[Bug libstdc++/58625] std::signbit always converts to double

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58625

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #1 from Paolo Carlini <paolo.carlini at oracle dot com> ---
Weird that nobody noticed for so much time. Are there actual testcases for
this? Or in practice it's just an optimization issue, not a correctness issue?

Now I suspect isinf and isnan too, for which we would not have available *f and
*l counterparts at all. Could you please check?

I suppose we, libstdc++-v3 people, simply assumed that the C++ front-end
provided a full set of intrinsics corresponding to C99 classification macros,
besides those for the comparison macros.

Adding Jakub in CC: the change per se is so easy, but maybe I'm missing some
background for this.