X-Authentication-Warning: delorie.com: mail set sender to djgpp-bounces using -f
From: "Martin Str|mberg (ams AT ludd DOT luth DOT se) [via djgpp AT delorie DOT com]" <djgpp AT delorie DOT com>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Fixing various bugs in frexp.S
Date: Tue, 9 Jan 2018 08:49:05 +0000 (UTC)
Organization: Aioe.org NNTP Server
Lines: 25
Message-ID: <p31vlv$nj2$1@gioia.aioe.org>
References: <5A4946EB DOT 3090500 AT gmx DOT de> <p2vmp0$14s4$1 AT gioia DOT aioe DOT org> <279ab41f-d898-4bcf-b5d4-ad9bf86e7a06 AT googlegroups DOT com>
NNTP-Posting-Host: gOdCqiuC+VwqSgvHBNC1tQ.user.gioia.aioe.org
X-Complaints-To: abuse AT aioe DOT org
User-Agent: tin/2.4.1-20161224 ("Daill") (UNIX) (Linux/4.10.0-32-generic (x86_64))
X-Notice: Filtered by postfilter v. 0.8.2
Bytes: 1675
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Reply-To: djgpp AT delorie DOT com

Juan Manuel Guerrero <juan DOT guerrero AT gmx DOT de> wrote:
> The goal is not to generate a negative NaN but to preserve the sign of the
> imput argument.  The implementation should no convert a -NaN in a +NaN IMO.

That is good. But why fiddle with it at all if it's a NaN? Just return
whatever you got.


( This part is only my own preferences/ideas:
> This way the implementation bahaves as the implementation on my linux box.

I don't like this conforming to <whatever>. _They_ do it like this,
thus _we_ must do it like this.

I _like_ when implementation differs (as long as they are compliant to
whatever standards there are), because then I can use these differing
platforms for testing portabilty.

If DJGPP is just like (e. g.) Linux (... yeah...), it won't potentially
tickle this or that bug (in my code).
)


-- 
MartinS