Date: Thu, 23 Apr 1998 13:16:17 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Paul Derbyshire <pderbysh AT usa DOT net>
cc: djgpp AT delorie DOT com
Subject: Re: Spotting NaNs
In-Reply-To: <353EF0E0.36E61729@usa.net>
Message-ID: <Pine.SUN.3.91.980423130834.9918A-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On Thu, 23 Apr 1998, Paul Derbyshire wrote:

> Impatient, I wrote my own only to get a warning about an ambiguous overload,
> which pointed me to an isnan in math.h. Oddly, it seems someone forgot to
> document this thing in the info files.

Well, it took DJGPP users about 4 years to find out, but finally someone 
said it.  Here's the scoop: libm.a has NO DOCS WHATSOEVER!  Why?  Because 
libm.a was lifted from Cygnus work, and they didn't provide any docs at 
the time.

So it's much more serious than just `isnan'.  The only math functions 
which are documented in libc.info are ANSI functions included with 
libc.a, and `isnan' isn't one of them.

The latest versions of Cygnus libm do have docs, but it isn't clear
whether DJGPP v2.02 will have libm upgraded (it needs some work to make it
compatible with DJGPP library build procedure). 

As usual, volunteers are welcome to sit down and contribute the missing 
work.