From: Greg Colvin (gcolvin_at_[hidden])
Date: 1999-07-29 08:59:39
From: Beman Dawes <beman_at_[hidden]>
> At 11:36 PM 7/28/99 -0500, Andy Glew wrote:
>
> >>I have read paragraphs 2 and 3 several times and can't see any
> >>difference, other that the ordering of sentances and "..." falling
> >>back to <...>. Oddly enough, the only problem I ever had with a
> real
> >>compiler was with "..." although that was clearly a compiler bug.
> >>
> >>What are you seeing that I am missing?
> >
> >
> >Well... back when I followed the original ANSI C meetings,
> >the intent was loosely described as follows:
> >
> >#include <header>
> >
> >was *NOT* required to include a file. <header> could be a
> >precompiled header file, or even could be an object hardwired
> >into the compiler. The implementation was *NOT* required to
> >disclose how to go from source.h to <header>, and was certainly
> >not required to implement the usual UNIX semantics of
> >a search path without an initial dot.
>
> Correct if <header> is a standard library header. But the wording of
> both standards seem clear that if <header> isn't a standard library
> header, then file inclusion is required for both <...> and "..."
> forms.
Actually, I think it is clear that neither form is required to
do file inclusion, or even that the system have any concept of
file.