Date: Mon, 28 Sep 1998 16:25:46 +0300 (IDT)
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
To: Boon van der RJ <rjvdboon AT sloep07 DOT cs DOT vu DOT nl>
cc: djgpp AT delorie DOT com
Subject: Re: Help - Floating Point Exception
In-Reply-To: <6unphg$ao8$1@star.cs.vu.nl>
Message-ID: <Pine.SUN.3.91.980928162356.735C-100000@is>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Precedence: bulk


On 28 Sep 1998, Boon van der RJ wrote:

> > I use standards... libm is faster than standard lib?
> 
> No on the contrary, but it is safer to use / sets errno when failing.
> (I hope I say it right, Eli Zaretskii (as always) knows more about
> this than I do)

That's correct: libm.a is slower than libc.a, but more accurate, and more 
standard-complying as far as setting errno is considered.

> > It's strange, I don't understand why it crashed under command.com
> > but it didn't under rhide, for example. 8-)
> 
> Coincidence? (Don't know about this one either.)

It could be that RHIDE itself masks FP exceptions, whereas stock DJGPP 
startup code masks only some of them.