http://bugs.winehq.org/show_bug.cgi?id=8974
------- Additional Comments From lhervier@ifrance.com 2007-30-07 10:45 ------- Second try gives:
cc90b25aef7ec8a357733c98b2e56105a1ee26e6 is first bad commit commit cc90b25aef7ec8a357733c98b2e56105a1ee26e6 Author: Alexandre Julliard julliard@winehq.org Date: Thu Jul 12 15:45:00 2007 +0200
user32: Check both A and W proc for previously allocated winprocs.
Some apps get a winproc as A and then set it again as W.
:040000 040000 f4b68e82c9577b7933bed7562a51b43ca1e9a493 a2ff25f1f42f4b3cd917e7c3c6cbf799794604eb M dlls
This commit is between 55d70e894cd70f69d313771e8f5dedd56cce4bc0 and 98934f20d8149ad93355e86d8a5f0a96b57678e9 as mentioned by Austin.
I really doubt that commit caused the regression. You can check by rebasing to origin and reverting that specific patch, recompiling, and trying again.
I never used git before, so it's time to learn ! I will read some documentation and try to revert this patch.