GH-101100: Resolve reference warnings in library/stdtypes.rst

[AA-Turner]

---

📚 Documentation preview 📚: https://cpython-previews--138420.org.readthedocs.build/

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

[serhiy-storchaka]

I think that the right fix for this is to add anchors instead of removing references.

[AA-Turner]

@serhiy-storchaka I've generally been hesitant to suppress cross-references, but I think all the instances here make sense, given their context. Are there some specific ones you would revert?

A

[serhiy-storchaka]

There is no frozenset.add().

[AA-Turner]

Because set and frozenset are documented together, there is no cross-reference target for set.add, it is labelled as frozenset.add instead. This link goes to the correct place.

[serhiy-storchaka]

But this is a wrong name for anchor. We should do something with this.

[AA-Turner]

I agree. I plan to fix it, but it requires changes (& a release) of Sphinx. This PR makes the reference link to the right place, and more importantly means that no future reference errors can be introduced into stdtypes.rst.

Once it is fixed, I plan to go back and update here & a few other places where we have used a different target to the intended one.

[serhiy-storchaka]

Now, I think that we should reorganize the set/frozenset documentation in the way similar to the sequence or bytes/bytearray documentation.

1. Dedent method descriptions and add class name prefixes. E.g.:

   .. method:: frozenset.isdisjoint(other, /)
               set.isdisjoint(other, /)

2. Move operators to separate tables. This may wait.

[serhiy-storchaka]

Are there some specific ones you would revert?

All of them? See my suggestion #101100 (comment) . After implementing it the references in footnotes will gone, as they will become the part of the method description, and many other references will became working. After that we can check what remains.

[AA-Turner]

See #138474, putting this in draft for now.

A