How to use abs and div with fixed-size integers

Question

In C99 we have fixed-size integer types, defined in stdint.h. In stdlib.h we have the abs and div functions which operate on ints, as well as their long int/long long int counterparts labs, llabs, ldiv, lldiv.

Since the size of int/long/long long varies based on platform and used compiler I wonder how to chose the right variant of abs/div when using fixed-size integers like int16_t, int32_t or int64_t?

Enable more warnings, and the compiler will tell you. And if you're worried, err on the safe side and always use the `long long` versions. — Some programmer dude, Aug 14 '15 at 09:21
There are also the `intmax_t` variants. And you could define your own set of functions if you want to use a function instead of directly writing the equivalent code. — cremno, Aug 14 '15 at 09:56

Grzegorz Szpetkowski · Accepted Answer · 2015-08-14T10:55:31.173

Assuming, that you are able to use a C11 compiler, you could use Generic macros feature and define wrapper, that covers fixed-size integers. An example might look like:

#include <math.h>
#include <inttypes.h>
#include <stdio.h>

#define safe_abs(n) _Generic((n), \
    signed char: abs(n), short: abs(n), int: abs(n), long: labs(n), long long: llabs(n))

int main(void)
{
    int8_t n1 = -123;
    printf("n1 = %" PRId8 "\n", safe_abs(n1));

    int16_t n2 = -1234;
    printf("n2 = %" PRId16 "\n", safe_abs(n2));

    int32_t n3 = -123456;
    printf("n3 = %" PRId32 "\n", safe_abs(n3));

    int64_t n4 = -12345678910;
    printf("n4 = %" PRId64 "\n", safe_abs(n4));

    return 0;
}

The trick is that these fixed-size integers are just type definitions for standard integer types.

How to use abs and div with fixed-size integers

1 Answers1