From owner-freebsd-current  Sat Mar 30 18:55:39 1996
Return-Path: owner-current
Received: (from root@localhost)
          by freefall.freebsd.org (8.7.3/8.7.3) id SAA06425
          for current-outgoing; Sat, 30 Mar 1996 18:55:39 -0800 (PST)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.7.3/8.7.3) with SMTP id SAA06418
          for <current@FreeBSD.org>; Sat, 30 Mar 1996 18:55:34 -0800 (PST)
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.12/8.6.9) id MAA06978; Sun, 31 Mar 1996 12:52:19 +1000
Date: Sun, 31 Mar 1996 12:52:19 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199603310252.MAA06978@godzilla.zeta.org.au>
To: current@FreeBSD.org, regnauld@tetard.frmug.fr.net
Subject: Re: Weird atof() behavior
Sender: owner-current@FreeBSD.org
X-Loop: FreeBSD.org
Precedence: bulk

>	A  friend of mine  has had  some problems with  Rayshade running on
>	2.0.5 (GCC   2.6.3)   -- it appeared   that  atof()  systematically
>	returned bogus values in certain conditions: if you omit to specify
>	#include <stdlib.h>,  gcc  does NOT issue  any sort  of warning for
>	atof().  Example:

>	#include <stdio.h>

>	main()
>	{
>		foo* char="3.1415926";
>		float bar=atof(foo);
>		printf("%f\n",bar);
>	}
>...
>	The warning does show up with -Wall:
>	
>	atof.c:5: warning: implicit declaration of function `atof'

This looks like a warning to me.  The behaviour is undefined because
atof is implicitly declared as returning `int' but actually returns
`double'.

Bruce