3 Oct
2025
3 Oct
'25
5:25 p.m.
One failure I know about is in kernel32:sync, I think it's enough of a corner case to not be worried about it. The test that fails was added specifically to check mutex ownership during the sync object refactor.
The owner thread gets terminated while every known handle for the owned mutex have been closed already, but while a thread is still waiting on it. From wineserver perspective the mutex doesn't exist anymore, and it has no way to know about the remaining waiters or owners, and this would need to be resolved on the kernel side.