You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Much of Ferrite's code is only tested for grids from the generate_grid function, but @koehlerson and I discussed that it would be valuable to have some grids available for testing that unstructured, stemming from e.g. Gmsh and Abaqus.
We recently found some code, not yet in Ferrite, that worked well on grids from generate_grid, but not for unstructured grids.
One suggestion would be to have a list of different types of relatively small grids for use with testing, both with the same cell type and mixed celltypes. One possibility could be to define something like
Maybe we can just dump some problematic unstructured grids in a test assets folder or hand-generate them tailored for specific corner cases to cover? Otherwise we might need to start thinking about randomization of the tests, if we want to test all combinations of every type of mesh parser+every type of element combination+every type of unstructuredness.
Much of Ferrite's code is only tested for grids from the
generate_grid
function, but @koehlerson and I discussed that it would be valuable to have some grids available for testing that unstructured, stemming from e.g. Gmsh and Abaqus.We recently found some code, not yet in Ferrite, that worked well on grids from
generate_grid
, but not for unstructured grids.One suggestion would be to have a list of different types of relatively small grids for use with testing, both with the same cell type and mixed celltypes. One possibility could be to define something like
just to make sure to cover some special cases during testing.
This would also serve as "reverse-ci" (cf. #797) for the mesh parsing.
Thoughts?
The text was updated successfully, but these errors were encountered: