Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: notesfiles
Path: utzoo!watmath!clyde!floyd!harpo!decvax!ucbvax!ucbcad!ucbesvax.turner
From: turner@ucbesvax.UUCP
Newsgroups: net.flame
Subject: Re: berkeley cretins
Message-ID: <2900057@ucbesvax.UUCP>
Date: Fri, 17-Feb-84 20:36:00 EST
Article-I.D.: ucbesvax.2900057
Posted: Fri Feb 17 20:36:00 1984
Date-Received: Wed, 29-Feb-84 11:22:33 EST
References: <4344@amd70.UUCP>
Organization: UC Berkeley, EE/SESM
Lines: 23
Nf-ID: #R:amd70:-434400:ucbesvax:2900057:000:1271
Nf-From: ucbesvax!turner    Feb 25 17:36:00 1984

> /***** ucbesvax:net.flame / amd70!phil / 12:49 am  Feb 25, 1984*/
> Has anyone ever told Berkeley that referencing the contents of a NULL
> pointer does not always yield 0?
> I know, this is a dead horse I'm beating. But damn, how many times do
> I have to get bitten by this??????

Much of Berkeley's Franz Lisp was predicated on the ability to read out
a zero from (cons *) NULL.  Keith Sklower was rather peeved that we (at
Dual Systems, when I was there) were unable to rearrange the memory map
of our systems to Franz's convenience.  (But he ported it anyway.)  This
is the only software I know of where this feature of VAXen was knowingly
incorporated in the design.

Phil will continue to be bitten by the dead horse he is beating here, simply
because the "read 0 from location 0" is usually a bug in the code where it
crops up.  It's just that it you never trip over it on VAXen, so those bugs
must wait for a different machine on which to manifest themselves.  The code
for string routines has not, to my knowledge, been adjusted.  (E.g., strlen
((char *) 0) returns zero, without complaint AND WITHOUT CHECKING FOR A NULL
ARGUMENT.  That's just stupid.)  But other cases will have be taken care of
as they come up.
---
Michael Turner (ucbvax!ucbesvax.turner)