Dear reviewers, I did create neither an issue on BPO nor a news entry as the fix is IMO a trivial one.

cc: @jaraco

Thanks for this. I agree the changes are beneficial. I need to figure out how these changes reconcile with the back port. I don’t think these Sphinx roles exist in the Sphinx environment in which the backport docs render. It might be time, however, to defer all significant docs to CPython (and remove them from the backport).

Is that something you would like to help with?

@jaraco Am I right about this being the corresponding backport's doc? I have tried to apply the patch from this PR and besides referring to a new label I have introduced in reference/import.rst , the rest of the changes applies fine.

Is that something you would like to help with?

Sure, I will gladly prepare a separate PR for the backport docs.

Thanks Oleg. Yes, that's the doc. I'd have provided a link but was on my mobile at the time.

I think the main problem you're going to find is that the Sphinx roles aren't configured for the docs builds on the backport. Perhaps they should be.

The good news is that importlib_metadata does run the docs as part of the CI, so if the GitLab merge request is passing, that should be sufficient.

@jaraco The pipeline looks good, so I've landed the PR. It has more changes than this one because I also went through the index.rst, so please proceed with caution :)

Thanks for the extra effort required of the backport!

@jaraco gladly, and thank you for a fast response and helpful review!