gh-101100: [docs] Fix Sphinx warnings in dis module

[sobolevn]

Original errors:

./cpython/Doc/library/dis.rst:46: WARNING: py:func reference target not found: myfunc
./cpython/Doc/library/dis.rst:51: WARNING: py:func reference target not found: myfunc
./cpython/Doc/library/dis.rst:795: WARNING: py:func reference target not found: builtins.__build_class__
./cpython/Doc/library/dis.rst:852: WARNING: py:attr reference target not found: co_names
./cpython/Doc/library/dis.rst:859: WARNING: py:attr reference target not found: co_names
./cpython/Doc/library/dis.rst:898: WARNING: py:attr reference target not found: co_names
./cpython/Doc/library/dis.rst:907: WARNING: py:attr reference target not found: co_names
./cpython/Doc/library/dis.rst:1426: WARNING: c:func reference target not found: PyObject_Format

A couple of notes for reviewers:

1. :func:myfunc was rendered incorrectly (invalid signature):

   

   Because :func: is reserved for functions defined with .. function:: directive.

2. I've silenced !builtins.__build_class__ reference, because as far as I understand, there's no desire in documenting this feature. So, now this reference renders the same way, but just does not error out

3. :attr:!codeobject.* are also just ignoring warnings, because right now there's no way to address these attributes documented here: https://github.com/python/cpython/blame/5ffdaf748d98da6065158534720f1996a45a0072/Doc/reference/datamodel.rst#L956 So, we just add an ignore and slightly more context

4. PyObject_Format will be found after gh-102595: Document PyObject_Format c-api function #102596 is merged

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

[CAM-Gerlach]

Seems otherwise good overall, but instead of disabling resolution and silencing the warnings, could co_names be properly documented instead, just like codeobject.co_positions currently is, especially since it's referenced nearly a dozen times in this document?

[sobolevn]

I don't know enough about the history of these fields. They might be not documented on purpose. We need a field expert here to decide.

[sobolevn]

@iritkatriel can you please take a look? :)

[AA-Turner]

A comment on the pre-fixing of codeobject. -- I think we should revert this for now.

It may also be better to add an entry to nitpick_ignore in conf.py with an explanatory note rather than supressing these warnings, if we intend to document the code object attributes at a later date.

[hugovk]

if we intend to document the code object attributes at a later date.

Or can we do it now?

[sobolevn]

Closed by #108476 by @erlend-aasland

[erlend-aasland]

Sorry, @sobolevn, I didn't know you had a PR open :(