From owner-freebsd-current  Thu Jul  6 22:03:05 1995
Return-Path: current-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id WAA16900
          for current-outgoing; Thu, 6 Jul 1995 22:03:05 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id WAA16889
          for <current@freebsd.org>; Thu, 6 Jul 1995 22:02:53 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id PAA05751; Fri, 7 Jul 1995 15:01:15 +1000
Date: Fri, 7 Jul 1995 15:01:15 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199507070501.PAA05751@godzilla.zeta.org.au>
To: gordon@sneaky.lonestar.org
Subject: Re: i386/596: \<ctype.h> and \<stdlib.h> conflict with _POSIX_SOURCE
Cc: current@freebsd.org
Sender: current-owner@freebsd.org
Precedence: bulk

>>Synopsis:       <ctype.h> and <stdlib.h> conflict with _POSIX_SOURCE

>>How-To-Repeat:

>	Compile the following 3-line program (cc -c x.c):

>	# define _POSIX_SOURCE
>	# include <stdlib.h>
>	# include <ctype.h>

>	and get a flood of error messages starting with complaints
>	about rune_t (a typedef apparently not defined).

>	however, this works:

>	# define _POSIX_SOURCE
>	# include <ctype.h>
>	# include <stdlib.h>

Oops.  They also work individually, so they should work in any order if
they are idempotent.  They aren't idempotent.  <runetype.h> bases its
decision to declare rune_t (and wchar_t) on the undefined'ness of
_BSD_WCHAR_T_, but <stdlib.h> uses _BSD_WCHAR_T_ to control only wchar_t
if either _POSIX_SOURCE or _ANSI_SOURCE is defined.

Fix: there should be a separate guard _BSD_RUNE_T_ for rune_t.  rune_t
should not be declared if _POSIX_SOURCE is defined and only the above
headers are included, but this is a different bug.  The separate guard
is required to make the following work:

	#define _POSIX_SOURCE
	#include <stdlib.h>
	#include <ctype.h>
	#include <runetype.h>

Bruce