Kyle Mandli

So for in-place installations we would create a symlink down into the package or just get rid of the top-level directory all-together?

+1 on removing the read functionality from solution. I would say that we should keep the ability to initialize a `solution` object by providing a path and frame, in effect...

Huh, was their a reason we did not do this in the past beyond it being inefficient?

I was suggested to take a look at https://github.com/conda-forge/mpich-feedstock/issues/55

@gradylemoine I don't suppose you could look at this and confirm? You have more recent experience with this code and could more quickly confirm the bug and bug fix.

Thanks Grady for getting back so quickly. After looking at this myself, I think that the current code is correct. Looking at the update formulas I think that the capacity...

That's a great question. I took a look through [1] and [2] but could not find a definitive answer. 1. LeVeque, R. J. Wave Propagation Algorithms for Multidimensional Hyperbolic Systems....

I would favor them being identical but also long. We should probably explicitly format the Python and make the Fortran longer then?

This would be a good issue to raise generally but label also as a good "first issue" as I don't think this would be difficult to do with some guidance.

I got it to work with 5.7, what seg fault were you seeing?