TUHS

tuhs@tuhs.org

6 participants
6532 discussions

Linux displacing UnixWare on the PC architecture

by Josh Good

On 2017 Mar 9, 21:26, Josh Good wrote: > > And by the way, the two user limit in the "Personal Edition" of UnixWare > 2.1 seems to be real: > > $ telnet 172.27.101.128 > Trying 172.27.101.128... > Connected to 172.27.101.128. > Escape character is '^]'. > > > UnixWare 2.1 (gollum1) (pts/2) > > login: jgood > Password: > UnixWare 2.1 > gollum1 > Copyright 1996 The Santa Cruz Operation, Inc. All Rights > Reserved. > Copyright 1984-1995 Novell, Inc. All Rights Reserved. > Copyright 1987, 1988 Microsoft Corp. All Rights Reserved. > U.S. Pat. No. 5,349,642 > Last login: Tue Mar 9 20:57:05 1999 on pts000 > telnetd: set_id() failed: Too many users > . > Connection closed by foreign host. > > > This thing was released in 1996. Obviously, with this limitation it could > not hold a candle to the emerging Linux tsunammi full of free source code. On the subject of Linux displacing UnixWare on the PC architecture in the mid-90's, I've found this most illuminating Usenet thread from 1994, whose participants include Alan Cox, Theo Tso, and some Novell Product Managers: http://tech-insider.org/linux/research/1994/1025.html And what came after that, as they say, is history. -- Josh Good

8 years, 3 months

Help request: uucp, mail on 4.2BSD

by Warren Toomey

Hi all, as part of my effort to recreate part of a simulated Usenet, I'm trying to bring up uucp, then mail, then C-news on 4.2BSD boxes. I've got a hardwired serial port between them, and I can see a basic uucp conversation when I do this: munnari.oz# /usr/lib/uucp/uucico -r1 -sseismo -x7 uucp seismo (3/6-8:04-132) DEBUG (ENABLED) . . . uucp seismo (3/6-8:04-132) SUCCEEDED (call to seismo ) imsg >\015\012\020< Shere\000imsg >\020< ROK\000msg-ROK Rmtname seismo, Role MASTER, Ifn - 5, Loginuser - uucp . . . I tried e-mail to seismo!wkt and wkt(a)seismo.UUCP but it's been deferred. I now need some help with the sendmail config. I did play around with sendmail.cf/mc way back, but it never involved uucp so I'm stuck. Anybody want to help (and dust out those cobwebs at the same time)? Thanks, Warren

8 years, 4 months

C-news is built

by Warren Toomey

OK, Geoff Collyer has built the C-News binaries for the 4.2 emulated systems. They are temporarily at http://minnie.tuhs.org/Y5/Cnews/ Does someone want to try and get them up and running on an emulated system? Also, I've build a 4.3BSD version of the emulated uucp systems. It's a separate branch at https://github.com/DoctorWkt/4bsd-uucp. You can get it by doing: git clone https://github.com/DoctorWkt/4bsd-uucp.git \ --branch 4.3BSD --single-branch Once it's solid enough I will make this the default branch, but I'll leave the 4.2BSD branch there as well. Thanks Geoff! Warren

8 years, 4 months

Re: [TUHS] "Open" Stanford University Network System, et al [was RT/PC-centric AIX history]

by Clem Cole

On Fri, Mar 10, 2017 at 8:15 AM, Jason Stevens < jsteve(a)superglobalmegacorp.com> wrote: > > That almost reminds me to ask about the whole "open" Stanford 68000 board > that became the Cisco AGS, and SUN 100.. and I think SGi 1000 > Jason -- I'm not sure what you are trying to say. It was a different time, different culture, different rules. Note: Please I'm not accusing you of this, but I worry you are getting dangerous close to an error that I see made by a lot of folks that grew in the time of the GPL and the "Open Source Culture." My apologies in advance if you think I'm going a little too far, but I want to make something clear that seems to have been lost in time and culture. I do not want to be see as harassing or "shaming" in anyway way. I want to make a point for everyone since the words we use do matter (and I realize I screw them up myself often enough).. I am fairly certain that the "SUN board" - aka the Stanford University Network 68000 board, like UNIX itself was licensed IP. You are correct that the schematics (like the UNIX sources) were well known at the time and "open" in the sense that all of the licenses had them. It was not hard to find papers with a much of the design described. In fact Andy had worked on a similar set of boards when he was a CMU a few years earlier for what we called the "distributed front-end" project (the earlier version was much weaker and had started as Intel chip of sometime which I have forgotten and switched to the 68000 at some point - Phil Karn might remember and even have a copy, I think my copy has been lost to time). Anyway, to build and sell a Multibus board based on Andy's design that he did at Stanford as a grad student, you needed a license from Stanford. You are correct a lot of firms, particularly Cisco, later VLSI Technology - ney Sun Micro Systems, Imagen, and host of took out licenses to build that board. Thus a lot of companies built "JAWS" (just another workstation - so called "3M systems" with a disk), or sometimes diskless terminals as Andy had imagined it in his papers, or purpose built boxes such the AGS router and the Imagen printers. But I flinch a little when I see people call the "SUN" an "open" design. It was "well know" but it was not what we might call "Free and Open" today. I admit you just said "open" in your reply to Charlie and may have meant something different; but so many people today leave the "free" off when they say "open." *i.e.* People often incorrect deny that Unix was open as it actually always was from the beginning -- if you had a license, it just was not "free" to get same. My point is that I believe a license for the "SUN" was from Stanford was not "free" either. Same with the the "MIPS" chip technology of a few years later also from Stanford. So, I would have been happier if you had said something that had included the words "licensed from Stanford." Anyway, Research Universities, such as MIT, Stanford and frankly my own CMU, have long been known for charging for licenses (not always mind you). In fact, I laud my other institution, because I have always said the real father of "free and open source" is my old thesis advisor, the late Don Pederson. In the late 1960s, he founded the UCB EE "Industrial Liaison Program" which was the auspicious institution that original "BSD" tape would be released years later. When he first released the first version of "Simulation Program for Integrate Circuit Evaluation" - aka SPICE, in approx 67 time frame "dop" said: *"I always have given away our work. It means we get to go in the back door and talk to the engineers. My colleagues at some of the other places license there work and they have go in the front door like any other salesman."* When the CS group was added to EE a few years later, their was history, mechanism, etc. Berkeley had been release source code for a lots of different project. The Berkeley Software Distribution for Unix V6 was just the the drop for UNIX - who knew at the time the life it wold spawn (although I note SPICE is still being used, so even with UNIX's success, SPICE still hold the record for the "longest" used" BSD release code). Anyway, " do p" used to love to remind the students of that mantra. And he came up with it 20-25 years before Eric Raymond ever wrote his book and started equating "open" with "Stallmanism." ;-) I hope have a great one, and I hope I did not offend. Clem

8 years, 4 months

uucp setup

by William Pechter

One note for those who've been away from 4.x for a while... If you're using a console window for editing and you just wonder why the full screen of the VT100 doesn't show up -- it's because the getty is set down at 1200 baud for the good old LA120 DECwriter III. Set /etc/ttys to 18console or 12console and it's expects 9600baud and then vi will let you use full screen to edit. Been a while since I ran a fake Vax under Unix. Bill

8 years, 4 months

Re: [TUHS] RT/PC-centric AIX history

by jnc＠mercury.lcs.mit.edu

> From: Jason Stevens > it also appears that AOS was the router backbone of the NSFNet once > they started to migrate off of the IMPs Say what? IMPs were only every used in the ARPANET (and networks built by BBN for private clients using that technology). The first routers used in the NSFNET were things called Fuzzballs - PDP-11's running software from Dave Mills, driving 56KB lines. They eventually decided they needed to step up a level, and a consortium involving IBM won, with IBM RT PC's running AIX driving T1 lines. Noel

8 years, 4 months

RT/PC-centric AIX history

by Charles H Sauer

I've refrained from jumping into AIX & RT/PC discussions on TUHS. It seems more appropriate to summarize AIX history than try to correct or clarify specifics out of context. I wrote about 5 pages, got feedback, revised accordingly, and posted at https://notes.technologists.com/notes/2017/03/08/lets-start-at-the-very-beg…. Charlie

8 years, 4 months

Re: [TUHS] UUCP: working systems, come and get them

by Warren Toomey

On Thu, Mar 09, 2017 at 01:57:05PM +0100, Lars Brinkhoff wrote: > Is it ok to do experimental testing with that host? I've never set up > uucp, so I do not yet know quit what I'm doing. Neither have I! But yes, feel free. In yur SimH .ini file, put (or change) this line to say: attach dz line=0,Connect=simh.tuhs.org:5000 which will connect /dev/tty00 to simh.tuhs.org port 5000. Then set up your L.sys file with a line that says: seismo Any;9 DIR 9600 tty00 "" "" ogin:--ogin:--ogin: uucp ssword: uucp so that the uucp site seismo can be contacted via /dev/tty00. Then you can try doing: # echo hello there | mail seismo\!root <wait a few seconds> # /usr/lib/uucp/uucico -r1 -sseismo -x7 and you should see the debug information with parts of the uucp conversation. Cheers, Warren

8 years, 4 months

Re: [TUHS] UUCP: working systems, come and get them

by Cory Smelosky

On Thu, Mar 9, 2017, at 18:02, Warren Toomey wrote: > On Thu, Mar 09, 2017 at 05:57:00PM -0800, Cory Smelosky wrote: > > # grep -i "tty00" /etc/ttys > > #tty00 "/etc/getty std.9600" unknown on secure > > > > # tip dialer > > /dev/tty00: Permission denied > > You need to chown uucp /dev/tty00 I think! > > Cheers, Warren > Email had 1 attachment: > + signature.asc > 1k (application/pgp-signature) That did it. -- Cory Smelosky b4(a)gewt.net

8 years, 4 months

Re: [TUHS] UUCP: working systems, come and get them

by Warren Toomey

On Thu, Mar 09, 2017 at 04:01:09PM -0700, John Floren wrote: > Well, I'm trying to set up lanl-a, it's at 199.180.255.235:6666 > (theoretically). I've set it up to point at seismo but uucico hangs > waiting for the login prompt. OK, try this: Edit your /etc/remote file to say this for dialer: dialer:dv=/dev/tty00:br#9600: Now try: # tip dialer which should connect out over /dev/tty00 to seismo via the TCP connection. Hit Return a few times to see if there is any response. On your host system, do netstat -a | grep ESTAB and see if there is a TCP connection to simh.tuhs.org:5000. I also forgot. To be able to send e-mail, you need to add seismo to the list of known remote sites in /usr/lib/sendmail.cf: CWseismo Cheers, Warren

8 years, 4 months

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

2003

2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

1992

1991

1990

TUHS