RFA: PATCH to initializer_constant_valid_p for c++/36912 (static initialization and -frounding-math)

[Richard Guenther]

On Thu, Nov 5, 2009 at 2:19 AM, Jason Merrill <jason@redhat.com> wrote:
> There's a lot of discussion in the PR about how we might want to support
> constant folding of floating point in the context of -frounding-math, but
> regardless of that,
>
> 1) An unfolded floating point expression isn't a valid constant initializer,
> and
> 2) output_constant shouldn't try to treat the expression as a REAL_CST after
> complaining that it isn't one.
>
> Tested x86_64-pc-linux-gnu.  OK for trunk?

Ok.

Thanks,
Richard.