Oliver Stieber <oliver_stieber <at> yahoo.co.uk> writes: > Your right the mipmap level should only ever need to be set once, just after > the texture is > created. The problem is that this doesn't seem to work, at least with ATIs > drivers this demo > (http://www.codesampler.com/dx9src/dx9src_3.htm#dx9_texture_filtering) fails, > along with many > games and demos without some weird code. This is one of the two regression > problems that are > preventing d3d8-wined3d wrapper and evict managed resources from being checked > in. >
I'd say check it in anyways... if you are worried about breakage on ATI machines, you could always check the gl_vendor and skip the code that causes problems on ATI machines. > Again, I'm fairly sure that there are some things that don't work without > reapplying the states > all the time. The last time I looked at this was just after the last big patch > I put out so I > can't quite remember which demos/games didn't work. I'll test these on my ATI machines and check to see if these work-arounds are still valid/needed. If there are any bugs or issues with ATI's driver, please let me know the details and I'll communicate them to ATI. > I'm just about to start tidying up directx.c, or at least the supported > formats bit of directx.c > and then I was going to tidyup drawprim a bit. What is the expected coding > style, I just coppied > that one that was being used. That would be great. If there is anything you'd like to delegate to me just let me know... maybe I'll start with a clean up of device.c. The only thing that really bothers me, style-wise, is the line length of some of the code. But wine doesn't seem to really have any style guidelines so I guess I'll have to live with this for now. Regards, Aric
