http://bugs.winehq.org/show_bug.cgi?id=31557
--- Comment #30 from Alpo Turunen alpoturunen@gmail.com 2013-05-13 12:06:29 CDT --- Created attachment 44461 --> http://bugs.winehq.org/attachment.cgi?id=44461 Traces of gdiplus when taking a screenshot in GW1
From reading the code, it seems the memcpy operations of wicformat have same
results on both pre-regression and post-regression code. The only difference it that old code left gdipformat to 0 when IsEqualGUID(&wicformat, wic_pixel_formats[i]) is false for all i, and new code sets gdipformat = PixelFormat32bppARGB in that case.
Pleaes correct me if I'm wrong, and please take a look at trace+gdiplus of the different versions if it helps give any idea what's wrong with the code.