This question is unlikely to help any future visitors; it is only relevant to a small geographic area, a specific moment in time, or an extraordinarily narrow situation that is not generally applicable to the worldwide audience of the internet. For help making this question more broadly applicable, visit the help center.
Closed 10 years ago.
I have a long running task method, using Sleep
public Task LongRunning() {
return Task.Factory.StartNew(
() => {
Trace.TraceInformation("Start Sleep");
Thread.Sleep(10000);
Trace.TraceInformation("End Sleep");
});
}
This is called by my test, and it works fine
[TestMethod]
public void SimpleContinueWith() {
Trace.TraceInformation("Start");
LongRunning()
.ContinueWith(
t => Trace.TraceInformation("End")
).Wait();
}
> QTAgent32.exe Information: 0 : Start
> QTAgent32.exe Information: 0 : Start Sleep
> QTAgent32.exe Information: 0 : End Sleep
> QTAgent32.exe Information: 0 : End
But using async/await the test falls straight through
[TestMethod]
public async void SimpleAwait() {
Trace.TraceInformation("Start");
await LongRunning();
Trace.TraceInformation("End");
}
> QTAgent32.exe Information: 0 : Start
> QTAgent32.exe Information: 0 : Start Sleep
Why is that then?
MSTest cannot (currently) handle asynchronous tests. I'm not sure if Microsoft is going to add this for the final release. Update: VS11 Beta did add support for async unit tests; see below.
You can unit test asynchronous methods by providing an async context yourself. There's some included in the Async CTP (Microsoft Visual Studio Async CTP\Samples\(C# Testing) Unit Testing\AsyncTestUtilities), or you can use one I wrote called AsyncContext.
Using AsyncContext, your test can be written as:
[TestMethod]
public void SimpleAwait() {
AsyncContext.Run(async () =>
{
Trace.TraceInformation("Start");
await LongRunning();
Trace.TraceInformation("End");
});
}
Update, 2012-02-05: Another option is the new AsyncUnitTests library. Install that NuGet package, change your TestClass to AsyncTestClass, and your async unit tests can be written much more naturally:
[TestMethod]
public async void SimpleAwait() {
Trace.TraceInformation("Start");
await LongRunning();
Trace.TraceInformation("End");
}
Update, 2012-06-06: If you update to VS2012 Beta, you can define asynchronous unit tests; they just have to return Task:
[TestMethod]
public async Task SimpleAwait() {
Trace.TraceInformation("Start");
await LongRunning();
Trace.TraceInformation("End");
}
Related
I need a periodic background job in Xamarin that runs once per day. I've landed on the Shiny library to handle this functionality. The tutorial I found by Allan Ritchie says that Run() needs to return true if network calls were made. Today there is no return value for Run(); only Task. Is this not still a necessity?
In my project I setup a Shiny job exactly as defined in the sample from Github. It runs great for 2 days and just stops running any more after that. This is on iOS. I basically copy/pasta'd the sample and the job kicks off fine for a few times and then just stops the next day.
What could I be missing or doing wrong?
Source for IJob: https://github.com/shinyorg/shiny/blob/v2.7.0/src/Shiny.Jobs/IJob.cs
Sample of Shiny Jobs: https://github.com/shinyorg/samples/blob/main/Jobs/Sample/SampleJob.cs
Code from the tutorial by Allan Ritchie: https://allanritchie.com/posts/shinyjobs
public class YourFirstJob : Shiny.Jobs.IJob
{
readonly IYourDepdendency depdency;
public YourFirstJob(IYourDependency dependency)
{
this.dependency = dependency;
}
public async Task<bool> Run(JobInfo jobInfo, CancellationToken cancelToken)
{
var id = jobInfo.GetValue("Id", 25); // we'll cover this in a minute
await this.dependency.SomeAsyncMethod(id);
return true; // this is for iOS - try not to lie about this - return true when you actually do receive new data from the remote method
}
}
Im trying to write a unit test that should validate if a consumer ran or not.
But for some reason the consumer never executes.
I have created a more simplified version of the problem using the Getting-Started-sample of MassTransit.
This is "my" code:
[TestClass]
public class UnitTest1
{
private ITestHarness _testHarness;
[TestInitialize]
public void Initialize()
{
var serviceCollection = new ServiceCollection();
serviceCollection.AddMassTransitTestHarness(busRegistrationConfigurator =>
{
busRegistrationConfigurator.AddConsumer<MessageConsumer>();
});
var serviceProvider = serviceCollection.BuildServiceProvider();
_testHarness = serviceProvider.GetRequiredService<ITestHarness>();
}
[TestMethod]
public async Task TestMethod1()
{
await _testHarness.Bus.Publish(new Message { Text = "Hello, world!" });
(await _testHarness.Published.Any<Message>()).Should().BeTrue();
(await _testHarness.Consumed.Any<Message>()).Should().BeTrue();
}
}
The following assertion:
(await _testHarness.Consumed.Any<Message>()).Should().BeTrue();
Fails, since it always returns False.
I guess that I somehow need to await the consumer to execute? And I might be missing even more..
Been having a look at the documentations of MassTransit regarding testing, but Im not sure if the sample provided in the docs apply to this scenario (?), since the sample provided involves a request and response.
Would very much appreciate any help!
You're missing one line:
await _testHarness.Start();
Of course, you'll need to be sure and Dispose the container to ensure the bus is stopped and cleaned up.
I have a 3rd party API IOS Binding which I am trying to test (more like an integration test) using TouchRunner.
An example API Method is this -
_client.AuthenticateWithUsername(username, token,
() => { // Success Callback },
() => { // NoConnection Callback },
(obj) => { // Other Error Callback });
The API when called goes off and does some work in the background then eventually makes one of the callbacks above, I would like to control the flow of the unit test using something like -
How can I unit test async methods on the UI Thread with Xamarin iOS TouchRunner
Unfortunately, when I insert the AutoResetEvent code, TouchRunner just hangs and never returns to the GUI.
I have also tried to use a TaskCompletionSource as follows -
public async Task<AuthResponse> AuthenticateUserAsync(string username, string password)
{
TaskCompletionSource<AuthResponse> tcs = new TaskCompletionSource<AuthResponse>();
AuthResponse response = new AuthResponse { Success = false };
LoginResponse loginResponse = await LoginUser(username, password);
_client.AuthenticateWithUsername(username, loginResponse.token,
() =>
{
response.Success = true;
Console.WriteLine("Auth");
tcs.SetResult(response);
},
() => { tcs.SetResult(response); },
obj => { tcs.SetResult(response); },
obj => { tcs.SetResult(response); });
return await tcs.Task;
}
[Test]
public async void AuthenticateUserAsyncTest()
{
var auth = await AuthenticateUserAsync(_username, _password);
Assert.IsTrue(auth.Success);
}
The debugger stepped through fine until the return await tcs.Task, but then results in a similar HUNG runner.
How can I work out why the hang is happening?
As this was not working, I then resorted to code like this -
_client.AuthenticateWithUsername(_username, loginResponse.token,
() =>
{
Assert.Pass("This crashes the runner");
Assert.True(true); // This DOES NOT!
},
() =>
{
// This will crash runner also
Assert.Fail("NoConnection");
},
(InvalidTokenError obj) =>
{
Assert.Fail("InvalidToken" + obj.Description);
},
(ClientError obj) =>
{
Assert.Fail("ClientError" + obj.Description);
});
As you can see, the flow ends up (understandably), run test, runs client call, end of test method completes which shows test as success, then the callback returns and the asserts get called, which crash the app, which we assume is because the runner has already completed the test, why one assert works and other crashes I do not know.
So,
Am I approaching this the right way?
Could something be happening in the 3rd Party API that will cause these approaches to hang? and how would I debug that?
Thanks #Nkosi, that is a good suggestion, I forgot to add that during my original testing that when I ran the code with async Task rather than void I got an immediate block from TouchRunner without even adding any other code other than the API call! This was a red flag I suppose, but using async void "seemed" to allow "standard" async testing, so I progressed and then ended up in the loop above.
As TouchRunner has not been updated in a very long time I have just spent time re-creating the test project using XUnit after various suggestions to try it in the forums and on stack.
https://github.com/xunit/devices.xunit - runners for Xamarin IOS + Android
https://xunit.github.io/docs/comparisons - to port NUnit syntax
Some other useful links are -
https://xunit.github.io/docs/getting-started-devices.html
https://gregshackles.com/testing-xamarin-apps-getting-started-with-xunit/
https://oren.codes/2014/07/10/getting-started-with-xunit-for-xamarin/
RESULT: I am very pleased to say all the above code now works for both the TaskCompletionSource and the AutoResetTask scenarios
I can now safely test my event based API :)
I just wanted to ensure other users are aware of this.
Thanks for your help.
One observation is that the test should be async Task and not async void ie
public async Task AuthenticateUserAsyncTest() {
//...code removed for brevity.
}
async void is a fire and forget so any exceptions thrown wont happen in the current context so they wont be caught.
Reference Async/Await - Best Practices in Asynchronous Programming
I want to link async method to a delegate command in prism framework in Xamarin.Forms and my question is how to do it?
Is below solution correct? Is there exist any pitfall? (deadlock, UI slow or freezing, bad practices, ...)
{ // My view model constructor
...
MyCommand = new DelegateCommand(async () => await MyJobAsync());
...
}
private async Task MyJobAsync()
{
... // Some await calls
... // Some UI element changed such as binded Observable collections
}
You can use async void directly. However, a few notes from my experience...
The structure of your code is: start asynchronous operation and then update UI with the results. This implies to me that you would be better served with a NotifyTask<T> kind of approach to asynchronous data binding, not commands. See my async MVVM data binding article for more about the design behind NotifyTask<T> (but note that the latest code has a bugfix and other enhancements).
If you really do need an asynchronous command (which is much more rare), you can use async void directly or build an async command type as I describe in my article on async MVVM commmands. I also have types to support this but the APIs for these are more in flux.
If you do choose to use async void directly:
Consider making your async Task logic public, or at least accessible to your unit tests.
Don't forget to handle exceptions properly. Just like a plain DelegateTask, any exceptions from your delegate must be properly handled.
Just have a look at this link if you're using Prism Library: https://prismlibrary.com/docs/commands/commanding.html#implementing-a-task-based-delegatecommand
In case you want to pass a CommandParameter to DelegateCommand, use in the DelegateCommand variable declaration this syntax
public DelegateCommand<object> MyCommand { get; set; }
In the constructor of the ViewModel initialize it this way:
MyCommand = new DelegateCommand<object>(HandleTap);
where HandleTap is declared as
private async void HandleTap(object param)
Hope it helps.
As has already been mentioned the way to handle async code with delegate command is to use async void. There has been a lot of discussion on this, far beyond just Prism or Xamarin Forms. The bottom line is that ICommand that both the Xamarin Forms Command and Prism DelegateCommand are limited by ICommand's void Execute(object obj). If you'd like to get more information on this I would encourage you to read the blog by Brian Lagunas explaining why DelegateCommand.FromAsync handler is obsolete.
Generally most concerns are handled very easily by updating the code. For example. I often hear complaints about Exceptions as "the reason" why FromAsync was necessary, only to see in their code they never had a try catch. Because async void is fire and forget, another complaint I've heard is that a command could execute twice. That also is easily fixed with DelegateCommands ObservesProperty and ObservesCanExecute.
I think the two main problems when calling an asynchronous method from one that executes synchronously (ICommand.Execute) are 1) denying to execute again while previous call is still running 2) handling of exceptions. Both can be tackled with an implementation like the following (prototype). This would be an async replacement for the DelegateCommand.
public sealed class AsyncDelegateCommand : ICommand
{
private readonly Func<object, Task> func;
private readonly Action<Exception> faultHandlerAction;
private int callRunning = 0;
// Pass in the async delegate (which takes an object parameter and returns a Task)
// and a delegate which handles exceptions
public AsyncDelegateCommand(Func<object, Task> func, Action<Exception> faultHandlerAction)
{
this.func = func;
this.faultHandlerAction = faultHandlerAction;
}
public bool CanExecute(object parameter)
{
return callRunning == 0;
}
public void Execute(object parameter)
{
// Replace value of callRunning with 1 if 0, otherwise return - (if already 1).
// This ensures that there is only one running call at a time.
if (Interlocked.CompareExchange(ref callRunning, 1, 0) == 1)
{
return;
}
OnCanExecuteChanged();
func(parameter).ContinueWith((task, _) => ExecuteFinished(task), null, TaskContinuationOptions.ExecuteSynchronously);
}
private void ExecuteFinished(Task task)
{
// Replace value of callRunning with 0
Interlocked.Exchange(ref callRunning, 0);
// Call error handling if task has faulted
if (task.IsFaulted)
{
faultHandlerAction(task.Exception);
}
OnCanExecuteChanged();
}
public event EventHandler CanExecuteChanged;
private void OnCanExecuteChanged()
{
// Raising this event tells for example a button to display itself as "grayed out" while async operation is still running
var handler = CanExecuteChanged;
if (handler != null) handler(this, EventArgs.Empty);
}
}
async void
I personally would avoid "async void" at all cost. It is impossible to know from the outside when the operation has finished and error handling becomes tricky. In regards to latter, for instance writing an "async Task" method which is called from an "async void" method almost needs to be aware of how its failing Task is propagated:
public async Task SomeLogic()
{
var success = await SomeFurtherLogic();
if (!success)
{
throw new DomainException(..); // Normal thing to do
}
}
And then someone writing on a different day:
public async void CommandHandler()
{
await SomeLogic(); // Calling a method. Normal thing to do but can lead to an unobserved Task exception
}
Is UI thread running DelegateCommand and background threads running await expression?
Yes, the UI thread runs the DelegateCommand. In case of an async one, it runs until the first await statement, and then resumes his regular UI thread work. If the awaiter is configured to capture the synchronization context (that is, you do not use .ConfigureAwait(false)) the UI thread will continue to run the DelegateCommand after the await.
Is UI thread running DelegateCommand and background threads running await expression?
Whether the "await expression" runs on a background thread, foreground thread, a threadpool thread or whatever depends on the api you call. For example, you can push cpu-bound work to the threadpool using Task.Run or you can wait for an i/o-operation without using any thread at all with methods like Stream.ReadAsync
public ICommand MyCommand{get;set;}
//constructor
public ctor()
{
MyCommand = new Xamarin.Forms.Command(CmdDoTheJob);
}
public async void DoTheJob()
{
await TheMethod();
}
public DelegateCommand MyCommand => new DelegateCommand(MyMethod);
private async void MyMethod()
{
}
There are no pitfalls. A void return type in async method was created especially for delegates. If you want to change something, that has reflected on UI, insert relevant code in this block:
Device.BeginOnMainThread(()=>
{
your code;
});
Actually, ICommand and DelegateCommand pretty similar, so an above answer is quite right.
I'm unsure about the best practice for obtaining and updating input received from a controller monitored using the GamePad class in UWP.
I've seen a couple of examples of people using Dispatch Timers and async loops inside the GamePadAdded event. In Win32 applications, I would have handled input in the WinMain update/message loop, but in UWP apps I don't know of anything similar.
Is there a loop in UWP apps that input should be collected/handled like in Win32 apps? What is the recommended protocol for polling for input from a input device (nominally a Xbox One controller)?
I'm happy to read more about UWP app development but I'm unsure of any guides that reference something like this.
Edit: It would be productive if, instead of downvoting and moving on, you shared thoughts on why this question deserved a downvote.
I've seen a couple of examples of people using Dispatch Timers and async loops inside the GamePadAdded event
This is the right way in UWP app to read Gamepad data. A little suggestion is, move the loop reading part on UI thread if you need to update UI frequently. See the solution in this blog
Is there a loop in UWP apps that input should be collected/handled like in Win32 apps
You may make a wrapper with custom event, see the open source implementation: XBoxGamepad
public class XBoxGamepad
{
private List<Gamepad> _controllers = new List<Gamepad>();
private bool _running = true;
Task backgroundWorkTask;
public event EventHandler<GamepadButtons> OnXBoxGamepadButtonPressA;
//omitted......
public XBoxGamepad()
{
Gamepad.GamepadAdded += Gamepad_GamepadAdded;
Gamepad.GamepadRemoved += Gamepad_GamepadRemoved;
backgroundWorkTask = Task.Run(() => PollGamepad());
}
//omitted......
private void Start()
{
_running = true;
}
public void Stop()
{
_running = false;
}
public async Task PollGamepad()
{
while (true)
{
if (_running)
{
foreach (Gamepad controller in _controllers)
{
if (controller.GetCurrentReading().Buttons == GamepadButtons.A)
{
OnXBoxGamepadButtonPressA(controller, controller.GetCurrentReading().Buttons);
}
//omitted......
}
}
await Task.Delay(50);
}
}
private void Gamepad_GamepadRemoved(object sender, Gamepad e)
{
_controllers.Remove(e);
}
private void Gamepad_GamepadAdded(object sender, Gamepad e)
{
_controllers.Add(e);
}
}