gh-101100: fix sphinx warnings in reference/datamodel.rst

[skirpichev]

• Issue: Fix all Sphinx reference warnings in the documentation #101100

---

📚 Documentation preview 📚: https://cpython-previews--114280.org.readthedocs.build/

[AlexWaygood]

I'd prefer to fix these warnings by creating a single canonical table in the datamodel docs, and removing duplicated documentation elsewhere: see #101100 (comment) (and see PRs where I did similar things for other lists like this in this document: #112781, #112735, #112832, #112933)

[AlexWaygood]

In my opinion, the attributes listed here should only be attributes such as __dict__ and __class__ that appear on all Python objects (I'm aware that's not currently the case, but that's a problem, in my opinion).

We already have a dedicated reference for the __module__ attribute on function objects; I think we should do the same for the __module__ attribute on class objects: https://docs.python.org/3/reference/datamodel.html#function.__module__ (ditto for __doc__)

[AlexWaygood]

Suggestion: you may want to consider separate PRs for tackling warnings relating to module attributes and those relating to class attributes. Fixing the Sphinx nitpicks in the datamodel docs in a principled way is not trivial -- I've been picking away at them for a while, in #112737, #112781, #112735, #112832, #112933, etc. :)

Keeping the PRs small will make it easier for us to review your PRs, and make it easier for you to address our feedback.

I haven't as much time to work on it right now, so it's great some other people are getting involved!

[serhiy-storchaka]

Directive currentclass (similar to currentmodule) would help us.

[skirpichev]

#114336 - is a set/frozenset issue.