This is the mail archive of the
libc-alpha@sources.redhat.com
mailing list for the glibc project.
Re: BSD signal blocking
- To: kaz at ashi dot footprints dot net
- Subject: Re: BSD signal blocking
- From: Paul Eggert <eggert at twinsun dot com>
- Date: Wed, 2 Aug 2000 13:38:03 -0700 (PDT)
- CC: drepper at cygnus dot com, haible at ilog dot fr, libc-alpha at sourceware dot cygnus dot com
- References: <Pine.LNX.4.10.10008021317470.19338-100000@ashi.FootPrints.net>
Date: Wed, 2 Aug 2000 13:20:09 -0700 (PDT)
From: Kaz Kylheku <kaz@ashi.footprints.net>
I use -ansi all the time, and then use various feature selection macros
to select additional interfaces that are not in ANSI C. That's how it
is supposed to work.
In the autoconf mailing list we've recently gone through a discussion
of feature-test macros like _ALL_SOURCE, __EXTENSIONS__, _GNU_SOURCE,
_HPUX_SOURCE, _MINIX, _POSIX_SOURCE, _POSIX_C_SOURCE, etc., and the
consensus is that we don't have time to configure them systematically
in the next autoconf release. Maybe later.
In the meantime, it's generally the builder's responsibility to futz
with these macros, but personally I would try to avoid them as they
will suck up your valuable time. For more discussion about this,
please see:
http://deja.com/=dnc/getdoc.xp?AN=614050311
Without the -D_BSD_SOURCE, the compiler must not add any non-ANSI-C
materials into standard ANSI C headers like <signal.h>, except
where permitted.
Yes, that's correct. ANSI C does not allow <signal.h> to define
"sigmask", so Ulrich is right to reject that change.