Mesh boundary flags are not set correctly in (possibly all?) submitted mesh files
Consider a 2D mesh file where:
- all faces have flag 0
- all inner edges have flag 0
- all boundary edges have flag 1
In most (all?) mesh files that are currently online, the mesh flags of the vertices are set to 0 and then, in the code, forced to either 1 or 2 depending on the neighboring edges (Dirichlet wins over Neumann).
With the new approach, the IDs do not anymore refer to a certain BC. Thus it is theoretically wrong to assume that 0 == Inner, 1 == Dirichlet etc.
I will (for now) solve this problem by keeping this priority. But we need to refactor all mesh files that are affected at some point in time.
A possible solution would be to write a GMSH serialization routine so that we cann modify the mesh via tinyhhg and output the improved version.
Edited by Nils Kohl