http://bugs.winehq.org/show_bug.cgi?id=10204
--- Comment #17 from Kai Blin kai.blin@gmail.com 2008-06-14 17:07:20 --- (In reply to comment #15)
(In reply to comment #14)
I didn't have time to look into this yet, but I'm not really sure the write event test cases are valid. They seem to fail on XP and above, at least.
Could you describe what do you mean by "seem to fail"? I've just checked my standalone test program on Windows XP and it behaves as described in this bug entry (which means that it behaves differently than in Wine). I'm going to run Wine test on Windows XP using MinGW, to see if there aren't any problems with the test itself.
Fail as in:
http://test.winehq.org/data/062d61a5a4ba2d7972a0011387ceda64c79dd4e3/xp_P4XP...
Which seems to be just the FD_WRITE event test.
Also fails on Win2k3 http://test.winehq.org/data/062d61a5a4ba2d7972a0011387ceda64c79dd4e3/2003_jh...
http://test.winehq.org/data/062d61a5a4ba2d7972a0011387ceda64c79dd4e3/2003_W2...
and Vista http://test.winehq.org/data/062d61a5a4ba2d7972a0011387ceda64c79dd4e3/vista_M...
So I'm talking about the wine test cases, not your standalone test.
See the interesting discussion on bug #9777 as well. Wine does something fishy there, but I didn't get a fix for this right yet.