On 13.08.2009 12:19, Michel Dänzer wrote: > On Wed, 2009-08-12 at 11:31 -0700, Eric Anholt wrote: >> Module: Mesa Branch: master Commit: >> e643bc5fc7afb563028f5a089ca5e38172af41a8 URL: >> http://cgit.freedesktop.org/mesa/mesa/commit/?id=e643bc5fc7afb563028f5a089ca5e38172af41a8 >> >> >> Author: Eric Anholt <e...@anholt.net> Date: Tue Aug 11 12:59:09 >> 2009 -0700 >> >> i965: Use _MaxElement instead of index-calculated min/max for VBO >> bounds. > > This change breaks things all over the place here. E.g. > progs/glsl/array and .../skinning are missing most of the geometry, > and a lot of the other glsl progs have weird lighting. The problem here is for vertex buffer elements which have zero stride count is 1. This is now used as bounds check, which will hence fail for most indices - according to docs hardware will simply return 0 in this case (and not the data at start index or something like that which would work). I'll fix this (should be ok to just disable bounds check for these cases as in fact any index is valid if stride is 0). (Looks like this isn't an issue for IGDNG (Ironlake I assume) since it appears like this one uses address range instead of maximum index to check against according to the code...).
Roland ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Mesa3d-dev mailing list Mesa3d-dev@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/mesa3d-dev