[Bug 53434] user32:msg - test_SendMessage_other_thread() sometimes gets a bad WM_USER message sequence on Windows

[WineHQ Bugzilla]

https://bugs.winehq.org/show_bug.cgi?id=53434

--- Comment #3 from Esme Povirk  ---
This pattern is masking a consistent failure on Windows 7 machines in
test_PeekMessage:

msg.c:13432: Test failed: wrong qstatus 00000000

The test doesn't have any obvious race condition (it correctly waits for
SendNotifyMessage in the other thread to complete), and the following
PeekMessage does process the notify message.

More strangely, this test has later checks doing almost the same thing (but
with more message types in the queue), and the queue status is set correctly
there.