I am really, really new to G+Smo. I am sorry if I say something silly. :-)

The method gsSurface<T>::toMesh() creates a gsGridIterator for the surface domain to addVertex to the mesh. Then it computes faces.

If after calling toMesh(), I want to compute the normals for each vertex. Of course, I could simply reproduce the internals of toMesh, and construct the same gsGridIterator and compute the mesh at the iterated points. However, I do not feel that depending on the internals of toMesh() is a good thing.

I would like to know (or propose) an elegant solution to this problem. Maybe toMesh() could return a copy of the iterator...

What do you think? Am I missing anything?

Reference:

PS: It is interesting that we do not want to use the four parameter version of addFace inside the double loop.