I'm running the latest CVS Mesa with the latest CVS glx in software mode and XFree86 3.3.4 under Linux. The application is Dive 3.3x, binaries available at http://www.sics.se/dive/distr/v3.3x/. I get a segmentation fault in fixup_first_4v. The problem is that the flag array is all 0s, so the loop starts going outside the array: 367 while ((flag[++i]&match) == 0) 368 COPY_4FV(data[i], dflt); Here's a backtrace: #0 0x40300d71 in fixup_first_4v (data=0x8414738, flag=0x83fa158, match=8388608, start=3, dflt=0x8414768) at vbxform.c:368 #1 0x4030160c in gl_fixup_input (ctx=0x838ff00, IM=0x83f9910) at vbxform.c:552 #2 0x401e1b0d in gl_cva_compile_cassette (ctx=0x838ff00) at cva.c:815 #3 0x402ff13b in gl_maybe_transform_vb (IM=0x83daaa0) at vbxform.c:76 #4 0x402ff19f in gl_flush_vb (ctx=0x838ff00, where=0x4032a72c "dlist") at vbxform.c:89 #5 0x401e8ac4 in save_ShadeModel (ctx=0x838ff00, mode=GL_SMOOTH) at dlist.c:1985 #6 0x401a842d in glShadeModel (mode=GL_SMOOTH) at api2.c:637 #7 0x4017655e in GLXDecodeShadeModel (length=4, data=0x83d65a4 "\001\035") at glxdecode.c:5672 #8 0x40167da0 in GLRender (client=0x836bbd0) at glxcmds.c:121 #9 0x401889de in ServerMesaDispatch (client=0x836bbd0) at glxmesa.c:330 #10 0x8175c45 in Dispatch () #11 0x8183e20 in main () #12 0x4005fcb3 in __libc_start_main (main=0x81839e0 <main>, argc=1, argv=0xbffff9e4, init=0x807f354 <_init>, fini=0x82b690c <_fini>, rtld_fini=0x4000a350 <_dl_fini>, stack_end=0xbffff9dc) at ../sysdeps/generic/libc-start.c:78 (gdb) up #1 0x4030160c in gl_fixup_input (ctx=0x838ff00, IM=0x83f9910) at vbxform.c:552 552 fixup_first_4v( IM->TexCoord[0], IM->Flag, 0, start, (gdb) print IM->Flag $2 = {0 <repeats 507 times>} (gdb) print IM->Count $3 = 483 I'm afraid I don't know the code well enough to know where to continue hunting this bug, but I'd be glad to give a hand in case there's anything I can do. Thanks! _______________________________________________ Mesa-dev maillist - [EMAIL PROTECTED] http://lists.mesa3d.org/mailman/listinfo/mesa-dev
