[Bug 54407] user32:msg - test_SendMessage_other_thread() sometimes gets a bad WM_USER message sequence in Wine

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=54407

--- Comment #1 from Esme Povirk  ---
Sounds like a synchronization failure? WM_USER+3 sent too early?

OK, so based on SendMessage_thread_2 being a todo_wine, we know we're dealing
with SendMessage_thread_1. We also know that the WM_USER+3 is processed inside
the GetMessageA call in test_SendMessage_other_thread. So it seems that after
WM_USER+2 is processed, the background thread is able to send out the WM_USER+3
before the main thread is able to remove the posted WM_USER from the queue and
return from GetMessageA.

I don't see anything that would prevent this from happening, so it's not clear
to me whether this is intended to test some subtle timing thing, or it's just a
race condition in the test.