Closed riaqn closed 6 years ago
Are you simply reporting a documentation error or are you really interested in having the code working on points from an exact construction kernel? In the latter case, what do you expect to get compared to using floating point?
Hi @sloriot thanks for the reply.
I don't know - I was using it to contract a mesh(for only once) and while the source mesh is free of self-intersection, the meso_skeleton() is not. So I assume it might have to do with the exact-ness.
The meso-skeleton is by construction not guaranteed to be free from self-intersect. The only guarantee you have is that it has the same topology as the input mesh.
I see. I know it's totally off-the-topic, but do you know if there is anything I can do to achieve the same goal? namely, contract a mesh such that it contracts less on vertices with higher mean curvature, and vice versa, and keep the output mesh free of self-intersection.
I don't think we have anything like that in CGAL for now.
Issue Details
Currently there are some uses of double in the code of Mesh_skeletonization, making it impossible to use kernel other than inexact ones.
Environment