Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site petrus.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gamma!epsilon!zeta!sabre!petrus!hammond
From: hammond@petrus.UUCP (Rich A. Hammond)
Newsgroups: net.bugs.usg
Subject: Re: Yet ANOTHER null pointer dereference problem (in "cat")
Message-ID: <627@petrus.UUCP>
Date: Mon, 7-Oct-85 07:01:23 EDT
Article-I.D.: petrus.627
Posted: Mon Oct  7 07:01:23 1985
Date-Received: Tue, 8-Oct-85 04:24:11 EDT
References: <2826@sun.uucp>, <2620@pegasus.UUCP> <6027@utzoo.UUCP>
Organization: Bell Communications Research, Inc
Lines: 18

> > The zero deference bug still exists in my Vr2 code. Does anyone think that
> > for Vr3 they'll turn on the loader option to disallow zero deferences?
> 
> Probably not until they run into the same situation that forced Berkeley
> to do something about it:  a machine which *cannot* support a readable
> location 0.  (Let there be no illusions that Berkeley fixed *NULL problems
> out of nobility or purity of heart; they fixed them because they had no
> choice if they wanted their code to run on the first Suns.)
> -- 
> 				Henry Spencer @ U of Toronto Zoology
> 				{allegra,ihnp4,linus,decvax}!utzoo!henry

The sad thing about all this is that I know of development groups in
AT&T Information Systems that have done the fixes, because they had
machines that didn't have readable location zero's.  However, getting
USG to buy back those modifications (and others for porting) never seemed
to work out.  After all, AT&T IS is a fully separated subsidiary.
Rich Hammond