This is the mail archive of the
libstdc++@sources.redhat.com
mailing list for the libstdc++ project.
Re: AIX status
- To: David Edelsohn <dje at watson dot ibm dot com>
- Subject: Re: AIX status
- From: Phil Edwards <pedwards at disaster dot jaj dot com>
- Date: Sat, 14 Oct 2000 18:53:29 -0400
- Cc: libstdc++ at sources dot redhat dot com
- References: <pedwards@disaster.jaj.com> <200010132046.QAA25186@mal-ach.watson.ibm.com>
On Fri, Oct 13, 2000 at 04:46:25PM -0400, David Edelsohn wrote:
> #undef _ANSI_C_SOURCE
> #undef _POSIX_SOURCE
> #undef _XOPEN_SOURCE
> #undef _H_STANDARDS
> #include <standards.h>
This is /exactly/ what I was talking about. Something similar happens in
the solaris headers, but I do an outright _XOPEN_SOURCE = 500 instead of
including anything. This causes all future inclusions from the /usr/include
tree to Do The Right Thing. And we know that the 500 is the Right Thing
here, from the POSIX docs and the Solaris man pages.
> The macro definitions which need to be unset require tracking the
> AIX standards.h header which is what I was trying to avoid.
Well, os_defines is /there/ for OS-specific stuff. It's even arguable
that your including a header is a more future-compatible solution than my
defining one symbol directly. (But I don't expect SUSv2/XNS5/XPG4v2 to
change that number anytime soon either. :-)
How far does the build go on AIX now?
Phil
--
pedwards at disaster dot jaj dot com | pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools. Fools are protected by more capable fools.