This is the mail archive of the
libc-hacker@sourceware.cygnus.com
mailing list for the glibc project.
Re: patch to clean up generic signum.h
- To: Zack Weinberg <zack@rabi.columbia.edu>
- Subject: Re: patch to clean up generic signum.h
- From: Andreas Schwab <schwab@issan.informatik.uni-dortmund.de>
- Date: 03 Nov 1998 10:50:07 +0100
- Cc: Roland McGrath <roland@frob.com>, libc-hacker@cygnus.com
- References: <199811022233.RAA27440@blastula.phys.columbia.edu>
Zack Weinberg <zack@rabi.columbia.edu> writes:
|> On Mon, 2 Nov 1998 17:27:05 -0500, Roland McGrath wrote:
|> >I think it is not "compulsive" but broken to omit signal numbers under
|> >_POSIX_SOURCE. A POSIX.1-compliant program must define _POSIX_SOURCE,
|> >and a well-written one would handle common non-POSIX signals by checking
|> >for them with #ifdef.
|>
|> The only things I put under __USE_MISC were the obsolete names SIGIOT
|> and SIGCLD.
|>
|> I did condition the POSIX signals on __USE_POSIX, which will affect
|> programs compiled with -ansi and nothing else that look for other
|> signals. If you think that's broken, it can be changed.
It is strictly ISO C compliant to define additional SIG[A-Z]* macros in
<signal.h>. So IMHO there should be no feature tests at all in here.
Andreas.
--
Andreas Schwab "And now for something
schwab@issan.cs.uni-dortmund.de completely different"
schwab@gnu.org