Newsgroups: comp.os.msdos.djgpp
From: Peter Berdeklis <peter AT atmosp DOT physics DOT utoronto DOT ca>
Subject: Re: Floating Point Exception 
Message-ID: <Pine.SGI.3.91.970827172731.22048A-100000@atmosp.physics.utoronto.ca>
Nntp-Posting-Host: chinook.physics.utoronto.ca
Sender: news AT info DOT physics DOT utoronto DOT ca (System Administrator)
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Organization: University of Toronto - Dept. of Physics
In-Reply-To: <5tspej$7gt$1@news.IAEhv.nl> 
Date: Wed, 27 Aug 1997 21:32:29 GMT
References: <01bcabc3$a01da100$5af382c1 AT damien> <33F8549A DOT 4CB5 AT cornell DOT edu> <01bcac2d$edf69fc0$LocalHost AT damien> <33FACBA4 DOT 1DFF7DBB AT Mathematik DOT TU-Chemnitz DOT DE> <33FAD7C8 DOT 72E7 AT chem DOT uw DOT edu DOT pl> <5tspej$7gt$1 AT news DOT IAEhv DOT nl> 
Lines: 25
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

On 25 Aug 1997, Jos Bergervoet wrote:

> There is actually no need to prevent division by zero. It should
> produce Inf or Nan, and often your computation can go on and 
> still deliver good results. DJGPP doesn't fail on this:
> 
>     double a=1, b=0, c;
>     c = a/b;      /* now c=Inf */
>     c = 1/c;      /* now c=1/Inf = 0, which is correct */
>     printf( "Hello: result = %10.4g\n", c );

On an INTEL PC a divide by zero will cause a hardware exception that will 
cause the processor to halt.  It will return Inf or NaN only if you 
install an exception handler that returns that value when the processor 
flags the exception.

On some other machines (namely the SGI that I work on) the processor will
return an Inf or NaN by itself, but the state of the processor is such 
that it will produce a steady stream of garbage afterwards.  Personally
I prefer a simple exception, thank you very much.

---------------
Peter Berdeklis
Dept. of Physics, Univ. of Toronto