scivision

Another thought is Catch2 is preparing a v3 release. The API has some changes, so maybe it's worth waiting for Catch2 v3?

I'm not sure what the CI failure here means. I have a CI workflow that passes with my h5fortran fpm.toml using fpm 0.5.0.

I didn't try the changes myself, but I agree with all these changes. Nice work! Note: before incorporating these changes into P4EST, etc. such parent projects need to have their...

There was conflicting CMake variable use, fixed by this patch [178.patch](https://github.com/cburstedde/libsc/files/15381488/178.patch) Note that CMAKE_BUILD_TYPE must be set before `project()` or it has no effect.

I commented on this in the xarray branch first, but just to summarize--I have had severe issues with the non-threadsafe of f2py's Fortran modules, such that multiple instances (say via...

It should be easy to make milestones work with NetCDF4 as well. Gemini allows one to load HDF5 and save NetCDF or vice versa. The main practical limitation I've seen...

I updated the linking and package and [added an HWM14 CI case](https://github.com/gemini3d/gemini3d/commit/8d553f544d87cc6743f4416937b6abed72f38ae0) I think HWM14 should build OK now, please reopen issue if not.

Paul I've repackaged how HWM14 works since then. This is no longer an issue when not using HWM14. I do need to reimplement how HWM14 works when it is used.

This is something to look at before Dec 16 meeting if possible. Or at least see if it's an issue as this is something Steve is interested in using.

there are procedures with that pattern in a lot of places; do you have particular files in mind?