http://bugs.winehq.org/show_bug.cgi?id=12948
--- Comment #9 from Anastasius Focht focht@gmx.net 2008-07-28 01:17:28 --- Created an attachment (id=15131) --> (http://bugs.winehq.org/attachment.cgi?id=15131) +tid,+server tracelogs for successful and failing run
Hello,
such race conditions do not manifest every time (by nature). Everything in system can change timing behaviour. There is only a small time frame where this problem manifests (milliseconds range).
I did my analysis and came to this conclusion. Attached are the server traces from failing and successful run.
You can grep the files for all mutexes, remove wine builtin ones and compare both files regarding order of mutex creation (by thread id). The key one is "<pid>:DAF".
For me this case is closed.
Regards