add a build variant that links Kwant against mumps-mpi

[basnijholt]

Checklist

• [x] Used a personal fork of the feedstock to propose changes
• [x] Bumped the build number (if the version is unchanged)
• [x] Reset the build number to 0 (if the version changed)
• [x] Re-rendered with the latest conda-smithy (Use the phrase code>@<space/conda-forge-admin, please rerender in a comment in this PR for automated rerendering)
• [x] Ensured the license file is being packaged.

[conda-forge-linter]

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) and found it was in an excellent condition.

[basnijholt]

@conda-forge-admin, please rerender

[jbweston]

@conda-forge-admin, please rerender

[jbweston]

All the tests now pass, but when I manually install kwant and fenics together in the same environment fenics gives all sorts of warnings on import about integer size mismatches.

[akhmerov]

Do you think these warnings indicate that things are broken?

[jbweston]

Do you think these warnings indicate that things are broken?

Yes, but I don't know what it is exactly. I suspect that it is due to a difference in compile-time and run-time versions of MPI.

[basnijholt]

Great work Joe!

I say, let's go ahead and merge 👍

[akhmerov]

A quick question. If a user writes mamba install kwant in a fresh env, which variant is installed? Can we select a default?

[basnijholt]

The default will be the nompi build, this is because we set the build number higher, so the default stays the same.

If someone prefers the mpi version, one should specify: conda install "kwant=*=*mpi".

[akhmerov]

Ah, I see that now. Neat! Please add a comment around that line to explain the purpose of the build number bump.

Given that I'm :+1: on merging, but I'd like @jbweston's confirmation.

[jbweston]

We can't merge because things are still broken!

For clarity here's what happens when I import fenics in an environment where the mpi variant of kwant is installed:

@basnijholt also, as our variants are named nompi and mpi I think this means if we want the mpi variant we have to specify kwant=*=*_mpi with a weird leading underscore. This is because kwant=*=*mpi will also match nompi.

[basnijholt]

@jbweston, this seems unrelated to the changes made here.

Try installing the current version of kwant in Python 3.9 to get the same warnings.

[conda-forge-linter]

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

I wanted to let you know that I linted all conda-recipes in your PR (recipe) and found some lint.

Here's what I've got...

For recipe:

• Failed to even lint the recipe, probably because of a conda-smithy bug :cry:. This likely indicates a problem in your meta.yaml, though. To get a traceback to help figure out what's going on, install conda-smithy and run conda smithy recipe-lint . from the recipe directory.

[conda-forge-linter]

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) and found it was in an excellent condition.

[basnijholt]

@conda-forge-admin, please rerender

[jbweston]

@jbweston, this seems unrelated to the changes made here.

Try installing the current version of kwant in Python 3.9 to get the same warnings.

Indeed, conda create -n fenics-alone -c conda-forge fenics; conda activate fenics-alone; python -c 'import fenics' gives these warnings, so it's not us.

[github-actions[bot]]

Hi! This is the friendly automated conda-forge-webservice. I tried to rerender for you, but it looks like there was nothing to do.

[github-actions[bot]]

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

• linter: passed
• azure: passed

Thus the PR was passing and merged! Have a great day!