Bug#153548: libc6: Results of math functions are sometimes wrong

[Lukas Geyer <lukas@debian.org>]

Vincent Lefevre wrote:
----------------------------<snip>-----------------------------------
#include <stdio.h>
#include <math.h>

int main (void)
{
  double x, y;
  x = 0.80190127184058835;
  y = cos (x);
  printf ("f(%.17g) = %.17g\n", x, y);
  return 0;
}

gives

f(0.80190127184058835) = 0.71867942238767868

on my PowerPC machine (Titanium). The result should be something like:

f(0.80190127184058835) = 0.69534156199418473
-----------------------------<snip>----------------------------------

The funny thing about this is that the result is _exactly_ sin(x)
instead of cos(x). Furthermore, it seems to be a special case, as
slightly changing x gives correct results. Could it be a typo in some
special case handling for cos()? I can confirm the error here, but I
do not have the glibc sources handy to track the error down.

Cheers,
Lukas