Re: gzip on PDP-11: not so simple

<Well, I've got uncompress working, but I thought having gunzip would <be good as it gives better compression results. The question is why? Generally compression is a diminishing returns for computational effort with 80% for the first 10% effort. I can see having it if needed to gain access to software and the current platform is the only one. For sim to hardware transfers simple works better... Allison Received: (from major@localhost) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id IAA23282 for pups-liszt; Thu, 26 Mar 1998 08:55:30 +1100 (EST) X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups(a)minnie.cs.adfa.oz.au using -f Received: from henry.cs.adfa.oz.au (henry.cs.adfa.oz.au [131.236.21.158]) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) with ESMTP id IAA23277 for &lt;pups(a)minnie.cs.adfa.oz.au&gt;au>; Thu, 26 Mar 1998 08:55:27 +1100 (EST) Received: (from wkt@localhost) by henry.cs.adfa.oz.au (8.7.5/8.7.3) id IAA03217 for pups(a)minnie.cs.adfa.oz.au; Thu, 26 Mar 1998 08:55:36 +1100 (EST) From: Warren Toomey &lt;wkt(a)henry.cs.adfa.oz.au&gt; Message-Id: &lt;199803252155.IAA03217(a)henry.cs.adfa.oz.au&gt; Subject: Re: Follow-up: oddball versions of Unix Date: Thu, 26 Mar 1998 08:55:36 +1100 (EST) Cc: pups(a)minnie.cs.adfa.oz.au In-Reply-To: &lt;199803252150.NAA10104(a)rainbow.Corp.Sun.COM&gt; from Chris Drake at "Mar 25, 98 01:50:07 pm" Reply-To: wkt(a)cs.adfa.oz.au X-Mailer: ELM [version 2.4ME+ PL22 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-pups(a)minnie.cs.adfa.oz.au Precedence: bulk In article by Chris Drake: Yep, it's in the archive! Warren Received: (from major@localhost) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id JAA23322 for pups-liszt; Thu, 26 Mar 1998 09:04:52 +1100 (EST) X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups(a)minnie.cs.adfa.oz.au using -f Received: from post.mail.demon.net (post-10.mail.demon.net [193.195.0.154]) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) with SMTP id JAA23317 for &lt;pups(a)minnie.cs.adfa.oz.au&gt;au>; Thu, 26 Mar 1998 09:04:44 +1100 (EST) Received: from falstaf.demon.co.uk ([158.152.152.109]) by post.mail.demon.net id aa1007927; 25 Mar 98 21:57 GMT Message-ID: &lt;0S+aPCA11XG1EwK5(a)falstaf.demon.co.uk&gt; Date: Wed, 25 Mar 1998 21:56:05 +0000 To: wkt(a)cs.adfa.oz.au Cc: PDP Unix Preservation &lt;pups(a)minnie.cs.adfa.oz.au&gt; From: Robin Birch &lt;robin(a)falstaf.demon.co.uk&gt; Subject: Re: gzip on PDP-11: not so simple In-Reply-To: &lt;199803252033.HAA03043(a)henry.cs.adfa.oz.au&gt; MIME-Version: 1.0 X-Mailer: Turnpike (32) Version 3.05 <YSkKJATYm0AxAbv6xUlJqScVOx> Sender: owner-pups(a)minnie.cs.adfa.oz.au Precedence: bulk In message &lt;199803252033.HAA03043(a)henry.cs.adfa.oz.au&gt;au>, Warren Toomey &lt;wkt(a)henry.cs.adfa.oz.au&gt; writes I looked at this several years ago and gave up at the save point as Warren. I looked at compress using 16 bits and hit the same sort of constructs. After a bit of thinking I believe there may be a way round it but at the time I didn't know the algorithms used in compress or gzip so didn't try playing. The problem is that the compression algorithm needs a 64k space to do all of its sums in, don't ask me why, if someone could tell us the algorithm them I would understand a lot better. These are defined as 64k address spaces which the data page isn't holding cos they don't fit. If you write a virtual mem system then this will work. This causes problems in the standalone world obviously but steve wrote a vm lookalike for 2.11 that uses files, yes a lump of real mem aka the partition concept with movable windows in RSX would be nice but we can't have everything, but compress and maybe gip should be able to be cooked into using such a system for vm. This would be slow but what are we after?, an all singing all dancing system or something that would work in the background whilst we get a beer and wait for the system to install?. Cheers Robin Robin Birch robin(a)falstaf.demon.co.uk M1ASU/2E0ARJ Old computers and radios always welcome Received: (from major@localhost) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id JAA23342 for pups-liszt; Thu, 26 Mar 1998 09:07:50 +1100 (EST) X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups(a)minnie.cs.adfa.oz.au using -f Received: from henry.cs.adfa.oz.au (henry.cs.adfa.oz.au [131.236.21.158]) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) with ESMTP id JAA23337 for &lt;pups(a)minnie.cs.adfa.oz.au&gt;au>; Thu, 26 Mar 1998 09:07:46 +1100 (EST) Received: (from wkt@localhost) by henry.cs.adfa.oz.au (8.7.5/8.7.3) id JAA03305 for pups(a)minnie.cs.adfa.oz.au; Thu, 26 Mar 1998 09:07:35 +1100 (EST) From: Warren Toomey &lt;wkt(a)henry.cs.adfa.oz.au&gt; Message-Id: &lt;199803252207.JAA03305(a)henry.cs.adfa.oz.au&gt; Subject: Re: gzip on PDP-11: not so simple To: pups(a)minnie.cs.adfa.oz.au (PDP Unix Preservation) Date: Thu, 26 Mar 1998 09:07:35 +1100 (EST) In-Reply-To: &lt;0S+aPCA11XG1EwK5(a)falstaf.demon.co.uk&gt; from Robin Birch at "Mar 25, 98 09:56:05 pm" Reply-To: wkt(a)cs.adfa.oz.au X-Mailer: ELM [version 2.4ME+ PL22 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-pups(a)minnie.cs.adfa.oz.au Precedence: bulk In article by Robin Birch: [ not being able to run gzip on a PDP-11 ] You're right I think. At least compress -b12 works, and as you say, a bit of extra wait isn't going to hurt too much. Peter Chubb seems interested in fitting gunzip into 64K. I'll see how he goes with it. Thanks all for your comments, Warren Received: (from major@localhost) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id JAA23554 for pups-liszt; Thu, 26 Mar 1998 09:29:54 +1100 (EST) X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups(a)minnie.cs.adfa.oz.au using -f Received: from henry.cs.adfa.oz.au (henry.cs.adfa.oz.au [131.236.21.158]) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) with ESMTP id JAA23549 for &lt;pups(a)minnie.cs.adfa.oz.au&gt;au>; Thu, 26 Mar 1998 09:29:50 +1100 (EST) Received: (from wkt@localhost) by henry.cs.adfa.oz.au (8.7.5/8.7.3) id JAA03395 for pups(a)minnie.cs.adfa.oz.au; Thu, 26 Mar 1998 09:30:00 +1100 (EST) From: Warren Toomey &lt;wkt(a)henry.cs.adfa.oz.au&gt; Message-Id: &lt;199803252230.JAA03395(a)henry.cs.adfa.oz.au&gt; Subject: Available: tool to write disk images to PDP-11 To: pups(a)minnie.cs.adfa.oz.au (PDP Unix Preservation) Date: Thu, 26 Mar 1998 09:30:00 +1100 (EST) Reply-To: wkt(a)cs.adfa.oz.au X-Mailer: ELM [version 2.4ME+ PL22 (25)] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-pups(a)minnie.cs.adfa.oz.au Precedence: bulk Ok, I debugged the thing yesterday, it works well. If you want to write a PDP-11 disk image to a real PDP-11, you might like to look in: ftp://minnie.cs.adfa.oz.au/pub/PDP-11/Vtserver and at the file zcat.README there. Current disk and tapes supported: hp: RP04, RP05 and RP06 disks. rp: RP03 disks. rk: RK05 disks. rl: RL01 and RL02 disks. ht: TU16 or TE16 tape drive. tm: TU10 tape drive. vt: The Virtual Tape drive. You can download from any tape to any disk. The Virtual Tape drive allows you to download the image over a KL11 at 9,600 baud. Any type of disk image can be downloaded, not just Unix ones. You will need compress(1). And a bit of patience. Let's hope someone tries this out! Ciao, Warren P.S I plan on migrating to the 2.11BSD standalone stuff, which supports more tape drives and disk drives. Sometime. Received: (from major@localhost) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id OAA24607 for pups-liszt; Thu, 26 Mar 1998 14:21:30 +1100 (EST) X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups(a)minnie.cs.adfa.oz.au using -f Received: from suede.sw.oz.au (firewall-user(a)gw.softway.com.au [203.31.96.1]) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) with ESMTP id OAA24602 for &lt;pups(a)minnie.cs.adfa.oz.au&gt;au>; Thu, 26 Mar 1998 14:21:23 +1100 (EST) Received: from bookworm.softway.com.au (root(a)bookworm.sw.oz.au [192.41.203.51]) by suede.sw.oz.au (8.8.8/8.8.8) with SMTP id OAA14866 for &lt;pups(a)minnie.adfa.oz.au&gt;au>; Thu, 26 Mar 1998 14:21:17 +1100 (EST) Received: by bookworm.softway.com.au (Smail3.1.29.1 #4) id m0yI3EB-000FlbC; Thu, 26 Mar 98 14:21 +1000 Message-Id: &lt;m0yI3EB-000FlbC(a)bookworm.softway.com.au&gt; Date: Thu, 26 Mar 98 14:21 +1000 From: Peter Chubb &lt;peterc(a)softway.com.au&gt; To: pups(a)minnie.cs.adfa.oz.au Subject: Progress on zcat 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 {#.&lt;}~(T^aJ~]-.C&#x27;p~saJ7qZXP-$AY==]7,9?WVSH5sQ}g3,8j&gt;u%(a)f$/Z6,WR7*E~BFY.Yjw,H6&lt; F.cEDj2$S:kO2+-5<]afj@kC!:uw\(<>lVpk)lPZs+2(=?=D/TZPG+P9LDN#1RRUPxdX Sender: owner-pups(a)minnie.cs.adfa.oz.au Precedence: bulk Well... my cut-down zcat now works under Linux, and compiles and links cleanly under v7 on the simulator. But the semantics are wrong! Big problem is the lack of unsigned char and unsigned long types. I'm gradually going through and finding places where left shifts, or sign extensions are happening, and masking them explicitly. I'm almost sure that at UNSW we had a C compiler on Unix V7 that had an unsigned long data type... Anyway, there's progress. And if it all goes OK, then on machines that have separate I&D spaces, the resulting zcat will be compatible with gzip everywhere. Peter C Received: (from major@localhost) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id OAA29445 for pups-liszt; Fri, 27 Mar 1998 14:51:52 +1100 (EST) X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups(a)minnie.cs.adfa.oz.au using -f Received: from renoir.op.net (root(a)renoir.op.net [209.152.193.4]) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) with ESMTP id OAA29440 for &lt;pups(a)minnie.cs.adfa.oz.au&gt;au>; Fri, 27 Mar 1998 14:51:43 +1100 (EST) Received: from goppelt.op.net (d-phlarc1-04.ppp.op.net [209.152.199.68]) by renoir.op.net (o1/$Revision: 1.15 $) with SMTP id WAA01451 for &lt;pups(a)minnie.cs.adfa.oz.au&gt;au>; Thu, 26 Mar 1998 22:51:32 -0500 (EST) Message-Id: &lt;199803270351.WAA01451(a)renoir.op.net&gt; Comments: Authenticated sender is &lt;edgee(a)cyberpass.net&gt; From: "Ed G." &lt;edgee(a)cyberpass.net&gt; To: pups(a)minnie.cs.adfa.oz.au Date: Thu, 26 Mar 1998 22:51:31 -0400 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: Bug in Bob Supnik's Emulator! Reply-to: edgee(a)cyberpass.net X-mailer: Pegasus Mail for Win32 (v2.54) Sender: owner-pups(a)minnie.cs.adfa.oz.au Precedence: bulk As you know, I wrote this list recently about a bug in Bob Supnik's emulator which manifests when running factor (1). I've had a chance to do some further research on this, and it seems to me that the bug is definitely in Supnik's emulator. What do you all think? Am I onto something? If so, what part of Supnik's code is probably to blame? Here's what I've learned so far: 1. factor on Supnik's emulator fails most of the time (see below for examples). 2. factor works fine on Ersatz-11 2. On the off-chance that I munged the disk images and somehow corrupted factor, I reextracted virgin images from the tar ball. factor still fails while running on Supnik's emulator. 3. Peter Turnbull wrote me that factor running on under uv7 on his PDP-11/23 runs the test case 'factor 6' without error. Here's what factor does on Supnik's emulator for a variety of values: factor 6 2 3 17 17 etc. factor 257 263 263 etc. factor 263 269 269 etc. factor 1009 (works correctly) 1009 Received: (from major@localhost) by minnie.cs.adfa.oz.au (8.8.5/8.8.5) id RAA29824 for pups-liszt; Fri, 27 Mar 1998 17:28:35 +1100 (EST) X-Authentication-Warning: minnie.cs.adfa.oz.au: major set sender to owner-pups(a)minnie.cs.adfa.oz.au using -f