Open Pro-Ly opened 1 month ago
Thanks; I'm traveling, so it will take a while for me to check this out in depth. This part of the pair tester involves some extremely subtle details if I remember correctly. I'm going to have to stare at it for a while. If you have a reliable reproduction case for the bug (e.g. a convexhull and a set of poses which shows the failure), that'd help me a lot.
There is a bug in convexhull tester: when one edge from shape B's face perform clipping with shape A'face find two edges parallel, it will override the right result to invalid.
In fact. I believe that there is no need to do anything with parallel edge cases. They will be filtered in the later "if (earliestExit >= latestEntry)" judgements.