fix for stub generation

[rykerfish]

Checklist

• Used a personal fork of the feedstock to propose changes
• Bumped the build number (if the version is unchanged)
• Ensured the license file is being packaged.

@RaulPPelaez I think this (finally) did it, but I'm still a tad confused as to why. The fix is to explicitly link both .so files against cblas, otherwise the stubgen call fails with undefined symbol: cblas_sgemv when importing a solver. I created the conda package locally with the build-locally.py script and was able to verify that the necessary .pyi files get created by stubgen. You can also see in the Azure dev ops logs that the errors where stubgen were crashing (around line 1400) are no longer there.

Should we avoid doing it this way since it wasn't something we had to link against before, or we were linking incorrectly before and it just happened to be satisfied by a closely related BLAS package?

[conda-forge-admin]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe/meta.yaml) and found it was in an excellent condition.

I do have some suggestions for making it better though...

For recipe/meta.yaml:

• ℹ️ The recipe is not parsable by parser conda-recipe-manager. The recipe can only be automatically migrated to the new v1 format if it is parseable by conda-recipe-manager.

_{This message was generated by GitHub Actions workflow run https://github.com/conda-forge/conda-forge-webservices/actions/runs/22115716483. Examine the logs at this URL for more detail.}

[RaulPPelaez]

I think this is fine, but if stochasticHydroTools/libMobility#73 includes this patch I would make a release and let the conda-forge infra put it here. Otherwise you have to do extra work to remove the patch in the next release.

[rykerfish]

Great- I'll close this and release via libmobility release once we have the CUDA 13 version working in there as well. Thanks!