This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: rel_ops issues
- To: Theodore Papadopoulo <Theodore dot Papadopoulo at sophia dot inria dot fr>
- Subject: Re: rel_ops issues
- From: Phil Edwards <pedwards at disaster dot jaj dot com>
- Date: Thu, 29 Mar 2001 23:29:01 -0500
- Cc: Joe Buck <jbuck at racerx dot synopsys dot com>, libstdc++ at gcc dot gnu dot org
- References: <jbuck@racerx.synopsys.com> <200103290857.f2T8vRS04951@mururoa.inria.fr>
On Thu, Mar 29, 2001 at 10:57:27AM +0200, Theodore Papadopoulo wrote:
> By the way, I always thought that the headers like
> map.h and friends were mandated by the standard with the meaning of
No, they aren't. Only the C headers are required to do that. Since there
has never been a previous C++ standard, those headers do not need to be
treated the same as the C headers.
Personally, I don't care /too/ much about supporting pre-standard STL usage.
Things not in the STL that were used much more heavily, such as iostream.h,
we have to be more careful about.
> I cannot find this in the standard ???
> Is this still true ?
> If no, shouldn't these headers be moved to the ext (or other)
> directory.
They are in the backward subdirectory, for backwards compatability.
And originally they were installed there too, but were changed to be in
the main include directory for easier backwards compatability.
I believe that, whatever the good reasons for more support of rel_ops may
be, kicking along things like map.h isn't one of them.
Phil
--
pedwards at disaster dot jaj dot com | pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools. Fools are protected by more capable fools.