Following up to myself...
On Tue, 2002-05-21 at 19:55, Michel D�nzer wrote: > > Patch against the trunk is at > http://penguinppc.org/~daenzer/DRI/radeon-ppc.diff > > Requests for comments before I commit: [...] > - Is this a good way to detect endianness? And there's probably a better > place for the BE_SWAP32 definitions? Probably extras/Mesa/src/macros.h? > The last outstanding arch-dependent issue is broken fog, as seen in > http://penguinppc.org/~daenzer/DRI/csmash.jpeg (the fog is supposed to > be almost invisible). I'd appreciate any pointers where to look for > this. Fixed. I've updated the patch, unless someone comes up with an argument against it I'll commit it soon. > I thought this was related to another issue, but it turns out that one > shows up on x86 as well: Wrong shading/lighting of (some) clipped > objects, as seen in http://penguinppc.org/~daenzer/DRI/fsv.jpeg (the > left box is supposed to be the same color as the right one). This is > fixed in the current r128 driver (but it draws lines which should be > invisible), the Mesa 3.4 based one was also broken. On the other hand, > http://penguinppc.org/~daenzer/DRI/gears.jpeg is still present in r128 > as well. > > Another arch-independent issue I found is that sometimes parts of > objects aren't rendered (correctly), as seen in > http://penguinppc.org/~daenzer/DRI/morph3d.jpeg and > http://penguinppc.org/~daenzer/DRI/armagetron.jpeg (both trails are > supposed to look like the right part of the red one). Dunno if it's > related but I saw a similar, if much worse effect until I fixed > EMIT_TWO_ELTS in lib/GL/mesa/src/drv/radeon/radeon_render.c. > > Last, but not least, there's an off-by-one error (?) with textures, as > seen in http://penguinppc.org/~daenzer/DRI/evas_test.jpeg and > http://penguinppc.org/~daenzer/DRI/celestia.jpeg . Surely others must have run across these as well? Any ideas? > I'm going to try the TCL branch to see if those are still there. Well, glxgears hangs with that both on an Athlon box at work with a VE and on this TiBook with an M6. And on a Cube with a QD (which actually has a TCL unit), the X server doesn't even start. Looks like there's more porting work to do. :/ -- Earthling Michel D�nzer (MrCooper)/ Debian GNU/Linux (powerpc) developer XFree86 and DRI project member / CS student, Free Software enthusiast _______________________________________________________________ Don't miss the 2002 Sprint PCS Application Developer's Conference August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
