>Number:         153
>Category:       lib
>Synopsis:       hypot(3) gives incorrect results when compiled with -O
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    gnats-admin (Library Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Thu Mar  3 15:20:04 1994
>Originator:     Matthieu Herrb
>Organization:
 CNRS - LAAS 
 7, avenue du Colonel Roche
 31077 Toulouse Cedex - France
>Release:        current of 26/02
>Environment:
	486DX33 16M memory
System: NetBSD cougar 0.9a COUGAR#8 i386


>Description:
	the __ieee754_hypot function returns the wrong answer in almost every 
	case when compiled with -O.
>How-To-Repeat:
The following program should print h = 1.00 for every i:
#include <stdio.h>
#include <math.h>

main()
{
    double a, s, c, h,;
    int i;
    
    printf("  i        a       h\n");
    for (i = 0; i <= 360; i++) {
	a = M_PI*i/180;
	s = sin(a1);
	c = cos(a1);
	h = hypot(s, c);
	
	printf("%3d %7.4f %7.4f\n", 
	       i, a, h);
    }
}
>Fix:
	Workaround: compile e_hypot.c without -O




>Audit-Trail:
>Unformatted:


------------------------------------------------------------------------------