keys"))
            .ProtectKeysWithDpapi();
        var services = serviceCollection.BuildServiceProvider();

        // get a protector and perform a protect operation
        var protector = services.GetDataProtector("Sample.DangerousUnprotect");
        Console.Write("Input: ");
        byte[] input = Encoding.UTF8.GetBytes(Console.ReadLine());
        var protectedData = protector.Protect(input);
        Console.WriteLine($"Protected payload: {Convert.ToBase64String(protectedData)}");

        // demonstrate that the payload round-trips properly
        var roundTripped = protector.Unprotect(protectedData);
        Console.WriteLine($"Round-tripped payload: {Encoding.UTF8.GetString(roundTripped)}");

        // get a reference to the key manager and revoke all keys in the key ring
        var keyManager = services.GetService<IKeyManager>();
        Console.WriteLine("Revoking all keys in the key ring...");
        keyManager.RevokeAllKeys(DateTimeOffset.Now, "Sample revocation.");

        // try calling Protect - this should throw
        Console.WriteLine("Calling Unprotect...");
        try
        {
            var unprotectedPayload = protector.Unprotect(protectedData);
            Console.WriteLine($"Unprotected payload: {Encoding.UTF8.GetString(unprotectedPayload)}");
        }
        catch (Exception ex)
        {
            Console.WriteLine($"{ex.GetType().Name}: {ex.Message}");
        }

        // try calling DangerousUnprotect
        Console.WriteLine("Calling DangerousUnprotect...");
        try
        {
            IPersistedDataProtector persistedProtector = protector as IPersistedDataProtector;
            if (persistedProtector == null)
            {
                throw new Exception("Can't call DangerousUnprotect.");
            }

            bool requiresMigration, wasRevoked;
            var unprotectedPayload = persistedProtector.DangerousUnprotect(
                protectedData: protectedData,
                ignoreRevocationErrors: true,
                requiresMigration: out requiresMigration,
                wasRevoked: out wasRevoked);
            Console.WriteLine($"Unprotected payload: {Encoding.UTF8.GetString(unprotectedPayload)}");
            Console.WriteLine($"Requires migration = {requiresMigration}, was revoked = {wasRevoked}");
        }
        catch (Exception ex)
        {
            Console.WriteLine($"{ex.GetType().Name}: {ex.Message}");
        }
    }
}

/*
 * SAMPLE OUTPUT
 *
 * Input: Hello!
 * Protected payload: CfDJ8LHIzUCX1ZVBn2BZ...
 * Round-tripped payload: Hello!
 * Revoking all keys in the key ring...
 * Calling Unprotect...
 * CryptographicException: The key {...} has been revoked.
 * Calling DangerousUnprotect...
 * Unprotected payload: Hello!
 * Requires migration = True, was revoked = True
 */

Configuration¶

Configuring Data Protection¶

When the data protection system is initialized it applies some default settings based on the operational environment. These settings are generally good for applications running on a single machine. There are some cases where a developer may want to change these (perhaps because his application is spread across multiple machines or for compliance reasons), and for these scenarios the data protection system offers a rich configuration API.

There is an extension method AddDataProtection which returns an IDataProtectionBuilder which itself exposes extension methods that you can chain together to configure various data protection options. For instance, to store keys at a UNC share instead of %LOCALAPPDATA% (the default), configure the system as follows:

public void ConfigureServices(IServiceCollection services)
{
    services.AddDataProtection()
        .PersistKeysToFileSystem(new DirectoryInfo(@"\\server\share\directory\"));

}

Warning

If you change the key persistence location, the system will no longer automatically encrypt keys at rest since it doesn’t know whether DPAPI is an appropriate encryption mechanism.

You can configure the system to protect keys at rest by calling any of the ProtectKeysWith* configuration APIs. Consider the example below, which stores keys at a UNC share and encrypts those keys at rest with a specific X.509 certificate.

public void ConfigureServices(IServiceCollection services)
  {
      services.AddDataProtection()
          .PersistKeysToFileSystem(new DirectoryInfo(@"\\server\share\directory\"))
          .ProtectKeysWithCertificate("thumbprint");
  }

See key encryption at rest for more examples and for discussion on the built-in key encryption mechanisms.

To configure the system to use a default key lifetime of 14 days instead of 90 days, consider the following example:

public void ConfigureServices(IServiceCollection services)
{
    services.AddDataProtection()
        .SetDefaultKeyLifetime(TimeSpan.FromDays(14));
}

By default the data protection system isolates applications from one another, even if they’re sharing the same physical key repository. This prevents the applications from understanding each other’s protected payloads. To share protected payloads between two different applications, configure the system passing in the same application name for both applications as in the below example:

public void ConfigureServices(IServiceCollection services)
{
    services.AddDataProtection()
        .SetApplicationName("my application");
}

Finally, you may have a scenario where you do not want an application to automatically roll keys as they approach expiration. One example of this might be applications set up in a primary / secondary relationship, where only the primary application is responsible for key management concerns, and all secondary applications simply have a read-only view of the key ring. The secondary applications can be configured to treat the key ring as read-only by configuring the system as below:

public void ConfigureServices(IServiceCollection services)
{
  services.AddDataProtection()
      .DisableAutomaticKeyGeneration();
}

Per-application isolation¶

When the data protection system is provided by an ASP.NET Core host, it will automatically isolate applications from one another, even if those applications are running under the same worker process account and are using the same master keying material. This is somewhat similar to the IsolateApps modifier from System.Web’s <machineKey> element.

The isolation mechanism works by considering each application on the local machine as a unique tenant, thus the IDataProtector rooted for any given application automatically includes the application ID as a discriminator. The application’s unique ID comes from one of two places.

If the application is hosted in IIS, the unique identifier is the application’s configuration path. If an application is deployed in a farm environment, this value should be stable assuming that the IIS environments are configured similarly across all machines in the farm.
If the application is not hosted in IIS, the unique identifier is the physical path of the application.

The unique identifier is designed to survive resets - both of the individual application and of the machine itself.

This isolation mechanism assumes that the applications are not malicious. A malicious application can always impact any other application running under the same worker process account. In a shared hosting environment where applications are mutually untrusted, the hosting provider should take steps to ensure OS-level isolation between applications, including separating the applications’ underlying key repositories.

If the data protection system is not provided by an ASP.NET Core host (e.g., if the developer instantiates it himself via the DataProtectionProvider concrete type), application isolation is disabled by default, and all applications backed by the same keying material can share payloads as long as they provide the appropriate purposes. To provide application isolation in this environment, call the SetApplicationName method on the configuration object, see the code sample above.

Changing algorithms¶

The data protection stack allows changing the default algorithm used by newly-generated keys. The simplest way to do this is to call UseCryptographicAlgorithms from the configuration callback, as in the below example.

services.AddDataProtection()
    .UseCryptographicAlgorithms(new AuthenticatedEncryptionSettings()
    {
        EncryptionAlgorithm = EncryptionAlgorithm.AES_256_CBC,
        ValidationAlgorithm = ValidationAlgorithm.HMACSHA256
    });

The default EncryptionAlgorithm and ValidationAlgorithm are AES-256-CBC and HMACSHA256, respectively. The default policy can be set by a system administrator via Machine Wide Policy, but an explicit call to UseCryptographicAlgorithms will override the default policy.

Calling UseCryptographicAlgorithms will allow the developer to specify the desired algorithm (from a predefined built-in list), and the developer does not need to worry about the implementation of the algorithm. For instance, in the scenario above the data protection system will attempt to use the CNG implementation of AES if running on Windows, otherwise it will fall back to the managed System.Security.Cryptography.Aes class.

The developer can manually specify an implementation if desired via a call to UseCustomCryptographicAlgorithms, as show in the below examples.

Tip

Changing algorithms does not affect existing keys in the key ring. It only affects newly-generated keys.

Specifying custom managed algorithms¶

To specify custom managed algorithms, create a ManagedAuthenticatedEncryptionSettings instance that points to the implementation types.

serviceCollection.AddDataProtection()
    .UseCustomCryptographicAlgorithms(new ManagedAuthenticatedEncryptionSettings()
    {
        // a type that subclasses SymmetricAlgorithm
        EncryptionAlgorithmType = typeof(Aes),

        // specified in bits
        EncryptionAlgorithmKeySize = 256,

        // a type that subclasses KeyedHashAlgorithm
        ValidationAlgorithmType = typeof(HMACSHA256)
    });

Generally the *Type properties must point to concrete, instantiable (via a public parameterless ctor) implementations of SymmetricAlgorithm and KeyedHashAlgorithm, though the system special-cases some values like typeof(Aes) for convenience.

Note

The SymmetricAlgorithm must have a key length of ≥ 128 bits and a block size of ≥ 64 bits, and it must support CBC-mode encryption with PKCS #7 padding. The KeyedHashAlgorithm must have a digest size of >= 128 bits, and it must support keys of length equal to the hash algorithm’s digest length. The KeyedHashAlgorithm is not strictly required to be HMAC.

Specifying custom Windows CNG algorithms¶

To specify a custom Windows CNG algorithm using CBC-mode encryption + HMAC validation, create a CngCbcAuthenticatedEncryptionSettings instance that contains the algorithmic information.

services.AddDataProtection()
    .UseCustomCryptographicAlgorithms(new CngCbcAuthenticatedEncryptionSettings()
    {
        // passed to BCryptOpenAlgorithmProvider
        EncryptionAlgorithm = "AES",
        EncryptionAlgorithmProvider = null,

        // specified in bits
        EncryptionAlgorithmKeySize = 256,

        // passed to BCryptOpenAlgorithmProvider
        HashAlgorithm = "SHA256",
        HashAlgorithmProvider = null
    });

Note

The symmetric block cipher algorithm must have a key length of ≥ 128 bits and a block size of ≥ 64 bits, and it must support CBC-mode encryption with PKCS #7 padding. The hash algorithm must have a digest size of >= 128 bits and must support being opened with the BCRYPT_ALG_HANDLE_HMAC_FLAG flag. The *Provider properties can be set to null to use the default provider for the specified algorithm. See the BCryptOpenAlgorithmProvider documentation for more information.

To specify a custom Windows CNG algorithm using Galois/Counter Mode encryption + validation, create a CngGcmAuthenticatedEncryptionSettings instance that contains the algorithmic information.

services.AddDataProtection()
    .UseCustomCryptographicAlgorithms(new CngGcmAuthenticatedEncryptionSettings()
    {
        // passed to BCryptOpenAlgorithmProvider
        EncryptionAlgorithm = "AES",
        EncryptionAlgorithmProvider = null,

        // specified in bits
        EncryptionAlgorithmKeySize = 256
    });
});

Note

The symmetric block cipher algorithm must have a key length of ≥ 128 bits and a block size of exactly 128 bits, and it must support GCM encryption. The EncryptionAlgorithmProvider property can be set to null to use the default provider for the specified algorithm. See the BCryptOpenAlgorithmProvider documentation for more information.

Specifying other custom algorithms¶

Though not exposed as a first-class API, the data protection system is extensible enough to allow specifying almost any kind of algorithm. For example, it is possible to keep all keys contained within an HSM and to provide a custom implementation of the core encryption and decryption routines. See IAuthenticatedEncryptorConfiguration in the core cryptography extensibility section for more information.

Default Settings¶

Key Management¶

The system tries to detect its operational environment and provide good zero-configuration behavioral defaults. The heuristic used is as follows.

If the system is being hosted in Azure Web Sites, keys are persisted to the “%HOME%\ASP.NET\DataProtection-Keys” folder. This folder is backed by network storage and is synchronized across all machines hosting the application. Keys are not protected at rest.
If the user profile is available, keys are persisted to the “%LOCALAPPDATA%\ASP.NET\DataProtection-Keys” folder. Additionally, if the operating system is Windows, they’ll be encrypted at rest using DPAPI.
If the application is hosted in IIS, keys are persisted to the HKLM registry in a special registry key that is ACLed only to the worker process account. Keys are encrypted at rest using DPAPI.
If none of these conditions matches, keys are not persisted outside of the current process. When the process shuts down, all generated keys will be lost.

The developer is always in full control and can override how and where keys are stored. The first three options above should good defaults for most applications similar to how the ASP.NET <machineKey> auto-generation routines worked in the past. The final, fall back option is the only scenario that truly requires the developer to specify configuration upfront if he wants key persistence, but this fall-back would only occur in rare situations.

Warning

If the developer overrides this heuristic and points the data protection system at a specific key repository, automatic encryption of keys at rest will be disabled. At rest protection can be re-enabled via configuration.

Key Lifetime¶

Keys by default have a 90-day lifetime. When a key expires, the system will automatically generate a new key and set the new key as the active key. As long as retired keys remain on the system you will still be able to decrypt any data protected with them. See key lifetime for more information.

Default Algorithms¶

The default payload protection algorithm used is AES-256-CBC for confidentiality and HMACSHA256 for authenticity. A 512-bit master key, rolled every 90 days, is used to derive the two sub-keys used for these algorithms on a per-payload basis. See subkey derivation for more information.

Machine Wide Policy¶

When running on Windows, the data protection system has limited support for setting default machine-wide policy for all applications which consume data protection. The general idea is that an administrator might wish to change some default setting (such as algorithms used or key lifetime) without needing to manually update every application on the machine.

Warning

The system administrator can set default policy, but he cannot enforce it. The application developer can always override any value with one of his own choosing. The default policy only affects applications where the developer has not specified an explicit value for some particular setting.

Setting default policy¶

To set default policy, an administrator can set known values in the system registry under the following key.

Reg key: HKLM\SOFTWARE\Microsoft\DotNetPackages\Microsoft.AspNetCore.DataProtection

If you’re on a 64-bit operating system and want to affect the behavior of 32-bit applications, remember also to configure the Wow6432Node equivalent of the above key.

The supported values are:

EncryptionType [string] - specifies which algorithms should be used for data protection. This value must be “CNG-CBC”, “CNG-GCM”, or “Managed” and is described in more detail below.
DefaultKeyLifetime [DWORD] - specifies the lifetime for newly-generated keys. This value is specified in days and must be ≥ 7.
KeyEscrowSinks [string] - specifies the types which will be used for key escrow. This value is a semicolon-delimited list of key escrow sinks, where each element in the list is the assembly qualified name of a type which implements IKeyEscrowSink.

Encryption types¶

If EncryptionType is “CNG-CBC”, the system will be configured to use a CBC-mode symmetric block cipher for confidentiality and HMAC for authenticity with services provided by Windows CNG (see Specifying custom Windows CNG algorithms for more details). The following additional values are supported, each of which corresponds to a property on the CngCbcAuthenticatedEncryptionSettings type:

EncryptionAlgorithm [string] - the name of a symmetric block cipher algorithm understood by CNG. This algorithm will be opened in CBC mode.
EncryptionAlgorithmProvider [string] - the name of the CNG provider implementation which can produce the algorithm EncryptionAlgorithm.
EncryptionAlgorithmKeySize [DWORD] - the length (in bits) of the key to derive for the symmetric block cipher algorithm.
HashAlgorithm [string] - the name of a hash algorithm understood by CNG. This algorithm will be opened in HMAC mode.
HashAlgorithmProvider [string] - the name of the CNG provider implementation which can produce the algorithm HashAlgorithm.

If EncryptionType is “CNG-GCM”, the system will be configured to use a Galois/Counter Mode symmetric block cipher for confidentiality and authenticity with services provided by Windows CNG (see Specifying custom Windows CNG algorithms for more details). The following additional values are supported, each of which corresponds to a property on the CngGcmAuthenticatedEncryptionSettings type:

EncryptionAlgorithm [string] - the name of a symmetric block cipher algorithm understood by CNG. This algorithm will be opened in Galois/Counter Mode.
EncryptionAlgorithmProvider [string] - the name of the CNG provider implementation which can produce the algorithm EncryptionAlgorithm.
EncryptionAlgorithmKeySize [DWORD] - the length (in bits) of the key to derive for the symmetric block cipher algorithm.

If EncryptionType is “Managed”, the system will be configured to use a managed SymmetricAlgorithm for confidentiality and KeyedHashAlgorithm for authenticity (see Specifying custom managed algorithms for more details). The following additional values are supported, each of which corresponds to a property on the ManagedAuthenticatedEncryptionSettings type:

EncryptionAlgorithmType [string] - the assembly-qualified name of a type which implements SymmetricAlgorithm.
EncryptionAlgorithmKeySize [DWORD] - the length (in bits) of the key to derive for the symmetric encryption algorithm.
ValidationAlgorithmType [string] - the assembly-qualified name of a type which implements KeyedHashAlgorithm.

If EncryptionType has any other value (other than null / empty), the data protection system will throw an exception at startup.

Warning

When configuring a default policy setting that involves type names (EncryptionAlgorithmType, ValidationAlgorithmType, KeyEscrowSinks), the types must be available to the application. In practice, this means that for applications running on Desktop CLR, the assemblies which contain these types should be GACed. For ASP.NET Core applications running on .NET Core, the packages which contain these types should be referenced in project.json.

Non DI Aware Scenarios¶

The data protection system is normally designed to be added to a service container and to be provided to dependent components via a DI mechanism. However, there may be some cases where this is not feasible, especially when importing the system into an existing application.

To support these scenarios the package Microsoft.AspNetCore.DataProtection.Extensions provides a concrete type DataProtectionProvider which offers a simple way to use the data protection system without going through DI-specific code paths. The type itself implements IDataProtectionProvider, and constructing it is as easy as providing a DirectoryInfo where this provider’s cryptographic keys should be stored.

For example:

using System;
using System.IO;
using Microsoft.AspNetCore.DataProtection;

public class Program
{
    public static void Main(string[] args)
    {
        // get the path to %LOCALAPPDATA%\myapp-keys
        string destFolder = Path.Combine(
            Environment.GetEnvironmentVariable("LOCALAPPDATA"),
            "myapp-keys");

        // instantiate the data protection system at this folder
        var dataProtectionProvider = DataProtectionProvider.Create(
            new DirectoryInfo(destFolder));

        var protector = dataProtectionProvider.CreateProtector("Program.No-DI");
        Console.Write("Enter input: ");
        string input = Console.ReadLine();

        // protect the payload
        string protectedPayload = protector.Protect(input);
        Console.WriteLine($"Protect returned: {protectedPayload}");

        // unprotect the payload
        string unprotectedPayload = protector.Unprotect(protectedPayload);
        Console.WriteLine($"Unprotect returned: {unprotectedPayload}");
    }
}

/*
 * SAMPLE OUTPUT
 *
 * Enter input: Hello world!
 * Protect returned: CfDJ8FWbAn6...ch3hAPm1NJA
 * Unprotect returned: Hello world!
 */

Warning

By default the DataProtectionProvider concrete type does not encrypt raw key material before persisting it to the file system. This is to support scenarios where the developer points to a network share, in which case the data protection system cannot automatically deduce an appropriate at-rest key encryption mechanism.

Additionally, the DataProtectionProvider concrete type does not isolate applications by default, so all applications pointed at the same key directory can share payloads as long as their purpose parameters match.

The application developer can address both of these if desired. The DataProtectionProvider constructor accepts an optional configuration callback which can be used to tweak the behaviors of the system. The sample below demonstrates restoring isolation via an explicit call to SetApplicationName, and it also demonstrates configuring the system to automatically encrypt persisted keys using Windows DPAPI. If the directory points to a UNC share, you may wish to distribute a shared certificate across all relevant machines and to configure the system to use certificate-based encryption instead via a call to ProtectKeysWithCertificate.

using System;
using System.IO;
using Microsoft.AspNetCore.DataProtection;

public class Program
{
    public static void Main(string[] args)
    {
        // get the path to %LOCALAPPDATA%\myapp-keys
        string destFolder = Path.Combine(
            Environment.GetEnvironmentVariable("LOCALAPPDATA"),
            "myapp-keys");

        // instantiate the data protection system at this folder
        var dataProtectionProvider = DataProtectionProvider.Create(
            new DirectoryInfo(destFolder),
            configuration =>
            {
                configuration.SetApplicationName("my app name");
                configuration.ProtectKeysWithDpapi();
            });

        var protector = dataProtectionProvider.CreateProtector("Program.No-DI");
        Console.Write("Enter input: ");
        string input = Console.ReadLine();

        // protect the payload
        string protectedPayload = protector.Protect(input);
        Console.WriteLine($"Protect returned: {protectedPayload}");

        // unprotect the payload
        string unprotectedPayload = protector.Unprotect(protectedPayload);
        Console.WriteLine($"Unprotect returned: {unprotectedPayload}");
    }
}

Tip

Instances of the DataProtectionProvider concrete type are expensive to create. If an application maintains multiple instances of this type and if they’re all pointing at the same key storage directory, application performance may be degraded. The intended usage is that the application developer instantiate this type once then keep reusing this single reference as much as possible. The DataProtectionProvider type and all IDataProtector instances created from it are thread-safe for multiple callers.

Extensibility APIs¶

Core cryptography extensibility¶

Warning

Types that implement any of the following interfaces should be thread-safe for multiple callers.

IAuthenticatedEncryptor¶

The IAuthenticatedEncryptor interface is the basic building block of the cryptographic subsystem. There is generally one IAuthenticatedEncryptor per key, and the IAuthenticatedEncryptor instance wraps all cryptographic key material and algorithmic information necessary to perform cryptographic operations.

As its name suggests, the type is responsible for providing authenticated encryption and decryption services. It exposes the following two APIs.

Decrypt(ArraySegment<byte> ciphertext, ArraySegment<byte> additionalAuthenticatedData) : byte[]
Encrypt(ArraySegment<byte> plaintext, ArraySegment<byte> additionalAuthenticatedData) : byte[]

The Encrypt method returns a blob that includes the enciphered plaintext and an authentication tag. The authentication tag must encompass the additional authenticated data (AAD), though the AAD itself need not be recoverable from the final payload. The Decrypt method validates the authentication tag and returns the deciphered payload. All failures (except ArgumentNullException and similar) should be homogenized to CryptographicException.

Note

The IAuthenticatedEncryptor instance itself doesn’t actually need to contain the key material. For example, the implementation could delegate to an HSM for all operations.

IAuthenticatedEncryptorDescriptor¶

The IAuthenticatedEncryptorDescriptor interface represents a type that knows how to create an IAuthenticatedEncryptor instance. Its API is as follows.

CreateEncryptorInstance() : IAuthenticatedEncryptor
ExportToXml() : XmlSerializedDescriptorInfo

Like IAuthenticatedEncryptor, an instance of IAuthenticatedEncryptorDescriptor is assumed to wrap one specific key. This means that for any given IAuthenticatedEncryptorDescriptor instance, any authenticated encryptors created by its CreateEncryptorInstance method should be considered equivalent, as in the below code sample.

// we have an IAuthenticatedEncryptorDescriptor instance
IAuthenticatedEncryptorDescriptor descriptor = ...;

// get an encryptor instance and perform an authenticated encryption operation
ArraySegment<byte> plaintext = new ArraySegment<byte>(Encoding.UTF8.GetBytes("plaintext"));
ArraySegment<byte> aad = new ArraySegment<byte>(Encoding.UTF8.GetBytes("AAD"));
var encryptor1 = descriptor.CreateEncryptorInstance();
byte[] ciphertext = encryptor1.Encrypt(plaintext, aad);

// get another encryptor instance and perform an authenticated decryption operation
var encryptor2 = descriptor.CreateEncryptorInstance();
byte[] roundTripped = encryptor2.Decrypt(new ArraySegment<byte>(ciphertext), aad);


// the 'roundTripped' and 'plaintext' buffers should be equivalent

XML Serialization¶

The primary difference between IAuthenticatedEncryptor and IAuthenticatedEncryptorDescriptor is that the descriptor knows how to create the encryptor and supply it with valid arguments. Consider an IAuthenticatedEncryptor whose implementation relies on SymmetricAlgorithm and KeyedHashAlgorithm. The encryptor’s job is to consume these types, but it doesn’t necessarily know where these types came from, so it can’t really write out a proper description of how to recreate itself if the application restarts. The descriptor acts as a higher level on top of this. Since the descriptor knows how to create the encryptor instance (e.g., it knows how to create the required algorithms), it can serialize that knowledge in XML form so that the encryptor instance can be recreated after an application reset.

The descriptor can be serialized via its ExportToXml routine. This routine returns an XmlSerializedDescriptorInfo which contains two properties: the XElement representation of the descriptor and the Type which represents an IAuthenticatedEncryptorDescriptorDeserializer which can be used to resurrect this descriptor given the corresponding XElement.

The serialized descriptor may contain sensitive information such as cryptographic key material. The data protection system has built-in support for encrypting information before it’s persisted to storage. To take advantage of this, the descriptor should mark the element which contains sensitive information with the attribute name “requiresEncryption” (xmlns “http://schemas.asp.net/2015/03/dataProtection”), value “true”.

Tip

There’s a helper API for setting this attribute. Call the extension method XElement.MarkAsRequiresEncryption() located in namespace Microsoft.AspNetCore.DataProtection.AuthenticatedEncryption.ConfigurationModel.

There can also be cases where the serialized descriptor doesn’t contain sensitive information. Consider again the case of a cryptographic key stored in an HSM. The descriptor cannot write out the key material when serializing itself since the HSM will not expose the material in plaintext form. Instead, the descriptor might write out the key-wrapped version of the key (if the HSM allows export in this fashion) or the HSM’s own unique identifier for the key.

IAuthenticatedEncryptorDescriptorDeserializer¶

The IAuthenticatedEncryptorDescriptorDeserializer interface represents a type that knows how to deserialize an IAuthenticatedEncryptorDescriptor instance from an XElement. It exposes a single method:

ImportFromXml(XElement element) : IAuthenticatedEncryptorDescriptor

The ImportFromXml method takes the XElement that was returned by IAuthenticatedEncryptorDescriptor.ExportToXml and creates an equivalent of the original IAuthenticatedEncryptorDescriptor.

Types which implement IAuthenticatedEncryptorDescriptorDeserializer should have one of the following two public constructors:

.ctor(IServiceProvider)
.ctor()

Note

The IServiceProvider passed to the constructor may be null.

IAuthenticatedEncryptorConfiguration¶

The IAuthenticatedEncryptorConfiguration interface represents a type which knows how to create IAuthenticatedEncryptorDescriptor instances. It exposes a single API.

CreateNewDescriptor() : IAuthenticatedEncryptorDescriptor

Think of IAuthenticatedEncryptorConfiguration as the top-level factory. The configuration serves as a template. It wraps algorithmic information (e.g., this configuration produces descriptors with an AES-128-GCM master key), but it is not yet associated with a specific key.

When CreateNewDescriptor is called, fresh key material is created solely for this call, and a new IAuthenticatedEncryptorDescriptor is produced which wraps this key material and the algorithmic information required to consume the material. The key material could be created in software (and held in memory), it could be created and held within an HSM, and so on. The crucial point is that any two calls to CreateNewDescriptor should never create equivalent IAuthenticatedEncryptorDescriptor instances.

The IAuthenticatedEncryptorConfiguration type serves as the entry point for key creation routines such as automatic key rolling. To change the implementation for all future keys, register a singleton IAuthenticatedEncryptorConfiguration in the service container.

Key management extensibility¶

Tip

Read the key management section before reading this section, as it explains some of the fundamental concepts behind these APIs.

Warning

Types that implement any of the following interfaces should be thread-safe for multiple callers.

Key¶

The IKey interface is the basic representation of a key in cryptosystem. The term key is used here in the abstract sense, not in the literal sense of “cryptographic key material”. A key has the following properties:

Activation, creation, and expiration dates
Revocation status
Key identifier (a GUID)

Additionally, IKey exposes a CreateEncryptorInstance method which can be used to create an IAuthenticatedEncryptor instance tied to this key.

Note

There is no API to retrieve the raw cryptographic material from an IKey instance.

IKeyManager¶

The IKeyManager interface represents an object responsible for general key storage, retrieval, and manipulation. It exposes three high-level operations:

Create a new key and persist it to storage.
Get all keys from storage.
Revoke one or more keys and persist the revocation information to storage.

Warning

Writing an IKeyManager is a very advanced task, and the majority of developers should not attempt it. Instead, most developers should take advantage of the facilities offered by the XmlKeyManager class.

XmlKeyManager¶

The XmlKeyManager type is the in-box concrete implementation of IKeyManager. It provides several useful facilities, including key escrow and encryption of keys at rest. Keys in this system are represented as XML elements (specifically, XElement).

XmlKeyManager depends on several other components in the course of fulfilling its tasks:

IAuthenticatedEncryptorConfiguration, which dictates the algorithms used by new keys.
IXmlRepository, which controls where keys are persisted in storage.
IXmlEncryptor [optional], which allows encrypting keys at rest.
IKeyEscrowSink [optional], which provides key escrow services.

Below are high-level diagrams which indicate how these components are wired together within XmlKeyManager.

Key Creation / CreateNewKey

In the implementation of CreateNewKey, the IAuthenticatedEncryptorConfiguration component is used to create a unique IAuthenticatedEncryptorDescriptor, which is then serialized as XML. If a key escrow sink is present, the raw (unencrypted) XML is provided to the sink for long-term storage. The unencrypted XML is then run through an IXmlEncryptor (if required) to generate the encrypted XML document. This encrypted document is persisted to long-term storage via the IXmlRepository. (If no IXmlEncryptor is configured, the unencrypted document is persisted in the IXmlRepository.)

Key Retrieval / GetAllKeys

In the implementation of GetAllKeys, the XML documents representing keys and revocations are read from the underlying IXmlRepository. If these documents are encrypted, the system will automatically decrypt them. XmlKeyManager creates the appropriate IAuthenticatedEncryptorDescriptorDeserializer instances to deserialize the documents back into IAuthenticatedEncryptorDescriptor instances, which are then wrapped in individual IKey instances. This collection of IKey instances is returned to the caller.

Further information on the particular XML elements can be found in the key storage format document.

IXmlRepository¶

The IXmlRepository interface represents a type that can persist XML to and retrieve XML from a backing store. It exposes two APIs:

GetAllElements() : IReadOnlyCollection<XElement>
StoreElement(XElement element, string friendlyName)

Implementations of IXmlRepository don’t need to parse the XML passing through them. They should treat the XML documents as opaque and let higher layers worry about generating and parsing the documents.

There are two built-in concrete types which implement IXmlRepository: FileSystemXmlRepository and RegistryXmlRepository. See the key storage providers document for more information. Registering a custom IXmlRepository would be the appropriate manner to use a different backing store, e.g., Azure Blob Storage. To change the default repository application-wide, register a custom singleton IXmlRepository in the service provider.

IXmlEncryptor¶

The IXmlEncryptor interface represents a type that can encrypt a plaintext XML element. It exposes a single API:

Encrypt(XElement plaintextElement) : EncryptedXmlInfo

If a serialized IAuthenticatedEncryptorDescriptor contains any elements marked as “requires encryption”, then XmlKeyManager will run those elements through the configured IXmlEncryptor’s Encrypt method, and it will persist the enciphered element rather than the plaintext element to the IXmlRepository. The output of the Encrypt method is an EncryptedXmlInfo object. This object is a wrapper which contains both the resultant enciphered XElement and the Type which represents an IXmlDecryptor which can be used to decipher the corresponding element.

There are four built-in concrete types which implement IXmlEncryptor: CertificateXmlEncryptor, DpapiNGXmlEncryptor, DpapiXmlEncryptor, and NullXmlEncryptor. See the key encryption at rest document for more information. To change the default key-encryption-at-rest mechanism application-wide, register a custom singleton IXmlEncryptor in the service provider.

IXmlDecryptor¶

The IXmlDecryptor interface represents a type that knows how to decrypt an XElement that was enciphered via an IXmlEncryptor. It exposes a single API:

Decrypt(XElement encryptedElement) : XElement

The Decrypt method undoes the encryption performed by IXmlEncryptor.Encrypt. Generally each concrete IXmlEncryptor implementation will have a corresponding concrete IXmlDecryptor implementation.

Types which implement IXmlDecryptor should have one of the following two public constructors:

.ctor(IServiceProvider)
.ctor()

Note

The IServiceProvider passed to the constructor may be null.

IKeyEscrowSink¶

The IKeyEscrowSink interface represents a type that can perform escrow of sensitive information. Recall that serialized descriptors might contain sensitive information (such as cryptographic material), and this is what led to the introduction of the IXmlEncryptor type in the first place. However, accidents happen, and keyrings can be deleted or become corrupted.

The escrow interface provides an emergency escape hatch, allowing access to the raw serialized XML before it is transformed by any configured IXmlEncryptor. The interface exposes a single API:

Store(Guid keyId, XElement element)

It is up to the IKeyEscrowSink implementation to handle the provided element in a secure manner consistent with business policy. One possible implementation could be for the escrow sink to encrypt the XML element using a known corporate X.509 certificate where the certificate’s private key has been escrowed; the CertificateXmlEncryptor type can assist with this. The IKeyEscrowSink implementation is also responsible for persisting the provided element appropriately.

By default no escrow mechanism is enabled, though server administrators can configure this globally. It can also be configured programmatically via the IDataProtectionBuilder.AddKeyEscrowSink method as shown in the sample below. The AddKeyEscrowSink method overloads mirror the IServiceCollection.AddSingleton and IServiceCollection.AddInstance overloads, as IKeyEscrowSink instances are intended to be singletons. If multiple IKeyEscrowSink instances are registered, each one will be called during key generation, so keys can be escrowed to multiple mechanisms simultaneously.

There is no API to read material from an IKeyEscrowSink instance. This is consistent with the design theory of the escrow mechanism: it’s intended to make the key material accessible to a trusted authority, and since the application is itself not a trusted authority, it shouldn’t have access to its own escrowed material.

The following sample code demonstrates creating and registering an IKeyEscrowSink where keys are escrowed such that only members of “CONTOSODomain Admins” can recover them.

Note

To run this sample, you must be on a domain-joined Windows 8 / Windows Server 2012 machine, and the domain controller must be Windows Server 2012 or later.

using System;
using System.IO;
using System.Xml.Linq;
using Microsoft.AspNetCore.DataProtection;
using Microsoft.AspNetCore.DataProtection.KeyManagement;
using Microsoft.AspNetCore.DataProtection.XmlEncryption;
using Microsoft.Extensions.DependencyInjection;

public class Program
{
    public static void Main(string[] args)
    {
        var serviceCollection = new ServiceCollection();
        serviceCollection.AddDataProtection()
            .PersistKeysToFileSystem(new DirectoryInfo(@"c:\temp-keys"))
            .ProtectKeysWithDpapi()
            .AddKeyEscrowSink(sp => new MyKeyEscrowSink(sp));
        var services = serviceCollection.BuildServiceProvider();

        // get a reference to the key manager and force a new key to be generated
        Console.WriteLine("Generating new key...");
        var keyManager = services.GetService<IKeyManager>();
        keyManager.CreateNewKey(
            activationDate: DateTimeOffset.Now,
            expirationDate: DateTimeOffset.Now.AddDays(7));
    }

    // A key escrow sink where keys are escrowed such that they
    // can be read by members of the CONTOSO\Domain Admins group.
    private class MyKeyEscrowSink : IKeyEscrowSink
    {
        private readonly IXmlEncryptor _escrowEncryptor;

        public MyKeyEscrowSink(IServiceProvider services)
        {
            // Assuming I'm on a machine that's a member of the CONTOSO
            // domain, I can use the Domain Admins SID to generate an
            // encrypted payload that only they can read. Sample SID from
            // https://technet.microsoft.com/en-us/library/cc778824(v=ws.10).aspx.
            _escrowEncryptor = new DpapiNGXmlEncryptor(
                "SID=S-1-5-21-1004336348-1177238915-682003330-512",
                DpapiNGProtectionDescriptorFlags.None,
                services);
        }

        public void Store(Guid keyId, XElement element)
        {
            // Encrypt the key element to the escrow encryptor.
            var encryptedXmlInfo = _escrowEncryptor.Encrypt(element);

            // A real implementation would save the escrowed key to a
            // write-only file share or some other stable storage, but
            // in this sample we'll just write it out to the console.
            Console.WriteLine($"Escrowing key {keyId}");
            Console.WriteLine(encryptedXmlInfo.EncryptedElement);

            // Note: We cannot read the escrowed key material ourselves.
            // We need to get a member of CONTOSO\Domain Admins to read
            // it for us in the event we need to recover it.
        }
    }
}

/*
 * SAMPLE OUTPUT
 *
 * Generating new key...
 * Escrowing key 38e74534-c1b8-4b43-aea1-79e856a822e5
 * <encryptedKey>
 *   <!-- This key is encrypted with Windows DPAPI-NG. -->
 *   <!-- Rule: SID=S-1-5-21-1004336348-1177238915-682003330-512 -->
 *   <value>MIIIfAYJKoZIhvcNAQcDoIIIbTCCCGkCAQ...T5rA4g==</value>
 * </encryptedKey>
 */

Miscellaneous APIs¶

Warning

Types that implement any of the following interfaces should be thread-safe for multiple callers.

ISecret¶

The ISecret interface represents a secret value, such as cryptographic key material. It contains the following API surface.

Length : int
Dispose() : void
WriteSecretIntoBuffer(ArraySegment<byte> buffer) : void

The WriteSecretIntoBuffer method populates the supplied buffer with the raw secret value. The reason this API takes the buffer as a parameter rather than returning a byte[] directly is that this gives the caller the opportunity to pin the buffer object, limiting secret exposure to the managed garbage collector.

The Secret type is a concrete implementation of ISecret where the secret value is stored in in-process memory. On Windows platforms, the secret value is encrypted via CryptProtectMemory.

Implementation¶

Authenticated encryption details.¶

Calls to IDataProtector.Protect are authenticated encryption operations. The Protect method offers both confidentiality and authenticity, and it is tied to the purpose chain that was used to derive this particular IDataProtector instance from its root IDataProtectionProvider.

IDataProtector.Protect takes a byte[] plaintext parameter and produces a byte[] protected payload, whose format is described below. (There is also an extension method overload which takes a string plaintext parameter and returns a string protected payload. If this API is used the protected payload format will still have the below structure, but it will be base64url-encoded.)

Protected payload format¶

The protected payload format consists of three primary components:

A 32-bit magic header that identifies the version of the data protection system.
A 128-bit key id that identifies the key used to protect this particular payload.
The remainder of the protected payload is specific to the encryptor encapsulated by this key. In the example below the key represents an AES-256-CBC + HMACSHA256 encryptor, and the payload is further subdivided as follows: * A 128-bit key modifier. * A 128-bit initialization vector. * 48 bytes of AES-256-CBC output. * An HMACSHA256 authentication tag.

A sample protected payload is illustrated below.

09 F0 C9 F0 80 9C 81 0C 19 66 19 40 95 36 53 F8
AA FF EE 57 57 2F 40 4C 3F 7F CC 9D CC D9 32 3E
84 17 99 16 EC BA 1F 4A A1 18 45 1F 2D 13 7A 28
79 6B 86 9C F8 B7 84 F9 26 31 FC B1 86 0A F1 56
61 CF 14 58 D3 51 6F CF 36 50 85 82 08 2D 3F 73
5F B0 AD 9E 1A B2 AE 13 57 90 C8 F5 7C 95 4E 6A
8A AA 06 EF 43 CA 19 62 84 7C 11 B2 C8 71 9D AA
52 19 2E 5B 4C 1E 54 F0 55 BE 88 92 12 C1 4B 5E
52 C9 74 A0

From the payload format above the first 32 bits, or 4 bytes are the magic header identifying the version (09 F0 C9 F0)

The next 128 bits, or 16 bytes is the key identifier (80 9C 81 0C 19 66 19 40 95 36 53 F8 AA FF EE 57)

The remainder contains the payload and is specific to the format used.

Warning

All payloads protected to a given key will begin with the same 20-byte (magic value, key id) header. Administrators can use this fact for diagnostic purposes to approximate when a payload was generated. For example, the payload above corresponds to key {0c819c80-6619-4019-9536-53f8aaffee57}. If after checking the key repository you find that this specific key’s activation date was 2015-01-01 and its expiration date was 2015-03-01, then it is reasonable to assume that the payload (if not tampered with) was generated within that window, give or take a small fudge factor on either side.

Subkey Derivation and Authenticated Encryption¶

Most keys in the key ring will contain some form of entropy and will have algorithmic information stating “CBC-mode encryption + HMAC validation” or “GCM encryption + validation”. In these cases, we refer to the embedded entropy as the master keying material (or KM) for this key, and we perform a key derivation function to derive the keys that will be used for the actual cryptographic operations.

Note

Keys are abstract, and a custom implementation might not behave as below. If the key provides its own implementation of IAuthenticatedEncryptor rather than using one of our built-in factories, the mechanism described in this section no longer applies.

Additional authenticated data and subkey derivation¶

The IAuthenticatedEncryptor interface serves as the core interface for all authenticated encryption operations. Its Encrypt method takes two buffers: plaintext and additionalAuthenticatedData (AAD). The plaintext contents flow unchanged the call to IDataProtector.Protect, but the AAD is generated by the system and consists of three components:

The 32-bit magic header 09 F0 C9 F0 that identifies this version of the data protection system.
The 128-bit key id.
A variable-length string formed from the purpose chain that created the IDataProtector that is performing this operation.

Because the AAD is unique for the tuple of all three components, we can use it to derive new keys from KM instead of using KM itself in all of our cryptographic operations. For every call to IAuthenticatedEncryptor.Encrypt, the following key derivation process takes place:

( K_E, K_H ) = SP800_108_CTR_HMACSHA512(K_M, AAD, contextHeader || keyModifier)

Here, we’re calling the NIST SP800-108 KDF in Counter Mode (see NIST SP800-108, Sec. 5.1) with the following parameters:

Key derivation key (KDK) = K_M
PRF = HMACSHA512
label = additionalAuthenticatedData
context = contextHeader || keyModifier

The context header is of variable length and essentially serves as a thumbprint of the algorithms for which we’re deriving K_E and K_H. The key modifier is a 128-bit string randomly generated for each call to Encrypt and serves to ensure with overwhelming probability that KE and KH are unique for this specific authentication encryption operation, even if all other input to the KDF is constant.

For CBC-mode encryption + HMAC validation operations, | K_E | is the length of the symmetric block cipher key, and | K_H | is the digest size of the HMAC routine. For GCM encryption + validation operations, | K_H | = 0.

CBC-mode encryption + HMAC validation¶

Once K_E is generated via the above mechanism, we generate a random initialization vector and run the symmetric block cipher algorithm to encipher the plaintext. The initialization vector and ciphertext are then run through the HMAC routine initialized with the key K_H to produce the MAC. This process and the return value is represented graphically below.

output:= keyModifier || iv || E_cbc (K_E,iv,data) || HMAC(K_H, iv || E_cbc (K_E,iv,data))

Note

The IDataProtector.Protect implementation will prepend the magic header and key id to output before returning it to the caller. Because the magic header and key id are implicitly part of AAD, and because the key modifier is fed as input to the KDF, this means that every single byte of the final returned payload is authenticated by the MAC.

Galois/Counter Mode encryption + validation¶

Once K_E is generated via the above mechanism, we generate a random 96-bit nonce and run the symmetric block cipher algorithm to encipher the plaintext and produce the 128-bit authentication tag.

output := keyModifier || nounce || E_gcm (K_E,nounce,data) || authTag

Note

Even though GCM natively supports the concept of AAD, we’re still feeding AAD only to the original KDF, opting to pass an empty string into GCM for its AAD parameter. The reason for this is two-fold. First, to support agility we never want to use K_M directly as the encryption key. Additionally, GCM imposes very strict uniqueness requirements on its inputs. The probability that the GCM encryption routine is ever invoked on two or more distinct sets of input data with the same (key, nonce) pair must not exceed 2³². If we fix K_E we cannot perform more than 2³² encryption operations before we run afoul of the 2^-32 limit. This might seem like a very large number of operations, but a high-traffic web server can go through 4 billion requests in mere days, well within the normal lifetime for these keys. To stay compliant of the 2^-32 probability limit, we continue to use a 128-bit key modifier and 96-bit nonce, which radically extends the usable operation count for any given K_M. For simplicity of design we share the KDF code path between CBC and GCM operations, and since AAD is already considered in the KDF there is no need to forward it to the GCM routine.

Context headers¶

Background and theory¶

In the data protection system, a “key” means an object that can provide authenticated encryption services. Each key is identified by a unique id (a GUID), and it carries with it algorithmic information and entropic material. It is intended that each key carry unique entropy, but the system cannot enforce that, and we also need to account for developers who might change the key ring manually by modifying the algorithmic information of an existing key in the key ring. To achieve our security requirements given these cases the data protection system has a concept of cryptographic agility, which allows securely using a single entropic value across multiple cryptographic algorithms.

Most systems which support cryptographic agility do so by including some identifying information about the algorithm inside the payload. The algorithm’s OID is generally a good candidate for this. However, one problem that we ran into is that there are multiple ways to specify the same algorithm: “AES” (CNG) and the managed Aes, AesManaged, AesCryptoServiceProvider, AesCng, and RijndaelManaged (given specific parameters) classes are all actually the same thing, and we’d need to maintain a mapping of all of these to the correct OID. If a developer wanted to provide a custom algorithm (or even another implementation of AES!), he’d have to tell us its OID. This extra registration step makes system configuration particularly painful.

Stepping back, we decided that we were approaching the problem from the wrong direction. An OID tells you what the algorithm is, but we don’t actually care about this. If we need to use a single entropic value securely in two different algorithms, it’s not necessary for us to know what the algorithms actually are. What we actually care about is how they behave. Any decent symmetric block cipher algorithm is also a strong pseudorandom permutation (PRP): fix the inputs (key, chaining mode, IV, plaintext) and the ciphertext output will with overwhelming probability be distinct from any other symmetric block cipher algorithm given the same inputs. Similarly, any decent keyed hash function is also a strong pseudorandom function (PRF), and given a fixed input set its output will overwhelmingly be distinct from any other keyed hash function.

We use this concept of strong PRPs and PRFs to build up a context header. This context header essentially acts as a stable thumbprint over the algorithms in use for any given operation, and it provides the cryptographic agility needed by the data protection system. This header is reproducible and is used later as part of the subkey derivation process. There are two different ways to build the context header depending on the modes of operation of the underlying algorithms.

CBC-mode encryption + HMAC authentication¶

The context header consists of the following components:

[16 bits] The value 00 00, which is a marker meaning “CBC encryption + HMAC authentication”.
[32 bits] The key length (in bytes, big-endian) of the symmetric block cipher algorithm.
[32 bits] The block size (in bytes, big-endian) of the symmetric block cipher algorithm.
[32 bits] The key length (in bytes, big-endian) of the HMAC algorithm. (Currently the key size always matches the digest size.)
[32 bits] The digest size (in bytes, big-endian) of the HMAC algorithm.
EncCBC(K_E, IV, “”), which is the output of the symmetric block cipher algorithm given an empty string input and where IV is an all-zero vector. The construction of K_E is described below.
MAC(K_H, “”), which is the output of the HMAC algorithm given an empty string input. The construction of K_H is described below.

Ideally we could pass all-zero vectors for K_E and K_H. However, we want to avoid the situation where the underlying algorithm checks for the existence of weak keys before performing any operations (notably DES and 3DES), which precludes using a simple or repeatable pattern like an all-zero vector.

Instead, we use the NIST SP800-108 KDF in Counter Mode (see NIST SP800-108, Sec. 5.1) with a zero-length key, label, and context and HMACSHA512 as the underlying PRF. We derive | K_E | + | K_H | bytes of output, then decompose the result into K_E and K_H themselves. Mathematically, this is represented as follows.

( K_E || K_H ) = SP800_108_CTR(prf = HMACSHA512, key = “”, label = “”, context = “”)

Example: AES-192-CBC + HMACSHA256¶

As an example, consider the case where the symmetric block cipher algorithm is AES-192-CBC and the validation algorithm is HMACSHA256. The system would generate the context header using the following steps.

First, let ( K_E || K_H ) = SP800_108_CTR(prf = HMACSHA512, key = “”, label = “”, context = “”), where | K_E | = 192 bits and | K_H | = 256 bits per the specified algorithms. This leads to K_E = 5BB6..21DD and K_H = A04A..00A9 in the example below:

5B B6 C9 83 13 78 22 1D 8E 10 73 CA CF 65 8E B0
61 62 42 71 CB 83 21 DD A0 4A 05 00 5B AB C0 A2
49 6F A5 61 E3 E2 49 87 AA 63 55 CD 74 0A DA C4
B7 92 3D BF 59 90 00 A9

Next, compute Enc_CBC (K_E, IV, “”) for AES-192-CBC given IV = 0* and K_E as above.

result := F474B1872B3B53E4721DE19C0841DB6F

Next, compute MAC(K_H, “”) for HMACSHA256 given K_H as above.

result := D4791184B996092EE1202F36E8608FA8FBD98ABDFF5402F264B1D7211536220C

This produces the full context header below:

00 00 00 00 00 18 00 00 00 10 00 00 00 20 00 00
00 20 F4 74 B1 87 2B 3B 53 E4 72 1D E1 9C 08 41
DB 6F D4 79 11 84 B9 96 09 2E E1 20 2F 36 E8 60
8F A8 FB D9 8A BD FF 54 02 F2 64 B1 D7 21 15 36
22 0C

This context header is the thumbprint of the authenticated encryption algorithm pair (AES-192-CBC encryption + HMACSHA256 validation). The components, as described above are:

the marker (00 00)
the block cipher key length (00 00 00 18)
the block cipher block size (00 00 00 10)
the HMAC key length (00 00 00 20)
the HMAC digest size (00 00 00 20)
the block cipher PRP output (F4 74 - DB 6F) and
the HMAC PRF output (D4 79 - end).

Note

The CBC-mode encryption + HMAC authentication context header is built the same way regardless of whether the algorithms implementations are provided by Windows CNG or by managed SymmetricAlgorithm and KeyedHashAlgorithm types. This allows applications running on different operating systems to reliably produce the same context header even though the implementations of the algorithms differ between OSes. (In practice, the KeyedHashAlgorithm doesn’t have to be a proper HMAC. It can be any keyed hash algorithm type.)

Example: 3DES-192-CBC + HMACSHA1¶

First, let ( K_E || K_H ) = SP800_108_CTR(prf = HMACSHA512, key = “”, label = “”, context = “”), where | K_E | = 192 bits and | K_H | = 160 bits per the specified algorithms. This leads to K_E = A219..E2BB and K_H = DC4A..B464 in the example below:

A2 19 60 2F 83 A9 13 EA B0 61 3A 39 B8 A6 7E 22
61 D9 F8 6C 10 51 E2 BB DC 4A 00 D7 03 A2 48 3E
D1 F7 5A 34 EB 28 3E D7 D4 67 B4 64

Next, compute Enc_CBC (K_E, IV, “”) for 3DES-192-CBC given IV = 0* and K_E as above.

result := ABB100F81E53E10E

Next, compute MAC(K_H, “”) for HMACSHA1 given K_H as above.

result := 76EB189B35CF03461DDF877CD9F4B1B4D63A7555

This produces the full context header which is a thumbprint of the authenticated encryption algorithm pair (3DES-192-CBC encryption + HMACSHA1 validation), shown below:

00 00 00 00 18 00 00 00 08 00 00 00 14 00 00
14 AB B1 00 F8 1E 53 E1 0E 76 EB 18 9B 35 CF
46 1D DF 87 7C D9 F4 B1 B4 D6 3A 75 55

The components break down as follows:

the marker (00 00)
the block cipher key length (00 00 00 18)
the block cipher block size (00 00 00 08)
the HMAC key length (00 00 00 14)
the HMAC digest size (00 00 00 14)
the block cipher PRP output (AB B1 - E1 0E) and
the HMAC PRF output (76 EB - end).

Galois/Counter Mode encryption + authentication¶

The context header consists of the following components:

[16 bits] The value 00 01, which is a marker meaning “GCM encryption + authentication”.
[32 bits] The key length (in bytes, big-endian) of the symmetric block cipher algorithm.
[32 bits] The nonce size (in bytes, big-endian) used during authenticated encryption operations. (For our system, this is fixed at nonce size = 96 bits.)
[32 bits] The block size (in bytes, big-endian) of the symmetric block cipher algorithm. (For GCM, this is fixed at block size = 128 bits.)
[32 bits] The authentication tag size (in bytes, big-endian) produced by the authenticated encryption function. (For our system, this is fixed at tag size = 128 bits.)
[128 bits] The tag of Enc_GCM (K_E, nonce, “”), which is the output of the symmetric block cipher algorithm given an empty string input and where nonce is a 96-bit all-zero vector.

K_E is derived using the same mechanism as in the CBC encryption + HMAC authentication scenario. However, since there is no K_H in play here, we essentially have | K_H | = 0, and the algorithm collapses to the below form.

K_E = SP800_108_CTR(prf = HMACSHA512, key = “”, label = “”, context = “”)

Example: AES-256-GCM¶

First, let K_E = SP800_108_CTR(prf = HMACSHA512, key = “”, label = “”, context = “”), where | K_E | = 256 bits.

K_E := 22BC6F1B171C08C4AE2F27444AF8FC8B3087A90006CAEA91FDCFB47C1B8733B8

Next, compute the authentication tag of Enc_GCM (K_E, nonce, “”) for AES-256-GCM given nonce = 096 and K_E as above.

result := E7DCCE66DF855A323A6BB7BD7A59BE45

This produces the full context header below:

00 01 00 00 00 20 00 00 00 0C 00 00 00 10 00 00
00 10 E7 DC CE 66 DF 85 5A 32 3A 6B B7 BD 7A 59
BE 45

The components break down as follows:

the marker (00 01)

the block cipher key length (00 00 00 20)

the nonce size (00 00 00 0C)

the block cipher block size (00 00 00 10)

the authentication tag size (00 00 00 10) and

the authentication tag from running the block cipher (E7 DC - end).

Key Management¶

The data protection system automatically manages the lifetime of master keys used to protect and unprotect payloads. Each key can exist in one of four stages.

Created - the key exists in the key ring but has not yet been activated. The key shouldn’t be used for new Protect operations until sufficient time has elapsed that the key has had a chance to propagate to all machines that are consuming this key ring.
Active - the key exists in the key ring and should be used for all new Protect operations.
Expired - the key has run its natural lifetime and should no longer be used for new Protect operations.
Revoked - the key is compromised and must not be used for new Protect operations.

Created, active, and expired keys may all be used to unprotect incoming payloads. Revoked keys by default may not be used to unprotect payloads, but the application developer can override this behavior if necessary.

Warning

The developer might be tempted to delete a key from the key ring (e.g., by deleting the corresponding file from the file system). At that point, all data protected by the key is permanently undecipherable, and there is no emergency override like there is with revoked keys. Deleting a key is truly destructive behavior, and consequently the data protection system exposes no first-class API for performing this operation.

Default key selection¶

When the data protection system reads the key ring from the backing repository, it will attempt to locate a “default” key from the key ring. The default key is used for new Protect operations.

The general heuristic is that the data protection system chooses the key with the most recent activation date as the default key. (There’s a small fudge factor to allow for server-to-server clock skew.) If the key is expired or revoked, and if the application has not disabled automatic key generation, then a new key will be generated with immediate activation per the key expiration and rolling policy below.

The reason the data protection system generates a new key immediately rather than falling back to a different key is that new key generation should be treated as an implicit expiration of all keys that were activated prior to the new key. The general idea is that new keys may have been configured with different algorithms or encryption-at-rest mechanisms than old keys, and the system should prefer the current configuration over falling back.

There is an exception. If the application developer has disabled automatic key generation, then the data protection system must choose something as the default key. In this fallback scenario, the system will choose the non-revoked key with the most recent activation date, with preference given to keys that have had time to propagate to other machines in the cluster. The fallback system may end up choosing an expired default key as a result. The fallback system will never choose a revoked key as the default key, and if the key ring is empty or every key has been revoked then the system will produce an error upon initialization.

Key expiration and rolling¶

When a key is created, it is automatically given an activation date of { now + 2 days } and an expiration date of { now + 90 days }. The 2-day delay before activation gives the key time to propagate through the system. That is, it allows other applications pointing at the backing store to observe the key at their next auto-refresh period, thus maximizing the chances that when the key ring does become active it has propagated to all applications that might need to use it.

If the default key will expire within 2 days and if the key ring does not already have a key that will be active upon expiration of the default key, then the data protection system will automatically persist a new key to the key ring. This new key has an activation date of { default key’s expiration date } and an expiration date of { now + 90 days }. This allows the system to automatically roll keys on a regular basis with no interruption of service.

There might be circumstances where a key will be created with immediate activation. One example would be when the application hasn’t run for a time and all keys in the key ring are expired. When this happens, the key is given an activation date of { now } without the normal 2-day activation delay.

The default key lifetime is 90 days, though this is configurable as in the following example.

services.AddDataProtection()
    // use 14-day lifetime instead of 90-day lifetime
    .SetDefaultKeyLifetime(TimeSpan.FromDays(14));

An administrator can also change the default system-wide, though an explicit call to SetDefaultKeyLifetime will override any system-wide policy. The default key lifetime cannot be shorter than 7 days.

Automatic keyring refresh¶

When the data protection system initializes, it reads the key ring from the underlying repository and caches it in memory. This cache allows Protect and Unprotect operations to proceed without hitting the backing store. The system will automatically check the backing store for changes approximately every 24 hours or when the current default key expires, whichever comes first.

Warning

Developers should very rarely (if ever) need to use the key management APIs directly. The data protection system will perform automatic key management as described above.

The data protection system exposes an interface IKeyManager that can be used to inspect and make changes to the key ring. The DI system that provided the instance of IDataProtectionProvider can also provide an instance of IKeyManager for your consumption. Alternatively, you can pull the IKeyManager straight from the IServiceProvider as in the example below.

Any operation which modifies the key ring (creating a new key explicitly or performing a revocation) will invalidate the in-memory cache. The next call to Protect or Unprotect will cause the data protection system to reread the key ring and recreate the cache.

The sample below demonstrates using the IKeyManager interface to inspect and manipulate the key ring, including revoking existing keys and generating a new key manually.

using System;
using System.IO;
using System.Threading;
using Microsoft.AspNetCore.DataProtection;
using Microsoft.AspNetCore.DataProtection.KeyManagement;
using Microsoft.Extensions.DependencyInjection;

public class Program
{
    public static void Main(string[] args)
    {
        var serviceCollection = new ServiceCollection();
        serviceCollection.AddDataProtection()
            // point at a specific folder and use DPAPI to encrypt keys
            .PersistKeysToFileSystem(new DirectoryInfo(@"c:\temp-keys"))
            .ProtectKeysWithDpapi();
        var services = serviceCollection.BuildServiceProvider();

        // perform a protect operation to force the system to put at least
        // one key in the key ring
        services.GetDataProtector("Sample.KeyManager.v1").Protect("payload");
        Console.WriteLine("Performed a protect operation.");
        Thread.Sleep(2000);

        // get a reference to the key manager
        var keyManager = services.GetService<IKeyManager>();

        // list all keys in the key ring
        var allKeys = keyManager.GetAllKeys();
        Console.WriteLine($"The key ring contains {allKeys.Count} key(s).");
        foreach (var key in allKeys)
        {
            Console.WriteLine($"Key {key.KeyId:B}: Created = {key.CreationDate:u}, IsRevoked = {key.IsRevoked}");
        }

        // revoke all keys in the key ring
        keyManager.RevokeAllKeys(DateTimeOffset.Now, reason: "Revocation reason here.");
        Console.WriteLine("Revoked all existing keys.");

        // add a new key to the key ring with immediate activation and a 1-month expiration
        keyManager.CreateNewKey(
            activationDate: DateTimeOffset.Now,
            expirationDate: DateTimeOffset.Now.AddMonths(1));
        Console.WriteLine("Added a new key.");

        // list all keys in the key ring
        allKeys = keyManager.GetAllKeys();
        Console.WriteLine($"The key ring contains {allKeys.Count} key(s).");
        foreach (var key in allKeys)
        {
            Console.WriteLine($"Key {key.KeyId:B}: Created = {key.CreationDate:u}, IsRevoked = {key.IsRevoked}");
        }
    }
}

/*
 * SAMPLE OUTPUT
 *
 * Performed a protect operation.
 * The key ring contains 1 key(s).
 * Key {1b948618-be1f-440b-b204-64ff5a152552}: Created = 2015-03-18 22:20:49Z, IsRevoked = False
 * Revoked all existing keys.
 * Added a new key.
 * The key ring contains 2 key(s).
 * Key {1b948618-be1f-440b-b204-64ff5a152552}: Created = 2015-03-18 22:20:49Z, IsRevoked = True
 * Key {2266fc40-e2fb-48c6-8ce2-5fde6b1493f7}: Created = 2015-03-18 22:20:51Z, IsRevoked = False
 */

Key storage¶

The data protection system has a heuristic whereby it tries to deduce an appropriate key storage location and encryption at rest mechanism automatically. This is also configurable by the app developer. The following documents discuss the in-box implementations of these mechanisms:

Key Storage Providers¶

By default the data protection system employs a heuristic to determine where cryptographic key material should be persisted. The developer can override the heuristic and manually specify the location.

Note

If you specify an explicit key persistence location, the data protection system will deregister the default key encryption at rest mechanism that the heuristic provided, so keys will no longer be encrypted at rest. It is recommended that you additionally specify an explicit key encryption mechanism for production applications.

The data protection system ships with two in-box key storage providers.

File system¶

We anticipate that the majority of applications will use a file system-based key repository. To configure this, call the PersistKeysToFileSystem configuration routine as demonstrated below, providing a DirectoryInfo pointing to the repository where keys should be stored.

sc.AddDataProtection()
    // persist keys to a specific directory
    .PersistKeysToFileSystem(new DirectoryInfo(@"c:\temp-keys\"));

The DirectoryInfo can point to a directory on the local machine, or it can point to a folder on a network share. If pointing to a directory on the local machine (and the scenario is that only applications on the local machine will need to use this repository), consider using Windows DPAPI to encrypt the keys at rest. Otherwise consider using an X.509 certificate to encrypt keys at rest.

Registry¶

Sometimes the application might not have write access to the file system. Consider a scenario where an application is running as a virtual service account (such as w3wp.exe’s app pool identity). In these cases, the administrator may have provisioned a registry key that is appropriate ACLed for the service account identity. Call the PersistKeysToRegistry configuration routine as demonstrated below to take advantage of this, providing a RegistryKey pointing to the location where cryptographic key material should be stored.

sc.AddDataProtection()
    // persist keys to a specific location in the system registry
    .PersistKeysToRegistry(Registry.CurrentUser.OpenSubKey(@"SOFTWARE\Sample\keys"));

If you use the system registry as a persistence mechanism, consider using Windows DPAPI to encrypt the keys at rest.

Custom key repository¶

If the in-box mechanisms are not appropriate, the developer can specify his own key persistence mechanism by providing a custom IXmlRepository.

Key Encryption At Rest¶

By default the data protection system employs a heuristic to determine how cryptographic key material should be encrypted at rest. The developer can override the heuristic and manually specify how keys should be encrypted at rest.

Note: If you specify an explicit key encryption at rest mechanism, the data protection system will deregister the default key storage mechanism that the heuristic provided. You must specify an explicit key storage mechanism, otherwise the data protection system will fail to start.

The data protection system ships with three in-box key encryption mechanisms.

Windows DPAPI¶

This mechanism is available only on Windows.

When Windows DPAPI is used, key material will be encrypted via CryptProtectData before being persisted to storage. DPAPI is an appropriate encryption mechanism for data that will never be read outside of the current machine (though it is possible to back these keys up to Active Directory; see DPAPI and Roaming Profiles). For example to configure DPAPI key-at-rest encryption.

sc.AddDataProtection()
    // only the local user account can decrypt the keys
    .ProtectKeysWithDpapi();

If ProtectKeysWithDpapi is called with no parameters, only the current Windows user account can decipher the persisted key material. You can optionally specify that any user account on the machine (not just the current user account) should be able to decipher the key material, as shown in the below example.

sc.AddDataProtection()
    // all user accounts on the machine can decrypt the keys
    .ProtectKeysWithDpapi(protectToLocalMachine: true);

X.509 certificate¶

This mechanism is not yet available on `.NET Core`.

If your application is spread across multiple machines, it may be convenient to distribute a shared X.509 certificate across the machines and to configure applications to use this certificate for encryption of keys at rest. See below for an example.

sc.AddDataProtection()
    // searches the cert store for the cert with this thumbprint
    .ProtectKeysWithCertificate("3BCE558E2AD3E0E34A7743EAB5AEA2A9BD2575A0");

Because this mechanism uses X509Certificate2 and EncryptedXml under the covers, this feature is currently only available on Desktop CLR. Additionally, due to .NET Framework limitations only certificates with CAPI private keys are supported. See Certificate-based encryption with Windows DPAPI-NG below for possible workarounds to these limitations.

Windows DPAPI-NG¶

This mechanism is available only on Windows 8 / Windows Server 2012 and later.

Beginning with Windows 8, the operating system supports DPAPI-NG (also called CNG DPAPI). Microsoft lays out its usage scenario as follows.

Cloud computing, however, often requires that content encrypted on one computer be decrypted on another. Therefore, beginning with Windows 8, Microsoft extended the idea of using a relatively straightforward API to encompass cloud scenarios. This new API, called DPAPI-NG, enables you to securely share secrets (keys, passwords, key material) and messages by protecting them to a set of principals that can be used to unprotect them on different computers after proper authentication and authorization.

From https://msdn.microsoft.com/en-us/library/windows/desktop/hh706794(v=vs.85).aspx

The principal is encoded as a protection descriptor rule. Consider the below example, which encrypts key material such that only the domain-joined user with the specified SID can decrypt the key material.

sc.AddDataProtection()
  // uses the descriptor rule "SID=S-1-5-21-..."
  .ProtectKeysWithDpapiNG("SID=S-1-5-21-...",
    flags: DpapiNGProtectionDescriptorFlags.None);

There is also a parameterless overload of ProtectKeysWithDpapiNG. This is a convenience method for specifying the rule “SID=mine”, where mine is the SID of the current Windows user account.

sc.AddDataProtection()
  // uses the descriptor rule "SID={current account SID}"
  .ProtectKeysWithDpapiNG();

In this scenario, the AD domain controller is responsible for distributing the encryption keys used by the DPAPI-NG operations. The target user will be able to decipher the encrypted payload from any domain-joined machine (provided that the process is running under their identity).

Certificate-based encryption with Windows DPAPI-NG¶

If you’re running on Windows 8.1 / Windows Server 2012 R2 or later, you can use Windows DPAPI-NG to perform certificate-based encryption, even if the application is running on .NET Core. To take advantage of this, use the rule descriptor string “CERTIFICATE=HashId:thumbprint”, where thumbprint is the hex-encoded SHA1 thumbprint of the certificate to use. See below for an example.

sc.AddDataProtection()
    // searches the cert store for the cert with this thumbprint
    .ProtectKeysWithDpapiNG("CERTIFICATE=HashId:3BCE558E2AD3E0E34A7743EAB5AEA2A9BD2575A0",
        flags: DpapiNGProtectionDescriptorFlags.None);

Any application which is pointed at this repository must be running on Windows 8.1 / Windows Server 2012 R2 or later to be able to decipher this key.

Custom key encryption¶

If the in-box mechanisms are not appropriate, the developer can specify their own key encryption mechanism by providing a custom IXmlEncryptor.

Key Immutability and Changing Settings¶

Once an object is persisted to the backing store, its representation is forever fixed. New data can be added to the backing store, but existing data can never be mutated. The primary purpose of this behavior is to prevent data corruption.

One consequence of this behavior is that once a key is written to the backing store, it is immutable. Its creation, activation, and expiration dates can never be changed, though it can revoked by using IKeyManager. Additionally, its underlying algorithmic information, master keying material, and encryption at rest properties are also immutable.

If the developer changes any setting that affects key persistence, those changes will not go into effect until the next time a key is generated, either via an explicit call to IKeyManager.CreateNewKey or via the data protection system’s own automatic key generation behavior. The settings that affect key persistence are as follows:

If you need these settings to kick in earlier than the next automatic key rolling time, consider making an explicit call to IKeyManager.CreateNewKey to force the creation of a new key. Remember to provide an explicit activation date ({ now + 2 days } is a good rule of thumb to allow time for the change to propagate) and expiration date in the call.

Tip

All applications touching the repository should specify the same settings with the IDataProtectionBuilder extension methods, otherwise the properties of the persisted key will be dependent on the particular application that invoked the key generation routines.

Key Storage Format¶

Objects are stored at rest in XML representation. The default directory for key storage is %LOCALAPPDATA%\ASP.NET\DataProtection-Keys\.

The <key> element¶

Keys exist as top-level objects in the key repository. By convention keys have the filename key-{guid}.xml, where {guid} is the id of the key. Each such file contains a single key. The format of the file is as follows.

<?xml version="1.0" encoding="utf-8"?>
<key id="80732141-ec8f-4b80-af9c-c4d2d1ff8901" version="1">
  <creationDate>2015-03-19T23:32:02.3949887Z</creationDate>
  <activationDate>2015-03-19T23:32:02.3839429Z</activationDate>
  <expirationDate>2015-06-17T23:32:02.3839429Z</expirationDate>
  <descriptor deserializerType="{deserializerType}">
    <descriptor>
      <encryption algorithm="AES_256_CBC" />
      <validation algorithm="HMACSHA256" />
      <enc:encryptedSecret decryptorType="{decryptorType}" xmlns:enc="...">
        <encryptedKey>
          <!-- This key is encrypted with Windows DPAPI. -->
          <value>AQAAANCM...8/zeP8lcwAg==</value>
        </encryptedKey>
      </enc:encryptedSecret>
    </descriptor>
  </descriptor>
</key>

The <key> element contains the following attributes and child elements:

The key id. This value is treated as authoritative; the filename is simply a nicety for human readability.
The version of the <key> element, currently fixed at 1.
The key’s creation, activation, and expiration dates.
A <descriptor> element, which contains information on the authenticated encryption implementation contained within this key.

In the above example, the key’s id is {80732141-ec8f-4b80-af9c-c4d2d1ff8901}, it was created and activated on March 19, 2015, and it has a lifetime of 90 days. (Occasionally the activation date might be slightly before the creation date as in this example. This is due to a nit in how the APIs work and is harmless in practice.)

The <descriptor> element¶

The outer <descriptor> element contains an attribute deserializerType, which is the assembly-qualified name of a type which implements IAuthenticatedEncryptorDescriptorDeserializer. This type is responsible for reading the inner <descriptor> element and for parsing the information contained within.

The particular format of the <descriptor> element depends on the authenticated encryptor implementation encapsulated by the key, and each deserializer type expects a slightly different format for this. In general, though, this element will contain algorithmic information (names, types, OIDs, or similar) and secret key material. In the above example, the descriptor specifies that this key wraps AES-256-CBC encryption + HMACSHA256 validation.

The <encryptedSecret> element¶

An <encryptedSecret> element which contains the encrypted form of the secret key material may be present if encryption of secrets at rest is enabled. The attribute decryptorType will be the assembly-qualified name of a type which implements IXmlDecryptor. This type is responsible for reading the inner <encryptedKey> element and decrypting it to recover the original plaintext.

As with <descriptor>, the particular format of the <encryptedSecret> element depends on the at-rest encryption mechanism in use. In the above example, the master key is encrypted using Windows DPAPI per the comment.

The <revocation> element¶

Revocations exist as top-level objects in the key repository. By convention revocations have the filename revocation-{timestamp}.xml (for revoking all keys before a specific date) or revocation-{guid}.xml (for revoking a specific key). Each file contains a single <revocation> element.

For revocations of individual keys, the file contents will be as below.

<?xml version="1.0" encoding="utf-8"?>
<revocation version="1">
  <revocationDate>2015-03-20T22:45:30.2616742Z</revocationDate>
  <key id="eb4fc299-8808-409d-8a34-23fc83d026c9" />
  <reason>human-readable reason</reason>
</revocation>

In this case, only the specified key is revoked. If the key id is “*”, however, as in the below example, all keys whose creation date is prior to the specified revocation date are revoked.

<?xml version="1.0" encoding="utf-8"?>
<revocation version="1">
  <revocationDate>2015-03-20T15:45:45.7366491-07:00</revocationDate>
  <!-- All keys created before the revocation date are revoked. -->
  <key id="*" />
  <reason>human-readable reason</reason>
</revocation>

The <reason> element is never read by the system. It is simply a convenient place to store a human-readable reason for revocation.

Ephemeral data protection providers¶

There are scenarios where an application needs a throwaway IDataProtectionProvider. For example, the developer might just be experimenting in a one-off console application, or the application itself is transient (it’s scripted or a unit test project). To support these scenarios the package Microsoft.AspNetCore.DataProtection includes a type EphemeralDataProtectionProvider. This type provides a basic implementation of IDataProtectionProvider whose key repository is held solely in-memory and isn’t written out to any backing store.

Each instance of EphemeralDataProtectionProvider uses its own unique master key. Therefore, if an IDataProtector rooted at an EphemeralDataProtectionProvider generates a protected payload, that payload can only be unprotected by an equivalent IDataProtector (given the same purpose chain) rooted at the same EphemeralDataProtectionProvider instance.

The following sample demonstrates instantiating an EphemeralDataProtectionProvider and using it to protect and unprotect data.

using System;
using Microsoft.AspNetCore.DataProtection;

public class Program
{
    public static void Main(string[] args)
    {
        const string purpose = "Ephemeral.App.v1";

        // create an ephemeral provider and demonstrate that it can round-trip a payload
        var provider = new EphemeralDataProtectionProvider();
        var protector = provider.CreateProtector(purpose);
        Console.Write("Enter input: ");
        string input = Console.ReadLine();

        // protect the payload
        string protectedPayload = protector.Protect(input);
        Console.WriteLine($"Protect returned: {protectedPayload}");

        // unprotect the payload
        string unprotectedPayload = protector.Unprotect(protectedPayload);
        Console.WriteLine($"Unprotect returned: {unprotectedPayload}");

        // if I create a new ephemeral provider, it won't be able to unprotect existing
        // payloads, even if I specify the same purpose
        provider = new EphemeralDataProtectionProvider();
        protector = provider.CreateProtector(purpose);
        unprotectedPayload = protector.Unprotect(protectedPayload); // THROWS
    }
}

/*
 * SAMPLE OUTPUT
 *
 * Enter input: Hello!
 * Protect returned: CfDJ8AAAAAAAAAAAAAAAAAAAAA...uGoxWLjGKtm1SkNACQ
 * Unprotect returned: Hello!
 * << throws CryptographicException >>
 */

Compatibility¶

Replacing <machineKey> in ASP.NET¶

The implementation of the <machineKey> element in ASP.NET is replaceable. This allows most calls to ASP.NET cryptographic routines to be routed through a replacement data protection mechanism, including the new data protection system.

Package installation¶

Note

The new data protection system can only be installed into an existing ASP.NET application targeting .NET 4.5.1 or higher. Installation will fail if the application targets .NET 4.5 or lower.

To install the new data protection system into an existing ASP.NET 4.5.1+ project, install the package Microsoft.AspNetCore.DataProtection.SystemWeb. This will instantiate the data protection system using the default configuration settings.

When you install the package, it inserts a line into Web.config that tells ASP.NET to use it for most cryptographic operations, including forms authentication, view state, and calls to MachineKey.Protect. The line that’s inserted reads as follows.

<machineKey compatibilityMode="Framework45" dataProtectorType="..." />

Tip

You can tell if the new data protection system is active by inspecting fields like __VIEWSTATE, which should begin with “CfDJ8” as in the below example. “CfDJ8” is the base64 representation of the magic “09 F0 C9 F0” header that identifies a payload protected by the data protection system.

<input type="hidden" name="__VIEWSTATE" id="__VIEWSTATE" value="CfDJ8AWPr2EQPTBGs3L2GCZOpk..." />

Package configuration¶

The data protection system is instantiated with a default zero-setup configuration. However, since by default keys are persisted to the local file system, this won’t work for applications which are deployed in a farm. To resolve this, you can provide configuration by creating a type which subclasses DataProtectionStartup and overrides its ConfigureServices method.

Below is an example of a custom data protection startup type which configured both where keys are persisted and how they’re encrypted at rest. It also overrides the default app isolation policy by providing its own application name.

using System;
using System.IO;
using Microsoft.AspNetCore.DataProtection;
using Microsoft.AspNetCore.DataProtection.SystemWeb;
using Microsoft.Extensions.DependencyInjection;

namespace DataProtectionDemo
{
    public class MyDataProtectionStartup : DataProtectionStartup
    {
        public override void ConfigureServices(IServiceCollection services)
        {
            services.AddDataProtection()
                .SetApplicationName("my-app")
                .PersistKeysToFileSystem(new DirectoryInfo(@"\\server\share\myapp-keys\"))
                .ProtectKeysWithCertificate("thumbprint");
        }
    }
}

Tip

You can also use <machineKey applicationName=”my-app” ... /> in place of an explicit call to SetApplicationName. This is a convenience mechanism to avoid forcing the developer to create a DataProtectionStartup-derived type if all he wanted to configure was setting the application name.

To enable this custom configuration, go back to Web.config and look for the <appSettings> element that the package install added to the config file. It will look like the below.

<appSettings>
  <!--
  If you want to customize the behavior of the ASP.NET Core Data Protection stack, set the
  "aspnet:dataProtectionStartupType" switch below to be the fully-qualified name of a
  type which subclasses Microsoft.AspNetCore.DataProtection.SystemWeb.DataProtectionStartup.
  -->
  <add key="aspnet:dataProtectionStartupType" value="" />
</appSettings>

Fill in the blank value with the assembly-qualified name of the DataProtectionStartup-derived type you just created. If the name of the application is DataProtectionDemo, this would look like the below.

<add key="aspnet:dataProtectionStartupType"
     value="DataProtectionDemo.MyDataProtectionStartup, DataProtectionDemo" />

The newly-configured data protection system is now ready for use inside the application.

Safe storage of app secrets during development¶

By Rick Anderson, Daniel Roth

This document shows how you can use the Secret Manager tool to keep secrets out of your code. The most important point is you should never store passwords or other sensitive data in source code, and you shouldn’t use production secrets in development and test mode. You can instead use the configuration system to read these values from environment variables or from values stored using the Secret Manager tool. The Secret Manager tool helps prevent sensitive data from being checked into source control. The configuration system can read secrets stored with the Secret Manager tool described in this article.

Sections:

Environment variables
Secret Manager
Accessing user secrets via configuration
How the Secret Manager tool works
Additional Resources

Environment variables ¶

To avoid storing app secrets in code or in local configuration files you store secrets in environment variables. You can setup the configuration framework to read values from environment variables by calling AddEnvironmentVariables. You can then use environment variables to override configuration values for all previously specified configuration sources.

For example, if you create a new ASP.NET Core web app with individual user accounts, it will add a default connection string to the appsettings.json file in the project with the key DefaultConnection. The default connection string is setup to use LocalDB, which runs in user mode and doesn’t require a password. When you deploy your application to a test or production server you can override the DefaultConnection key value with an environment variable setting that contains the connection string (potentially with sensitive credentials) for a test or production database server.

Warning

Environment variables are generally stored in plain text and are not encrypted. If the machine or process is compromised then environment variables can be accessed by untrusted parties. Additional measures to prevent disclosure of user secrets may still be required.

Secret Manager ¶

The Secret Manager tool provides a more general mechanism to store sensitive data for development work outside of your project tree. The Secret Manager tool is a project tool that can be used to store secrets for a .NET Core project during development. With the Secret Manager tool you can associate app secrets with a specific project and share them across multiple projects.

Warning

The Secret Manager tool does not encrypt the stored secrets and should not be treated as a trusted store. It is for development purposes only. The keys and values are stored in a JSON configuration file in the user profile directory.

Installing the Secret Manager tool¶

Add SecretManager.Tools to the tools section of the project.json file and run dotnet restore.

"tools": {
  "Microsoft.AspNetCore.Razor.Tools": "1.0.0-preview2-final",
  "Microsoft.Extensions.SecretManager.Tools": "1.0.0-preview2-final"
},

Test the Secret Manager tool by running the following command:
```
dotnet user-secrets -h
```

Note

When any of the tools are defined in the project.json file, you must be in the same directory in order to use the tooling commands.

The Secret Manager tool will display usage, options and command help.

The Secret Manager tool operates on project specific configuration settings that are stored in your user profile. To use user secrets the project must specify a userSecretsId value in its project.json file. The value of userSecretsId is arbitrary, but is generally unique to the project.

Add a userSecretsId for your project in its project.json file:

{
"userSecretsId": "aspnet-WebApp1-c23d27a4-eb88-4b18-9b77-2a93f3b15119",

"dependencies": {

Use the Secret Manager tool to set a secret. For example, in a command window from the project directory enter the following:
```
dotnet user-secrets set MySecret ValueOfMySecret
```

You can run the secret manager tool from other directories, but you must use the --project option to pass in the path to the project.json file:

dotnet user-secrets set MySecret ValueOfMySecret --project c:\work\WebApp1

You can also use the Secret Manager tool to list, remove and clear app secrets.

Accessing user secrets via configuration ¶

You access Secret Manager secrets through the configuration system. Add the Microsoft.Extensions.Configuration.UserSecrets as a dependency in your project.json file and run dotnet restore.

  "Microsoft.Extensions.Configuration.UserSecrets": "1.0.0-rc2-final",

Add the user secrets configuration source to the Startup method:

public Startup(IHostingEnvironment env)
{
    var builder = new ConfigurationBuilder()
        .SetBasePath(env.ContentRootPath)
        .AddJsonFile("appsettings.json", optional: true, reloadOnChange: true)
        .AddJsonFile($"appsettings.{env.EnvironmentName}.json", optional: true);

    if (env.IsDevelopment())
    {
        // For more details on using the user secret store see http://go.microsoft.com/fwlink/?LinkID=532709
        builder.AddUserSecrets();
    }

    builder.AddEnvironmentVariables();
    Configuration = builder.Build();
}

You can now access user secrets via the configuration API:

string testConfig = Configuration["MySecret"];

How the Secret Manager tool works ¶

The secret manager tool abstracts away the implementation details, such as where and how the values are stored. You can use the tool without knowing these implementation details. In the current version, the values are stored in a JSON configuration file in the user profile directory:

Windows: %APPDATA%\microsoft\UserSecrets\<userSecretsId>\secrets.json
Linux: ~/.microsoft/usersecrets/<userSecretsId>/secrets.json
Mac: ~/.microsoft/usersecrets/<userSecretsId>/secrets.json

The value of userSecretsId comes from the value specified in project.json.

You should not write code that depends on the location or format of the data saved with the secret manager tool, as these implementation details might change. For example, the secret values are currently not encrypted today, but could be someday.

Additional Resources ¶

Configuration.

🔧 Enforcing SSL¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

🔧 Anti-Request Forgery¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

🔧 Preventing Open Redirect Attacks¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

🔧 Preventing Cross-Site Scripting¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Enabling Cross-Origin Requests (CORS)¶

By Mike Wasson

Browser security prevents a web page from making AJAX requests to another domain. This restriction is called the same-origin policy, and prevents a malicious site from reading sensitive data from another site. However, sometimes you might want to let other sites make cross-origin requests to your web app.

Cross Origin Resource Sharing (CORS) is a W3C standard that allows a server to relax the same-origin policy. Using CORS, a server can explicitly allow some cross-origin requests while rejecting others. CORS is safer and more flexible than earlier techniques such as JSONP. This topic shows how to enable CORS in your ASP.NET Core application.

Sections:

What is “same origin”?
Setting up CORS
Enabling CORS with middleware
Enabling CORS in MVC
CORS policy options
How CORS works

What is “same origin”?¶

Two URLs have the same origin if they have identical schemes, hosts, and ports. (RFC 6454)

These two URLs have the same origin:

http://example.com/foo.html
http://example.com/bar.html

These URLs have different origins than the previous two:

http://example.net - Different domain
http://example.com:9000/foo.html - Different port
https://example.com/foo.html - Different scheme
http://www.example.com/foo.html - Different subdomain

Note

Internet Explorer does not consider the port when comparing origins.

Setting up CORS ¶

To setup CORS for your application you use the Microsoft.AspNetCore.Cors package. In your project.json file, add the following:

  "dependencies": {
    "Microsoft.AspNet.Cors": "6.0.0-rc1-final",
  },

Add the CORS services in Startup.cs:

public void ConfigureServices(IServiceCollection services)
{
    services.AddCors();
}

Enabling CORS with middleware ¶

To enable CORS for your entire application add the CORS middleware to your request pipeline using the UseCors extension method. Note that the CORS middleware must precede any defined endpoints in your app that you want to support cross-origin requests (ex. before any call to UseMvc).

You can specify a cross-origin policy when adding the CORS middleware using the CorsPolicyBuilder class. There are two ways to do this. The first is to call UseCors with a lambda:

public void Configure(IApplicationBuilder app)
{
    app.UseCors(builder =>
        builder.WithOrigins("http://example.com"));
}

The lambda takes a CorsPolicyBuilder object. I’ll describe all of the configuration options later in this topic. In this example, the policy allows cross-origin requests from “http://example.com” and no other origins.

Note that CorsPolicyBuilder has a fluent API, so you can chain method calls:

app.UseCors(builder =>
    builder.WithOrigins("http://example.com")
           .AllowAnyHeader()
    );

The second approach is to define one or more named CORS policies, and then select the policy by name at run time.

public void ConfigureServices(IServiceCollection services)
{
    services.AddCors(options =>
    {
        options.AddPolicy("AllowSpecificOrigin",
            builder => builder.WithOrigins("http://example.com"));
    });
}

public void Configure(IApplicationBuilder app)
{
    app.UseCors("AllowSpecificOrigin");
    app.Run(async (context) =>
    {
        await context.Response.WriteAsync("Hello World!");
    });
}

This example adds a CORS policy named “AllowSpecificOrigin”. To select the policy, pass the name to UseCors.

Enabling CORS in MVC ¶

You can alternatively use MVC to apply specific CORS per action, per controller, or globally for all controllers. When using MVC to enable CORS the same CORS services are used, but the CORS middleware is not.

Per action¶

To specify a CORS policy for a specific action add the [EnableCors] attribute to the action. Specify the policy name.

public class HomeController : Controller
{
    [EnableCors("AllowSpecificOrigin")] 
    public IActionResult Index()
    {
        return View();
    }

Per controller¶

To specify the CORS policy for a specific controller add the [EnableCors] attribute to the controller class. Specify the policy name.

[EnableCors("AllowSpecificOrigin")]
public class HomeController : Controller
{

Globally¶

You can enable CORS globally for all controllers by adding the CorsAuthorizationFilterFactory filter to the global filter collection:

public void ConfigureServices(IServiceCollection services)
{
    services.AddMvc();
    services.Configure<MvcOptions>(options =>
    {
        options.Filters.Add(new CorsAuthorizationFilterFactory("AllowSpecificOrigin"));
    });
}

The precedence order is: Action, controller, global. Action-level policies take precedence over controller-level policies, and controller-level policies take precedence over global policies.

Disable CORS¶

To disable CORS for a controller or action, use the [DisableCors] attribute.

    [DisableCors]
    public IActionResult About()
    {
        return View();
    }

CORS policy options ¶

This section describes the various options that you can set in a CORS policy.

Set the allowed origins
Set the allowed HTTP methods
Set the allowed request headers
Set the exposed response headers
Credentials in cross-origin requests
Set the preflight expiration time

For some options it may be helpful to read How CORS works first.

Set the allowed origins¶

To allow one or more specific origins:

options.AddPolicy("AllowSpecificOrigins",
builder =>
{
    builder.WithOrigins("http://example.com", "http://www.contoso.com");
});

To allow all origins:

options.AddPolicy("AllowAllOrigins",
    builder =>
    {
        builder.AllowAnyOrigin();
    });

Consider carefully before allowing requests from any origin. It means that literally any website can make AJAX calls to your app.

Set the allowed HTTP methods¶

To specify which HTTP methods are allowed to access the resource.

options.AddPolicy("AllowSpecificMethods",
    builder =>
    {
        builder.WithOrigins("http://example.com")
               .WithMethods("GET", "POST", "HEAD");
    });

To allow all HTTP methods:

options.AddPolicy("AllowAllMethods",
    builder =>
    {
        builder.WithOrigins("http://example.com")
               .AllowAnyMethod();
    });

This affects pre-flight requests and Access-Control-Allow-Methods header.

Set the allowed request headers¶

A CORS preflight request might include an Access-Control-Request-Headers header, listing the HTTP headers set by the application (the so-called “author request headers”).

To whitelist specific headers:

options.AddPolicy("AllowHeaders",
    builder =>
    {
        builder.WithOrigins("http://example.com")
               .WithHeaders("accept", "content-type", "origin", "x-custom-header");
    });

To allow all author request headers:

options.AddPolicy("AllowAllHeaders",
    builder =>
    {
        builder.WithOrigins("http://example.com")
               .AllowAnyHeader();
    });

Browsers are not entirely consistent in how they set Access-Control-Request-Headers. If you set headers to anything other than “*”, you should include at least “accept”, “content-type”, and “origin”, plus any custom headers that you want to support.

Set the exposed response headers¶

By default, the browser does not expose all of the response headers to the application. (See http://www.w3.org/TR/cors/#simple-response-header.) The response headers that are available by default are:

Cache-Control
Content-Language
Content-Type
Expires
Last-Modified
Pragma

The CORS spec calls these simple response headers. To make other headers available to the application:

options.AddPolicy("ExposeResponseHeaders",
    builder =>
    {
        builder.WithOrigins("http://example.com")
               .WithExposedHeaders("x-custom-header");
    });

Credentials in cross-origin requests¶

Credentials require special handling in a CORS request. By default, the browser does not send any credentials with a cross-origin request. Credentials include cookies as well as HTTP authentication schemes. To send credentials with a cross-origin request, the client must set XMLHttpRequest.withCredentials to true.

Using XMLHttpRequest directly:

var xhr = new XMLHttpRequest();
xhr.open('get', 'http://www.example.com/api/test');
xhr.withCredentials = true;

In jQuery:

$.ajax({
    type: 'get',
    url: 'http://www.example.com/home',
    xhrFields: {
        withCredentials: true
    }

In addition, the server must allow the credentials. To allow cross-origin credentials:

options.AddPolicy("AllowCredentials",
    builder =>
    {
        builder.WithOrigins("http://example.com")
               .AllowCredentials();
    });

Now the HTTP response will include an Access-Control-Allow-Credentials header, which tells the browser that the server allows credentials for a cross-origin request.

If the browser sends credentials, but the response does not include a valid Access-Control-Allow-Credentials header, the browser will not expose the response to the application, and the AJAX request fails.

Be very careful about allowing cross-origin credentials, because it means a website at another domain can send a logged-in user’s credentials to your app on the user’s behalf, without the user being aware. The CORS spec also states that setting origins to “*” (all origins) is invalid if the Access-Control-Allow-Credentials header is present.

Set the preflight expiration time¶

The Access-Control-Max-Age header specifies how long the response to the preflight request can be cached. To set this header:

options.AddPolicy("SetPreflightExpiration",
    builder =>
    {
        builder.WithOrigins("http://example.com")
               .SetPreflightMaxAge(TimeSpan.FromSeconds(2520));
    });

How CORS works ¶

This section describes what happens in a CORS request, at the level of the HTTP messages. It’s important to understand how CORS works, so that you can configure the your CORS policy correctly, and troubleshoot if things don’t work as you expect.

The CORS specification introduces several new HTTP headers that enable cross-origin requests. If a browser supports CORS, it sets these headers automatically for cross-origin requests; you don’t need to do anything special in your JavaScript code.

Here is an example of a cross-origin request. The “Origin” header gives the domain of the site that is making the request:

GET http://myservice.azurewebsites.net/api/test HTTP/1.1
Referer: http://myclient.azurewebsites.net/
Accept: */*
Accept-Language: en-US
Origin: http://myclient.azurewebsites.net
Accept-Encoding: gzip, deflate
User-Agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)
Host: myservice.azurewebsites.net

If the server allows the request, it sets the Access-Control-Allow-Origin header. The value of this header either matches the Origin header, or is the wildcard value “*”, meaning that any origin is allowed.:

HTTP/1.1 200 OK
Cache-Control: no-cache
Pragma: no-cache
Content-Type: text/plain; charset=utf-8
Access-Control-Allow-Origin: http://myclient.azurewebsites.net
Date: Wed, 20 May 2015 06:27:30 GMT
Content-Length: 12

Test message

If the response does not include the Access-Control-Allow-Origin header, the AJAX request fails. Specifically, the browser disallows the request. Even if the server returns a successful response, the browser does not make the response available to the client application.

Preflight Requests¶

For some CORS requests, the browser sends an additional request, called a “preflight request”, before it sends the actual request for the resource. The browser can skip the preflight request if the following conditions are true:

The request method is GET, HEAD, or POST, and
The application does not set any request headers other than Accept, Accept-Language, Content-Language, Content-Type, or Last-Event-ID, and
The Content-Type header (if set) is one of the following:
- application/x-www-form-urlencoded
- multipart/form-data
- text/plain

The rule about request headers applies to headers that the application sets by calling setRequestHeader on the XMLHttpRequest object. (The CORS specification calls these “author request headers”.) The rule does not apply to headers the browser can set, such as User-Agent, Host, or Content-Length.

Here is an example of a preflight request:

OPTIONS http://myservice.azurewebsites.net/api/test HTTP/1.1
Accept: */*
Origin: http://myclient.azurewebsites.net
Access-Control-Request-Method: PUT
Access-Control-Request-Headers: accept, x-my-custom-header
Accept-Encoding: gzip, deflate
User-Agent: Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)
Host: myservice.azurewebsites.net
Content-Length: 0

The pre-flight request uses the HTTP OPTIONS method. It includes two special headers:

Access-Control-Request-Method: The HTTP method that will be used for the actual request.
Access-Control-Request-Headers: A list of request headers that the application set on the actual request. (Again, this does not include headers that the browser sets.)

Here is an example response, assuming that the server allows the request:

HTTP/1.1 200 OK
Cache-Control: no-cache
Pragma: no-cache
Content-Length: 0
Access-Control-Allow-Origin: http://myclient.azurewebsites.net
Access-Control-Allow-Headers: x-my-custom-header
Access-Control-Allow-Methods: PUT
Date: Wed, 20 May 2015 06:33:22 GMT

The response includes an Access-Control-Allow-Methods header that lists the allowed methods, and optionally an Access-Control-Allow-Headers header, which lists the allowed headers. If the preflight request succeeds, the browser sends the actual request, as described earlier.

Performance¶

🔧 Measuring Application Performance¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Caching¶

In Memory Caching¶

Caching involves keeping a copy of data in a location that can be accessed more quickly than the source data. ASP.NET Core has rich support for caching in a variety of ways, including keeping data in memory on the local server, which is referred to as in memory caching.

Sections:

Caching Basics
Configuring In Memory Caching
Reading and Writing to a Memory Cache
Cache Dependencies and Callbacks
Other Resources

Caching Basics ¶

Caching can dramatically improve the performance and scalability of ASP.NET applications, by eliminating unnecessary requests to external data sources for data that changes infrequently.

Note

Caching in all forms (in-memory or distributed, including session state) involves making a copy of data in order to optimize performance. The copied data should be considered ephemeral - it could disappear at any time. Apps should be written to not depend on cached data, but use it when available.

ASP.NET supports several different kinds of caches, the simplest of which is represented by the IMemoryCache interface, which represents a cache stored in the memory of the local web server.

You should always write (and test!) your application such that it can use cached data if it’s available, but otherwise will work correctly using the underlying data source.

An in-memory cache is stored in the memory of a single server hosting an ASP.NET app. If an app is hosted by multiple servers in a web farm or cloud hosting environment, the servers may have different values in their local in-memory caches. Apps that will be hosted in server farms or on cloud hosting should use a distributed cache to avoid cache consistency problems.

Tip

A common use case for caching is data-driven navigation menus, which rarely change but are frequently read for display within an application. Caching results that do not vary often but which are requested frequently can greatly improve performance by reducing round trips to out of process data stores and unnecessary computation.

Configuring In Memory Caching ¶

To use an in memory cache in your ASP.NET application, add the following dependencies to your project.json file:

  "dependencies": {
    "Microsoft.AspNetCore.Server.IISIntegration": "1.0.0-rc2-final",
    "Microsoft.AspNetCore.Server.Kestrel": "1.0.0-rc2-final",
    "Microsoft.Extensions.Caching.Memory": "1.0.0-rc2-final",
    "Microsoft.Extensions.Logging": "1.0.0-rc2-final",
    "Microsoft.Extensions.Logging.Console": "1.0.0-rc2-final"
  },

Caching in ASP.NET Core is a service that should be referenced from your application by Dependency Injection. To register the caching service and make it available within your app, add the following line to your ConfigureServices method in Startup:

public void ConfigureServices(IServiceCollection services)
{
    services.AddMemoryCache();

You utilize caching in your app by requesting an instance of IMemoryCache in your controller or middleware constructor. In the sample for this article, we are using a simple middleware component to handle requests by returning customized greeting. The constructor is shown here:

public GreetingMiddleware(RequestDelegate next,
    IMemoryCache memoryCache,
    ILogger<GreetingMiddleware> logger,
    IGreetingService greetingService)
{
    _next = next;
    _memoryCache = memoryCache;
    _greetingService = greetingService;
    _logger = logger;
}

Reading and Writing to a Memory Cache ¶

The middleware’s Invoke method returns the cached data when it’s available.

There are two methods for accessing cache entries:

Get: Get will return the value if it exists, but otherwise returns null.
TryGet: TryGet will assign the cached value to an out parameter and return true if the entry exists. Otherwise it returns false.

Use the Set method to write to the cache. Set accepts the key to use to look up the value, the value to be cached, and a set of MemoryCacheEntryOptions. The MemoryCacheEntryOptions allow you to specify absolute or sliding time-based cache expiration, caching priority, callbacks, and dependencies. These options are detailed below.

The sample code (shown below) uses the SetAbsoluteExpiration method on MemoryCacheEntryOptions to cache greetings for one minute.

public Task Invoke(HttpContext httpContext)
{
    string cacheKey = "GreetingMiddleware-Invoke";
    string greeting;

    // try to get the cached item; null if not found
    // greeting = _memoryCache.Get(cacheKey) as string;

    // alternately, TryGet returns true if the cache entry was found
    if(!_memoryCache.TryGetValue(cacheKey, out greeting))
    {
        // fetch the value from the source
        greeting = _greetingService.Greet("world");

        // store in the cache
        _memoryCache.Set(cacheKey, greeting,
            new MemoryCacheEntryOptions()
            .SetAbsoluteExpiration(TimeSpan.FromMinutes(1)));
        _logger.LogInformation($"{cacheKey} updated from source.");
    }
    else
    {
        _logger.LogInformation($"{cacheKey} retrieved from cache.");
    }

    return httpContext.Response.WriteAsync(greeting);
}

In addition to setting an absolute expiration, a sliding expiration can be used to keep frequently requested items in the cache:

// keep item in cache as long as it is requested at least
// once every 5 minutes
new MemoryCacheEntryOptions()
  .SetSlidingExpiration(TimeSpan.FromMinutes(5))

To avoid having frequently-accessed cache entries growing too stale (because their sliding expiration is constantly reset), you can combine absolute and sliding expirations:

// keep item in cache as long as it is requested at least
// once every 5 minutes...
// but in any case make sure to refresh it every hour
new MemoryCacheEntryOptions()
  .SetSlidingExpiration(TimeSpan.FromMinutes(5))
  .SetAbsoluteExpiration(TimeSpan.FromHours(1))

By default, an instance of MemoryCache will automatically manage the items stored, removing entries when necessary in response to memory pressure in the app. You can influence the way cache entries are managed by setting their CacheItemPriority when adding the item to the cache. For instance, if you have an item you want to keep in the cache unless you explicitly remove it, you would use the NeverRemove priority option:

// keep item in cache indefinitely unless explicitly removed
new MemoryCacheEntryOptions()
  .SetPriority(CacheItemPriority.NeverRemove))

When you do want to explicitly remove an item from the cache, you can do so easily using the Remove method:

cache.Remove(cacheKey);

Cache Dependencies and Callbacks ¶

You can configure cache entries to depend on other cache entries, the file system, or programmatic tokens, evicting the entry in response to changes. You can register a callback, which will run when a cache item is evicted.

{
    var pause = new ManualResetEvent(false);

    _memoryCache.Set(_cacheKey, _cacheItem,
        new MemoryCacheEntryOptions()
        .RegisterPostEvictionCallback(
            (key, value, reason, substate) =>
            {
                _result = $"'{key}':'{value}' was evicted because: {reason}";
                pause.Set();
            }
        ));

    _memoryCache.Remove(_cacheKey);

    Assert.True(pause.WaitOne(500));

    Assert.Equal("'key':'value' was evicted because: Removed", _result);
}

The callback is run on a different thread from the code that removes the item from the cache.

Warning

If the callback is used to repopulate the cache it is possible other requests for the cache will take place (and find it empty) before the callback completes, possibly resulting in several threads repopulating the cached value.

Possible eviction reasons are:

None: No reason known.
Removed: The item was manually removed by a call to Remove()
Replaced: The item was overwritten.
Expired: The item timed out.
TokenExpired: The token the item depended upon fired an event.
Capacity: The item was removed as part of the cache’s memory management process.

You can specify that one or more cache entries depend on a CancellationTokenSource by adding the expiration token to the MemoryCacheEntryOptions object. When a cached item is invalidated, call Cancel on the token, which will expire all of the associated cache entries (with a reason of TokenExpired). The following unit test demonstrates this:

public void CancellationTokenFiresCallback()
{
    var cts = new CancellationTokenSource();
    var pause = new ManualResetEvent(false);
    _memoryCache.Set(_cacheKey, _cacheItem,
        new MemoryCacheEntryOptions()
        .AddExpirationToken(new CancellationChangeToken(cts.Token))
        .RegisterPostEvictionCallback(
            (key, value, reason, substate) =>
            {
                _result = $"'{key}':'{value}' was evicted because: {reason}";
                pause.Set();
            }
        ));

    // trigger the token
    cts.Cancel();

    Assert.True(pause.WaitOne(500));

    Assert.Equal("'key':'value' was evicted because: TokenExpired", _result);
}

Using a CancellationTokenSource allows multiple cache entries to all be expired without the need to create a dependency between cache entries themselves (in which case, you must ensure that the source cache entry exists before it is used as a dependency for other entries).

Cache entries will inherit triggers and timeouts from other entries accessed while creating the new entry. This approach ensures that subordinate cache entries expire at the same time as related entries.

[Fact]
public void CacheEntryDependencies()
{
    var cts = new CancellationTokenSource();
    var pause = new ManualResetEvent(false);

    using (var cacheEntry = _memoryCache.CreateEntry(_cacheKey))
    {
        _memoryCache.Set("master key", "some value",
            new MemoryCacheEntryOptions()
            .AddExpirationToken(new CancellationChangeToken(cts.Token)));

        cacheEntry.SetValue(_cacheItem)
            .RegisterPostEvictionCallback(
                (key, value, reason, substate) =>
                {
                    _result = $"'{key}':'{value}' was evicted because: {reason}";
                    pause.Set();
                }
            );
    }

    // trigger the token to expire the master item
    cts.Cancel();

    Assert.True(pause.WaitOne(500));

    Assert.Equal("'key':'value' was evicted because: TokenExpired", _result);
}

Note

When one cache entry is used to create another, the new one copies the existing entry’s expiration tokens and time-based expiration settings, if any. It is not expired in response to manual removal or updating of the existing entry.

Other Resources ¶

Working with a Distributed Cache

Working with a Distributed Cache¶

Distributed caches can improve the performance and scalability of ASP.NET Core apps, especially when hosted in a cloud or server farm environment. This article explains how to work with ASP.NET Core’s built-in distributed cache abstractions and implementations.

Sections:

What is a Distributed Cache
The IDistributedCache Interface
Using a Redis Distributed Cache
Using a SQL Server Distributed Cache
Recommendations

What is a Distributed Cache ¶

A distributed cache is shared by multiple app servers (see Caching Basics). The information in the cache is not stored in the memory of individual web servers, and the cached data is available to all of the app’s servers. This provides several advantages:

Cached data is coherent on all web servers. Users don’t see different results depending on which web server handles their request
Cached data survives web server restarts and deployments. Individual web servers can be removed or added without impacting the cache.
The source data store has fewer requests made to it (than with multiple in-memory caches or no cache at all).

Note

If using a SQL Server Distributed Cache, some of these advantages are only true if a separate database instance is used for the cache than for the app’s source data.

Like any cache, a distributed cache can dramatically improve an app’s responsiveness, since typically data can be retrieved from the cache much faster than from a relational database (or web service).

Cache configuration is implementation specific. This article describes how to configure both Redis and SQL Server distributed caches. Regardless of which implementation is selected, the app interacts with the cache using a common IDistributedCache interface.

The IDistributedCache Interface ¶

The IDistributedCache interface includes synchronous and asynchronous methods. The interface allows items to be added, retrieved, and removed from the distributed cache implementation. The IDistributedCache interface includes the following methods:

Get, GetAsync: Takes a string key and retrieves a cached item as a byte[] if found in the cache.
Set, SetAsync: Adds an item (as byte[]) to the cache using a string key.
Refresh, RefreshAsync: Refreshes an item in the cache based on its key, resetting its sliding expiration timeout (if any).
Remove, RemoveAsync: Removes a cache entry based on its key.

To use the IDistributedCache interface:

Specify the dependencies needed in project.json.

Configure the specific implementation of IDistributedCache in your Startup class’s ConfigureServices method, and add it to the container there.

From the app’s Middleware or MVC controller classes, request an instance of IDistributedCache from the constructor. The instance will be provided by Dependency Injection (DI).

Note

There is no need to use a Singleton or Scoped lifetime for IDistributedCache instances (at least for the built-in implementations). You can also create an instance wherever you might need one (instead of using Dependency Injection), but this can make your code harder to test, and violates the Explicit Dependencies Principle.

The following example shows how to use an instance of IDistributedCache in a simple middleware component:

using Microsoft.AspNetCore.Builder;
using Microsoft.AspNetCore.Http;
using Microsoft.Extensions.Caching.Distributed;
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.Threading.Tasks;

namespace DistCacheSample
{
    public class StartTimeHeader
    {
        private readonly RequestDelegate _next;
        private readonly IDistributedCache _cache;

        public StartTimeHeader(RequestDelegate next,
           IDistributedCache cache)
        {
            _next = next;
            _cache = cache;
        }

        public async Task Invoke(HttpContext httpContext)
        {
            string startTimeString = "Not found.";
            var value = await _cache.GetAsync("lastServerStartTime");
            if (value != null)
            {
                startTimeString = Encoding.UTF8.GetString(value);
            }

            httpContext.Response.Headers.Append("Last-Server-Start-Time", startTimeString);

            await _next.Invoke(httpContext);
        }
    }


    // Extension method used to add the middleware to the HTTP request pipeline.
    public static class StartTimeHeaderExtensions
    {
        public static IApplicationBuilder UseStartTimeHeader(this IApplicationBuilder builder)
        {
            return builder.UseMiddleware<StartTimeHeader>();
        }
    }
}

In the code above, the cached value is read, but never written. In this sample, the value is only set when a server starts up, and doesn’t change. In a multi-server scenario, the most recent server to start will overwrite any previous values that were set by other servers. The Get and Set methods use the byte[] type. Therefore, the string value must be converted using Encoding.UTF8.GetString (for Get) and Encoding.UTF8.GetBytes (for Set).

The following code from Startup.cs shows the value being set:

        public void Configure(IApplicationBuilder app,
            IDistributedCache cache)
        {
            var serverStartTimeString = DateTime.Now.ToString();
            byte[] val = Encoding.UTF8.GetBytes(serverStartTimeString);
            cache.Set("lastServerStartTime", val);
            
            app.UseStartTimeHeader();

Note

Since IDistributedCache is configured in the ConfigureServices method, it is available to the Configure method as a parameter. Adding it as a parameter will allow the configured instance to be provided through DI.

Using a Redis Distributed Cache ¶

Redis is an open source in-memory data store, which is often used as a distributed cache. You can use it locally, and you can configure an Azure Redis Cache for your Azure-hosted ASP.NET Core apps. Your ASP.NET Core app configures the cache implementation using a RedisDistributedCache instance.

You configure the Redis implementation in ConfigureServices and access it in your app code by requesting an instance of IDistributedCache (see the code above).

In the sample code, a RedisCache implementation is used when the server is configured for a Staging environment. Thus the ConfigureStagingServices method configures the RedisCache:

        /// <summary>
        /// Use Redis Cache in Staging
        /// </summary>
        /// <param name="services"></param>
        public void ConfigureStagingServices(IServiceCollection services)
        {

            services.AddDistributedRedisCache(options =>
            {
                options.Configuration = "localhost";
                options.InstanceName = "SampleInstance";
            });
        }

Note

To install Redis on your local machine, install the chocolatey package http://chocolatey.org/packages/redis-64/ and run redis-server from a command prompt.

Using a SQL Server Distributed Cache ¶

The SqlServerCache implementation allows the distributed cache to use a SQL Server database as its backing store. To create SQL Server table you can use sql-cache tool, the tool creates a table with the name and schema you specify.

To use sql-cache tool add SqlConfig.Tools to the tools section of the project.json file and run dotnet restore.

  "tools": {
    "Microsoft.AspNetCore.Server.IISIntegration.Tools": {
      "version": "1.0.0-preview2-final",
      "imports": "portable-net45+win8+dnxcore50"
    },
    "Microsoft.Extensions.Caching.SqlConfig.Tools": "1.0.0-preview2-final"
  },

Test SqlConfig.Tools by running the following command

C:\DistCacheSample\src\DistCacheSample>dotnet sql-cache create --help

sql-cache tool will display usage, options and command help, now you can create tables into sql server, running “sql-cache create” command :

C:\DistCacheSample\src\DistCacheSample>dotnet sql-cache create "Data Source=(localdb)\v11.0;Initial Catalog=DistCache;Integrated Security=True;" dbo TestCache
info: Microsoft.Extensions.Caching.SqlConfig.Tools.Program[0]
    Table and index were created successfully.

The created table have the following schema:

Like all cache implementations, your app should get and set cache values using an instance of IDistributedCache, not a SqlServerCache. The sample implements SqlServerCache in the Production environment (so it is configured in ConfigureProductionServices).

        /// Use SQL Server Cache in Production
        /// </summary>
        /// <param name="services"></param>
        public void ConfigureProductionServices(IServiceCollection services)
        {
          
            services.AddDistributedSqlServerCache(options =>
            {
                options.ConnectionString = @"Data Source=(localdb)\v11.0;Initial Catalog=DistCache;Integrated Security=True;";
                options.SchemaName = "dbo";
                options.TableName = "TestCache";
            });
            
        }

Note

The ConnectionString (and optionally, SchemaName and TableName) should typically be stored outside of source control (such as UserSecrets), as they may contain credentials.

Recommendations ¶

When deciding which implementation of IDistributedCache is right for your app, choose between Redis and SQL Server based on your existing infrastructure and environment, your performance requirements, and your team’s experience. If your team is more comfortable working with Redis, it’s an excellent choice. If your team prefers SQL Server, you can be confident in that implementation as well. Note that A traditional caching solution stores data in-memory which allows for fast retrieval of data. You should store commonly used data in a cache and store the entire data in a backend persistent store such as SQL Server or Azure Storage. Redis Cache is a caching solution which gives you high throughput and low latency as compared to SQL Cache. Also, you should avoid using the in-memory implementation (MemoryCache) in multi-server environments.

Azure Resources:

Tip

The in-memory implementation of IDistributedCache should only be used for testing purposes or for applications that are hosted on just one server instance.

Response Caching¶

Steve Smith

Sections:

What is Response Caching
ResponseCache Attribute

What is Response Caching ¶

Response caching refers to specifying cache-related headers on HTTP responses made by ASP.NET Core MVC actions. These headers specify how you want client and intermediate (proxy) machines to cache responses to certain requests (if at all). This can reduce the number of requests a client or proxy makes to the web server, since future requests for the same action may be served from the client or proxy’s cache. In this case, the request is never made to the web server.

The primary HTTP header used for caching is Cache-Control. The HTTP 1.1 specification details many options for this directive. Three common directives are:

public: Indicates that the response may be cached.
private: Indicates the response is intended for a single user and must not be cached by a shared cache. The response could still be cached in a private cache (for instance, by the user’s browser).
no-cache: Indicates the response must not be used by a cache to satisfy any subsequent request (without successful revalidation with the origin server).

Note

Response caching does not cache responses on the web server. It differs from output caching, which would cache responses in memory on the server in earlier versions of ASP.NET and ASP.NET MVC. Output caching middleware is planned to be added to ASP.NET Core in a future release.

Additional HTTP headers used for caching include Pragma and Vary, which are described below. Learn more about Caching in HTTP from the specification.

ResponseCache Attribute ¶

The ResponseCacheAttribute is used to specify how a controller action’s headers should be set to control its cache behavior. The attribute has the following properties, all of which are optional unless otherwise noted.

Duration int: The maximum duration (in seconds) the response should be cached. Required unless NoStore is true.
Location ResponseCacheLocation: The location where the response may be cached. May be Any, None, or Client. Default is Any.
NoStore bool: Determines whether the value should be stored or not, and overrides other property values. When true, Duration is ignored and Location is ignored for values other than None.
VaryByHeader string: When set, a vary response header will be written with the response.
CacheProfileName string: When set, determines the name of the cache profile to use.
Order int: The order of the filter (from IOrderedFilter).

The ResponseCacheAttribute is used to configure and create (via IFilterFactory) a ResponseCacheFilter, which performs the work of writing the appropriate HTTP headers to the response. The filter will first remove any existing headers for Vary, Cache-Control, and Pragma, and then will write out the appropriate headers based on the properties set in the ResponseCacheAttribute.

The Vary Header¶

This header is only written when the VaryByHeader propery is set, in which case it is set to that property’s value.

NoStore and Location.None¶

NoStore is a special property that overrides most of the other properties. When this property is set to true, the Cache-Control header will be set to “no-store”. Additionally, if Location is set to None, then Cache-Control will be set to “no-store, no-cache” and Pragma is likewise set to no-cache. (If NoStore is false and Location is None, then both Cache-Control and Pragma will be set to no-cache).

A good scenario in which to set NoStore to true is error pages. It’s unlikely you would want to respond to a user’s request with the error response a different user previously generated, and such responses may include stack traces and other sensitive information that shouldn’t be stored on intermdiate servers. For example:

[ResponseCache(Location = ResponseCacheLocation.None, NoStore = true)]
public IActionResult Error()
{
    return View();
}

This will result in the following headers:

Cache-Control: no-store,no-cache
Pragma: no-cache

Location and Duration¶

To enable caching, Duration must be set to a positive value and Location must be either Any (the default) or Client. In this case, the Cache-Control header will be set to the location value followed by the “max-age” of the response.

Note

Location‘s options of Any and Client translate into Cache-Control header values of public and private, respectively. As noted previously, setting Location to None will set both Cache-Control and Pragma headers to no-cache.

Below is an example showing the headers produced by setting Duration and leaving the default Location value.

[ResponseCache(Duration = 60)]
public IActionResult Contact()
{
    ViewData["Message"] = "Your contact page.";

    return View();
}

Produces the following headers:

Cache-Control: public,max-age=60

Cache Profiles¶

Instead of duplicating ResponseCache settings on many controller action attributes, cache profiles can be configured as options when setting up MVC in the ConfigureServices method in Startup. Values found in a referenced cache profile will be used as the defaults by the ResponseCache attribute, and will be overridden by any properties specified on the attribute.

Setting up a cache profile:

// This method gets called by the web host. Use this method to configure the HTTP request pipeline.
public void Configure(IApplicationBuilder app)
{
    app.UseMvc(routes =>
    {
        routes.MapRoute(
            name: "default",
            template: "{controller=Home}/{action=Index}/{id?}");
    });
}

// Entry point for the application.
public static void Main(string[] args)
{
    var host = new WebHostBuilder()
        .UseKestrel()

Referencing a cache profile:

[ResponseCache(Duration = 30)]
public class HomeController : Controller
{
    [ResponseCache(CacheProfileName = "Default")]
    public IActionResult Index()
    {
        return View();
    }
}

Tip

The ResponseCache attribute can be applied both to actions (methods) as well as controllers (classes). Method-level attributes will override the settings specified in class-level attributes.

In the above example, a class-level attribute specifies a duration of 30 seconds while a method-level attributes references a cache profile with a duration set to 60 seconds.

The resulting header:

Cache-Control: public,max-age=60

🔧 Output Caching¶

Note

We are currently working on this topic.

We welcome your input to help shape the scope and approach. You can track the status and provide input on this issue at GitHub.

If you would like to review early drafts and outlines of this topic, please leave a note with your contact information in the issue.

Learn more about how you can contribute on GitHub.

Migration¶

Migrating From ASP.NET MVC to ASP.NET Core MVC¶

By Rick Anderson, Daniel Roth, Steve Smith and Scott Addie

This article shows how to get started migrating an ASP.NET MVC project to ASP.NET Core MVC. In the process, it highlights many of the things that have changed from ASP.NET MVC. Migrating from ASP.NET MVC is a multiple step process and this article covers the initial setup, basic controllers and views, static content, and client-side dependencies. Additional articles cover migrating configuration and identity code found in many ASP.NET MVC projects.

Sections:

Create the starter ASP.NET MVC project
Create the ASP.NET Core project
Configure the site to use MVC
Add a controller and view
Controllers and views
Static content
Gulp
NPM
Migrate the layout file
Configure Bundling
Solving HTTP 500 errors
Additional Resources

Create the starter ASP.NET MVC project ¶

To demonstrate the upgrade, we’ll start by creating a ASP.NET MVC app. Create it with the name WebApp1 so the namespace will match the ASP.NET Core project we create in the next step.

_images/new-project-select-mvc-template.png

Optional: Change the name of the Solution from WebApp1 to Mvc5. Visual Studio will display the new solution name (Mvc5), which will make it easier to tell this project from the next project.

Create the ASP.NET Core project ¶

Create a new empty ASP.NET Core web app with the same name as the previous project (WebApp1) so the namespaces in the two projects match. Having the same namespace makes it easier to copy code between the two projects. You’ll have to create this project in a different directory than the previous project to use the same name.

_images/new-project-select-empty-aspnet5-template.png

Optional: Create a new ASP.NET Core app named WebApp1 with authentication set to Individual User Accounts. Rename this app FullAspNetCore. Creating this project will save you time in the conversion. You can look at the template generated code to see the end result or to copy code to the conversion project. It’s also helpful when you get stuck on a conversion step to compare with the template generated project.

Configure the site to use MVC ¶

Open the project.json file.

Add Microsoft.AspNetCore.Mvc and Microsoft.AspNetCore.StaticFiles to the dependencies property:
Add the “prepublish” line to the “scripts” section:

"scripts": {
   "prepublish": [ "npm install", "bower install", "gulp clean", "gulp min" ],
       "postpublish": [ "dotnet publish-iis
         --publish-folder %publish:OutputPath%
         --framework %publish:FullTargetFramework%" ]
}

Microsoft.AspNetCore.Mvc Installs in the ASP.NET Core MVC framework package
Microsoft.AspNetCore.StaticFiles is the static file handler. The ASP.NET runtime is modular, and you must explicitly opt in to serve static files (see Working with Static Files).
The “scripts”/”prepublish” line is needed for Gulp, Bower and NPM. We’ll talk about that later.
Open the Startup.cs file and change the code to match the following:

public class Startup
{
    // This method gets called by the runtime. Use this method to add services to the container.
    // For more information on how to configure your application, visit http://go.microsoft.com/fwlink/?LinkID=398940
    public void ConfigureServices(IServiceCollection services)
    {
        services.AddMvc();
    }

    // This method gets called by the runtime. Use this method to configure the HTTP request pipeline.
    public void Configure(IApplicationBuilder app)
    {
        app.UseStaticFiles();

        app.UseMvc(routes =>
        {
            routes.MapRoute(
                name: "default",
                template: "{controller=Home}/{action=Index}/{id?}");
        });
    }
}

UseStaticFiles adds the static file handler. As mentioned previously, the ASP.NET runtime is modular, and you must explicitly opt in to serve static files. app.UseMvc(routes => adds routing. For more information, see Application Startup and Routing.

Add a controller and view ¶

In this section, you’ll add a minimal controller and view to serve as placeholders for the ASP.NET MVC controller and views you’ll migrate in the next section.

Add a Controllers folder.
Add an MVC controller class with the name HomeController.cs to the Controllers folder.

Add a Views folder.
Add a Views/Home folder.
Add an Index.cshtml MVC view page to the Views/Home folder.

The project structure is shown below:

_images/project-structure-controller-view.png

Replace the contents of the Views/Home/Index.cshtml file with the following:

<h1>Hello world!</h1>

Run the app.

See Controllers and Views for more information.

Now that we have a minimal working ASP.NET Core project, we can start migrating functionality from the ASP.NET MVC project. We will need to move the following:

client-side content (CSS, fonts, and scripts)
controllers
views
models
bundling
filters
Log in/out, identity (This will be done in the next tutorial.)

Controllers and views ¶

Copy each of the methods from the ASP.NET MVC HomeController to the new HomeController. Note that in ASP.NET MVC, the built-in template’s controller action method return type is ActionResult; in ASP.NET Core MVC the actions return IActionResult instead. ActionResult implements IActionResult, so there is no need to change the return type of your action methods.
Copy the About.cshtml, Contact.cshtml, and Index.cshtml Razor view files from the ASP.NET MVC project to the ASP.NET Core project.
Run the ASP.NET Core app and test each method. We haven’t migrated the layout file or styles yet, so the rendered views will only contain the content in the view files. You won’t have the layout file generated links for the About and Contact views, so you’ll have to invoke them from the browser (replace 2468 with the port number used in your project).
- http://localhost:2468/home/about
- http://localhost:2468/home/contact

Note the lack of styling and menu items. We’ll fix that in the next section.

Static content ¶

In previous versions of ASP.NET MVC, static content was hosted from the root of the web project and was intermixed with server-side files. In ASP.NET Core, static content is hosted in the wwwroot folder. You’ll want to copy the static content from your old ASP.NET MVC app to the wwwroot folder in your ASP.NET Core project. In this sample conversion:

Copy the favicon.ico file from the old MVC project to the wwwroot folder in the ASP.NET Core project.

The old ASP.NET MVC project uses Bootstrap for its styling and stores the Bootstrap files in the Content and Scripts folders. The template-generated old ASP.NET MVC project references Bootstrap in the layout file (Views/Shared/_Layout.cshtml). You could copy the bootstrap.js and bootstrap.css files from the ASP.NET MVC project to the wwwroot folder in the new project, but that approach doesn’t use the improved mechanism for managing client-side dependencies in ASP.NET Core.

In the new project, we’ll add support for Bootstrap (and other client-side libraries) using Bower:

Add a Bower configuration file named bower.json to the project root (Right-click on the project, and then Add > New Item > Bower Configuration File). Add Bootstrap and jQuery to the file ^[1]: (see the highlighted lines below).

{
  "name": "asp.net",
  "private": true,
  "dependencies": {
    "bootstrap": "3.3.6",
    "jquery": "2.2.0"
  }
}

Upon saving the file, Bower will automatically download the dependencies to the wwwroot/lib folder. You can use the Search Solution Explorer box to find the path of the assets.

Note

bower.json is not visible in Solution Explorer. You can display the hidden .json files by selecting the project in Solution Explorer and then tapping the Show All Files icon. You won’t see Show All Files unless the project is selected.

See Manage Client-Side Packages with Bower for more information.

Gulp ¶

When you create a new web app using the ASP.NET Core Web Application template, the project is setup to use Gulp. Gulp is a streaming build system for client-side code (HTML, LESS, SASS, etc.). The gulpfile.js contains JavaScript that defines a set of gulp tasks that you can set to run automatically on build events or you can run manually using Task Runner Explorer. In this section we’ll show how to use gulpfile.js to bundle and minify the projects JavaScript and CSS files.

If you created the optional FullAspNetCore project (a new ASP.NET Core web app with Individual User Accounts), add gulpfile.js from that project to the project we are updating. In Solution Explorer, right-click the web app project and choose Add > Existing Item.

Navigate to gulpfile.js from the new ASP.NET Core web app with Individual User Accounts and add the add gulpfile.js file. Alternatively, right-click the web app project and choose Add > New Item. Select Gulp Configuration File, and name the file gulpfile.js. The gulpfile.js file:

/// <binding Clean='clean' />
"use strict";

var gulp = require("gulp"),
    rimraf = require("rimraf"),
    concat = require("gulp-concat"),
    cssmin = require("gulp-cssmin"),
    uglify = require("gulp-uglify");

var webroot = "./wwwroot/";

var paths = {
    js: webroot + "js/**/*.js",
    minJs: webroot + "js/**/*.min.js",
    css: webroot + "css/**/*.css",
    minCss: webroot + "css/**/*.min.css",
    concatJsDest: webroot + "js/site.min.js",
    concatCssDest: webroot + "css/site.min.css"
};

gulp.task("clean:js", function (cb) {
    rimraf(paths.concatJsDest, cb);
});

gulp.task("clean:css", function (cb) {
    rimraf(paths.concatCssDest, cb);
});

gulp.task("clean", ["clean:js", "clean:css"]);

gulp.task("min:js", function () {
    return gulp.src([paths.js, "!" + paths.minJs], { base: "." })
        .pipe(concat(paths.concatJsDest))
        .pipe(uglify())
        .pipe(gulp.dest("."));
});

gulp.task("min:css", function () {
    return gulp.src([paths.css, "!" + paths.minCss])
        .pipe(concat(paths.concatCssDest))
        .pipe(cssmin())
        .pipe(gulp.dest("."));
});

gulp.task("min", ["min:js", "min:css"]);

The code above performs these functions:

Cleans (deletes) the target files.
Minifies the JavaScript and CSS files.
Bundles (concatenates) the JavaScript and CSS files.

See Using Gulp.

NPM ¶

NPM (Node Package Manager) is a package manager which is used to acquire tooling such as Bower and Gulp; and, it is fully supported in Visual Studio. We’ll use NPM to manage Gulp dependencies.

If you created the optional FullAspNetCore project, add the package.json NPM file from that project to the project we are updating. The package.json NPM file lists the dependencies for the client-side build processes defined in gulpfile.js.

Note

package.json is not visible in Solution Explorer.

Right-click the web app project, choose Add > Existing Item, and add the package.json NPM file. Alternatively, you can add a new NPM configuration file as follows:

In Solution Explorer, right-click the project
Select Add > New Item
Select NPM Configuration File
Leave the default name: package.json
Tap Add

The package.json file ^[1]:

{
  "name": "asp.net",
  "version": "0.0.0",
  "private": true,
  "devDependencies": {
    "gulp": "3.8.11",
    "gulp-concat": "2.5.2",
    "gulp-cssmin": "0.1.7",
    "gulp-uglify": "1.2.0",
    "rimraf": "2.2.8"
  }
}

Right-click on gulpfile.js and select Task Runner Explorer. Double-click on a task to run it.

For more information, see Client-Side Development.

Migrate the layout file ¶

Copy the _ViewStart.cshtml file from the old ASP.NET MVC project’s Views folder into the ASP.NET Core project’s Views folder. The _ViewStart.cshtml file has not changed in ASP.NET Core MVC.
Create a Views/Shared folder.
Optional: Copy _ViewImports.cshtml from the old MVC project’s Views folder into the ASP.NET Core project’s Views folder. Remove any namespace declaration in the _ViewImports.cshtml file. The _ViewImports.cshtml file provides namespaces for all the view files and brings in doc:Tag Helpers </mvc/views/tag-helpers/index>. Tag Helpers are used in the new layout file. The _ViewImports.cshtml file is new for ASP.NET Core
Copy the _Layout.cshtml file from the old ASP.NET MVC project’s Views/Shared folder into the ASP.NET Core project’s Views/Shared folder.

Open _Layout.cshtml file and make the following changes (the completed code is shown below):

Replace @Styles.Render(“~/Content/css”) with a <link> element to load bootstrap.css (see below)

Remove @Scripts.Render(“~/bundles/modernizr”)

Comment out the @Html.Partial(“_LoginPartial”) line (surround the line with @*...*@) - we’ll return to it in a future tutorial

Replace @Scripts.Render(“~/bundles/jquery”) with a <script> element (see below)

Replace @Scripts.Render(“~/bundles/bootstrap”) with a <script> element (see below)

The replacement CSS link:

<link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />

The replacement script tags:

<script src="~/lib/jquery/dist/jquery.js"></script>
<script src="~/lib/bootstrap/dist/js/bootstrap.js"></script>

The updated _Layout.cshtml file is shown below:

<!DOCTYPE html>
<html>
<head>
    <meta charset="utf-8" />
    <meta name="viewport" content="width=device-width, initial-scale=1.0">
    <title>@ViewBag.Title - My ASP.NET Application</title>
    <link rel="stylesheet" href="~/lib/bootstrap/dist/css/bootstrap.css" />
</head>
<body>
    <div class="navbar navbar-inverse navbar-fixed-top">
        <div class="container">
            <div class="navbar-header">
                <button type="button" class="navbar-toggle" data-toggle="collapse" data-target=".navbar-collapse">
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                    <span class="icon-bar"></span>
                </button>
                @Html.ActionLink("Application name", "Index", "Home", new { area = "" }, new { @class = "navbar-brand" })
            </div>
            <div class="navbar-collapse collapse">
                <ul class="nav navbar-nav">
                    <li>@Html.ActionLink("Home", "Index", "Home")</li>
                    <li>@Html.ActionLink("About", "About", "Home")</li>
                    <li>@Html.ActionLink("Contact", "Contact", "Home")</li>
                </ul>
                @*@Html.Partial("_LoginPartial")*@
            </div>
        </div>
    </div>
    <div class="container body-content">
        @RenderBody()
        <hr />
        <footer>
            <p>&copy; @DateTime.Now.Year - My ASP.NET Application</p>
        </footer>
    </div>

    <script src="~/lib/jquery/dist/jquery.js"></script>
    <script src="~/lib/bootstrap/dist/js/bootstrap.js"></script>
    @RenderSection("scripts", required: false)
</body>
</html>

View the site in the browser. It should now load correctly, with the expected styles in place.

Optional: You might want to try using the new layout file. For this project you can copy the layout file from the FullAspNetCore project. The new layout file uses Tag Helpers and has other improvements.

Configure Bundling ¶

The ASP.NET MVC starter web template utilized the ASP.NET Web Optimization for bundling. In ASP.NET Core, this functionality is performed as part of the build process using Gulp. We’ve previously configured bundling and minification; all that’s left is to change the references to Bootstrap, jQuery and other assets to use the bundled and minified versions. You can see how this is done in the layout file (Views/Shared/_Layout.cshtml) of the full template project. See Bundling and Minification for more information.

Solving HTTP 500 errors ¶

There are many problems that can cause a HTTP 500 error message that contain no information on the source of the problem. For example, if the Views/_ViewImports.cshtml file contains a namespace that doesn’t exist in your project, you’ll get a HTTP 500 error. To get a detailed error message, add the following code:

 public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
 {
     if (env.IsDevelopment())
     {
          app.UseDeveloperExceptionPage();
     }

     app.UseStaticFiles();

     app.UseMvc(routes =>
     {
         routes.MapRoute(
             name: "default",
             template: "{controller=Home}/{action=Index}/{id?}");
     });
 }

See Using the Developer Exception Page in Error Handling for more information.

Additional Resources ¶

[1] The version numbers in the samples might not be current. You may need to update your projects accordingly.

Migrating Configuration¶

By Steve Smith and Scott Addie

In the previous article, we began migrating an ASP.NET MVC project to ASP.NET Core MVC. In this article, we migrate configuration.

Sections:

Setup Configuration
Migrate Configuration Settings from web.config
Summary

Setup Configuration ¶

ASP.NET Core no longer uses the Global.asax and web.config files that previous versions of ASP.NET utilized. In earlier versions of ASP.NET, application startup logic was placed in an Application_StartUp method within Global.asax. Later, in ASP.NET MVC, a Startup.cs file was included in the root of the project; and, it was called when the application started. ASP.NET Core has adopted this approach completely placing all startup logic in the Startup.cs file.

The web.config file has also been replaced in ASP.NET Core. Configuration itself can now be configured, as part of the application startup procedure described in Startup.cs. Configuration can still utilize XML files, but typically ASP.NET Core projects will place configuration values in a JSON-formatted file, such as appsettings.json. ASP.NET Core’s configuration system can also easily access environment variables, which can provide a more secure and robust location for environment-specific values. This is especially true for secrets like connection strings and API keys that should not be checked into source control. See Configuration to learn more about configuration in ASP.NET Core.

For this article, we are starting with the partially-migrated ASP.NET Core project from the previous article. To setup configuration add the following constructor and property to the Startup.cs class located in the root of the project:

public Startup(IHostingEnvironment env)
{
    // Set up configuration sources.
    var builder = new ConfigurationBuilder()
        .AddJsonFile("appsettings.json")
        .AddEnvironmentVariables();
    Configuration = builder.Build();
}

public IConfigurationRoot Configuration { get; set; }

Note that at this point the Startup.cs file will not compile, as we still need to add the following using statement:

using Microsoft.Extensions.Configuration;

Add an appsettings.json file to the root of the project using the appropriate item template:

Migrate Configuration Settings from web.config ¶

Our ASP.NET MVC project included the required database connection string in web.config, in the <connectionStrings> element. In our ASP.NET Core project, we are going to store this information in the appsettings.json file. Open appsettings.json, and note that it already includes the following:

{
	"Data": {
		"DefaultConnection": {
			"ConnectionString": "Server=(localdb)\\MSSQLLocalDB;Database=_CHANGE_ME;Trusted_Connection=True;"
		}
	}
}

In the highlighted line depicted above, change the name of the database from _CHANGE_ME to the name of your database.

Summary ¶

ASP.NET Core places all startup logic for the application in a single file, in which the necessary services and dependencies can be defined and configured. It replaces the web.config file with a flexible configuration feature that can leverage a variety of file formats, such as JSON, as well as environment variables.

Migrating Authentication and Identity¶

In the previous article we migrated configuration from an ASP.NET MVC project to ASP.NET Core MVC. In this article, we migrate the registration, login, and user management features.

Sections:

Configure Identity and Membership
Migrate Registration and Login Logic
Summary

Configure Identity and Membership ¶

In ASP.NET MVC, authentication and identity features are configured using ASP.NET Identity in Startup.Auth.cs and IdentityConfig.cs, located in the App_Start folder. In ASP.NET Core MVC, these features are configured in Startup.cs. Before pulling in the required services and configuring them, we should add the required dependencies to the project. Open project.json and add Microsoft.AspNetCore.Identity.EntityFramework and Microsoft.AspNetCore.Identity.Cookies to the list of dependencies:

"dependencies": {
  "Microsoft.AspNetCore.Mvc": "1.0.0",
  "Microsoft.AspNetCore.Identity.EntityFramework": "1.0.0",
  "Microsoft.AspNetCore.Security.Cookies": "1.0.0"
},

Now, open Startup.cs and update the ConfigureServices() method to use Entity Framework and Identity services:

public void ConfigureServices(IServiceCollection services)
{
  // Add EF services to the services container.
  services.AddEntityFramework(Configuration)
    .AddSqlServer()
    .AddDbContext<ApplicationDbContext>();

  // Add Identity services to the services container.
  services.AddIdentity<ApplicationUser, IdentityRole>(Configuration)
    .AddEntityFrameworkStores<ApplicationDbContext>();

  services.AddMvc();
}

At this point, there are two types referenced in the above code that we haven’t yet migrated from the ASP.NET MVC project: ApplicationDbContext and ApplicationUser. Create a new Models folder in the ASP.NET Core project, and add two classes to it corresponding to these types. You will find the ASP.NET MVC versions of these classes in /Models/IdentityModels.cs, but we will use one file per class in the migrated project since that’s more clear.

ApplicationUser.cs:

using Microsoft.AspNetCore.Identity.EntityFrameworkCore;

namespace NewMvc6Project.Models
{
  public class ApplicationUser : IdentityUser
  {
  }
}

ApplicationDbContext.cs:

using Microsoft.AspNetCore.Identity.EntityFramework;
using Microsoft.Data.Entity;

namespace NewMvc6Project.Models
{
  public class ApplicationDbContext : IdentityDbContext<ApplicationUser>
  {
    private static bool _created = false;
    public ApplicationDbContext()
    {
      // Create the database and schema if it doesn't exist
      // This is a temporary workaround to create database until Entity Framework database migrations
      // are supported in ASP.NET Core
      if (!_created)
      {
        Database.AsMigrationsEnabled().ApplyMigrations();
        _created = true;
      }
    }

    protected override void OnConfiguring(DbContextOptions options)
    {
      options.UseSqlServer();
    }
  }
}

The ASP.NET Core MVC Starter Web project doesn’t include much customization of users, or the ApplicationDbContext. When migrating a real application, you will also need to migrate all of the custom properties and methods of your application’s user and DbContext classes, as well as any other Model classes your application utilizes (for example, if your DbContext has a DbSet<Album>, you will of course need to migrate the Album class).

With these files in place, the Startup.cs file can be made to compile by updating its using statements:

using Microsoft.Framework.ConfigurationModel;
using Microsoft.AspNetCore.Hosting;
using NewMvc6Project.Models;
using Microsoft.AspNetCore.Identity;

Our application is now ready to support authentication and identity services - it just needs to have these features exposed to users.

Migrate Registration and Login Logic ¶

With identity services configured for the application and data access configured using Entity Framework and SQL Server, we are now ready to add support for registration and login to the application. Recall that earlier in the migration process we commented out a reference to _LoginPartial in _Layout.cshtml. Now it’s time to return to that code, uncomment it, and add in the necessary controllers and views to support login functionality.

Update _Layout.cshtml; uncomment the @Html.Partial line:

      <li>@Html.ActionLink("Contact", "Contact", "Home")</li>
    </ul>
    @*@Html.Partial("_LoginPartial")*@
  </div>
</div>

Now, add a new MVC View Page called _LoginPartial to the Views/Shared folder:

Update _LoginPartial.cshtml with the following code (replace all of its contents):

@using System.Security.Principal

@if (User.Identity.IsAuthenticated)
{
    using (Html.BeginForm("LogOff", "Account", FormMethod.Post, new { id = "logoutForm", @class = "navbar-right" }))
    {
        @Html.AntiForgeryToken()
        <ul class="nav navbar-nav navbar-right">
            <li>
                @Html.ActionLink("Hello " + User.Identity.GetUserName() + "!", "Manage", "Account", routeValues: null, htmlAttributes: new { title = "Manage" })
            </li>
            <li><a href="javascript:document.getElementById('logoutForm').submit()">Log off</a></li>
        </ul>
    }
}
else
{
    <ul class="nav navbar-nav navbar-right">
        <li>@Html.ActionLink("Register", "Register", "Account", routeValues: null, htmlAttributes: new { id = "registerLink" })</li>
        <li>@Html.ActionLink("Log in", "Login", "Account", routeValues: null, htmlAttributes: new { id = "loginLink" })</li>
    </ul>
}

At this point, you should be able to refresh the site in your browser.

Summary ¶

ASP.NET Core introduces changes to the ASP.NET Identity features. In this article, you have seen how to migrate the authentication and user management features of an ASP.NET Identity to ASP.NET Core.

Migrating from ASP.NET Web API¶

By Steve Smith and Scott Addie

Web APIs are HTTP services that reach a broad range of clients, including browsers and mobile devices. ASP.NET Core MVC includes support for building Web APIs providing a single, consistent way of building web applications. In this article, we demonstrate the steps required to migrate a Web API implementation from ASP.NET Web API to ASP.NET Core MVC.

Sections::

Review ASP.NET Web API Project
Create the Destination Project
Migrate Configuration
Migrate Models and Controllers
Summary

Review ASP.NET Web API Project ¶

This article uses the sample project, ProductsApp, created in the article Getting Started with ASP.NET Web API as its starting point. In that project, a simple ASP.NET Web API project is configured as follows.

In Global.asax.cs, a call is made to WebApiConfig.Register:

using System;
using System.Collections.Generic;
using System.Linq;
using System.Web;
using System.Web.Http;
using System.Web.Routing;

namespace ProductsApp
{
    public class WebApiApplication : System.Web.HttpApplication
    {
        protected void Application_Start()
        {
            GlobalConfiguration.Configure(WebApiConfig.Register);
        }
    }
}

WebApiConfig is defined in App_Start, and has just one static Register method:

using System;
using System.Collections.Generic;
using System.Linq;
using System.Web.Http;

namespace ProductsApp
{
    public static class WebApiConfig
    {
        public static void Register(HttpConfiguration config)
        {
            // Web API configuration and services

            // Web API routes
            config.MapHttpAttributeRoutes();

            config.Routes.MapHttpRoute(
                name: "DefaultApi",
                routeTemplate: "api/{controller}/{id}",
                defaults: new { id = RouteParameter.Optional }
            );
        }
    }
}

This class configures attribute routing, although it’s not actually being used in the project. It also configures the routing table which is used by ASP.NET Web API. In this case, ASP.NET Web API will expect URLs to match the format /api/{controller}/{id}, with {id} being optional.

The ProductsApp project includes just one simple controller, which inherits from ApiController and exposes two methods:

using ProductsApp.Models;
using System;
using System.Collections.Generic;
using System.Linq;
using System.Net;
using System.Web.Http;

namespace ProductsApp.Controllers
{
    public class ProductsController : ApiController
    {
        Product[] products = new Product[] 
        { 
            new Product { Id = 1, Name = "Tomato Soup", Category = "Groceries", Price = 1 }, 
            new Product { Id = 2, Name = "Yo-yo", Category = "Toys", Price = 3.75M }, 
            new Product { Id = 3, Name = "Hammer", Category = "Hardware", Price = 16.99M } 
        };

        public IEnumerable<Product> GetAllProducts()
        {
            return products;
        }

        public IHttpActionResult GetProduct(int id)
        {
            var product = products.FirstOrDefault((p) => p.Id == id);
            if (product == null)
            {
                return NotFound();
            }
            return Ok(product);
        }
    }
}

Finally, the model, Product, used by the ProductsApp, is a simple class:

namespace ProductsApp.Models
{
    public class Product
    {
        public int Id { get; set; }
        public string Name { get; set; }
        public string Category { get; set; }
        public decimal Price { get; set; }
    }
}

Now that we have a simple project from which to start, we can demonstrate how to migrate this Web API project to ASP.NET Core MVC.

Create the Destination Project ¶

Using Visual Studio, create a new, empty solution, and name it WebAPIMigration. Add the existing ProductsApp project to it, then, add a new ASP.NET Core Web Application Project to the solution. Name the new project ProductsCore.

Next, choose the Web API project template. We will migrate the ProductsApp contents to this new project.

Delete the Project_Readme.html file from the new project. Your solution should now look like this:

Migrate Configuration ¶

ASP.NET Core no longer uses Global.asax, web.config, or App_Start folders. Instead, all startup tasks are done in Startup.cs in the root of the project (see Application Startup). In ASP.NET Core MVC attribute-based routing is now included by default when UseMvc() is called and this is the recommended approach for configuring Web API routes (and is how the Web API starter project handles routing).

using System;
using System.Collections.Generic;
using System.Linq;
using System.Threading.Tasks;
using Microsoft.AspNet.Builder;
using Microsoft.AspNet.Hosting;
using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Logging;

namespace ProductsCore
{
    public class Startup
    {
        public Startup(IHostingEnvironment env)
        {
            // Set up configuration sources.
            var builder = new ConfigurationBuilder()
                .AddJsonFile("appsettings.json")
                .AddEnvironmentVariables();
            Configuration = builder.Build();
        }

        public IConfigurationRoot Configuration { get; set; }

        // This method gets called by the runtime. Use this method to add services to the container.
        public void ConfigureServices(IServiceCollection services)
        {
            // Add framework services.
            services.AddMvc();
        }

        // This method gets called by the runtime. Use this method to configure the HTTP request pipeline.
        public void Configure(IApplicationBuilder app, IHostingEnvironment env, ILoggerFactory loggerFactory)
        {
            loggerFactory.AddConsole(Configuration.GetSection("Logging"));
            loggerFactory.AddDebug();

            app.UseIISPlatformHandler();

            app.UseStaticFiles();

            app.UseMvc();
        }

        // Entry point for the application.
        public static void Main(string[] args) => WebApplication.Run<Startup>(args);
    }
}

Assuming you want to use attribute routing in your project going forward, no additional configuration is needed. Simply apply the attributes as needed to your controllers and actions, as is done in the sample ValuesController class that is included in the Web API starter project:

using System;
using System.Collections.Generic;
using System.Linq;
using Microsoft.AspNet.Mvc;

namespace ProductsCore.Controllers
{
    [Route("api/[controller]")]
    public class ValuesController : Controller
    {
        // GET: api/values
        [HttpGet]
        public IEnumerable<string> Get()
        {
            return new string[] { "value1", "value2" };
        }

        // GET api/values/5
        [HttpGet("{id}")]
        public string Get(int id)
        {
            return "value";
        }

        // POST api/values
        [HttpPost]
        public void Post([FromBody]string value)
        {
        }

        // PUT api/values/5
        [HttpPut("{id}")]
        public void Put(int id, [FromBody]string value)
        {
        }

        // DELETE api/values/5
        [HttpDelete("{id}")]
        public void Delete(int id)
        {
        }
    }
}

Note the presence of [controller] on line 8. Attribute-based routing now supports certain tokens, such as [controller] and [action]. These tokens are replaced at runtime with the name of the controller or action, respectively, to which the attribute has been applied. This serves to reduce the number of magic strings in the project, and it ensures the routes will be kept synchronized with their corresponding controllers and actions when automatic rename refactorings are applied.

To migrate the Products API controller, we must first copy ProductsController to the new project. Then simply include the route attribute on the controller:

[Route("api/[controller]")]

You also need to add the [HttpGet] attribute to the two methods, since they both should be called via HTTP Get. Include the expectation of an “id” parameter in the attribute for GetProduct():

// /api/products
[HttpGet]
...

// /api/products/1
[HttpGet("{id}")]

At this point, routing is configured correctly; however, we can’t yet test it. Additional changes must be made before ProductsController will compile.

Migrate Models and Controllers ¶

The last step in the migration process for this simple Web API project is to copy over the Controllers and any Models they use. In this case, simply copy Controllers/ProductsController.cs from the original project to the new one. Then, copy the entire Models folder from the original project to the new one. Adjust the namespaces to match the new project name (ProductsCore). At this point, you can build the application, and you will find a number of compilation errors. These should generally fall into the following categories:

ApiController does not exist
System.Web.Http namespace does not exist
IHttpActionResult does not exist
NotFound does not exist
Ok does not exist

Fortunately, these are all very easy to correct:

Change ApiController to Controller (you may need to add using Microsoft.AspNetCore.Mvc)
Delete any using statement referring to System.Web.Http
Change any method returning IHttpActionResult to return a IActionResult
Change NotFound to HttpNotFound
Change Ok(product) to new ObjectResult(product)

Once these changes have been made and unused using statements removed, the migrated ProductsController class looks like this:

using Microsoft.AspNet.Mvc;
using ProductsCore.Models;
using System.Collections.Generic;
using System.Linq;

namespace ProductsCore.Controllers
{
    [Route("api/[controller]")]
    public class ProductsController : Controller
    {
        Product[] products = new Product[] 
        { 
            new Product { Id = 1, Name = "Tomato Soup", Category = "Groceries", Price = 1 }, 
            new Product { Id = 2, Name = "Yo-yo", Category = "Toys", Price = 3.75M }, 
            new Product { Id = 3, Name = "Hammer", Category = "Hardware", Price = 16.99M } 
        };

        // /api/products
        [HttpGet]
        public IEnumerable<Product> GetAllProducts()
        {
            return products;
        }

        // /api/products/1
        [HttpGet("{id}")]
        public IActionResult GetProduct(int id)
        {
            var product = products.FirstOrDefault((p) => p.Id == id);
            if (product == null)
            {
                return HttpNotFound();
            }
            return new ObjectResult(product);
        }
    }
}

You should now be able to run the migrated project and browse to /api/products; and, you should see the full list of 3 products. Browse to /api/products/1 and you should see the first product.

Summary ¶

Migrating a simple ASP.NET Web API project to ASP.NET Core MVC is fairly straightforward, thanks to the built-in support for Web APIs in ASP.NET Core MVC. The main pieces every ASP.NET Web API project will need to migrate are routes, controllers, and models, along with updates to the types used by controllers and actions.

Migrating HTTP Modules to Middleware¶

By Matt Perdeck

This article shows how to migrate existing ASP.NET HTTP modules and handlers to ASP.NET Core middleware.

Sections:

Handlers and modules revisited
From handlers and modules to middleware
Migrating module code to middleware
Migrating module insertion into the request pipeline
Migrating handler code to middleware
Migrating handler insertion into the request pipeline
Loading middleware options using the options pattern
Loading middleware options through direct injection
Migrating to the new HttpContext
Additional Resources

Handlers and modules revisited ¶

Before proceeding to ASP.NET Core middleware, let’s first recap how HTTP modules and handlers work:

Handlers are:

Classes that implement IHttpHandler
Used to handle requests with a given file name or extension, such as .report
Configured in Web.config

Modules are:

Classes that implement IHttpModule
Invoked for every request
Able to short-circuit (stop further processing of a request)
Able to add to the HTTP response, or create their own
Configured in Web.config

The order in which modules process incoming requests is determined by:

The application life cycle, which is a series events fired by ASP.NET: BeginRequest, AuthenticateRequest, etc. Each module can create a handler for one or more events.

For the same event, the order in which they are configured in Web.config.

In addition to modules, you can add handlers for the life cycle events to your Global.asax.cs file. These handlers run after the handlers in the configured modules.

From handlers and modules to middleware ¶

Middleware are simpler than HTTP modules and handlers:

Modules, handlers, Global.asax.cs, Web.config (except for IIS configuration) and the application life cycle are gone
The roles of both modules and handlers have been taken over by middleware
Middleware are configured using code rather than in Web.config
Pipeline branching lets you send requests to specific middleware, based on not only the URL but also on request headers, query strings, etc.

Middleware are very similar to modules:

Invoked in principle for every request
Able to short-circuit a request, by not passing the request to the next middleware
Able to create their own HTTP response

Middleware and modules are processed in a different order:

Order of middleware is based on the order in which they are inserted into the request pipeline, while order of modules is mainly based on application life cycle events
Order of middleware for responses is the reverse from that for requests, while order of modules is the same for requests and responses
See Creating a middleware pipeline with IApplicationBuilder

Note how in the image above, the authentication middleware short-circuited the request.

Migrating module code to middleware ¶

An existing HTTP module will look similar to this:

// ASP.NET 4 module

using System;
using System.Web;

namespace MyApp.Modules
{
    public class MyModule : IHttpModule
    {
        public void Dispose()
        {
        }

        public void Init(HttpApplication application)
        {
            application.BeginRequest += (new EventHandler(this.Application_BeginRequest));
            application.EndRequest += (new EventHandler(this.Application_EndRequest));
        }

        private void Application_BeginRequest(Object source, EventArgs e)
        {
            HttpContext context = ((HttpApplication)source).Context;

            // Do something with context near the beginning of request processing.
        }

        private void Application_EndRequest(Object source, EventArgs e)
        {
            HttpContext context = ((HttpApplication)source).Context;

            // Do something with context near the end of request processing.
        }
    }
}

As shown in the Middleware page, an ASP.NET Core middleware is simply a class that exposes an Invoke method taking an HttpContext and returning a Task. Your new middleware will look like this:

// ASP.NET 5 middleware

using Microsoft.AspNet.Builder;
using Microsoft.AspNet.Http;
using System.Threading.Tasks;

namespace MyApp.Middleware
{
    public class MyMiddleware
    {
        private readonly RequestDelegate _next;

        public MyMiddleware(RequestDelegate next)
        {
            _next = next;
        }

        public async Task Invoke(HttpContext context)
        {
            // Do something with context near the beginning of request processing.

            await _next.Invoke(context);

            // Clean up.
        }
    }

    public static class MyMiddlewareExtensions
    {
        public static IApplicationBuilder UseMyMiddleware(this IApplicationBuilder builder)
        {
            return builder.UseMiddleware<MyMiddleware>();
        }
    }
}

The above middleware template was taken from the section on writing middleware.

The MyMiddlewareExtensions helper class makes it easier to configure your middleware in your Startup class. The UseMyMiddleware method adds your middleware class to the request pipeline. Services required by the middleware get injected in the middleware’s constructor.

Your module might terminate a request, for example if the user is not authorized:

// ASP.NET 4 module that may terminate the request

private void Application_BeginRequest(Object source, EventArgs e)
{
    HttpContext context = ((HttpApplication)source).Context;

    // Do something with context near the beginning of request processing.

    if (TerminateRequest())
    {
        context.Response.End();
        return;
    }
}

A middleware handles this by simply not calling Invoke on the next middleware in the pipeline. Keep in mind that this does not fully terminate the request, because previous middlewares will still be invoked when the response makes its way back through the pipeline.

// ASP.NET 5 middleware that may terminate the request

public async Task Invoke(HttpContext context)
{
    // Do something with context near the beginning of request processing.

    if (!TerminateRequest())
        await _next.Invoke(context);

    // Clean up.
}

When you migrate your module’s functionality to your new middleware, you may find that your code doesn’t compile because the HttpContext class has significantly changed in ASP.NET Core. Later on, you’ll see how to migrate to the new ASP.NET Core HttpContext.

Migrating module insertion into the request pipeline ¶

HTTP modules are typically added to the request pipeline using Web.config:

<?xml version="1.0" encoding="utf-8"?>
<!--ASP.NET 4 web.config-->
<configuration>
  <system.webServer>
    <modules>
      <add name="MyModule" type="MyApp.Modules.MyModule"/>
    </modules>
  </system.webServer>
</configuration>

Convert this by adding your new middleware to the request pipeline in your Startup class:

// ASP.NET 5 Startup class

namespace Asp.Net5
{
    public class Startup
    {
        public void Configure(IApplicationBuilder app, IHostingEnvironment env, 
            ILoggerFactory loggerFactory)
        {
            // ...

            app.UseMyMiddleware();

            // ...
        }
    }
}

The exact spot in the pipeline where you insert your new middleware depends on the event that it handled as a module (BeginRequest, EndRequest, etc.) and its order in your list of modules in Web.config.

As previously stated, there is no more application life cycle in ASP.NET Core and the order in which responses are processed by middleware differs from the order used by modules. This could make your ordering decision more challenging.

If ordering becomes a problem, you could split your module into multiple middleware that can be ordered independently.

Migrating handler code to middleware ¶

An HTTP handler looks something like this:

// ASP.NET 4 handler

using System.Web;

namespace MyApp.HttpHandlers
{
    public class MyHandler : IHttpHandler
    {
        public bool IsReusable { get { return true; } }

        public void ProcessRequest(HttpContext context)
        {
            string response = GenerateResponse(context);

            context.Response.ContentType = GetContentType();
            context.Response.Output.Write(response);
        }

        // ...
    }
}

In your ASP.NET Core project, you would translate this to a middleware similar to this:

// ASP.NET 5 middleware migrated from a handler

using Microsoft.AspNet.Builder;
using Microsoft.AspNet.Http;
using System.Threading.Tasks;

namespace MyApp.Middleware
{
    public class MyHandlerMiddleware
    {

        // Must have constructor with this signature, otherwise exception at run time
        public MyHandlerMiddleware(RequestDelegate next)
        {
            // This is an HTTP Handler, so no need to store next
        }

        public async Task Invoke(HttpContext context)
        {
            string response = GenerateResponse(context);

            context.Response.ContentType = GetContentType();
            await context.Response.WriteAsync(response);
        }

        // ...
    }

    public static class MyHandlerExtensions
    {
        public static IApplicationBuilder UseMyHandler(this IApplicationBuilder builder)
        {
            return builder.UseMiddleware<MyHandlerMiddleware>();
        }
    }
}

This middleware is very similar to the middleware corresponding to modules. The only real difference is that here there is no call to _next.Invoke(context). That makes sense, because the handler is at the end of the request pipeline, so there will be no next middleware to invoke.

Migrating handler insertion into the request pipeline ¶

Configuring an HTTP handler is done in Web.config and looks something like this:

<?xml version="1.0" encoding="utf-8"?>
<!--ASP.NET 4 web.config-->
<configuration>
  <system.webServer>
    <handlers>
      <add name="MyHandler" verb="*" path="*.report" type="MyApp.HttpHandlers.MyHandler" resourceType="Unspecified" preCondition="integratedMode"/>
    </handlers>
  </system.webServer>
</configuration>

You could convert this by adding your new handler middleware to the request pipeline in your Startup class, similar to middleware converted from modules. The problem with that approach is that it would send all requests to your new handler middleware. However, you only want requests with a given extension to reach your middleware. That would give you the same functionality you had with your HTTP handler.

One solution is to branch the pipeline for requests with a given extension, using the MapWhen extension method. You do this in the same Configure method where you add the other middleware:

// ASP.NET 5 Startup class

namespace Asp.Net5
{
    public class Startup
    {
        public void Configure(IApplicationBuilder app, IHostingEnvironment env, 
            ILoggerFactory loggerFactory)
        {
            // ...

            app.MapWhen(
                context => context.Request.Path.ToString().EndsWith(".report"),
                appBranch => {
                    // ... optionally add more middleware to this branch
                    appBranch.UseMyHandler();
                });
        }
    }
}

MapWhen takes these parameters:

A lambda that takes the HttpContext and returns true if the request should go down the branch. This means you can branch requests not just based on their extension, but also on request headers, query string parameters, etc.
A lambda that takes an IApplicationBuilder and adds all the middleware for the branch. This means you can add additional middleware to the branch in front of your handler middleware.

Middleware added to the pipeline before the branch will be invoked on all requests; the branch will have no impact on them.

Loading middleware options using the options pattern ¶

Some modules and handlers have configuration options that are stored in Web.config. However, in ASP.NET Core a new configuration model is used in place of Web.config.

The new configuration system gives you these options to solve this:

Directly inject the options into the middleware, as shown in the next section.
Use the options pattern:

Create a class to hold your middleware options, for example:

public class MyMiddlewareOptions
{
    public string Param1 { get; set; }
    public string Param2 { get; set; }
}

Store the option values
The new configuration system allows you to essentially store option values anywhere you want. However, most sites use appsettings.json, so we’ll take that approach:
1 2 3 4 5 6
{ "MyMiddlewareOptionsSection": { "Param1": "Param1Value", "Param2": "Param2Value" } }
MyMiddlewareOptionsSection here is simply a section name. It doesn’t have to be the same as the name of your options class.

Associate the option values with the options class

The options pattern uses ASP.NET Core’s dependency injection framework to associate the options type (such as MyMiddlewareOptions) with an MyMiddlewareOptions object that has the actual options.

Update your Startup class:

If you’re using appsettings.json, add it to the configuration builder in the Startup constructor:

public class Startup
{
    public Startup(IHostingEnvironment env)
    {
        // Set up configuration sources.
        var builder = new ConfigurationBuilder()
            .AddJsonFile("appsettings.json")
            .AddEnvironmentVariables();
        Configuration = builder.Build();
    }

}

Configure the options service:

public class Startup
{
    public void ConfigureServices(IServiceCollection services)
    {
        services.AddOptions();

        // ...
    }
}

Associate your options with your options class:

public class Startup
{
    public void ConfigureServices(IServiceCollection services)
    {
        services.AddOptions();

        services.Configure<MyMiddlewareOptions>(
            Configuration.GetSection("MyMiddlewareOptionsSection"));

        // ...
    }
}

Inject the options into your middleware constructor. This is similar to injecting options into a controller.

namespace MyApp.Middleware
{

    public class MyMiddlewareWithParams
    {
        private readonly RequestDelegate _next;
        private readonly MyMiddlewareOptions _myMiddlewareOptions;

        public MyMiddlewareWithParams(RequestDelegate next,
            IOptions<MyMiddlewareOptions> optionsAccessor)
        {
            _next = next;
            _myMiddlewareOptions = optionsAccessor.Value;
        }

        public async Task Invoke(HttpContext context)
        {
            // Do something with context near the beginning of request processing
            // using configuration in _myMiddlewareOptions

            await _next.Invoke(context);

            // Do something with context near the end of request processing
            // using configuration in _myMiddlewareOptions
        }
    }

The UseMiddleware extension method that adds your middleware to the IApplicationBuilder takes care of dependency injection.

This is not limited to IOptions objects. Any other object that your middleware requires can be injected this way.

Loading middleware options through direct injection ¶

The options pattern has the advantage that it creates loose coupling between options values and their consumers. Once you’ve associated an options class with the actual options values, any other class can get access to the options through the dependency injection framework. There is no need to pass around options values.

This breaks down though if you want to use the same middleware twice, with different options. For example an authorization middleware used in different branches allowing different roles. You can’t associate two different options objects with the one options class.

The solution is to get the options objects with the actual options values in your Startup class and pass those directly to each instance of your middleware.

Add a second key to appsettings.json

To add a second set of options to the appsettings.json file, simply use a new key to uniquely identify it:

{
  "MyMiddlewareOptionsSection2": {
    "Param1": "Param1Value2",
    "Param2": "Param2Value2"
  },
  "MyMiddlewareOptionsSection": {
    "Param1": "Param1Value",
    "Param2": "Param2Value"
  }
}

Retrieve options values. The Get method on the Configuration property lets you retrieve options values:

// ASP.NET 5 Startup class

namespace Asp.Net5
{
    public class Startup
    {
        public void Configure(IApplicationBuilder app, IHostingEnvironment env, 
            ILoggerFactory loggerFactory)
        {
            // ...

            var myMiddlewareOptions = 
                Configuration.Get<MyMiddlewareOptions>("MyMiddlewareOptionsSection");

            var myMiddlewareOptions2 = 
                Configuration.Get<MyMiddlewareOptions>("MyMiddlewareOptionsSection2");

            // ...

        }
    }
}

Pass options values to middleware. The Use... extension method (which adds your middleware to the pipeline) is a logical place to pass in the option values:

// ASP.NET 5 Startup class

namespace Asp.Net5
{
    public class Startup
    {
        public void Configure(IApplicationBuilder app, IHostingEnvironment env, 
            ILoggerFactory loggerFactory)
        {
            // ...

            var myMiddlewareOptions = 
                Configuration.Get<MyMiddlewareOptions>("MyMiddlewareOptionsSection");

            var myMiddlewareOptions2 = 
                Configuration.Get<MyMiddlewareOptions>("MyMiddlewareOptionsSection2");

            app.UseMyMiddlewareWithParams(myMiddlewareOptions);

            // ...

            app.UseMyMiddlewareWithParams(myMiddlewareOptions2);
        }
    }
}

Enable middleware to take an options parameter. Provide an overload of the Use... extension method (that takes the options parameter and passes it to UseMiddleware). When UseMiddleware is called with parameters, it passes the parameters to your middleware constructor when it instantiates the middleware object.

using Microsoft.AspNet.Builder;
using Microsoft.AspNet.Http;
using Microsoft.Extensions.OptionsModel;
using System.Threading.Tasks;

namespace MyApp.Middleware
{

    public static class MyMiddlewareWithParamsExtensions
    {
        public static IApplicationBuilder UseMyMiddlewareWithParams(
            this IApplicationBuilder builder)
        {
            return builder.UseMiddleware<MyMiddlewareWithParams>();
        }

        public static IApplicationBuilder UseMyMiddlewareWithParams(
            this IApplicationBuilder builder, MyMiddlewareOptions myMiddlewareOptions)
        {
            return builder.UseMiddleware<MyMiddlewareWithParams>(
                new OptionsWrapper<MyMiddlewareOptions>(myMiddlewareOptions));
        }
    }
}

Note how this wraps the options object in an OptionsWrapper object. This implements IOptions, as expected by the middleware constructor:

    // Remove this when Microsoft.Extensions.Options becomes available from NuGet
    public class OptionsWrapper<TOptions> : IOptions<TOptions> where TOptions : class, new()
    {
        public OptionsWrapper(TOptions options)
        {
            Value = options;
        }

        public TOptions Value { get; }
    }

Migrating to the new HttpContext ¶

You saw earlier that the Invoke method in your middleware takes a parameter of type HttpContext:

public async Task Invoke(HttpContext context)

HttpContext has significantly changed in ASP.NET Core. This section shows how to translate the most commonly used properties of System.Web.HttpContext to the new Microsoft.AspNetCore.Http.HttpContext.

HttpContext¶

HttpContext.Items translates to:

IDictionary<object, object> items = httpContext.Items;

Unique request ID (no System.Web.HttpContext counterpart)

Gives you a unique id for each request. Very useful to include in your logs.

string requestId = httpContext.TraceIdentifier;

HttpContext.Request¶

HttpContext.Request.HttpMethod translates to:

string httpMethod = httpContext.Request.Method;

HttpContext.Request.QueryString translates to:

IReadableStringCollection queryParameters = httpContext.Request.Query;

// If no query parameter "key" used, values will have 0 items
// If single value used for a key (...?key=v1), values will have 1 item ("v1")
// If key has multiple values (...?key=v1&key=v2), values will have 2 items ("v1" and "v2")
IList<string> values = queryParameters["key"];

// If no query parameter "key" used, value will be ""
// If single value used for a key (...?key=v1), value will be "v1"
// If key has multiple values (...?key=v1&key=v2), value will be "v1,v2"
string value = queryParameters["key"].ToString();

HttpContext.Request.Url and HttpContext.Request.RawUrl translate to:

// using Microsoft.AspNet.Http.Extensions;
var url = httpContext.Request.GetDisplayUrl();

HttpContext.Request.IsSecureConnection translates to:

var isSecureConnection = httpContext.Request.IsHttps;

HttpContext.Request.UserHostAddress translates to:

var userHostAddress = httpContext.Connection.RemoteIpAddress?.ToString();

HttpContext.Request.Cookies translates to:

IReadableStringCollection cookies = httpContext.Request.Cookies;
string unknownCookieValue = cookies["unknownCookie"]; // will be null (no exception)
string knownCookieValue = cookies["cookie1name"];     // will be actual value

HttpContext.Request.Headers translates to:

// using Microsoft.AspNet.Http.Headers;
// using Microsoft.Net.Http.Headers;

IHeaderDictionary headersDictionary = httpContext.Request.Headers;

// GetTypedHeaders extension method provides strongly typed access to many headers
var requestHeaders = httpContext.Request.GetTypedHeaders();
CacheControlHeaderValue cacheControlHeaderValue = requestHeaders.CacheControl;

// For unknown header, unknownheaderValues has zero items and unknownheaderValue is ""
IList<string> unknownheaderValues = headersDictionary["unknownheader"];
string unknownheaderValue = headersDictionary["unknownheader"].ToString();

// For known header, knownheaderValues has 1 item and knownheaderValue is the value
IList<string> knownheaderValues = headersDictionary[HeaderNames.AcceptLanguage];
string knownheaderValue = headersDictionary[HeaderNames.AcceptLanguage].ToString();

HttpContext.Request.UserAgent translates to:

string userAgent = headersDictionary[HeaderNames.UserAgent].ToString();

HttpContext.Request.UrlReferrer translates to:

string urlReferrer = headersDictionary[HeaderNames.Referer].ToString();

HttpContext.Request.ContentType translates to:

// using Microsoft.Net.Http.Headers;

MediaTypeHeaderValue mediaHeaderValue = requestHeaders.ContentType;
string contentType = mediaHeaderValue?.MediaType;   // ex. application/x-www-form-urlencoded
string contentMainType = mediaHeaderValue?.Type;    // ex. application
string contentSubType = mediaHeaderValue?.SubType;  // ex. x-www-form-urlencoded

System.Text.Encoding requestEncoding = mediaHeaderValue?.Encoding;

HttpContext.Request.Form translates to:

if (httpContext.Request.HasFormContentType)
{
    IFormCollection form;

    form = httpContext.Request.Form; // sync
    // Or
    form = await httpContext.Request.ReadFormAsync(); // async

    string firstName = form["firstname"];
    string lastName = form["lastname"];
}

Caution

Read form values only if the content sub type is x-www-form-urlencoded or form-data.

HttpContext.Request.InputStream translates to:

string inputBody;
using (var reader = new System.IO.StreamReader(
    httpContext.Request.Body, System.Text.Encoding.UTF8))
{
    inputBody = reader.ReadToEnd();
}

Caution

Use this code only in a handler type middleware, at the end of a pipeline.

You can read the raw body as shown above only once per request. Middleware trying to read the body after the first read will read an empty body.

This does not apply to reading a form as shown earlier, because that is done from a buffer.

HttpContext.Request.RequestContext.RouteData

RouteData is not available in middleware in RC1.

HttpContext.Response¶

HttpContext.Response.Status and HttpContext.Response.StatusDescription translate to:

// using Microsoft.AspNet.Http;
httpContext.Response.StatusCode = StatusCodes.Status200OK;

HttpContext.Response.ContentEncoding and HttpContext.Response.ContentType translate to:

// using Microsoft.Net.Http.Headers;
var mediaType = new MediaTypeHeaderValue("application/json");
mediaType.Encoding = System.Text.Encoding.UTF8;
httpContext.Response.ContentType = mediaType.ToString();

HttpContext.Response.ContentType on its own also translates to:

httpContext.Response.ContentType = "text/html";

HttpContext.Response.Output translates to:

string responseContent = GetResponseContent();
await httpContext.Response.WriteAsync(responseContent);

HttpContext.Response.TransmitFile

Serving up a file is discussed here.

HttpContext.Response.Headers

Sending response headers is complicated by the fact that if you set them after anything has been written to the response body, they will not be sent.

The solution is to set a callback method that will be called right before writing to the response starts. This is best done at the start of the Invoke method in your middleware. It is this callback method that sets your response headers.

The following code sets a callback method called SetHeaders:

public async Task Invoke(HttpContext httpContext)
{
    // ...
    httpContext.Response.OnStarting(SetHeaders, state: httpContext);

The SetHeaders callback method would look like this:

// using Microsoft.AspNet.Http.Headers;
// using Microsoft.Net.Http.Headers;

private Task SetHeaders(object context)
{
    var httpContext = (HttpContext)context;

    // Set header with single value
    httpContext.Response.Headers["ResponseHeaderName"] = "headerValue";

    // Set header with multiple values
    string[] responseHeaderValues = new string[] { "headerValue1", "headerValue1" };
    httpContext.Response.Headers["ResponseHeaderName"] = responseHeaderValues;

    // Translating ASP.NET 4's HttpContext.Response.RedirectLocation  
    httpContext.Response.Headers[HeaderNames.Location] = "http://www.example.com";
    // Or
    httpContext.Response.Redirect("http://www.example.com");

    // GetTypedHeaders extension method provides strongly typed access to many headers
    var responseHeaders = httpContext.Response.GetTypedHeaders();

    // Translating ASP.NET 4's HttpContext.Response.CacheControl 
    responseHeaders.CacheControl = new CacheControlHeaderValue
    {
        MaxAge = new System.TimeSpan(365, 0, 0, 0)
        // Many more properties available 
    };

    // If you use .Net 4.6+, Task.CompletedTask will be a bit faster
    return Task.FromResult(0);
}

HttpContext.Response.Cookies

Cookies travel to the browser in a Set-Cookie response header. As a result, sending cookies requires the same callback as used for sending response headers:

public async Task Invoke(HttpContext httpContext)
{
    // ...
    httpContext.Response.OnStarting(SetCookies, state: httpContext);
    httpContext.Response.OnStarting(SetHeaders, state: httpContext);

The SetCookies callback method would look like the following:

private Task SetCookies(object context)
{
    var httpContext = (HttpContext)context;

    IResponseCookies responseCookies = httpContext.Response.Cookies;

    responseCookies.Append("cookie1name", "cookie1value");
    responseCookies.Append("cookie2name", "cookie2value",
        new CookieOptions { Expires = System.DateTime.Now.AddDays(5), HttpOnly = true });

    // If you use .Net 4.6+, Task.CompletedTask will be a bit faster
    return Task.FromResult(0); 
}

Additional Resources ¶

Migrating from ASP.NET 5 RC1 to ASP.NET Core 1.0¶

By Cesar Blum Silveira, Rachel Appel, Rick Anderson

Sections:

Update Target Framework Monikers (TFMs)
Namespace and package ID changes
Commands and tools
Hosting
Kestrel
ASP.NET 5 MVC compile views
Configuration
Logging
Identity
Working with IIS
Updating Launch Settings in Visual Studio
Server garbage collection (GC)

ASP.NET 5 RC1 apps were based on the .NET Execution Environment (DNX) and made use of DNX specific features. ASP.NET Core 1.0 is based on .NET Core, so you must first migrate your application to the new .NET Core project model. See migrating from DNX to .NET Core CLI for more information.

See the following resources for a list of some of the most significant changes, announcements and migrations information:

Update Target Framework Monikers (TFMs)¶

If your app targeted dnx451 or dnxcore50 in the frameworks section of project.json, you must make the following changes:

DNX	.NET Core
`dnx451`	`net451`
`dnxcore50`	`netcoreapp1.0`

.NET Core apps must add a dependency to the Microsoft.NETCore.App package:

"dependencies": {
  "Microsoft.NETCore.App": {
    "version": "1.0.0",
    "type": "platform"
  },

Namespace and package ID changes ¶

ASP.NET 5 has been renamed to ASP.NET Core 1.0
ASP.NET MVC and Identity are now part of ASP.NET Core
ASP.NET MVC 6 is now ASP.NET Core MVC
ASP.NET Identity 3 is now ASP.NET Core Identity
ASP.NET Core 1.0 package versions are 1.0.0
ASP.NET Core 1.0 tool package versions are 1.0.0-preview2-final

Namespace and package name changes:

ASP.NET 5 RC1	ASP.NET Core 1.0
`Microsoft.AspNet.*`	`Microsoft.AspNetCore.*`
`EntityFramework.*`	`Microsoft.EntityFrameworkCore.*`
`Microsoft.Data.Entity.*`	`Microsoft.EntityFrameworkCore.*`

The EntityFramework.Commands package is no longer available. The ef command is now available as a tool in the Microsoft.EntityFrameworkCore.Tools package.

The following packages have been renamed:

ASP.NET 5 RC1	ASP.NET Core 1.0
EntityFramework.MicrosoftSqlServer	Microsoft.EntityFrameworkCore.SqlServer
Microsoft.AspNet.Diagnostics.Entity	Microsoft.AspNetCore.Dianostics.EntityFrameworkCore
Microsoft.AspNet.Identity.EntityFramework	Microsoft.AspNetCore.Identity.EntityFrameworkCore
Microsoft.AspNet.Tooling.Razor	Microsoft.AspNetCore.Razor.Tools

Commands and tools ¶

The commands section of the project.json file is no longer supported. Use dotnet run or dotnet <DLL name> instead.

.NET Core CLI has introduced the concept of tools. project.json now supports a tools section where packages containing tools can be specified. Some important functionality for ASP.NET Core 1.0 applications has been moved to tools.

See .NET Core CLI extensibility model for more information on .NET Core CLI tools.

Publishing to IIS¶

IIS publishing is now provided by the publish-iis tool in the Microsoft.AspNetCore.Server.IISIntegration.Tools package. If you intend to run your app behind IIS, add the publish-iis tool to your project.json:

{
  "tools": {
    "Microsoft.AspNetCore.Server.IISIntegration.Tools": "1.0.0-preview2-final"
  }
}

The publish-iis tool is commonly used in the postpublish script in project.json:

{
  "postpublish": [ "dotnet publish-iis --publish-folder %publish:OutputPath% --framework %publish:FullTargetFramework%" ]
}

Entity Framework commands¶

The ef tool is now provided in the Microsoft.EntityFrameworkCore.Tools package:

{
  "tools": {
    "Microsoft.EntityFrameworkCore.Tools": "1.0.0-preview2-final"
  }
}

For more information, see .NET Core CLI.

Razor tools¶

Razor tooling is now provided in the Microsoft.AspNetCore.Razor.Tools package:

{
  "tools": {
    "Microsoft.AspNetCore.Razor.Tools": "1.0.0-preview2-final"
  }
}

SQL cache tool¶

The sqlservercache command, formerly provided by the Microsoft.Extensions.Caching.SqlConfig package, has been replaced by the sql-cache tool, available through the Microsoft.Extensions.Caching.SqlConfig.Tools package:

{
  "tools": {
    "Microsoft.Extensions.Caching.SqlConfig.Tools": "1.0.0-preview2-final"
  }
}

User secrets manager¶

The user-secret command, formerly provided by the Microsoft.Extensions.SecretManager package, has been replaced by the user-secrets tool, available through the Microsoft.Extensions.SecretManager.Tools package:

{
  "tools": {
    "Microsoft.Extensions.SecretManager.Tools": "1.0.0-preview2-final"
  }
}

File watcher¶

The watch command, formerly provided by the Microsoft.Dnx.Watcher package, has been replaced by the watch tool, available through the Microsoft.DotNet.Watcher.Tools package:

{
  "tools": {
    "Microsoft.DotNet.Watcher.Tools": "1.0.0-preview2-final"
  }
}

For more information on the file watcher, see Dotnet watch in Tutorials.

Hosting ¶

Creating the web application host¶

ASP.NET Core 1.0 apps are console apps; you must define an entry point for your app that sets up a web host and runs it. Below is an example from the startup code for one of the Web Application templates in Visual Studio:

public class Program
{
    public static void Main(string[] args)
    {
        var host = new WebHostBuilder()
            .UseKestrel()
            .UseContentRoot(Directory.GetCurrentDirectory())
            .UseIISIntegration()
            .UseStartup<Startup>()
            .Build();

        host.Run();
    }
}

You must add the emitEntryPoint to the buildOptions section of your application’s project.json:

{
  "buildOptions": {
    "emitEntryPoint": true
  }
}

Class and interface renames¶

All classes and interfaces prefixed with WebApplication have been renamed to start with WebHost:

ASP.NET 5 RC1	ASP.NET Core 1.0
IWebApplicationBuilder	IWebHostBuilder
WebApplicationBuilder	WebHostBuilder
IWebApplication	IWebHost
WebApplication	WebHost
WebApplicationOptions	WebHostOptions
WebApplicationDefaults	WebHostDefaults
WebApplicationService	WebHostService
WebApplicationConfiguration	WebHostConfiguration

Content root and web root¶

The application base path is now called the content root.

The web root of your application is no longer specified in your project.json file. It is defined when setting up the web host and defaults to wwwroot. Call the UseWebRoot extension method to specify a different web root folder. Alternatively, you can specify the web root folder in configuration and call the UseConfiguration extension method.

Server address binding¶

The server addresses that your application listens on can be specified using the UseUrls extension method or through configuration.

Specifying only a port number as a binding address is no longer supported. The default binding address is http://localhost:5000

Hosting configuration¶

The UseDefaultHostingConfiguration method is no longer available. The only configuration values read by default by WebHostBuilder are those specified in environment variables prefixed with ASPNETCORE_*. All other configuration sources must now be added explicitly to an IConfigurationBuilder instance. See Configuration for more information.

The environment key is set with the ASPNETCORE_ENVIRONMENT environment variable. ASPNET_ENV and Hosting:Environment are still supported, but generate a deprecated message warning.

Hosting service changes¶

Dependency injection code that uses IApplicationEnvironment must now use IHostingEnvironment. For example, in your Startup class, change:

public Startup(IApplicationEnvironment applicationEnvironment)

To:

public Startup(IHostingEnvironment hostingEnvironment)

Kestrel ¶

Kestrel configuration has changed. This GitHub announcement outlines the changes you must make to configure Kestrel if you are not using default settings.

Controller and action results renamed¶

The following Controller methods have been renamed and moved to ControllerBase:

ASP.NET 5 RC1	ASP.NET Core 1.0
HttpUnauthorized	Unauthorized
HttpNotFound (and its overloads)	NotFound
HttpBadRequest (and its overloads)	BadRequest

The following action result types have also been renamed:

ASP.NET 5 RC1	ASP.NET Core 1.0
Microsoft.AspNet.Mvc.HttpOkObjectResult	Microsoft.AspNetCore.Mvc.OkObjectResult
Microsoft.AspNet.Mvc.HttpOkResult	Microsoft.AspNetCore.Mvc.OkResult
Microsoft.AspNet.Mvc.HttpNotFoundObjectResult	Microsoft.AspNetCore.Mvc.NotFoundObjectResult
Microsoft.AspNet.Mvc.HttpNotFoundResult	Microsoft.AspNetCore.Mvc.NotFoundResult
Microsoft.AspNet.Mvc.HttpStatusCodeResult	Microsoft.AspNetCore.Mvc.StatusCodeResult
Microsoft.AspNet.Mvc.HttpUnauthorizedResult	Microsoft.AspNetCore.Mvc.UnauthorizedResult

ASP.NET 5 MVC compile views ¶

To compile views, set the preserveCompilationContext option in project.json to preserve the compilation context, as shown here:

{
  "buildOptions": {
    "preserveCompilationContext": true
  }
}

Changes in views¶

Views now support relative paths.

The Validation Summary Tag Helper asp-validation-summary attribute value has changed. Change:

<div asp-validation-summary="ValidationSummary.All"></div>

To:

<div asp-validation-summary="All"></div>

Changes in ViewComponents¶

The sync APIs have been removed
Component.Render(), Component.RenderAsync(), and Component.Invoke() have been removed
To reduce ambiguity in View Component method selection, we’ve modified the selection to only allow exactly one Invoke() or InvokeAsync() per View Component
InvokeAsync() now takes an anonymous object instead of separate parameters
To use a view component, call @Component.InvokeAsync("Name of view component", <parameters>) from a view. The parameters will be passed to the InvokeAsync() method. The following example demonstrates the InvokeAsync() method call with two parameters:

ASP.NET 5 RC1:

@Component.InvokeAsync("Test", "MyName", 15)

ASP.NET Core 1.0:

@Component.InvokeAsync("Test", new { name = "MyName", age = 15 })
@Component.InvokeAsync("Test", new Dictionary<string, object> {
                       ["name"] = "MyName", ["age"] = 15 })
@Component.InvokeAsync<TestViewComponent>(new { name = "MyName", age = 15})

Updated controller discovery rules¶

There are changes that simplify controller discovery:

The new ControllerAttribute can be used to mark a class (and it’s subclasses) as a controller. A class whose name doesn’t end in Controller and derives from a base class that ends in Controller is no longer considered a controller. In this scenario, ControllerAttribute must be applied to the derived class itself or to the base class.

A type is considered a controller if all the following conditions are met:

The type is a public, concrete, non-open generic class
NonControllerAttribute is not applied to any type in its hierarchy
The type name ends with Controller, or ControllerAttribute is applied to the type or one of its ancestors.

Note

If NonControllerAttribute is applied anywhere in the type hierarchy, the discovery conventions will never consider that type or its descendants to be a controller. In other words, NonControllerAttribute takes precedence over ControllerAttribute.

Configuration ¶

The IConfigurationSource interface has been introduced to represent the configuration used to build an IConfigurationProvider. It is no longer possible to access the provider instances from IConfigurationBuilder, only the sources. This is intentional, and may cause loss of functionality as you can no longer do things like call Load on the provider instances.

File-based configuration providers support both relative and absolute paths to configuration files. If you want to specify file paths relative to your application’s content root, you must call the SetBasePath extension method on IConfigurationBuilder:

public Startup(IHostingEnvironment env)
{
    var builder = new ConfigurationBuilder()
        .SetBasePath(env.ContentRootPath)
        .AddJsonFile("appsettings.json");
}

Automatic reload on change¶

The IConfigurationRoot.ReloadOnChanged extension method is no longer available. File-based configuration providers now provide extension methods to IConfigurationBuilder that allow you to specify whether configuration from those providers should be reloaded when there are changes in their files. See AddJsonFile, AddXmlFile and AddIniFile for details.

Logging ¶

LogLevel.Verbose has been renamed to Trace and is now considered less severe than Debug.

The MinimumLevel property has been removed from ILoggerFactory. Each logging provider now provides extension methods to ILoggerFactory that allow specifying a minimum logging level. See AddConsole, AddDebug, and AddEventLog for details.

Identity ¶

The signatures for the following methods or properties have changed:

ASP.NET 5 RC1	ASP.NET Core 1.0
ExternalLoginInfo.ExternalPrincipal	ExternalLoginInfo.Principal
User.IsSignedIn()	SignInManager.IsSignedIn(User)
UserManager.FindByIdAsync(HttpContext.User.GetUserId())	UserManager.GetUserAsync(HttpContext.User)
User.GetUserId()	UserManager.GetUserId(User)

To use Identity in a view, add the following:

@using Microsoft.AspNetCore.Identity
@inject SignInManager<TUser> SignInManager
@inject UserManager<TUser> UserManager

Working with IIS ¶

The package Microsoft.AspNetCore.IISPlatformHandler has been replaced by Microsoft.AspNetCore.Server.IISIntegration.

HttpPlatformHandler has been replaced by the ASP.NET Core Module (ANCM). The web.config file created by the Publish to IIS tool now configures IIS to the ANCM instead of HttpPlatformHandler to reverse-proxy requests.

The ASP.NET Core Module must be configured in web.config:

<configuration>
  <system.webServer>
    <handlers>
      <add name="aspNetCore" path="*" verb="*" modules="AspNetCoreModule" resourceType="Unspecified"/>
    </handlers>
    <aspNetCore processPath="%LAUNCHER_PATH%" arguments="%LAUNCHER_ARGS%"
                stdoutLogEnabled="false" stdoutLogFile=".\logs\stdout"
                forwardWindowsAuthToken="false"/>
  </system.webServer>
</configuration>

The Publish to IIS tool generates a correct web.config. See Publishing to IIS for more details.

IIS integration middleware is now configured when creating the Microsoft.AspNetCore.Hosting.WebHostBuilder, and is no longer called in the Configure method of the Startup class:

var host = new WebHostBuilder()
    .UseIISIntegration()
    .Build();

Web Deploy changes¶

Delete any <app name> - Web Deploy-publish.ps1 scripts created with Visual Studio web deploy using ASP.NET 5 RC1. The ASP.NET 5 RC1 scripts (which are DNX based) are not compatible with dotnet based scripts. Use Visual Studio to generate new web deploy scripts.

applicationhost.config changes¶

An applicationhost.config file created with ASP.NET 5 RC1 will point ASP.NET Core to an invalid content root location. With such a applicationhost.config file, ASP.NET Core will be configured with content root/web root as the content root folder and therefore look for web.config in Content root/wwwroot. The web.config file must be in the content root folder. When configured like this, the app will terminate with an HTTP 500 error.

Updating Launch Settings in Visual Studio ¶

Update launchSettings.json to remove the web target and add the following:

{
  "WebApplication1": {
    "commandName": "Project",
    "launchBrowser": true,
    "launchUrl": "http://localhost:5000",
    "environmentVariables": {
      "ASPNETCORE_ENVIRONMENT": "Development"
    }
  }
}

Server garbage collection (GC)¶

You must turn on server garbage collection in project.json or app.config when running ASP.NET projects on the full .NET Framework:

 {
   "runtimeOptions": {
     "configProperties": {
       "System.GC.Server": true
     }
   }
 }

Migrating from ASP.NET Core RC2 to ASP.NET Core 1.0¶

By Cesar Blum Silveira

Sections:

Overview
Tools
Hosting
ASP.NET MVC Core
Security

Overview ¶

This migration guide covers migrating an ASP.NET Core RC2 application to ASP.NET Core 1.0.

There weren’t many significant changes to ASP.NET Core between the RC2 and 1.0 releases. For a complete list of changes, see the ASP.NET Core 1.0 announcements.

Install the new tools from https://dot.net/core and follow the instructions.

Update the global.json to

{
  "projects": [ "src", "test" ],
  "sdk": {
      "version": "1.0.0-preview2-003121"
  }
}

Tools ¶

For the tools we ship, you no longer need to use imports in project.json. For example:

{
  "tools": {
    "Microsoft.AspNetCore.Server.IISIntegration.Tools": {
      "version": "1.0.0-preview1-final",
      "imports": "portable-net45+win8+dnxcore50"
    }
  }
}

Becomes:

{
  "tools": {
    "Microsoft.AspNetCore.Server.IISIntegration.Tools": "1.0.0-preview2-final"
  }
}

Hosting ¶

The UseServer is no longer available for IWebHostBuilder. You must now use UseKestrel or UseWebListener.

ASP.NET MVC Core ¶

The HtmlEncodedString class has been replaced by HtmlString (contained in the Microsoft.AspNetCore.Html.Abstractions package).

Security ¶

The AuthorizationHandler<TRequirement> class now only contains an asynchronous interface.

Contribute¶

ASP.NET Docs Style Guide¶