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

[dickinsm at gmail dot com]

------- Comment #2 from dickinsm at gmail dot com  2008-10-02 14:34 -------
Thanks for the response!

It does appear to be true that cexp doesn't follow
Annex G of C99, on OS X 10.5.5.  I agree that this
is undesirable, but I can't see why it should
be considered a bug.

Annex G of C99 is merely informative, and as far
as I can tell __STDC_IEC_559_COMPLEX__ is not
defined on OS X 10.5.5.

In other words, it seems to me that there's no
basis for gcc's assumption that sin(-0.0) is
interchangeable with the imaginary part of
cexp(-0.0*I).


-- 


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