Received: (from major@localhost)
	by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id RAA20787
	for pups-liszt; Wed, 25 Mar 1998 17:44:22 +1100 (EST)
X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups@minnie.cs.adfa.oz.au using -f
Received: from suede.sw.oz.au (firewall-user@gw.softway.com.au [203.31.96.1])
	by minnie.cs.adfa.oz.au (8.8.5/8.8.5) with ESMTP id RAA20782
	for <pups@minnie.cs.adfa.oz.au>; Wed, 25 Mar 1998 17:44:12 +1100 (EST)
Received: from bookworm.softway.com.au (root@bookworm.sw.oz.au [192.41.203.51])
	by suede.sw.oz.au (8.8.8/8.8.8) with SMTP id RAA25023;
	Wed, 25 Mar 1998 17:43:20 +1100 (EST)
Received: by bookworm.softway.com.au (Smail3.1.29.1 #4)
	id m0yHjuC-000FlVC; Wed, 25 Mar 98 17:43 +1000
Message-Id: <m0yHjuC-000FlVC@bookworm.softway.com.au>
Date: Wed, 25 Mar 98 17:43 +1000
From: Peter Chubb <peterc@softway.com.au>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: Greg Lehey <grog@lemis.com>
Cc: PDP UNIX Preservation Society <pups@minnie.cs.adfa.oz.au>
Subject: Re: gzip on PDP-11: not so simple
In-Reply-To: <19980325161754.63486@freebie.lemis.com>
References: <19980325161754.63486@freebie.lemis.com>
X-Mailer: VM 6.35 under Emacs 20.2.2
Comments: Hyperbole mail buttons accepted, v04.023.
X-Face:  .slVUC18R`%{j(W3ztQe~*ATzet;h`*Wv33MZ]*M,}9AP<`+C=U)c#NzI5vK!0^d#6:<_`a
 {#.<}~(T^aJ~]-.C'p~saJ7qZXP-$AY==]7,9?WVSH5sQ}g3,8j>u%@f$/Z6,WR7*E~BFY.Yjw,H6<
 F.cEDj2$S:kO2+-5<]afj@kC!:uw\(<>lVpk)lPZs+2(=?=D/TZPG+P9LDN#1RRUPxdX
Sender: owner-pups@minnie.cs.adfa.oz.au
Precedence: bulk

>>>>> "Greg" == Greg Lehey <grog@lemis.com> writes:

Greg> OK, I've found the problems with gzip, and they're not
Greg> encouraging.  It would appear that the undefined references are
Greg> undefined because they refer to data which is too large.  Here's
Greg> the preprocessor output:

Greg>   uch inbuf[ 0x8000 + 64 ]; uch outbuf[ 16384 +2048 ]; ush
Greg> d_buf[ 0x8000 ]; uch window[ 2*0x8000 ]; # 194 "gzip.c"

You need to decrease the window size -- try setting it to 8k (instead
of 32k)

There should be a 
#define WSIZE 0x8000
somewhere.

It may be worth playing with a decompress only version -- compression
will take more space than decompression (you need two windows rather
than one, for a start).  inbuf can be smaller, too.  Try 512 bytes to
match the disc record size.

Peter C

