MPI: Handle sanity checks during initialization of unstructured boundary conditions differently

[ArseniyKholod]

Fix #1047:

During initialization of the boundary conditions for unstructured meshes (UnstructuredMesh and P4estMesh), there is a check if the boundary names given by the user via a dictionary are valid. This check compares the names given by the user in the dictionary to the names found in the boundary container. While this works fine for serial runs, this check can't be done in parallel runs because the local domain might not contain all boundaries and thus not all valid boundary names are present in the boundary container. In that case the check would fail even if all boundary names given by the user are valid.

[ArseniyKholod]

I cannot see the reasons for checks crashing. Github seem to not work correctly now.

[ranocha]

Yes, GitHub has some problems today...

[codecov]

Codecov Report

Merging #1447 (ac9ae41) into main (f581b9d) will decrease coverage by 0.00%.
The diff coverage is 86.67%.

@@            Coverage Diff             @@
##             main    #1447      +/-   ##
==========================================
- Coverage   95.72%   95.71%   -0.00%     
==========================================
  Files         358      358              
  Lines       29643    29657      +14     
==========================================
+ Hits        28374    28386      +12     
- Misses       1269     1271       +2     

Flag	Coverage Δ
unittests	95.71% <86.67%> (-<0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
...ers/dgsem_unstructured/sort_boundary_conditions.jl	91.11% <86.67%> (-2.44%)	⬇️

[sloede]

Thanks for taking a stab at this! I have left a few comments and suggestions

[sloede]

LGTM! Please also request a review from @lchristm. If he approves as well, this can be merged.

[lchristm]

Looks good to me, thanks!