Will Pazner
Will Pazner
Yes, the arrays would have to be made mutable. I don't have a strong preference one way or the other.
Yes, that sounds fine to me. `BuildDofToBdrArrays` would be made private, `BuildDofToArrays` remains public but deprecated, and an additional "internal" `BuildDofToArrays_` would be added (called by `GetElementForDof` and `GetLocalDofForDof`).
Thanks for the fix. I would suggest the following changes (to simplify the test case, and also add a surface mesh to the existing test case): ```diff diff --git a/tests/unit/mesh/test_mesh.cpp...
@tzanio, I can address the review comments on this PR if you want. It would be a good fix to have.
@hughcars, does e9ca0960f74b2e69d3f4026a832358bbe97142ee look OK to you? (Related to #4857).
The most efficient way to do this is probably with `QuadratureFunctionCoefficient`. You can create the `QuadratureFunctionCoefficient` ```cpp QuadratureSpace qs(mesh, ir); QuadratureFunction qf(qs); QuadratureFunctionCoefficient jh_coeff(qf); ``` Note that the integration rule...
Ah yes, good point. H(curl) is not yet supported by `QuadratureInterpolator`. You might be able to get a bit better performance by using `GetVectorValues` on each element, but the difference...
Good point @rcarson3, addressed in 59a95cc8e6a689eca49d6df5940ccf260d864c40 and 63d983ed6a365520bcf68448065aec2f453109d1.
@tarikdzanic, can you take a look when you get a chance? Thanks!
As you noted, this is because MFEM does not currently support gmsh v4 mesh files. (You can still output v2.2 format files with the latest version of gmsh). The other...