gh-91118: Fix docstrings that do not honor --without-doc-strings

[arhadthedev]

To support --without-doc-strings, all docstrings must be wrapped into PyDoc_STRVAR or PyDoc_STR (PEP 7). However, there are 18 occurrences in code and 10 in C API documentation that do not follow this rule. The documentation is important too because it should not teach people the wrong things.

To find the occurrences I searched for (?:^\s*.tp_doc = "|" \/\* tp_doc \*\/$) and^(?:static\s+)?const\s+char\s+[^=]+=\s*".

Fixes GH-91118.

[JelleZijlstra]

There are a bunch of others, mostly in ctypes:

% git grep '".*tp_doc\s*'
Doc/c-api/typeobj.rst:       "My objects",                   /* tp_doc */
Modules/_ctypes/_ctypes.c:    "deletes a key from a dictionary",          /* tp_doc */
Modules/_ctypes/_ctypes.c:    "metatype for the CData Objects",           /* tp_doc */
Modules/_ctypes/_ctypes.c:    "metatype for the CData Objects",           /* tp_doc */
Modules/_ctypes/_ctypes.c:    "metatype for the Pointer Objects",         /* tp_doc */
Modules/_ctypes/_ctypes.c:    "metatype for the Array Objects",           /* tp_doc */
Modules/_ctypes/_ctypes.c:    "metatype for the PyCSimpleType Objects",           /* tp_doc */
Modules/_ctypes/_ctypes.c:    "metatype for C function pointers",         /* tp_doc */
Modules/_ctypes/_ctypes.c:    "XXX to be provided",                       /* tp_doc */
Modules/_ctypes/_ctypes.c:    "Function Pointer",                         /* tp_doc */
Modules/_ctypes/_ctypes.c:    "Structure base class",                     /* tp_doc */
Modules/_ctypes/_ctypes.c:    "Union base class",                         /* tp_doc */
Modules/_ctypes/_ctypes.c:    "XXX to be provided",                       /* tp_doc */
Modules/_ctypes/_ctypes.c:    "XXX to be provided",                       /* tp_doc */
Modules/_ctypes/_ctypes.c:    "XXX to be provided",                       /* tp_doc */
Modules/_ctypes/callbacks.c:    "CThunkObject",                             /* tp_doc */
Modules/_ctypes/cfield.c:    "Structure/Union member",                   /* tp_doc */
Modules/_testcapimodule.c:    "Instantiating this exception starts infinite recursion.", /* tp_doc */

I'm a bit skeptical about how useful this is, though—does anyone actually use the configure option to strip docstrings?

[JelleZijlstra]

Suggested change

	:class:`~_ctypes.PyCPointerType:, :class:`~_ctypes.PyCSimpleType:, and
	:class:`~_ctypes.PyCPointerType`, :class:`~_ctypes.PyCSimpleType`, and

[JelleZijlstra]

Suggested change

	:class:`~_ctypes.PyCStructType:.
	:class:`~_ctypes.PyCStructType`.

[arhadthedev]

I'm a bit skeptical about how useful this is, though—does anyone actually use the configure option to strip docstrings?

I think that if a flag exists, it should work fully. Also it's enabled by default so strings that should not be in the binary are no longer there.

[miss-islington]

Thanks @arhadthedev for the PR, and @JelleZijlstra for merging it 🌮🎉.. I'm working now to backport this PR to: 3.9.
🐍🍒⛏🤖

[miss-islington]

Thanks @arhadthedev for the PR, and @JelleZijlstra for merging it 🌮🎉.. I'm working now to backport this PR to: 3.10.
🐍🍒⛏🤖

[miss-islington]

Sorry, @arhadthedev and @JelleZijlstra, I could not cleanly backport this to 3.9 due to a conflict.
Please backport using cherry_picker on command line.
cherry_picker a573cb2fec664c645ab744658d7e941d72e1a398 3.9

[miss-islington]

Sorry @arhadthedev and @JelleZijlstra, I had trouble checking out the 3.10 backport branch.
Please backport using cherry_picker on command line.
cherry_picker a573cb2fec664c645ab744658d7e941d72e1a398 3.10

[JelleZijlstra]

@arhadthedev do you want to do the backports? I feel like the docs changes are most important to backport.

[bedevere-bot]

GH-91662 is a backport of this pull request to the 3.10 branch.

[bedevere-bot]

GH-91663 is a backport of this pull request to the 3.9 branch.

[arhadthedev]

That backport to 3.9 was targeted to main instead of 3.9.

The new, correct entry is GH-91664.