I figured it would be about time to let you all know how I've been going with the updates to the GLU polygon tessellator. Yes, it is good news :) On Tuesday night, Australian time, I tracked down the last of my known bugs that was causing some triangles to be left out of the tessellation. This means the tessellation code is complete and correct enough that I feel comfortable merging it back into CVS when it is back up again. However, in fixing this bug (I was not updating my reflex vertex list correctly, which meant the earity tests were giving strange results) I saw an optimization that could be done that involved re-implementing the reflex vertex linked list as a hash table to allow faster inserts/removals. I'm working on that while the CVS server is down. So, I'm planning to merge my new code in on the weekend. I'm sure I'll make a mess of some of the build config (I'll make sure the makefiles work, and try for Win32 as well), so I appreciate your patience. Any help with this would be great. I've tested the polygon tessellator pretty thoroughly, but I'm less sure of the winding rule stuff (in terms of correctness). If you see behaviour that is unexpected when using multiple contours with the various winding rules, please let me know. I'll make another post when the code is merged. -- Gareth PS - Thanks for the mention at the OpenGL/Linux BOF meeting, Brian :) ============================================== Gareth Hughes mailto:[EMAIL PROTECTED] DEFINITYŽ Site Administration Project Lucent Technologies, Bell Labs Australia ph: +61 2 9352 8608 _______________________________________________ Mesa-dev maillist - [EMAIL PROTECTED] http://lists.mesa3d.org/mailman/listinfo/mesa-dev
