Sir,
I have been using your program to cross check my work in 3D subdivision and I believe I may have found a bug in your implementation of the linear subdivision module.
I am using a 12 sided di-pyramid (attached as Two Pyramids.OFF) and using the
Filters->Remeshing, Simplification and Restructing->Subdivision Surfaces : Midpoint tool.
Using this it appears that the first four iterations complete successfully but the fifth iteration generates two distinct sets of problems.
One of the sides appears to be creating vertices with either valence 8 or valence 4 instead of uniform valence 6.
The second side appears to have issues with the recreation of a symmetrical mesh
I have recreated the bug on two systems both using MeshLab_64Bit v1.3.3 (Apr 3 2014). I have attached the exported OFF files for examination and also attached a set of OFF files which I have generated which do not seem to exhibit the same problems.