[Bug c/37714] Sign of sin(-0.0) depends on optimization level

[rguenth at gcc dot gnu dot org]

------- Comment #1 from rguenth at gcc dot gnu dot org  2008-10-02 11:42 -------
Works on Linux where it replaces it with a call to sincos.  Also works with
the Linux glibc if explicitly calling cexp like

static my_complex
c_rect(double phi)
{
  my_complex z;
  z.imag = __imag cexp (I * phi);
  return z;
}

which is only with -ffast-math replaced by a call to sin.

So, this is a Darwin libm bug.


-- 

rguenth at gcc dot gnu dot org changed:

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37714