Understanding void* against intptr_t and uintptr_t

Question

This is the code I'm testing:

int value = 0;
void* addyvoid = static_cast<void*>(&value); // C++-style cast.

it works perfectly, but I could use uintptr_t / intptr_t. But they are not good for holding pointers as people said here because they are too big. So, is this true? If yes, however, for holding pointers using void* would be better, but will there be a loss of data?

Answer 1

The purpose of intptr_t and uintptr_t is that in some applications, you actually do need to do some sort of numeric computation on pointer values, perhaps by flipping individual bits, perhaps by XORing them, etc. In those cases, when you need to work with the numeric value of a pointer, intptr_t and uintptr_t are integer types that (if they exist) are guaranteed to be large enough to hold any pointer. This is not true of, say, int, since int's size relative to pointer sizes isn't specified.

Because it's fundamentally unsafe to do these conversions, C++ requires that you use reinterpret_cast to convert to and from intptr_t and uintptr_t and pointer types.

If all that you're doing is storing "a pointer to something," and provided that pointer isn't a function pointer or a member function pointer, you can just cast it to void*. That cast is guaranteed to work and the conversion from void* back to the original type only requires a static_cast and is guaranteed to be safe.

The size of intptr_t and uintptr_t isn't a good reason to avoid them. They're just for different applications. If you need to do numeric computations on pointers, use those types. Otherwise, if you just need to store "a pointer to something," use a void*.