Rémi Delaporte-Mathurin

I think our problems will all be non-linear, I'm happy to help streamline the newton solver integration into dolfinx if I can! 😄

Next step is to make this work only on a component of a vector function space: ```python def define_interior_eq(mesh, degree, submesh, submesh_to_mesh, value): # Compute map from parent entity to...

@jorgensd your code runs without segfaults on the nightly docker image. I fixed a few things (forgot to set the Dirichlet BC on the first component only) and it now...

This would simplify the parameters logic inside FESTIM.Trap

The tests would greatly benefit from refactoring and tidying up. Mostly I think the directory architecture needs rework and instead of having unit, integration, system etc, we have just files...

Yes that's fine with me!

In this specific case this is fine I think, these two assert statements are effectively testing the same thing.

> For these [tests](https://github.com/festim-dev/FESTIM/blob/fdfcb17465e12ac74bff76668d646aeaa6f2d30a/test/unit/test_helpers.py#L29-L63), can we define `CustomExpr` separately and call it in tests, parametrised with pytest? Yes > Also, I think [these ](https://github.com/festim-dev/FESTIM/blob/fdfcb17465e12ac74bff76668d646aeaa6f2d30a/test/unit/test_sources.py#L51-L93) tests can be parametrised Yes >...

The titles are not super consistent from one derived quantity to another. I think @jhdark has a PR to fix it (#711) that he needs to reopen. Yes, parametrise them...

I think we can close this!