PMP: todo: Fix slow compute_vertex_normal()
Summary of Changes
For the testcase provided in this PR compute_vertex_normal() is particularly slow. A vertex has a fan of incident faces with identical normal. As the first fix tests with operator==(Vector_3,Vector_3) it is still slow when the object is slightly rotated or perturbed. With a bound on the angle between normals it is fixed, but is that acceptable? For double probably. Do we have to specialize for exact number types?
Release Management
- Affected package(s): PMP
- License and copyright ownership: GF
where to learn the basics for contributing
Successfully tested in CGAL-6.1-Ic-189
If I remember correctly there is still something todo by @LeoValque .
This pull-request was previously marked with the label Tested, but has been modified with new commits. That label has been removed.
Successfully tested in CGAL-6.1-Ic-208