Strange struct syntax; type name:number

Bowsing the source for glibc-2.4 I found this strange stuff in ieee754.h:

Code:

---

union ieee754_float
  {
    float f;

    /* This is the IEEE 754 single-precision format.  */
    struct
      {
#if        __BYTE_ORDER == __BIG_ENDIAN
        unsigned int negative:1;
        unsigned int exponent:8;
        unsigned int mantissa:23;
#endif                                /* Big endian.  */
#if        __BYTE_ORDER == __LITTLE_ENDIAN
        unsigned int mantissa:23;
        unsigned int exponent:8;
        unsigned int negative:1;
#endif                                /* Little endian.  */
      } ieee;

    /* This format makes it easier to see if a NaN is a signalling NaN.  */
    struct
      {
#if        __BYTE_ORDER == __BIG_ENDIAN
        unsigned int negative:1;
        unsigned int exponent:8;
        unsigned int quiet_nan:1;
        unsigned int mantissa:22;
#endif                                /* Big endian.  */
#if        __BYTE_ORDER == __LITTLE_ENDIAN
        unsigned int mantissa:22;
        unsigned int quiet_nan:1;
        unsigned int exponent:8;
        unsigned int negative:1;
#endif                                /* Little endian.  */
      } ieee_nan;
  };

---

What is this :number thing? Can't find much information about it. Some kind of initializer?

Bit fields. It's the number of bits that variable is to take up. Read up on bit fields for more info.


Quzah.