X-Authentication-Warning: acp3bf.physik.rwth-aachen.de: broeker owned process doing -bs
Date: Tue, 28 Mar 2000 13:41:28 +0200 (MET DST)
From: Hans-Bernhard Broeker <broeker AT physik DOT rwth-aachen DOT de>
X-Sender: broeker AT acp3bf
To: djgpp-workers AT delorie DOT com
Subject: Re: Unnormals???
In-Reply-To: <200003272014.WAA11659@father.ludd.luth.se>
Message-ID: <Pine.LNX.4.10.10003281335360.2940-100000@acp3bf>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Reply-To: djgpp-workers AT delorie DOT com
Errors-To: dj-admin AT delorie DOT com
X-Mailing-List: djgpp-workers AT delorie DOT com
X-Unsubscribes-To: listserv AT delorie DOT com
Precedence: bulk

On Mon, 27 Mar 2000, Martin Str|mberg wrote:

> > in situations like printf("%Lf"). But in the definition space provided by
> > C99, we have only limited choices what to fpclassify() an unnormal as:
> > 
> > 	infinite
> > 	NaN
> > 	normal
> > 	subnormal
> > 	zero
> 
> That's not true. In 7.12, about <math.h> it says:
> The macros
> 	FP_INFINITE
> 	FP_NAN
> 	FP_NORMAL
> 	FP_SUBNORMAL
> 	FP_ZERO
> are for number classification. ... Additional implementation-defined
> floating-point classifications, ..., may also be specified by the
> implementation."

Seems they must have changed that really late in the process of 
drafting the C99 standard. There is no such passage in the C99 draft
I have here, dated January 18, 1999. To quote it verbatim:

       [#6] The macros

               FP_INFINITE
               FP_NAN
               FP_NORMAL
               FP_SUBNORMAL
               FP_ZERO

       are for number classification.  They represent the  mutually
       exclusive  kinds  of  floating-point values.  They expand to
       integer constant expressions with distinct values.

       [#7] The macro

               FP_FAST_FMA
[...]

No mention of additional implementation-defined classes. Neither here,
nor in the paragraph on fpclassify() itself, in 7.12.3.1:

       [#2]  The  fpclassify macro classifies its argument value as
       NaN,  infinite,  normal,  subnormal,  or  zero.   First,  an
       argument  represented  in  a  format wider than its semantic
       type is converted to its semantic type.  Then classification
       is based on the type of the argument.184)

What's the date and status of the C99 document you're citing? Is it by any
chance the final official one, already? I would have got one, myself, but
without a credit card, it's hard to shop overseas via the internet...

Hans-Bernhard Broeker (broeker AT physik DOT rwth-aachen DOT de)
Even if all the snow were burnt, ashes would remain.