Re: when is a double float nearest infinity ?

[Christian Bruel <chrbr at free dot fr>]

I found the answer to my second question : I just needed to include 
<fenv.h>. Attached a fixed version of the example. But I'm still puzzled 
by the first question (why mdmax+dmax) gives infinity ?

attached a modified version of the test program, compiled with
gcc testinf.c -pedantic -ansi -mieee-fp -lm -Wall -O2

I also tried to compile with -moft-soft to check the libgcc behavior but 
that lead to undefined `__eqdf2' `__adddf3' references.

thanks for any help

Christian

#include <stdio.h>
#include <float.h>
#include <math.h>
#include <fenv.h>

const double infd=(double)1.0/(double)0.0;

void testinf(double v) 
{
  if (v == infd)
    puts("infinity");

  else if (v == DBL_MAX)
    puts("ok nearest");

  else
    puts("error?");
}

int
main()
{
  volatile double max=DBL_MAX;

#if 0
  fesetround (FE_UPWARD);   
#endif

  if (fegetround() == FE_TONEAREST)
    puts ("rounding to nearest");

  testinf(max+(double)1.0);
  testinf(max+max);

  return 0;
}