[Bug libc/10355] New: cabs(z) == nan if z = inf + j * nan

[Bug libc/10355] New: cabs(z) == nan if z = inf + j * nan

[cournape at gmail dot com]

Hi,

I came across the following when using cabs and the glibc:

#define _ISOC99_SOURCE

#include <complex.h>
#include <stdio.h>
#include <math.h>

int main(void)
{
        complex z;
        double x, y;

        x = INFINITY;
        y = NAN;

        z = x + I * y;

        printf("%f\n", cabs(z));
        printf("%f\n", hypot(x, y));
        return 0;
}

This prints:
nan
inf

I believe that cabs behavior is incorrect in this case if I read the C99
standard correctly (Section G 6.6 from the committee draft of september 2007):

"
Each of the functions cabs and carg is specified by a formula in terms of a real
function (whose special cases are covered in annex F):
       cabs(x + iy) = hypot(x, y)
       carg(x + iy) = atan2(y, x)
"

glibc version: 2.8
gcc version: 4.3.2

-- 
           Summary: cabs(z) == nan if z = inf + j * nan
           Product: glibc
           Version: unspecified
            Status: NEW
          Severity: normal
          Priority: P2
         Component: libc
        AssignedTo: drepper at redhat dot com
        ReportedBy: cournape at gmail dot com
                CC: glibc-bugs at sources dot redhat dot com


http://sourceware.org/bugzilla/show_bug.cgi?id=10355

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.

[Bug libc/10355] cabs(z) == nan if z = inf + j * nan

[drepper at redhat dot com]

------- Additional Comments From drepper at redhat dot com  2009-10-30 06:29 -------
Have you even looked at the variable z?  It's value is (NaN,Nan), not (Inf,NaN).
 If you'd use correct code to initialize the variable it would work as expected.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |INVALID


http://sourceware.org/bugzilla/show_bug.cgi?id=10355

------- You are receiving this mail because: -------
You are on the CC list for the bug, or are watching someone who is.