> So, is it the existence of degenerate triangles that is the cause of the
> 'bug'?

	Actually -unfortunately- not.
	When adding an "offset 1e-10" to the tesselate{} block as
mentionned earlier (but with a much lower value), the holes in the
mesh disappear while there are still a lot of "degenerate" triangles
(that are not rendered anyway).
	So the bug is more likely a problem of accuracy in the
tesselation calc rather than in actual mesh construction (where the
code is the same as in Parse_Mesh, but reorganized in a more readable
style by Warp). Could it be a problem of finding the Intersection in
very particular cases ? Didn't find it yet...


*** Nicolas Calimet
*** http://pov4grasp.free.fr

Post a reply to this message