I need a both a Web API project and a function app project which share the same services / class libraries etc. A few questions
Is it bad practice to place them both in the same solution (or should they be in seperate repositories
Should the services be registered in the Startup class of the Web Api or the function app, or a mixture of the 2? Is it okay to register services only used by the function app in the web api startup class as this would make life easier? The services will do similar things and in some cases the same services will be used by both.
I am likely to only require a few timer trigger functions, most of the application will be build on the web api.
Specifically, the rest API will serve a frontend UI and the function app will provide updates to products on timer triggers.
Thanks for any guidance in advance and apologise if these are silly questions!
For #1 - I think it's totally fine to have them both in the same (git) repository - you might be following a monorepo strategy, OR these projects might be closely related e.g. part of the same business domain/capability where you might have something like a a core Web API project (ASP.NET Core WebApi) as well as some Function App event handlers (in a Function App project) that respond to external async events and invoke the core Web API. Alternatively, if they're completely unrelated repos and you're not following a monorepo strategy then maybe they should be in separate repos.
For #2 - some services might be exclusive to the ASP.NET Core WebApi, some exclusive to the Function App, and some might be shared. Assuming all the code is in the same repo; for the shared dependencies, you might consider having a third 'class library' project in the same git repo that can be referenced by both the ASP.NET Core WebApi project as well as the Function app project. This 'class library' project could contain all the shared dependencies. Dependencies should be registered in the startup of both the Web Api AND the function app.
I have a backend .NET 6 Web API RESTful service solution that exposes a complex class which is consumed by a Blazor Server App. Whenever the backend Web API's instantiated class changes values, I need to transmit that changed class to the Blazor app. Currently the Blazor app makes all of the calls to the backend service as would be expected.
I have been attempting to get an EventCallback working to overcome this challenge but it does not work. All of the examples I have found, such as https://blazor-university.com/components/component-events/, include their example within the same Blazor app.
Can somebody please point me in the right direction?
Thank you!
The EventCallback is a struct specifically created to be used in Blazor. It's purpose is to determine the type of the delegate of your event handler.
What you really need is a real time Notification System. It is best implemented through SignalR. Start working on it, and let us know of your difficulties.
Start here
Here's a complete working code sample of SignalR used with Blazor Server App with authentication.
A simpler solution, however, is to create a service that implements a Timer object, to query the database at given intervals.
I'm still struggling to understand the difference between ASP.NET Core Hosted and Server-side Blazor. I know same question already exists, but it's not satisfying. In fact, I couldn't find the satisfying answer anywhere - the answers were more or less the same.
If hosted option uses server (IIS, Kestrel), then why server-side? Confusing... It's a shame that official documentation didn't shed the light either...
UPDATE
The confusion stems from the fact that we have THREE options to create Blazor application. After executing dotnew new --list I get:
dotnet new blazorserver (Blazor Server App)
dotnet blazorwasm (Blazor WebAssembly App)
However, there's a third option:
dotnet blazorwasm --hosted (or dotnet blazor --hosted)
It's the same as check box in Visual Studio when creating application:
The documentation says:
you have the option of configuring the app to use an ASP.NET Core
backend by selecting the ASP.NET Core hosted check box
But no explanation was provided what does it mean...
Re this part of your question:
However, there's a third option:
dotnet blazorwasm --hosted (or dotnet blazor --hosted)
It's the same as check box in Visual Studio when creating application:
The
documentation
says:
you have the option of configuring the app to use an ASP.NET Core
backend by selecting the ASP.NET Core hosted check box
But no explanation was provided what does it mean...
TL;DR
'Hosted' is used where you want the back-end of your site and the Blazor client using that back-end to both be hosted on the same website.
In Detail
I agree, the documentation really isn't terribly clear about all of this, but the explanation is actually simpler than it seems:
The Blazor app has to be 'hosted' somewhere
The first thing to remember is that the Blazor WebAssembly 'app' is not a standalone website, it's an app that's embedded in a website. In a lot of cases it will behave like a website, because it will be used as a Single Page Application, but that is by no means required.
Essentially the Blazor WebAssembly app is a series of files and a JavaScript file that are created by compiling/publishing your Blazor project.
Those files then need to be put on a website somewhere and the combination of the name of a div tag and the Blazor JS file produced for your site deals with wiring your app files into the WebAssembly part of the browser so that it's then rendered on the page.
The key here is that the website 'hosting' your Blazor app does not have to be an ASP.NET Core site. It could be any site, pure HTML, Drupal, whatever, it just needs to be shown on a browser that handles WebAssembly and JavaScript correctly.
However, if you're also writing the backend of your site in ASP.NET Core, you can reuse that site
So, your Blazor project doesn't have to be hosted in a website written in ASP.NET Core, but it does have to be hosted somewhere (so the user can see it).
If you're also writing the back-end of the site at the same time, e.g. if you're writing an API or SignalR hub to send and receive data from your Blazor client, and if you're writing that back-end in ASP.NET Core, then you can reuse that same site to also host your Blazor client.
This scenario is what the 'Hosted' option is for.
If you create a project using the template in the screenshot above, with the 'hosted' option ticked, you'll see that the [YourProjectName].Server project that's created is the Start Up project, but the index.html page that's shown when you run that project has come from the [YourProjectName].Client project.
This approach means you only have one site running on your server (which could be good or bad) and also means you won't run across any CORS issues.
But you don't have to have an ASP.NET Core site at all
If your Blazor site is a standalone site that doesn't read/write from any server, or if it only talks to 3rd party APIs or an existing Web API running on the older .NET Framework, then you don't actually need an ASP.NET Core site at all.
In that case you don't use the 'hosted' option.
Instead, you can simply publish your Blazor project and then take the files from the release folder and host them in any site.
Update
I see where you are coming from now. The confusion stems from the fact that you have an option called --hosted when using the client-hosted Blazor. This options means having Blazor to include ASP.NET Core runtime.
Why this option? Because you can write an offline app (e.g. calculator app) that does not need any kind of connection to external services, making ASP.NET Core irrelevant. However, you might want to write an online app that accesses online DB, external APIs, do verification, etc. For these kind of apps, you will need an ASP.NET Core stack to support your app.
Check this FAQ: https://github.com/aspnet/Blazor/wiki/FAQ#q-can-i-use-blazor-with-aspnet-core-on-the-server
Original answer
They are two hosting models: server-hosted, and client-hosted.
The difference is whether the app is hosted in server, or in client. Server hosting means your app logic runs in the server (you can think of it similar to what Web Forms is), you click on a button, an "Ajax" call sends the request, the server receives the request, and sends back the updated page. However, here it uses SignalR not Ajax, which is a low level socket communication (read efficient). And instead of updating a whole page, it updates only the relevant parts (thus it is a single page application).
On the other hand, client hosting means your logic runs within the browser. Think of it as if your C# logic is converted into JS, and it is embedded in the page. So the logic runs in the browser. This is possible after the introduction of WebAssembly which you might want to read about.
Let's say you want to create a calculator app. Your server hosted app will then need to communicate with the server to calculate and get the result for each calculation, while the client hosted does not need, and calculates the result in browser.
You might wonder, why we have two options. The reason being that support for WebAssembly (which a client hosted app relies on) is either incomplete or non-existant in many browsers, and performance differs widely too.
https://caniuse.com/#feat=wasm
The question is about the option "ASP.NET Core hosted" in Visual Studio, while creating a new Blazor Project with Blazor WebAssembly App.
Selecting this option scaffolds out 3 Projects for you (vs 1 WebAssembly project, if this option is not selected)
Server Side Blazor Project
Client Side Blazor Project (WebAssembly)
Shared project for shared entities between Server and Client Projects.
With this option you can have
Blazor WebAssembly only option for sections of your project where the logic can execute in browser.
Server hosted option where all the processing is done on server and only HTML is rendered onto the browser.
When deployed, all these projects go to the same location. So if you have a requirement for both Server Side as well as Client side options, then go for it.
I too had the same confusion while reading the documentation. But it was explained in Blazor webassembly section
dotnet blazorwasm - is a standalone project
dotnet blazorwasm --hosted (or dotnet blazor --hosted) - is a projected hosted with ASP.NET Core backing API's.
A hosted deployment serves the Blazor WebAssembly app to browsers from an ASP.NET Core app that runs on a web server.
The client Blazor WebAssembly app is published into the /bin/Release/{TARGET FRAMEWORK}/publish/wwwroot folder of the server app, along with any other static web assets of the server app. The two apps are deployed together. A web server that is capable of hosting an ASP.NET Core app is required. For a hosted deployment, Visual Studio includes the Blazor WebAssembly App project template (blazorwasm template when using the dotnet new command) with the Hosted option selected (-ho|--hosted when using the dotnet new command).
Here are my two cents.
Create a Blazor WebAssembly App with ASP.NET Core Hosted check box selected.
Once created observe the project references.
Note that the server project references client project.
I had this same confusion as others, that the client calls the server, and that they are two independent Visual Studio projects. So this project references puzzled me.
The fact is client project is hosted by and served by the server project. When we want to run the app, we need to run the server project and not the client project.
So when you want to run the Solution(App), you must ensure to set the Server project as the startup project and not both the server and client. Due to my initial misunderstanding, I was setting multiple startup project, thinking that the client(something like React App) calls the server(WebApi), and that they both should be running simultaneously, for the client to call the server.
The above is correct, but the following is INCORRECT.
If you do that, you get the following error, and this stung me.
Microsoft Visual Studio - One or more errors occurred. Failed to launch debug adapter error
And finally if you are thinking to dockarize by adding docker files and docker-compose files for different kinds of Blazor apps, do take a look at this github repo.
Specifically for this kind of apps that we are talking about take a look at this and this folder in that repo.
Blazor-WASM
The application can work offline after the first loading. As an Example If you make Loan calculator with Blazor WASM, you can send all the data, business logic and validation to the client side in the first load like loan types, loan rates, max repayment period like that then because of the logic also in client side, it don't need do the communication with the server again and again when you change the loan type or repayment period because business logic in client side using the user selected data Blazor can render the UI and show the result.
Blazor-server
Application use server to render the UI so as on the Loan application if user change the loan type blazor send that user change value to server using SignalR and in the server has data and the logic for get the result, after that server render UI and send UI changes to the client side using SignalR and redraw the UI in clients screen.
Blazor-ASP.NET Core Hosted
Visual studio generate 3 projects in the core hosted template
ClientProject - this is a Blazor-WASM application
ServerProject - this is a .NET CORE API
SharedClassProject - this is a class library that holds shared
objects between client and server
As the loan example in the first load loan types and calculation logic are loaded in client side.
If the user selects the loan type Blazor creates an http call to the API application to get the loan rates according to the selected loan type if the interest rate is in the db.
Then the user enters the repayment period because of calculation logic in the client side. Now the application can calculate the result and render the UI to show the result. So with this model you can keep some data or logic in the server and other data or logic on the client side.
Apologies for the wordiness, I'm new to Web APIs. Allow me to provide a little background. I have a solution with three projects: Repository with SQL entity framework, Console app, and MVC Web API. My repository queries SQL, feeds that data to the console app, which outputs a list of calculations from the raw SQL data. I now want to be able to push that list to my localhost via Web API. I have the Web API pushing data to a localhost browser from a test repository with an array of test data within (this data isn't calculated, I just made a static array to see if I could pull anything), and it works just fine. But when I try to take the output list from the console app and feed it into my test repository and then push that data up the data doesn't push through. I made the output list available throughout the entire assembly so I can access the console app list from the Web API.
Any ideas?
Thanks
It sounds to me like you need to restructure your projects a bit. I recommend refactoring your code so that the projects are as follows:
Repository project: this will talk to your database and is responsible for retrieving data from your database.
Application project: this references your repository project and performs the calculations that are currently being done in your console app.
Console app project: this is your console app code. It references your Application project and does nothing more than request the results of the calculation from your Application project, then outputs the results to the console.
MVC WebAPI project: this also just references your Application project. Your WebAPI web service call will request the results of the calculation from your Application project, then return it to the caller.
This will be much simpler for you to manage. Also, you can easily unit test just your Application project class methods that perform the calculation.
I have a .NET solution which has two projects:
WPF project for the WPF application
ASP.NET WebApi project with a few controllers
I want to host the WebApi inside my WPF application (using OWIN). Any idea how can I do this?
If I create my controllers in the WPF project itself then I can host them using WebApp.Start (and pass a config class).
Here's an example of how to make a desktop application with OWIN and WebApi http://www.codeproject.com/Articles/869223/ASP-NET-Web-Api-Create-a-Self-Hosted-OWIN-Based-We
It's true that it's a console based application but you could use that in the bootstrap of your application to start the OWIN server when the app starts.
I would also advise trying it in a new test project until you get a hang of it before starting to create api controllers and mapping out routes
Yeah. Start reading documentation. Then "just go". This (self-hosting) is exactly what OWIN was designed to do and it is all documented.
IN your app at one point you will configure and start an OWIN based server. Contrary to how you do normal web apps, you can abandon the WebApi project - all controllers etc. can live in an assembly. Reflection is used to find them.