http://bugs.winehq.org/show_bug.cgi?id=421
--- Comment #366 from max@veneto.com 2010-05-25 17:00:04 ---
Well, I understand you did a dirty hack, so I was hoping I could help you (or someone trying to locate the problem) to find it. It might be a bug with wine, isn't it true?
Are you sure this is not the bad commit? I don't see what I have done wrong. I did a git reset --hard f2b993d414517037cf064e9aed954cdae96646f5, patched, compiled, and tested, it crashes. The commit before that is 5532927ed33f3682c48bdc2a2528488eeefa5bad (according to git log), so I reseted to it, patched, compiled, tested and it works.
And I've just tested it again.
Here is the sequence of commands (careful, it deletes WINEPREFIX):
export WINEPREFIX=/home/gugamilare/.wine-test/ rm -rf dlls/winedib.drv/ git checkout -f make clean git reset --hard f2b993d414517037cf064e9aed954cdae96646f5 sh patches/series # I transformed the file "series" of the tarball into a script autoconf CC="ccache gcc" ./configure && make depend && make rm -rf $WINEPREFIX ./wine notepad # => works rm -rf $WINEPREFIX WINEDIB=ON ./wine notepad # => crashes
rm -rf dlls/winedib.drv/ git checkout -f make clean git reset --hard 5532927ed33f3682c48bdc2a2528488eeefa5bad sh patches/series autoconf CC="ccache gcc" ./configure && make depend && make rm -rf $WINEPREFIX WINEDIB=ON ./wine notepad # => works
uhmmmm.... it would be weird. I'll check it on week end.... but the patch seems to me unrelated to the bug.... or it may expose a previous bug in dib engine. Your test seems correct to me, indeed. Could you add a trace of StretchDIBits() parameters in both cases (with and without the offending patch) ?
Max
Max