GitLab

Fixed Documentation stating unsigned int instead of size_t for n_entity() functions



See merge request !59

Some fixes for min gw (Thanks to Xan for the patch)

@xan: Your patches are pending for inclusion after test builds

See merge request !58

Copy per-halfedge and per-face properties to new elments during triangulation

This is an attempt to fix an issue we encountered when loading a textured quad mesh and triangulating it:
In the triangulated version, the respective values of the per-halfedge texcoords and face texture index properties  were not assigned to the new triangles that were created during triangulation.

See merge request !57

Fix internal compiler error on VS2015



See merge request !56

(cherry picked from commit dd91a7ea)

don't emit an error message for every face that could not be added (due to complex elements)



See merge request !55

- Always emit warning

Fix double swap



See merge request !54

add factory functions for creating PropertyManagers without restating the mesh type in C++11



See merge request !53

Not collecting failed faces but directly processing them

**Problem description:**

I encountered severe problems in my application because OpenMesh does not load attributes, such as face texture index, for failed faces (non-manifold geometry).
Instead of directly handling them, the Importer class defers the handling of failed faces to a post process - but at that point, the attribute information is not available any more, since it is read and applied in the reader, which only deals with the non-failed faces.


**Solution:**

My simple solution was to remove the list of failed faces and instead directly handle failed faces inside the "add_face" function of the Importer class.


**Discussion:**

I didn't see any drawback, compared to the previous code. It would be interesting to know why the list of failed faces was maintained - was this legacy code, or was there some idea to extend this concept, which was just  never realized? The only functional point that has changed is that I removed the code that adds degenerate faces to the mesh (line 129), since, in my opinion, this makes things just more complicated for some applications, and the benefit of being able to import degenerate faces seems marginal. But that's a design decision, it will be very easy to encapsulate the code that adds a separate face (line 140) into a separate function and use it also for degenerate faces.

It would be great to receive your feedback on this issue, and to know if it is possible to include this change (or something similar).

See merge request !51

Added check for property iterators



See merge request !52