What is the point of ":" (two points) in the statement of a member of a structure?

Question

I saw this:

typedef struct {
    unsigned char a : 1;
    unsigned char b : 7;
} Tipo;

What does this statement mean? What is this 1 and 7?

Answer 1

Let’s run this code to better understand:

#include <stdio.h>

typedef struct {
    unsigned char a : 1;
    unsigned char b : 7;
} Tipo;

int main(void) {
    Tipo x = { .a = 1, .b = 64 };
    printf("%d e %d\n", x.a, x.b);
    printf("%zd\n", sizeof(Tipo));
}

Behold working in the ideone. And in the repl it.. Also put on the Github for future reference.

Try to put on the limb a a value greater than 1. It doesn’t work, it can only be 0 or 1. Already caught why?

In the second print we see that the size of this structure is 1 byte. And now, killed?

The two points are used when the structure is used to mount a bit string and each member is called bit field. So what comes after the : is the amount of bits that member contains. Hence the a can only have 0 or 1. It only has 1 bit. The same is true if you try to put more than 127 in b that only has 7 bits.

Ideally these structures should have a total multiple size of 8 to fit with a byte, but if there is no alignment according to compiler rules.

Possible members types are specified as _Bool, signed int, unsigned int and others specified by the compiler, which is what was used. It is usually best to use universally defined sized types. So it is best to have a int32_t than a int.

It is very common that this type of structure is within a larger structure with other members.

It is also very common to use with union. So you can access a union member as a single die and the other union member as a bit structure.

#include <stdio.h>

typedef union {
    char valor;
    struct {
        unsigned char a : 1;
        unsigned char b : 1;
        unsigned char c : 1;
        unsigned char d : 1;
        unsigned char e : 1;
        unsigned char f : 1;
        unsigned char g : 1;
        unsigned char : 1;
    } bits;
} Tipo;

int main(void) {
    Tipo x = { .valor = 87 };
    printf("%d %d %d %d %d %d %d", x.bits.a, x.bits.b, x.bits.c, x.bits.d, x.bits.e, x.bits.f, x.bits.g);
}

Behold working in the ideone. And in the repl it.. Also put on the Github for future reference.

Note that I showed you another point. You are not required to identify all bits. Of course, not putting a name on some bit, or set of bits, will not be able to access its value directly and named.