This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: cpp discrepancies
- To: Marc Espie <Marc dot Espie at liafa dot jussieu dot fr>
- Subject: Re: cpp discrepancies
- From: Zack Weinberg <zack at rabi dot columbia dot edu>
- Date: Wed, 28 Apr 1999 10:37:59 -0400
- cc: egcs at egcs dot cygnus dot com
On Wed, 28 Apr 1999 16:32:06 +0200, Marc Espie wrote:
>Just to put things in context, I didn't try to find all those problems.
>It's simply that OpenBSD features a lint that invokes cpp directly,
>hence noticing that -undef was gone from the internal pre-processor,
>I proceeded to use the new cpp, and ran into ALL the problems I mentioned.
>
>On Wed, Apr 28, 1999 at 12:28:10AM -0400, Zack Weinberg wrote:
>> Specifically referring to cpp, the discrepancies you note are because
>> cpp.texi/cpp.1 document the interface to the internal binary, and they
>> don't get it exactly right either.
>That cpp.1 document that simply source cccp.1 doesn't help.
>
>There should be a definite warning in the source that cccp is *internal*
>and not meant to be used... having a man page for an internal program
>is an error, imo, unless it is quite clearly labelled as internal.
When those documents were written, it was possible to use cccp
directly. I would rather make the wrapper executable match the
documented interface (which is not hard - see post to egcs-patches)
than change the documented interface. cccp.1/cpp.texi can be read as
referring to /usr/bin/cpp; the fact that it is a wrapper around an
internal executable is not paricularly relevant to the user.
zw