Re: [boost] Formal Review for Interval Library

From: Fernando Cacciola (fcacciola_at_[hidden])
Date: 2002-09-10 17:21:14

----- Original Message -----
From: "Kevin Lynch" <krlynch_at_[hidden]>
To: <boost_at_[hidden]>
Sent: Tuesday, September 10, 2002 7:00 PM
Subject: Re: [boost] Formal Review for Interval Library

> >>>rounded_transc.hpp
> >>>~~~~~~~~~~~~~~~~
> >>>
> >>>1) Why are these hyperbolic functions over T unqualified?
> >>>
> >>>#ifdef BOOST_HAVE_INV_HYPERBOLIC
> >>>    T asinh_down(const T& x) { return ::asinh(x); }
> >>>    T asinh_up  (const T& x) { return ::asinh(x); }
> >>>    T acosh_down(const T& x) { return ::acosh(x); }
> >>>    T acosh_up  (const T& x) { return ::acosh(x); }
> >>>    T atanh_down(const T& x) { return ::atanh(x); }
> >>>    T atanh_up  (const T& x) { return ::atanh(x); }
> >>>#endif
> >>
> >>Somebody else will probably answer. But I seem to remember the Standard
> >>doesn't require these functions of <cmath> to be defined in the
namespace
> >>'std'. It was also discussed on gcc mailing-list (I hope my memory
> >>doesn't fail me).
> >>
> >
> > I see.
> > I don't know about that, -though I thought ALL std functions go in
std::-.
> > Perhaps someone else can inform us.
> >
>
> The inverse hyperbolics didn't appear in math.h until C99; in other
> words, they aren't technically part of C89/C++98.  Since std is (more or
> less) closed by fiat, they really shouldn't appear in std (<cmath> is
> section 26.5, I think ... but it might be 25.6)
>
Now I now... Thank you!

Fernando Cacciola
Sierra s.r.l.
fcacciola_at_[hidden]
www.gosierra.com