gcc -ffast-math defect with tan(x)

[<eric_backus at agilent dot com>]

Hello fellow cygwinners,

I'm seeing incorrect behavior when I use the "-ffast-math" option to gcc on cygwin.  Here is a Short Test Case:


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

int main(void)
{
    double d1 = 0.0;
    double d2 = 0.0;
    d1 = tan(d1);
    d2 = tan(d2);
    (void) printf("d1 = %lg, expecting 0 (or -0)\n", d1);
    (void) printf("d2 = %lg, expecting 0 (or -0)\n", d2);
    return 0;
}


Compile this with "gcc -ffast-math testprog.c -o testprog", then run "./testprog".  Correct output would be:

d1 = 0, expecting 0 (or -0)
d2 = 0, expecting 0 (or -0)

Some time ago, on cygwin 1.5 and presumably with an older version of gcc, this code (well, actually, a more complicated version of this in a library we use) worked correctly.  It worked correctly both with and without "-mno-cygwin", and both with and without -ffast-math.

Today, on cygwin 1.7, I get this output:

d1 = -0, expecting 0 (or -0)
d2 = nan, expecting 0 (or -0)

I get this output from both gcc-3 (which is 3.4.4) and from gcc-4 (which is 4.3.4), when -ffast-math is used.  If I remove -ffast-math, I get the expected output of 0 for both d1 and d2.  If I compile with -mno-cygwin on gcc-3, either with or without -ffast-math, I get the expected output of 0 for both d1 and d2.  So the problem seems to be limited to -ffast-math, and to the cygwin (non-mingw32) platform, and perhaps to relatively recent versions of gcc.

My completely uninformed guess is that this is a cygwin-specific defect in gcc.  But I really don't know.

Any ideas how I should deal with this would be helpful!

-- 
Eric



--
Problem reports:       http://cygwin.com/problems.html
FAQ:                   http://cygwin.com/faq/
Documentation:         http://cygwin.com/docs.html
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple