From: Noel Yap (yap_noel_at_[hidden])
Date: 2002-07-20 15:38:40
--- Beman Dawes <bdawes_at_[hidden]> wrote:
> At 02:09 PM 7/20/2002, Gennaro Prota wrote:
>
> >On Sat, 20 Jul 2002 10:03:23 -0400, "David
> Abrahams"
> ><david.abrahams_at_[hidden]> wrote:
> >>...
> >>What really determines whether a library header
> file is a "header" or a
> >>"source file"?
> >>
> >
> >Well, since you are one of the standard experts
> there must be
> >something wrong in what I know. So far I thought
> that in the standard
> >terminology 'header' is what we colloquially call
> 'standard header'.
> >In other words, there's no 'user header' like
> foo.hpp or
> >my_library.hpp. They are simply source files.
> >
> >In fact, the wording of 16.2 confirms this belief
> (it speaks of
> >'header' for the #include <...> form and of
> 'source file' for the
> >#include "..." form). Moreover, a brute search of
> the expression
> >"standard header" in the whole standard yielded no
> occurrence in
> >normative text.
> >
> >Opinion?
>
> That's my understanding also. To the standard, a
> header is a standard
> library header. There is no such thing as a user
> header, only a user
> source file.
>From what I understand from _C: A Complete Reference_,
the standard differentiates between the two so that
#include <stdio.h> will work even on systems where
stdio.h is an invalid filename. Can anyone confirm
this?
Noel
__________________________________________________
Do You Yahoo!?
Yahoo! Health - Feel better, live better
http://health.yahoo.com