[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: sys/time.h: no more -ansi timercmp ?
From: |
David Morse |
Subject: |
Re: sys/time.h: no more -ansi timercmp ? |
Date: |
Sun, 24 Jun 2001 12:10:18 -0700 |
User-agent: |
Mutt/1.3.18i |
Andreas Jaeger <address@hidden> wrote:
> Dave Morse <address@hidden> writes:
>
> > If gcc is invoked with "-ansi" on a file that includes <sys/time.h>, it
> > doesn't pick up timercmp, but it does pick up gettimeofday.
> >
> > In glibc 2.0 (or whatever came on Red Hat 6.2), it did pick it up.
> > In glibc 2.1 (or whatever came on Woody), its gone.
>
> Add the appropriate feature test macros (details are in the libc
> manual), e.g. _BSD_SOURCE.
Ok, read the entry in the manual. Is the cpp symbol _BSD_SOURCE itself a
gnu extension, or part of some portable standard? I.e. which of these is
correct:
#ifndef _BSD_SOURCE
#define _BSD_SOURCE
#include <sys/time.h>
#undef _BSD_SOURCE
#else
#include <sys/time.h>
#endif
or
#ifdef *GLIBC*
#ifndef _BSD_SOURCE
#define _BSD_SOURCE
#include <sys/time.h>
#undef _BSD_SOURCE
#else
#include <sys/time.h>
#endif
#else
#include <sys/time.h>
#endif
where *GLIBC* is some CPP symbol that I don't know yet, that that tells if
using glibc.
> > The man page for gettimeofday hasn't changed.
> Then tell it the man page maintainer (CC'ed).
> > Could the removal be a mistake?
>
> No - it's a way to stricter standard conformance.
>
> > If its intentional, how should I have figured that out by myself? :)
>
> Did you look at /usr/include/sys/time.h at all?