Re: Re: Tips for debugging D3D Textures & DirectSound?
Ove Kaaven <[email protected]> Wed, 13 Oct 2004 13:31:45 -0400
| Newsgroups | gmane.comp.emulators.winex.devel |
|---|---|
| Organization | TransGaming Technologies Inc |
| Message-ID | <1097688704.1850.413.camel@renegade> |
lør, 09.10.2004 kl. 00.09 skrev Adam Luchjenbroers: > The if block was setting gl's error # (returned via glGetError) to 0x0502 > (GL_INVALID_OPERATION). I was working off the assumption that this was > causing some textures to fail to load (got most of them loading, just the > loading screens to go and a rendering glitch in the unit-painter where a > texture is rendered WAY oversize). Sounds like you may have made some wrong assumptions, leading you to believe getting NPOT working would actually be easy. If it was, we would have done it ourselves long ago. Getting the textures loaded and used is easy if you know how (though it's not clear that you do), getting them rendered properly is the hard part. > The other thing I notice is the use of variables State and state in > D3D_GL_texture_map. As C is case-sensitive these are two seperate variables > yet I don't see state defined anywhere... is it a global or something? No. It's a local variable declared by the _DWTSS() macro that's used in there. > Also got started on finding out why sound doesn't work. It seems to be > independant of whether OSS or ALSA is used and is also unaffected by mmap. > There is however, nothing that really stands out in a trace log of +dsound, > +wave except for the following: > warn:dsound:DSOUND_MixCancelAt (0x553a5f70, 4096), buf_mixpos=4176 This message does not represent a real problem. It just warns about an inefficiency we still have in the dsound code, causing us to spend a little bit more CPU time than a perfect implementation would. Nothing serious.