NAME

copysign, isnan, finite, fp_class, nextafter, unordered  − appendix and related miscellaneous functions for IEEE arithmetic

LIBRARY

Math Library (libm.a)

SYNOPSIS

#include <math.h>

double copysign   (double x, double y);
float  copysignf  (float  x, float  y);
int    finite     (double x);
int    finitef    (float  x);
int    fp_class   (double x);
int    fp_classf  (float  x);
int    isnan      (double x);
int    isnanf     (float  x);
double nextafter  (double x, double y);
float  nextafterf (float  x, float  y);
int    unordered  (double x, double y);
int    unorderedf (float  x, float  y);

DESCRIPTION

copysign() and copysignf() return x with the same sign as y.  IEEE 754 requires copysign(x,NaN) and copysignf(x,NaN) = +x  or -x. 

finite() and finitef() return the integer value 1 (True) or 0 (False).  finite(x) and finitef(x) = 1 when -infinity < x < +infinity.  finite(x) and finitef(x) = 0 when |x| = infinity or x is a NaN. 

fp_class() and fp_classf() determine the class of IEEE floating-point values.  They return one of the constants in the file <fp_class.h> and never cause an exception, even for signaling NaNs.  These routines are to implement the recommended function class(x) in the appendix of the IEEE 754-1985 standard for binary floating-point arithmetic.  The constants in <fp_class.h> refer to the following classes of values:

isnan() and isnanf() return 1 (True) if x  is NaN (the IEEE floating point reserved not-a-number value) and 0 (False) otherwise. 

nextafter() and nextafterf() return the next machine representable number from x  in the direction of y. 

unordered() and unorderedf() return the value 1 (True) if either x  or y  or both is a NaN, and returns the value 0 (False) otherwise. 

RELATED INFORMATION

drem(3), logb(3), scalb(3), rint(3), sqrt(3), and ANSI/IEEE Std 754-1985, IEEE Standard for Binary Floating-Point Arithmetic