[Gmsh] Nonsensical physical surface groupings when using opencascade boolean operations.

[Christophe Geuzaine]

> On 28 Jun 2019, at 05:45, jacobp7 <jacobp7 at protonmail.ch> wrote:
> 
> I am trying out the Gmsh cascade boolean operations and have potentially found a bug trying to intersect two offset unit cubes (see test.geo attachment).
> 
> The resulting groupings of the resulting physical surfaces make no sense, as instead of grouping the surfaces according to the input physical surfaces, the output groupings make no sense. Maybe I'm missing a flag or something, but it would be nice to get a surface groupings that strays true to the input.

Gmsh attempts to  preserve the numbering of the geometrical entities specified in the boolean operations (if Geometry.OCCBooleanPreserveNumbering is set) by checking the parent/child relationships across the operations. For a boolean operation on volumes (your case), if a volume gets transformed into a single volume, then we reassign the parent number to the child and the physical group definition can be preserved across the operation. But nothing is done on the lower dimensional entities.

We haven't found how to do actual "feature-based" operations in OpenCASCADE, which could preserve all attributes across operations. All help is welcome in that regard!

Christophe



> <test.geo>_______________________________________________
> gmsh mailing list
> gmsh at onelab.info
> http://onelab.info/mailman/listinfo/gmsh

— 
Prof. Christophe Geuzaine
University of Liege, Electrical Engineering and Computer Science 
http://www.montefiore.ulg.ac.be/~geuzaine