My C++ recursion function is not calling itself or doing anything after it enters the first if statement (if (perm.empty()). I was confused so I converted the same logic to python and it works just fine. Am I missing something? Thank you all

C++ code:

bool permMaker(vector<int> choices, vector<int> & perm, int n, int i ){
	// remove last added element from chocies
	if (!perm.empty()) {
		choices.erase(remove(all(choices), perm.back()), perm.end());
		cout << "in" << endl;
	}
	if (perm.size() == n) {
		return true;
	}	
	rap(num, choices){
		if (perm.empty()) {
			perm.push_back(num);
			bool ans = permMaker(choices, perm, n, i +1 );
			if (ans) {
				return true;
			}
			perm.pop_back();
		}
		else {
			if (abs(num - perm.back()) > 1) {
				perm.push_back(num);		
				bool ans = permMaker(choices, perm, n, i + 1);
				if (ans) {
					return true;
				}
				perm.pop_back();
			}
		}
	}
	return false;
}

Here is the same logic in python which is working

def permMaker(choices, perm, length):
    choices = choices[:]
    if len(perm) >0:
        choices.remove(perm[-1])
    
    if len(perm) == length:
        return True
    
    for num in choices:
        if len(perm) == 0:
            perm.append(num)
            ans = permMaker(choices, perm, length)
            if ans:
                return True
            perm.pop()
        else:
            if abs(num - perm[-1]) > 1:
                perm.append(num)
                ans = permMaker(choices, perm, length)
                if ans:
                    return True
                perm.pop()
    
    return False

I am rather new to CP, and am curious that now that there is PyPy 64-bit on Codeforces, how much more competitive does that make python? I have seen in other blog posts on here that one of the main things slowing down python (PyPy) was the fact that >32 bit integers were being stored as big ints. Does this change make Python more competitive and actually usable? I only ask because I know c++ and python (I know python better than c++ as I've only recently been using c++ for cp), but can code solutions to problems much faster and "cleaner" in python.