Hendrik Ranocha

If we switch to allowing only a length-based version, it will be a breaking change (fine for me, but needs to be handled correctly). Might be worth benchmarking a bit...

Yeah, experimenting with different measures and settling on one of them as default sounds reasonable. If possible, it would of course be nice to have something that's somewhat "dimension independent".

There's no perfect solution for this. Dropping the `module` makes it impossible to change `struct`s defined there for development. However, that's the best solution we have right now, I think....

The issue is that we assume that `solution_variables` will have the same number of variables as the conserved variables when creating a `PlotData2D` object. - Is this reasonable? If so,...

The problem I see with `ScalarPlotData2D` right now is that you need to know the memory layout to generate the scalar data, which is kind of hacky for non-`DGMulti` solvers......

Let's turn this issue into a note on a possibly interesting new feature for our visualization pipeline: It would be nice to allow different numbers of variables when processing the...

We could work on it but it would be breaking. Unless we have a really good reason to release a breaking version right now with this change, I would prefer...

@sloede If you ask because you have some time to spend, please look at #985

We should also add a test using AD with MPI when doing this.

- Which version of Trixi.jl are you using? Did you try to update to the latest version? - Could you please post the complete error message including the full stacktrace?