7.5 Pointer to Pointer

A pointer is itself a variable, so it also has its own address. A pointer that points to a pointer is called a pointer to pointer, written int **.

int data = 5;
int *p = &data;     // first-level pointer, points to data
int **q = &p;       // pointer to pointer, points to p

std::cout << data << std::endl; // 5
std::cout << *p << std::endl;   // 5
std::cout << **q << std::endl;  // 5

q stores the address of p. So *q is p, and **q dereferences once more to reach data. Each extra * follows one more address.

Letting a function modify the caller's pointer

The most common use of a pointer to pointer is to let a function modify the caller's pointer. If you pass only an int *, the function changes a copy of the pointer; to change the caller's pointer itself, pass an int **.

void allocate(int **out) {
    *out = new int;
    **out = 7;
}

int main() {
    int *q = nullptr;
    allocate(&q);              // q now points to fresh memory
    std::cout << *q << std::endl;   // 7
    delete q;
    return 0;
}

argv is a pointer to pointer

In main(int argc, char* argv[]), argv behaves like a char**: it points to a set of string pointers, where each argv[i] is a string (see 6.3 for command-line arguments).