fpclassify()

Classifies floating-point numbers.

  
#include <math.h>
int fpclassify(x);
 

x

A value of type float, double, or long double.

This macro returns the type of a floating-point value as an integral value:

• FP_NAN: NaN ( "Not a Number" )
• FP_INFINITE: positive or negative infinity
• FP_ZERO: zero
• FP_NORMAL: a normalized floating point value
• FP_SUBNORMAL: a denormalized floating point value

Use this macro to check for errors in floating point computations. This facility may not be available on configurations of the EWL that run on platforms that do not have floating-point math capabilities.

#include <math>

#include <stdio.h>

void mypow(double x, double y)

{

double p;

p = pow(x, y);

switch (fpclassify(p))

{

case FP_ZERO:

case FP_NORMAL:

case FP_SUBNORMAL:

printf("%f", p);

break;

default:

printf("error in pow()");

break;

}

}