+- XBMC Community Forum (http://forum.xbmc.org)
+-- Forum: Development (/forumdisplay.php?fid=32)
+--- Forum: Development (/forumdisplay.php?fid=93)
+--- Thread: Developers wanted to help port XBMC for Mac to PPC (PowerPC) processor architecture (/showthread.php?tid=36229)
- Beenje - 2009-03-17 08:46
davilla Wrote:Well, it was a shot, OpenGL 1.3 might be too old, I forget the min we support.
Are you running 10.4 or 10.5 on the MacMini?
I'm running 10.5.6
So we can't expect updated drivers from apple...
I don't know much about OpenGL...
Can't the SW render be fixed at least?
Is the color pb a PPC bug?
- waldo22 - 2009-03-17 11:43
davilla Wrote:Well, it was a shot, OpenGL 1.3 might be too old, I forget the min we support.
Are you running 10.4 or 10.5 on the MacMini?
...it looks like it's 1.4 + GLSL
http://forum.xbmc.org/showthread.php?tid=42783
http://wiki.xbmc.org/?title=XBMC_for_Windows_specific_FAQ#XBMC_for_Windows_minimum_requirements
Hope you can get it to work in software...
I'm really excited following this thread.
-Wes
- Beenje - 2009-03-17 13:12
waldo22 Wrote:...it looks like it's 1.4 + GLSL
-Wes
Actually I saw on this page http://xbmc.org/download/ that even though the graphics device drivers only support OpenGL 1.2 on Mac with Intel chipset, the OpenGL extensions needed are present in those drivers on Mac OS X.
So I thought it might be the same on the PPC Mac mini (with OpenGL 1.3).
But it seems some needed extensions are not supported in those PPC graphic drivers on the Mac mini.
Will try to look at the SW mode.
- davilla - 2009-03-17 13:30
I'll see if I can vector d4rk into this, he's one of our our gl experts.
- Beenje - 2009-03-18 00:21
After reading a bit about OpenGL and endianness (here is a good link: OpenGL Universal Binary Programming Guidelines) my understanding is that when using GL_BGRA format, one should use GL_UNSIGNED_INT_8_8_8_8_REV type (and not GL_UNSIGNED_BYTE).
With GL_BGRA format, GL_UNSIGNED_BYTE gives the same as GL_UNSIGNED_INT_8_8_8_8_REV on little-endian system.
But that's not the case on big-endian system!
I tried this fix on my Mac mini and the colors are fine when playing video :-)
Can someone who really knows about OpenGL confirm that it's the right thing to do?
- monkeyman - 2009-03-18 03:06
Could you submit the diff as a patch? I'd like to review it.
Thanks.
- Beenje - 2009-03-18 08:42
monkeyman Wrote:Could you submit the diff as a patch? I'd like to review it.
Thanks.
Trac ticket created: http://trac.xbmc.org/ticket/6105
- d4rk - 2009-03-18 09:20
The is likely due to the difference in the way ffmpeg's swscale converts YUV to RGB. We instruct it to use PIX_FMT_RGB32, which is endian specific. On PPC, this becomes ARGB instead of BGRA.
Beenje, could you test by replacing PIX_FMT_RGB32 with PIX_FMT_BGR32_1 at line 810 in LinuxRendereGL.cpp, while maintaining GL_UNSIGNED_BYTE in the texture load operations? Thanks.
- Beenje - 2009-03-21 15:20
d4rk Wrote:Beenje, could you test by replacing PIX_FMT_RGB32 with PIX_FMT_BGR32_1 at line 810 in LinuxRendereGL.cpp, while maintaining GL_UNSIGNED_BYTE in the texture load operations? Thanks.
I tried and it works fine. Thanks.
I've proposed a new patch (see http://trac.xbmc.org/ticket/6105).
Have you had a look at the GL extensions available on the PPC Mac mini (http://pastebin.com/m50ca3da9)?
Are we stuck with SW mode?
- davilla - 2009-03-21 23:24
If Macports ever finishes installing, I'll have a dev box up
