Michael Hoisie
Michael Hoisie
@jongerrish could this be test pollution caused by threading? Such as a thread executing some logic in a SQLite database after the tests ends?
I believe this is only a problem with legacy SQLite, not native SQLite.
As you have probably discovered calling SystemClock.uptimeMillis and SystemClock.elapsedRealtime (and System.currentTimeMillis from framework code) give relatively meaningless values that exist in order to let Robolectric have fine-grained control over the...
If there is a testing use case here then I'd be fine with looking into supporting it. Can you give a potential sample test code that you would like to...
RE ```java @Test fun foo() { ShadowSystemClock.setCurrentTimeMillis(500) assertThat(System.currentTimeMillis()).isEqualTo(500) // I'm observing it to be the real time from the JVM, not 500 } ``` The expected behavior is to print...
What changed between 1.2 and 1.3 that made it stateful? Is there more state than SELECT and AUTH ?
When `@LooperMode(LooperMode.Mode.PAUSED)` is removed, is the test running in LEGACY looper mode? What is the value of ShadowLooper.looperMode()?
Have you also tried ShadowLooper.runUiThreadTasksIncludingDelayedTasks?
Sounds like a race condition. Maybe compose effects are occurring on a background thread, and it's possible that the result are seeing non-deterministic given that some processing is being done...
Do you have a sample app that demonstrates the problem? Is there a stack trace that you're seeing? I am unfortunately not too experienced with Compose, but I can try...