Note 1: there are cases in which the previous algorithm may have produced 'better' triangles, but my big hope is that poly2tri's CDT implementation will implement more advanced refinement algorithms over time.
Note 2: This issue http://code.google.com/p/poly2tri/issues/detail?id=34 is relevant, first versions of my poly2tri embedding would indeed stackoverflow or assert. I somehow avoided this by using Clipper as prepass and scaling the entire polygon to 0..1 range (as recommended).
git-svn-id: https://assimp.svn.sourceforge.net/svnroot/assimp/trunk@1118 67173fc5-114c-0410-ac8e-9d2fd5bffc1f
- IFC: rate all available representations by a simple ranking system and take the one that is easiest to load. This should avoid loading the same geometry twice. Also it might result in quality improvements when BRep geometry is avoided in favour of extrusion geometry.
# IFC: Various bugfixes related to geometry loading.
git-svn-id: https://assimp.svn.sourceforge.net/svnroot/assimp/trunk@1033 67173fc5-114c-0410-ac8e-9d2fd5bffc1f
aramis_acg
klickverbot