This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH] Fix PR35202, errno and (float)mathfn((double)x) folding
On Sun, 8 Feb 2009, Joseph S. Myers wrote:
> On Sun, 8 Feb 2009, Richard Guenther wrote:
>
> > On Sun, Jan 25, 2009 at 8:37 PM, Richard Guenther <rguenther@suse.de> wrote:
> > >
> > > This fixes said PR. I went over the C99 spec and recent glibc
> > > manpages for non-C99 functions. I explicitly excluded the
> > > possibility of extra domain errors from domains that are bounded
> > > by values that are exactly representable.
> > >
> > > Bootstrapped and tested on x86_64-unknown-linux-gnu, ok?
> >
> > Joseph, ping?
>
> This patch is outside the areas of the compiler I maintain. The only
> comment I would make is that I would expect expm1 to be handled the same
> as exp (GCC should not optimize on the basis of glibc bug 6788).
I was asking because you filed that bug and I am never sure if
convert.c is frontend or middle-end ;) You are right about expm1,
I'll adjust behavior.
Thanks,
Richard.