On 1/24/06, Keith Whitwell <[EMAIL PROTECTED]> wrote: > Jerome Glisse wrote: > > Trigger will likely fail on all hardware here is the dump > > of program generated by mesa : > > > > 0: TXP TEMP[0], INPUT[4], texture[0], 2D > > 1: TXP TEMP[1], INPUT[5], texture[1], 2D > > 2: MUL_SAT TEMP[2].xyz, TEMP[0], INPUT[1]; > > 3: MOV_SAT TEMP[2].w, INPUT[1]; > > 4: ADD, TEMP[1], TEMP[2]; > > 5: SUB OUTPUT[0].xyz, !unkown![255].xxxx, TEMP[0].xxxx; > > 6: MUL OUTPUT[0].w, TEMP[1], TEMP[2]; > > 7: END; > > > > Line 5 !unkown! shouldn't src/mesa/shader/program.c fail when > > it encounter !unkown! ? Keith do you have an idea on that ? > > You can try enabling the disassembly in texenvprogram.c to doublecheck > the above. > > My guess is that the unknown value should really be TEMP[1], which looks > like the result of the first unit's texture combine. I don't know what > swizzle it should have. > > It would be useful to know what combine state is being requested - that > should narrow things down a fair bit. Can you try and find that out?
Found it (thx to Aet :)): in emit_combine texenvprogram.c case MODE_ADD_SIGNED use tmp without asking for one. I will commit a patch latter, checking if there is no other case like that. best, Jerome Glisse ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid3432&bid#0486&dat1642 -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
