I would like to use NaN's. It appears that isnan()
works as advertised so I can check for NaN. However,
there is a man page for infnan() but no such function
appears to be available. Does anyone know the proper
way to generate a NaN value? (I am using NaN as an
initialization value so that I can check that a valid
value is used later.) For the moment I am using NAN
as defined in nan.h but it requires the use of
"#define _GNU_SOURCE 1". I would like to keep the code
somewhat portable.
/* HUGE_VAL is traditionally defined as positive infinity, or
alternatively, DBL_MAX. */
#if !HAVE_ISINF
#define isinf(X) \
(fabs (X) == HUGE_VAL)
#endif
/* A Not a Number is not equal to itself. */
#if !HAVE_ISNAN
#define isnan(X) \
((X) != (X))
#endif
/* Finite numbers are not infinities or NaNs. */
#if !HAVE_FINITE
#define finite(X) \
(!isinf (X) && !isnan (X))
#elif HAVE_IEEEFP_H
#include <ieeefp.h> /* Declares finite() under Solaris. */
#endif
The definition of infnan() is left as an exercise to the reader :-)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
