GitLab

#1053 Test vertex matching: introduce a post processing which loads all the dofs information. Final step will be to check we find the same interface vertex content before and after the interpolation.

#1053 Start filling the test vertex matching. I introduced a new unknown named 'chaos' which is there only to ensure we don't get identity in parallel: if we considered only my two unknowns with same shape functions we would actually get he exact same dof repartition on both meshes, thus potentially hiding a bug or an issue.

#1022 Interpolation holder: add interpolation fluid velocity on solid -> fluid monolithic. It doesn't work at the moment, due to the fact fluid velocity is P1b and interpolator used (FromVertexMatching) works currently only for P1.

#1022 Poro: change a matrix into which operator is assembled. Runtime still fails, but the conversion into the monolithic has not yet bee reformed, so it's not much of a surprise.

#1050 Improve ScalarParameterFromFile by disabling entirely the time dependency of the internal parameter, to avoid any ambiguity.

#1049 - #1050 Fix ScalarParameterFromFile and Parameter/TimeDependency test (the latter wasn't working even prior to the ticket #1049, hence the associated #1050 in this commit).

#1049 Fix active stress in SecondPiolaKirchhoffStressTensor operator and cardiac mechanics pre-stress model.

#1049 Actually the mesh argument in Parameter is superfluous; remove it as Domain is much more sensible. Same remark as previous commit: only hyperelastic case is guaranteed to compile and run.

#1049 [DOES NOT COMPILE EVERYTHING] Add a Domain constructor argument in parameter. Currently only hyperelastic model (and its 2D demo Lua file) are guaranteed to work with this new convention.