Fix #11713 FN constParameterPointer/Reference with unused parameter

[chrchr-github]

No description provided.

[chrchr-github]

I wonder how much noise this will cause.
But we already warn for

int f(int& r) {
	int i = r;
#ifdef MACRO
	r = 0;
#endif
	return i;
}

[danmar]

But we already warn for

I also classify that as a false positive. I would prefer that we fix such FPs... if there are ifdefs and stuff in the scope that could hide code then do not warn.

However I would suggest that the "possible FP" warnings are still reported in such cases with a separate ID (maybe enabled by --inconclusive).

[firewave]

I also classify that as a false positive. I would prefer that we fix such FPs... if there are ifdefs and stuff in the scope that could hide code then do not warn.

I do not agree. IMO that is the point of analyzing multiple configurations.

int f() {
	int i; // unassigned
#ifdef MACRO
	i = 0;
#endif
	return i; // uninitialized
}

In that case we correctly detect the unassigned variable and the uninitialized one. Not reporting it in the case this PR addresses would be inconsistent and thus unexpected.