Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!ll-xn!cit-vax!mangler
From: mangler@cit-vax.Caltech.Edu (System Mangler)
Newsgroups: comp.unix.wizards,comp.arch
Subject: Re: brk's zero-fill behavior on VAXen (useful undefined checks)
Message-ID: <1185@cit-vax.Caltech.Edu>
Date: Sat, 15-Nov-86 04:28:01 EST
Article-I.D.: cit-vax.1185
Posted: Sat Nov 15 04:28:01 1986
Date-Received: Sat, 15-Nov-86 23:53:30 EST
References: <7208@elsie.UUCP> <5142@brl-smoke.ARPA> <2447@hcr.UUCP> <160@haddock.UUCP>
Organization: California Institute of Technology
Lines: 12
Summary: MOVD checks for NaN
Xref: mnetor comp.unix.wizards:72 comp.arch:21

In article <160@haddock.UUCP>, karl@haddock.UUCP (Karl Heuer) writes:
> P.S. I like the way VAXen do this with floating point (the result of some
> operation is NaN, which triggers an FPE on the *next* usage, giving the user
> a chance to test for it first).

I inherited an interpreter that uses this trick.  Checking for a NaN was
done by a call-by-reference assembly language routine; when I rewrote it
as a call-by-value C routine, I got floating point exceptions.	Why?  The
MOVD instruction that pushes the value on the stack will raise an FPE if
asked to push a NaN!  You have to be very careful how you look at those...

Don Speck   speck@vlsi.caltech.edu  {seismo,rutgers,ames}!cit-vax!speck