further marching cubes work
- separating sculpting data from marching cubes view through a sampling interface
- pruning pass after initial generation to merge duplicate vertex data (21k verts to 3k verts in current test setup)
- some vertex generation changes that would let us interpret non-boolean datasets better