I'm systematically triggering this assertion failure with a simple test drawing textured quads using VBOs. The code looks like: glEnable(GL_VERTEX_ARRAY); glEnable(GL_TEXTURE_COORD_ARRAY); glEnable(tex->target); if (quad->vbo == 0) { Vertex vertices[4]; pf_glGenBuffersARB(1, &quad->vbo); pf_glBindBufferARB(GL_ARRAY_BUFFER, quad->vbo); vertices[0].x = quad->x; vertices[0].y = quad->y; vertices[0].tex_x = 0.0f; vertices[0].tex_y = 0.0f; vertices[1].x = quad->x + quad->width; vertices[1].y = quad->y; vertices[1].tex_x = tex->max_x; vertices[1].tex_y = 0.0f; vertices[2].x = quad->x + quad->width; vertices[2].y = quad->y + quad->height; vertices[2].tex_x = tex->max_x; vertices[2].tex_y = tex->max_y; vertices[3].x = quad->x; vertices[3].y = quad->y + quad->height; vertices[3].tex_x = 0.0f; vertices[3].tex_y = tex->max_y; pf_glBufferDataARB(GL_ARRAY_BUFFER, sizeof(vertices), vertices, GL_STATIC_DRAW_ARB); } else { pf_glBindBufferARB(GL_ARRAY_BUFFER, quad->vbo); } glBindTexture(tex->target, tex->handle); glVertexPointer(2, GL_FLOAT, sizeof(Vertex), (const GLvoid *) 0); glTexCoordPointer(2, GL_FLOAT, sizeof(Vertex), (const GLvoid *) (2 * sizeof(GLfloat))); glDrawArrays(GL_QUADS, 0, 4); This is on a debian unstable running the intel 2.6, linux 2.6.28, running UXA and DRI2. mesa is 154a9e5317f890618932cea0129ef887e16baf84.
(gdb) bt #0 0xb805a7f2 in _dl_sysinfo_int80 () from /lib/ld-linux.so.2 #1 0xb7c7f640 in raise () from /lib/i686/cmov/libc.so.6 #2 0xb7c81018 in abort () from /lib/i686/cmov/libc.so.6 #3 0xb7c785be in __assert_fail () from /lib/i686/cmov/libc.so.6 #4 0xb79e5d92 in _tnl_draw_prims (ctx=0x95eb938, arrays=0x961dee8, prim=0xbfd74940, nr_prims=1, ib=0x0, min_index=0, max_index=3) at tnl/t_draw.c:203 #5 0xb79dd5f7 in vbo_exec_DrawArrays (mode=7, start=0, count=4) at vbo/vbo_exec_array.c:267 #6 0xb79d0878 in neutral_DrawArrays (mode=7, start=0, count=4) at main/vtxfmt_tmp.h:328 #7 0x080499c2 in big_benchmark_draw_quad_with_vbo (quad=0xbfd74b7c) at benchmarks/benchmark.c:213 #8 0x0804932b in draw_scene (scene_data=0xbfd74b7c) at benchmarks/some-quads-vbo.c:34
Could you possibly create a small glut test program? Perhaps modify mesa/progs/tests/bufferobj.c? Which driver are you using?
Created attachment 33234 [details] flybye winedbg output broken again on i945gm with libdrm git, mesa git, intel-xorg git and kernel 2.6.32 (ubuntu karmic + xorg-edgers + ppa-kernel) when using wined3d (eg: http://pouet.net/prod.php?which=4484) (but the messed up gun in sauerbraten with enabled shaders is now fixed, maybe this is related to this bug) i also tried mesa 7.7 and 7.6, now there it is broken, too attached winedbg output
Created attachment 118216 [details] a version that can be compiled - may be need updates I used this compiled version of proposal defect. The assertion is gone so I propose to close this defect.
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.