Simple types

A boolean type GAN_BOOL is defined:

      typedef enum { GAN_FALSE=0, GAN_TRUE=1 } Gan_Bool;

The boolean type is the standard type returned by a Gandalf function, where a return type of GAN_TRUE indicates success, GAN_FALSE failure. Another use for the Gan_Bool type is with bit arrays and binary images. See Sections 2.3 and 4.4.

The Gan_Type enumerated type is used extensively to indicated different kinds of simple objects:

      /// labels for simple types used throughout Gandalf
      typedef enum
      {
         GAN_CHAR,       /**< signed character */
         GAN_UCHAR,      /**< unsigned character */
         GAN_SHORT,      /**< signed short integer */
         GAN_USHORT,     /**< unsigned short integer */
         GAN_INT,        /**< signed integer */
         GAN_UINT,       /**< unsigned integer */
         GAN_LONG,       /**< signed long integer */
         GAN_ULONG,      /**< unsigned long integer */
      #if (SIZEOF_LONG_LONG != 0)
         GAN_LONGLONG,   /**< signed extra-long integer */
      #endif
         GAN_FLOAT,      /**< single precision floating point */
         GAN_DOUBLE,     /**< double precision floating point */
         GAN_LONGDOUBLE, /**< long double precision floating point */
         GAN_STRING,     /**< string (array of characters) */
         GAN_BOOL,       /**< boolean */
         GAN_POINTER     /**< generic pointer */
      } Gan_Type;

Note that the GAN_LONGLONG value is only defined if the configure program finds the long long C type, and is able to determine its size. The array gan_type_sizes[] holds the sizes of each Gan_Type value:

      /// array of sizeof()'s of each Gandalf type, one for each value in a Gan_Type
      extern const size_t gan_type_sizes[];

gan_type_sizes and the gan_debug boolean flag (see below) are the only global variables in Gandalf.
gan_type_sizes[] is a constant array, so it is thread-safe.

Gandalf also provides single and double precision floating point versions of the integer limit values found in <limits.h>. For instance GAN_INT_MAXF and GAN_INT_MAXD are the float and double versions of INT_MAX.