Cristian Le

> Interesting part is I dont libpython.so anywhere here, should this be linking with python .so? You should find it when you run the `readelf` or equivalent on the `python...

Well, at this point I have no more ideas. Best to take it up with your cluster provider and/or whoever configured the python environment there.

🤷 wouldn't know how to read it, but it still feels like its an issue with your python environment. You could still build python from source

> [@LecrisUT](https://github.com/LecrisUT) is it possible to specify the location of the libpython librariy header files manually in cmake 3.30 Yes, you have many options[^1], the most basic of them is...

If it is an issue with `dnf install`ed packages than yes raise it with IBM, they need to know about this. Can you run `dnf info python3.12` to check which...

Ok, can you make one more issue at https://gitlab.kitware.com/cmake/cmake and cross-link with the IBM issue (both ways). CMake wise, you are building it as a `MODULE` through ```cmake python_add_library(MODULE) ```...

Ok, basic functionality is in. Should we refine the API here, or do that in a follow-up addressing other suggestions like #924. There is definitely some duplication of the API...

> Since the implementation is only for GitHub now, can you please update COMPATIBILITY.md accordingly? Yeah, would want to nail down more what we want to do with this interface...

Extracted from [comment](https://github.com/teemtee/fmf/pull/271#discussion_r1905352732) with @psss: Rename `Module` to something more standard, `Code`/`API Reference`/`Python (API) Reference`

I believe we had this issue before especially with `pip` and/or conda. There are a few configurations you can try: - Install all `build-system.requires` dependencies locally and try with `pip...