momchil-flex

I think the only time a feature PR should be explicitly moving the submodule state is if there really is an associated notebook, and even then it might be easier...

I think you may have hit a case which is strongly affected by the main source of inaccuracy in GME: the incompleteness of the basis. Specifically, we only use *fully*...

I think the issue might then only arise if someone puts a TFSF in a 1D simulation, i.e. simulation size 0 along both of the TFSF tangential directions.

> Oh yeah, that was a doctest change. Nice, we should move it to `scripts/` then. I can include this in the docs PR I'm trying to finish polishing atm....

I can leave this open for future improvements, but I think nothing else is planned for now?