Compare the `f->stackpointer` to the result of `_PyFrame_Stackbase(f)`

[nybblista]

I think we should compare f->stackpointer to the result of the _PyFrame_Stackbase(f) in the _PyFrame_StackPeek and in the _PyFrame_StackPop rather than repeat the offset calculation in both of them.

cpython/Include/internal/pycore_interpframe.h

Lines 46 to 60 in c492ac7

	static inline _PyStackRef *_PyFrame_Stackbase(_PyInterpreterFrame *f) {
	return (f->localsplus + _PyFrame_GetCode(f)->co_nlocalsplus);
	}
	static inline _PyStackRef _PyFrame_StackPeek(_PyInterpreterFrame *f) {
	assert(f->stackpointer > f->localsplus + _PyFrame_GetCode(f)->co_nlocalsplus);
	assert(!PyStackRef_IsNull(f->stackpointer[-1]));
	return f->stackpointer[-1];
	}
	static inline _PyStackRef _PyFrame_StackPop(_PyInterpreterFrame *f) {
	assert(f->stackpointer > f->localsplus + _PyFrame_GetCode(f)->co_nlocalsplus);
	f->stackpointer--;
	return *f->stackpointer;
	}

Linked PRs

• gh-133713: Compare the f->stackpointer to the result of _PyFrame_Stackbase(f) #133714

[picnixz]

Those are assertions. We don't really need to be efficient in DEUBG builds so I don't think it's worth the change. OTOH it could be a cleanup. Up to @markshannon then

[ZeroIntensity]

Moreover: I'd like to avoid these kinds of issues/PRs in the future. This sort of thing should happen in a PR review, but once the change has already landed, it's just needless refactoring.