What's new in Visual Studio for ASP.NET and web developers
Blog|by Mary Branscombe|13 September 2017
At the Build conference, Microsoft showed off new tools for developers working with both Microsoft’s own web technologies like ASP.NET as well as HTML, CSS, JavaScript and TypeScript. Visual Studio for Mac is a good example of that, supporting ASP.NET Core for creating web backends, HTML5, CSS3 and JavaScript as well as the mobile development that Xamarin Studio has always let you do.
The 15.3 preview of Visual Studio 2017 includes the ASP.NET Core 2.0 preview; this now includes the SignalR library for building real-time web functionality with Web Sockets (and older techniques for older browsers) and the new Razor Pages app model. This is a lightweight syntax for combining C# server code with HTML for building web applications that don’t need the power or complexity of full MVC. Instead of having a model, and a view and a controller all in separate folders, Razor pages puts all three onto the same page. It’s still the MVC model, so tag handlers and model bindings and dependency injections all work and you can use Razor pages in existing MVC apps for the simpler parts of the UI. If you need to add more functionality to a page later on, you can turn it into a view and move the code into a controller and make it a full MVC.
While it’s possible that there will be VB templates for ASP.NET Core 2, don’t expect a VB version of Razor; that will stay as HTML and C#.
There are also a lot of improvements that .NET and ASP.NET program manager Scott Hanselman refers to as “removing paper cuts” from the developer experience; from templates that set up things you used to have to do manually when you create a new project the templates, to 25% faster throughout and faster startup because views are precompiled, but all in a runtime that’s only 3MB. That’s because ASP.NET Core and Entity Framework Core are now included in .NET Core, so your ASP.NET Core app doesn’t need to publish the whole of ASP.NET every time you deploy it. .NET Core 2 and ASP.NET Core 2 no longer reference all the individual packages that make up .NET (that proved to be complex and fragile), just a superset that’s the package of packages; only the packages you need are deployed and if you prefer you can de-reference that superset and pick the individual packages you want.
If you run your ASP.NET Core 2 code on Azure you can automatically get App Insights diagnostics and monitoring (based on tag helpers and the logging infrastructure of ASP.NET Core, with JavaScript injected for you that covers client-side dependencies, so you don’t have to write extra code to enable it) – but you can also get that locally, using the App Insights SDK.
If you don’t want to run your own identity system for a web app, you can now use any Open ID Connect identity provider with ASP.NET Core 2, including Azure AD B2C; that’s an Azure service that can manage accounts and sign ins for a web site using the same secure identity management as Azure AD, but for any consumer email address.
You can now set up continuous deployment pipelines for ASP.NET and ASP.NET Core projects on Azure Web Apps directly from Visual Studio, with the Continuous Delivery Tools extension; that will soon support Node.js, PHP and Python apps (as a preview). The preview release of Visual Studio 2017 also now lets you containerise ASP.NET apps, just by right-clicking on them and adding Docker support; that generates the Docker Compose file for an existing project, not just a new project. That automatically switches .NET applications that need IIS and ASP.NET to deploying into a Docker container, making it easier to extend existing applications with new microservices.
Containerising ASP.NET applications also gives you new options for hosting them on Azure Service Fabric. Alaska Airlines moved the Alaska.com site from IIS to a Service Fabric cluster, hosting it in Docker containers so they can scale up and down. Integrating that with Visual Studio for debugging has been complicated, because it’s meant going through Service Fabric reverse proxies as a gateway to forward requests to new services. Service Fabric plans to support isolated networks for communicating between multiple containers deployed as an app and fixed IP addresses for each application, which will simplify this.
ASP.NET Core 2.0 lets you write complex web apps with MVC and simpler pages with Razor Pages – which still use the MVC patterns (Image: Microsoft)
The new Service Fabric API Management (due to launch this summer) will also help, by supporting service discovery, routing and partition resolution, so you can make calls directly to your back-end services and even to a specific service instance. “The way this works is when a request comes in to API Management, API Management is now aware of Service Fabric’s naming service and so it does the resolution, it does the lookup and then it can intelligently route that call to the right place,” program manage Mark Fussell explained at Build. That uses a new HTTP client library that will be available as NuGet packages, so you’ll be able to do Service Fabric management without installing the entire Service Fabric runtime and SDK on your machine.
At the other end of the scale, if you want to run code online without thinking about how that’s hosted, Azure Functions, Microsoft’s serverless platform, lets you write that in C#, F# or JavaScript (through Node.js), plus limited support for PHP, Python and PowerShell, and in future, Java. If you don’t want to do that in the portal, you can do some of that locally in Visual Studio, but for some of it, you’ll have to use Visual Studio Code instead.
The Visual Studio 2015 Tools for Azure Functions let you write both C# and JavaScript but they won’t be getting updated; the Visual Studio 2017 Tools for Azure Functions extension needs the Visual Studio 2017 15.3 preview, but it only lets you create pre-compiled C# functions.
The advantage of that is that they start more quickly on the Functions service than functions written as scripts. Functions run when the trigger you create is detected (triggers can be anything from a new entry in a database or a record in Salesforce getting updated or an HTTP request), and they use bindings to connect to the data your code operates on (which can be on Azure or an external service), but all the code of your function has to be loaded before it runs.
With the new Visual Studio Functions tools, you can define bindings in your C# code using Azure App Service WebJobs attributes that get automatically converted to the function.json file that the bindings are stored in, instead of creating them there directly. You can include functions written in JavaScript in the same project, but you wont be able to debug them.
ASP.NET Core lets you call Open ID Connect identity services like Azure AD B2C instead of writing your own account management and making users create a new username and password.
If you want to write Functions code locally in JavaScript and be able to debug them as well, you need to do that in Visual Studio Code and use the Azure CLI to deploy; the Azure Tools extension gives you access to the CLI and the Azure Functions Visual Studio Code extension gives you JSON IntelliSense for the host, function and proxies JSON files you need for a function. That makes sense, because many JavaScript developers who use Visual Studio have adopted Visual Studio Code, and when Functions supports Java there will be tooling for the IDEs Javs developers use, like Eclipse and IntelliJ.
And if you’re using Visual Studio Team Services, you can use that to deploy your C# and JavaScript Functions code to Azure.
Contact Grey Matter
If you have any questions or want some extra information, complete the form below and one of the team will be in touch ASAP. If you have a specific use case, please let us know and we'll help you find the right solution faster.
By submitting this form you are agreeing to our Privacy Policy and Website Terms of Use.
