On Monday 19 July 2010 21:41:26 Henri Verbeet wrote: > 2010/7/19 Oldřich Jedlička <oldium....@seznam.cz>: > > I will enhance the log message, no problem here. Should I consider the > > referenced patch as being applied (so that I should base my patch on it)? > > Yeah.
Ok, I will update it according to your changes :-) > > The tests will be a problem, because I don't have the Windows machine as > > a reference. Also `make test` fails on DirectX tests for me because of > > r600 driver bug... > > > > When I write CreateSurface tests (for various interfaces), is there a way > > for me to run the tests on Windows? I have a Windows XP machine at work. > > As Austin mentioned you can use "make crosstest" if you have a wine32 > cross compiler installed. You can also ask people on that bug, > wine-devel, or IRC to run a test for you on native Windows. I will have a look at this possibility, I've found a nice hint on cross- compiling for Gentoo at http://wiki.winehq.org/JohnKlehm. It looks like I can do it myself (for now). > Wrt the patch itself, can't you just check for the appropriate flag in > IDirectDraw7Impl_CreateSurface(), and allow it in > IDirectDrawImpl_CreateSurface()? (I.e., avoid adding the > AllowBackBuffer parameter.) I just wanted to keep the debugging messages (dumping DDSD) where they are and fail afterwards. If I moved the check to the interface-specific function, there would be no such DDSD dump. But yes, this can be work-around to dump it just when the back-buffer test fails. I will send the updated version (without tests for now) tomorrow for more comments :-) Thanks, Oldrich.