gh-112213: Update _weakref module to use new AC feature

[corona10]

• Issue: Support specifying object to lock in Argument Clinic's @critical_section directive #112213

[colesbury]

Great! Much nicer than manually writing the Py_BEGIN/END_CRITICAL_SECTION

[corona10]

@colesbury
One thing needs to be discussed:
Do we have to care about the thread-safe for is_dead_weakref?

cpython/Modules/_weakref.c

Lines 45 to 56 in 1a969b4

	is_dead_weakref(PyObject*value)
	{
	if (!PyWeakref_Check(value)){
	PyErr_SetString(PyExc_TypeError, "not a weakref");
	return-1;
	}
	intis_dead;
	Py_BEGIN_CRITICAL_SECTION(value);
	is_dead=_PyWeakref_IS_DEAD(value);
	Py_END_CRITICAL_SECTION();
	returnis_dead;
	}

Or should it be handled from _PyDict_DelItemIf since it is only used for _PyDict_DelItemIf?

cpython/Modules/_weakref.c

Line 74 in 1a969b4

if (_PyDict_DelItemIf(dct, key, is_dead_weakref) <0){

cpython/Objects/dictobject.c

Lines 2062 to 2069 in 1a969b4

	/* This function promises that the predicate -> deletion sequence is atomic
	* (i.e. protected by the GIL), assuming the predicate itself doesn't
	* release the GIL.
	*/
	int
	_PyDict_DelItemIf(PyObject*op, PyObject*key,
	int (*predicate)(PyObject*value))
	{

If we decide not to handle from is_dead_weakref, we can remove critical section related code fully.

[colesbury]

@corona10, we'll need to handle the thread-safety issues in _PyWeakref_IS_DEAD().

[corona10]

@corona10, we'll need to handle the thread-safety issues in _PyWeakref_IS_DEAD().

Okay let's handle it with upcoming PR :)