Add python_abi to run_constrained (for 3.6)

[jakirkham]

Checklist

• Used a 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-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.

[isuruf]

While I agree with this change, this wouldn't fix your issue.

[jakirkham]

Hmm...ok. What should we be doing instead? 🙂

[isuruf]

It's a bug/feature of conda solver. We should update this to 3.6.10. Can you click the allow edits from maintainers button?

[jakirkham]

Ack! Ok that makes sense. Should we be hot-fixes old python packages then?

I would, but GitHub doesn't show it 😞

[jakirkham]

Are there other changes we need?

[isuruf]

We need https://github.com/conda-forge/python-feedstock/blob/3.7/recipe/patches/0020-Use-ranlib-from-env-if-env-variable-is-set.patch backported to avoid a bug introduced between 3.6.7 and 3.6.10

[mbargull]

It's a bug/feature of conda solver.

Can you elaborate and/or point me to CI logs?

[isuruf]

@mbargull, pypy python version was 3.6.9 and cpython version was 3.6.7. In current_repodata.json, only pypy 3.6.9 was there. This means that conda install python=3.6 was resolved with the pypy python version 3.6.9 even though it has a track_features entry.

[mbargull]

Ah, that current_repodata.json thing... Makes sense, thank you!

[jakirkham]

How do we avoid this issue cropping up in the future?

[jakirkham]

Handled in PR ( #323 ).