Stephen Edwards

I think this is still focused on one test whereas the other is more 'nebulous' flakiness

@qwwdfsad Is the behaviour for A: ``` someFlow.onEach { updateValue() }.launchIn(scope + Dispatchers.Unconfined) ``` equivalent to B: ``` scope.launch(start = CoroutineStart.UNDISPATCHED, context = Dispatchers.Unconfined) { someFlow.collect { updateValue() } ```...

@MilanJovic92 can you give more context on why we should do this or what it accomplishes? Can you link this to an issue?

For example, #1105 added a new test class but did not invoke it. Making no changes to code that would be executed in CI. But yet the instrumentation shards all...

Renaming `headlessIntegrationTest` to something less focused makes sense as well.

CC: @zach-klippenstein this was the reason for the problem I referred to.

Would be very happy to see this library move to Kotlin! This is exciting!

Note that using a dispatcher hooked up to the frame clock - like Compose's `AndroidUiDispatcher.Main`- to dispatch the collection of the renderings from the Workflow runtime will have this effect...

I think this likely looks like a WorkflowTracer API that includes these.

The reason we could not use `yield()` in the middle of a runtime loop update is we want to keep the semantics of "1 main thread message for 1 workflow...