X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f
From: Andris Pavenis <pavenis AT latnet DOT lv>
To: djgpp AT delorie DOT com
Subject: Re: Cross Platform Incompatabilites?
Date: Mon, 16 Feb 2004 13:02:18 +0200
User-Agent: KMail/1.6
References: <3 DOT 0 DOT 1 DOT 16 DOT 20040215122529 DOT 34ef6022 AT earthlink DOT net> <2914-Sun15Feb2004193301+0200-eliz AT elta DOT co DOT il>
In-Reply-To: <2914-Sun15Feb2004193301+0200-eliz@elta.co.il>
MIME-Version: 1.0
Content-Disposition: inline
Content-Type: text/plain;
  charset="iso-8859-1"
Content-Transfer-Encoding: 7bit
Message-Id: <200402161302.19040.pavenis@latnet.lv>
Reply-To: djgpp AT delorie DOT com

On Sunday 15 February 2004 19:33, Eli Zaretskii wrote:
> > Date: Sun, 15 Feb 2004 12:25:29 -0500
> > From: Ethan Rosenberg <ethros AT earthlink DOT net>
> >
> > I thought that the problem could be from not including float.h [is this
> > needed anytime a float is used??]
>
> I cannot imagine that including float.h could have any influence on
> the results.  Apart of defining some parameters of the floating-point
> representation and the macros used by _control87, it doesn't do
> anything.  A program that doesn't use _control87 needs float.h only if
> it wants to know the likes of the maximum and minimum FP numbers, the
> machine epsilon, the maximum representable exponent, etc.

GCC (at least 3.3.X) has it own float.h

Some time ago I found this problem (no _control87() available) and put 
	#ifdef __DJGPP__
	#include_text <float.h>
	#endif
in GCC float.h (for DJGPP port only, it's not in FSF sources). It is so in 
both ports of gcc-3.3.1 and gcc-3.3.2.

Andris