use importlib_metadata instead of importlib.metadata for python<3.10

[mboisson]

Thanks for contributing, make sure you address all the checklists (for details on how see

development documentation)!

This PR is pretty minimal and is meant to address #2303
importlib.metadata does not provide unique results for entry_points when duplicate packages are found until python >= 3.10. To allow local installation of virtualenv to superseed global installations, we need to use importlib_metadata instead.

• ran the linter to address style issues (tox -e fix_lint)
• wrote descriptive pull request text
• ensured there are test(s) validating the fix
• added news fragment in docs/changelog folder
• updated/extended the documentation

[gaborbernat]

Can we find out why using the native importlib.metadata does not work?

[mboisson]

Can we find out why using the native importlib.metadata does not work?

That is explained in issue #2303, but here is a extract from the issue:

I believe this issue is caused by the switch from importlib_metadata to importlib.metadata in PR #2246

In importlib_metadata's version of entry_points, there is a step to make the imported entry points unique:
https://github.com/python/importlib_metadata/blob/main/importlib_metadata/__init__.py#L995

This ensures that only the highest priority entry_points.txt from virtualenv*.dist_info is considered.

This is also the case in importlib.metadata's latest version:
https://github.com/python/cpython/blob/main/Lib/importlib/metadata/__init__.py

but it was not the case in:
Python 3.8: https://github.com/python/cpython/blob/3.8/Lib/importlib/metadata.py#L541
Python 3.9: https://github.com/python/cpython/blob/3.9/Lib/importlib/metadata.py#L569
It was not until Python 3.10 that this was taken up:
https://github.com/python/cpython/blob/3.10/Lib/importlib/metadata/__init__.py#L978

I believe importlib_metadata should be used for Python < 3.10.

[gaborbernat]

Closing per #2303 (comment) Please consider using instead https://pypa.github.io/pipx via https://pypi.org/project/pipx-in-pipx/