delorie.com/archives/browse.cgi   search  
Mail Archives: djgpp-workers/1999/01/20/02:07:16

Date: Wed, 20 Jan 1999 09:05:52 +0200 (IST)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
X-Sender: eliz AT is
To: Stephen L Moshier <moshier AT mediaone DOT net>
cc: djgpp-workers AT delorie DOT com
Subject: Re: Bug when printing long doubles
In-Reply-To: <Pine.LNX.4.05.9901191949300.16986-100000@moshier.ne.mediaone.net>
Message-ID: <Pine.SUN.3.91.990120090527.2569D-100000@is>
MIME-Version: 1.0
Reply-To: djgpp-workers AT delorie DOT com

On Tue, 19 Jan 1999, Stephen L Moshier wrote:

> Yes, IEEE says for single and double precision a NaN has the maximum
> exponent value and any nonzero fraction.

Exactly.  And the bit pattern in the example we had didn't match this
definition.

> Extended formats are required to have at least one quiet and one
> signaling NaN, but the bit patterns are not specified.  IEEE further says
> "An implementation may also reserve some bit strings for purposes
> beyond the scope of this standard" (IEEE 754-1985 para 3.3).

I think, ultimately, we must be compatible with the processor that
runs the code (without violating standards, of course).  For this
reason, it is IMHO wrong to print "NaN" for something the processor
doesn't consider as such.  We need to find a different string.

- Raw text -


  webmaster     delorie software   privacy  
  Copyright © 2019   by DJ Delorie     Updated Jul 2019