Fix dependency on correct pybind11-abi package and add test to check that the correct abi version is used

[traversaro]

Fix #96
Fix part P1 of #95 (see #95 (comment)).

As now the abi version is no uniquely determined by the pybind11 version, the test on pybind11-abi package just check if the set abi is among the one defined in the header (see AnacondaRecipes#15), while the actual tested is moved to the pybind11 package, where it is indeed checked that the value of the macro PYBIND11_INTERNALS_VERSION actually defined matches the value used in the run_export .

Checklist

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

[conda-forge-webservices]

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.

[traversaro]

@conda-forge-admin, please rerender

[henryiii]

This isn’t true anymore, the latest versions of the last three releases don’t require matching ABI versions anymore.

[henryiii]

Only the platform part needs to match, the internals ABI version does not matter anymore. Though both extensions need to be compiled with the last release of 2.11, 2.12, or 2.13.

[traversaro]

Thanks for the input! Is this related to pybind/pybind11#5296 ?

[henryiii]

Yes, that’s it. We probably should make branches and build 2.11.2 and 2.12.1. And I’m assuming we would like to try to get conda-forge packages to rebuild with the latest pybind11 releases.

[traversaro]

And I’m assuming we would like to try to get conda-forge packages to rebuild with the latest pybind11 releases.

Probably we need anyhow to distinguish somehow packages built before and after pybind/pybind11#5296 . Perhaps we could bump pybind11-abi to some arbitrary number (like 10) to signal that pybind11-abi stopped tracking PYBIND11_INTERNALS_VERSION and is not used instead to distinguish between before (pybind-abi<=4) and after (pybind-abi==10) pybind/pybind11#5296 .

[traversaro]

Let's move the discussion back on the issues, for now this can be closed.