(inspired by #95302 (comment))

In Objects/weakrefobject.c, the clear_weakref() function is used to:

• pop the PyWeakReference object off its referenced object's tp_weaklist (a linked list)
• clear (but not call) the callback, if any

However, in the places it is used, it isn't clear what it's doing. When adding a new use of clear_weakref() in one of my PRs it was particularly confusing to a reviewer what the relationship was between clear_weakref() and an object's weakref list (i.e. tp_weaklist).

While we could simply rename it to "remove_and_clear_weakref", it may make sense to do the following:

• move the linked-list part to a separate remove_weakref() that has an explicit PyWeakReference ** parameter
• (hence, remove the linked list parts from clear_weakref())
• add remove_weakref_from_referent() which calls GET_WEAKREFS_LISTPTR(self->wr_object) and calls remove_weakref() on it
• add a call to it everywhere we already call clear_weakref()

Other observations:

• currently we aren't reseting self->hash or self->vectorcall to NULL in clear_weakref()
• we may be leaking weakref objects PyObject_ClearWeakRefs() when the weakref doesn't have a callback