Dispose of ThreadRental at conclusion of UITheory test

[AArnott]

Fixes #44 hang at end of test run.

[AArnott]

@weltkante Would you review?

[weltkante]

The change looks sensible. I've downloaded the artifact and can confirm both the repro and winforms no longer hang.

[weltkante]

I've started seeing following exception. Not sure if its related to this change. I doubt its related to the other PR (I'm using this PR's artifact). Also I'm not entirely sure if its a new issue or existed before and just wasn't surfacing due to the hang.

The stack trace is a bit weird, with a failed debug assert mixed into it, I'm still trying to figure out how to diagnose this, doesn't happen when debugging tests from inside VS [edit] turns out I was debugging the wrong tests

Unhandled exception. System.InvalidOperationException: Invoke oder BeginInvoke kann für ein Steuerelement erst aufgerufen werden, wenn das Fensterhandle erstellt wurde.
   at System.Windows.Forms.Control.MarshaledInvoke(Control caller, Delegate method, Object[] args, Boolean synchronous) in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\Control.cs:line 6896
   at System.Windows.Forms.Control.BeginInvoke(Delegate method, Object[] args) in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\Control.cs:line 4707
   at System.Windows.Forms.WindowsFormsSynchronizationContext.Post(SendOrPostCallback d, Object state) in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\WindowsFormsSynchronizationContext.cs:line 90
   at Xunit.Sdk.Utilities.SyncContextAwaiter.OnCompleted(Action continuation)
   at System.Runtime.CompilerServices.AsyncTaskMethodBuilder`1.AwaitUnsafeOnCompleted[TAwaiter](TAwaiter& awaiter, IAsyncStateMachineBox box)
--- End of stack trace from previous location ---
   at System.Threading.Tasks.Task.<>c.<ThrowAsync>b__139_1(Object state)
   at System.Threading.QueueUserWorkItemCallbackDefaultContext.Execute()
   at System.Threading.ThreadPoolWorkQueue.Dispatch()
   at System.Threading._ThreadPoolWaitCallback.PerformWaitCallback()
Process terminated. Assertion Failed
Marshaling control should have created its handle in its ctor.
   at System.Windows.Forms.WindowsFormsSynchronizationContext..ctor() in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\WindowsFormsSynchronizationContext.cs:line 35
   at System.Windows.Forms.WindowsFormsSynchronizationContext.InstallIfNeeded() in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\WindowsFormsSynchronizationContext.cs:line 143
   at System.Windows.Forms.Application.ThreadContext.RunMessageLoopInner(msoloop reason, ApplicationContext context) in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\Application.ThreadContext.cs:line 1119
   at System.Windows.Forms.Application.ThreadContext.RunMessageLoop(msoloop reason, ApplicationContext context) in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\Application.ThreadContext.cs:line 992
   at System.Windows.Forms.Application.DoEvents() in D:\projects\winforms\src\System.Windows.Forms\src\System\Windows\Forms\Application.cs:line 811
   at Xunit.Sdk.WinFormsSynchronizationContextAdapter.PumpTill(SynchronizationContext synchronizationContext, Task task)
   at Xunit.Sdk.ThreadRental.<>c__DisplayClass11_0.<CreateAsync>b__0()
   at System.Threading.ThreadHelper.ThreadStart_Context(Object state)
   at System.Threading.ExecutionContext.RunInternal(ExecutionContext executionContext, ContextCallback callback, Object state)
   at System.Threading.ThreadHelper.ThreadStart()

[weltkante]

Yeah can confirm this is not the fault of this PR, I rolled back to 1.0.31-beta and skipped the hanging test assembly and I was already getting the same crash. Just never seen it before because the hanging test assembly runs before the crashing one.

[AArnott]

Thanks for testing. I trust you'll file another bug if you find the crash you saw is due to xunit.stafact.

[weltkante]

if you find the crash you saw is due to xunit.stafact

Looks like this needs to be adressed in the WinForms unit tests, they are disposing the SC early. Probably shouldn't be WinFormsFacts if they setup/terminate the infrastructure manually.

[AArnott]

@weltkante I'm not sure whether you're blaming the WinFormsFactAttribute behavior or the unit tests in the WinForms repo. Can you clarify?

[weltkante]

Not blaming stafact, continuing discussion on dotnet/winforms#3297