From owner-linux-activists@joker.cs.hut.fi Tue Nov 24 01:57:33 1992
Status: RO
X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil]
["6091" "Tue" "24" "November" "1992" "01:45:16" "+0200" "Dan Shearer"
"ccdps@lux.levels.unisa.edu.au " nil "125" "Pathworks Server for Unix available"
"^From:" nil nil "11"])
Received: from joker.cs.hut.fi by hutcs.cs.hut.fi with SMTP id AA14953
(5.65c8/HUTCS-S 1.4 for ); Tue, 24 Nov 1992 01:57:28 +0200
Received: from joker.cs.hut.fi by niksula.hut.fi id <61726-2>;
Tue, 24 Nov 1992 01:57:12 +0200
Received: from Levels.UniSA.Edu.Au ([130.220.16.10]) by niksula.hut.fi with SMTP id
<61786-2>; Tue, 24 Nov 1992 01:56:41 +0200
Received: from lux.levels.unisa.edu.au by Levels.UniSA.Edu.Au (PMDF #2428 ) id
<01GRIYH18O289S3REI@Levels.UniSA.Edu.Au>; Tue, 24 Nov 1992 10:16:08 +1030
Received: by lux.levels.unisa.edu.au (4.1/SMI-4.1) id AA08287 for
linux-activists@niksula.hut.fi
Sender: owner-linux-activists@joker.cs.hut.fi
X-Note1: Remember to put 'X-Mn-Key: normal' to your mail body or header
Illegal-Object: Syntax error in To: address found on niksula.hut.fi:
To: linux-activists@niksula.hut.fi(Linux (Remember X-Mn-Key !)))
^-missing closing ')' in token
Message-Id: <9211232345.AA08287@lux.levels.unisa.edu.au>
X-Envelope-To: linux-activists@niksula.hut.fi
Content-Type: text
Content-Transfer-Encoding: 7BIT
X-Mailer: ELM [version 2.4 PL6]
Content-Length: 6089
From: ccdps@lux.levels.unisa.edu.au (Dan Shearer)
To: linux-activists@joker.cs.hut.fi
Subject: Pathworks Server for Unix available
Date: Tue, 24 Nov 1992 01:45:16 +0200
X-Mn-Key: NET
I have had a lot of enquiries about the free pathworks server after I followed
up a comment on the list. Alan Cox has tracked
down the author and I have got the full story from him. He is Andrew Tridgell,
at the Australian National University.
Summary
- A free Pathworks server exists, and I have made it available for anon ftp
in the directory lux.levels.unisa.edu.au:pub/pathworks.
- It implements NetBIOS, on top of sockets.
- It only took about two weeks to write.
- He no longer has a need for it, and so has stopped development.
--
Dan Shearer email: Dan.Shearer@UniSA.edu.au
Information Technology Branch Phone: +61 8 302 3479
University of South Australia Fax : +61 8 302 3385
In Andrew's words (compiled from several messages to me) :
Yes, I did write a pathworks server for unix. I wrote it because
we were interested in the PC X server offered by Dec but it required
using Pathworks as the net. We used PC-NFS and couldn't get the same
functionality from pathworks because we couldn't mount from non Ultrix
boxes (such as suns). So I wrote a server for a general unix box. I
didn't have a spec so I looked at the packets to and from a dec server
and worked out the structures they used. It works but has a few minor
bugs. I have never fixed them because the dec X server now works with
PC-NFS so our local site is no longer interested in pathworks at all!
Anyway - you are very welcome to the code, I'm never going to
commercialise it so I don't mind who does what with it. I'm not going
to be doing anything with it but I would hate to see all the work go to
a total waste. It would be particularly interesting to write a unix
client (not very hard I think) so you have a complete replacement for
NFS. It should (in theory) work with DECNET as well and should talk to
some other clients that are netmanager compatible (again I haven't tried
as I don't have one).
[another message - Dan]
>From start to finish it took under 2 weeks. Most of that time was spent
staring at packets and trying to find out what the protocol was. I wrote
a short socket program that grabs all packets (to disk) then passes them
on to the destination. It's like a simple sniffer I suppose and it means
I only have to modify /etc/inetd.conf to look at any programs socket
communications. I got the idea from the dec gateway sample code.
As you have probably noticed the actual code is fairly simple. The hard
bit is knowing things like that when answering a dfree request the 43rd
and 44th bytes are the byte swapped result and the 38th byte has to be a
'%'. If you had a copy of the spec I'm sure my code would be hilarious.
That '%' has got to mean something - I'm just buggered if I can figure
out what. By the end of it I got the feeling that the spec was generated
with a random number generator.
[another message - Dan]
> 1. Did you not have a spec because no-one would give you one or because
> you couldn't find one easy enough? (ie, if someone wanted to make sure
> it really was legal to put on say, a Decnet network, would it be possible
> to say "yes, because it conforms to XYZ.")
The real problem was that I didn't know there was a spec! I had assumed
that it was a proprietry DEC protocol and only after I released it did
someone say "hey! you've implemented netbios!". I dealt only with the
bytes I saw in the dec server, not knowing it was copying a microsoft
spec.
I have checked with dec and they said what I did was legal - because
there was a published spec (even if I didn't have it).
>
> 2. What lots and *lots* of people want is a way of doing printer/file
> sharing for PCs, most commonly Macs and IBM clones. How do you see your
> efforts fitting in with this aim? In particular people/institutions
> would like to have royalty-free servers accessible with their existing
> workstations. We already have a free, reliable Unix working on cheap
> hardware (Linux) and its ethernet interface is coming together. The
> concept of putting Lan Manager (=Pathworks, almost) on top of this is
> very attractive indeed.
I've no idea about print serving, but it wouldn't be hard to write a dos
client for this for file sharing. All you need is a socket library and a
few dos type changes (\ for / etc). The NCSA socket lib would probably
do the trick. I have used my server for quite a while on a Sun, using
the DEC PC client and mounting serveral remote disks.
My server should certainly work with mininal changes (if any) on almost
any unix implementation that has sockets. The headache with writing a
dos client would be the memory resident stuff, as it needs to reliably
access the disk. Of course this wouldn't be a problem with Linux.
> 3. I understand (from Alan Cox in the UK, who has run your server) that
> you have implemented NetBIOS, what other protocols can/do you implement
> in your code? How do they interface with Unix?
>
If only I knew it was netbios when I wrote it! I might have been able to
get a spec. I suspect that I'm missing many calls of the spec - I only
implemented the ones I saw the dec server using. Adding new calls is
pretty easy though. I use sockets as the communication method. This
allows decnet to be used as it uses basically the same syntax (for
simple things like I do). I've no idea about other transports or
protocols. One problem is that I know next to nothing about networking,
I just know that a socket is like a file and write to it like a file.
> However I'd like to know which is the best ftp site. Once I understand
> a little bit more about it I will let other potential developers/users know.
> We'll see what happens. If it catches the imagination of enough people
> some development may well take place.
I just checked and the machine I was distributing it on has gone! I
attach a uuencoded version below.
[which I have made available for anonymous ftp - Dan]
Cheers,
Andrew
From owner-linux-activists@joker.cs.hut.fi Tue Nov 24 12:57:30 1992
Status: RO
X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil]
["486" "Tue" "24" "November" "1992" "12:39:55" "+0200" "Alan Cox"
"iiitac@pyr.swan.ac.uk" nil "12" "Re: Pathworks Server for Unix available"
"^From:" nil nil "11"])
Received: from joker.cs.hut.fi by hutcs.cs.hut.fi with SMTP id AA18373
(5.65c8/HUTCS-S 1.4 for ); Tue, 24 Nov 1992 12:57:26 +0200
Received: from joker.cs.hut.fi by niksula.hut.fi id <61768-4>;
Tue, 24 Nov 1992 12:56:54 +0200
Received: from santra.hut.fi ([130.233.224.1]) by niksula.hut.fi with SMTP id
<61756-2>; Tue, 24 Nov 1992 12:56:43 +0200
Received: from finhutc.hut.fi by santra.hut.fi
(5.65c/8.0/TeKoLa) id AA04413; Tue, 24 Nov 1992 12:56:50 +0200
Received: from Finhutc.HUT.FI by FINHUTC.hut.fi (IBM VM SMTP R1.2.2MX)
with BSMTP id 0211; Tue, 24 Nov 92 12:57:09 EET
Received: from UKACRL.BITNET by Finhutc.HUT.FI (Mailer R2.08 R208004) with
BSMTP id 0210; Tue, 24 Nov 92 12:57:08 EET
Received: from RL.IB by UKACRL.BITNET (Mailer R2.07) with BSMTP id 9370; Tue,
24 Nov 92 10:56:11 GMT
Received: from RL.IB by UK.AC.RL.IB (Mailer R2.07) with BSMTP id 2494; Tue, 24
Nov 92 10:55:57 GMT
Via: UK.AC.SWAN.PYR; 24 NOV 92 10:37:39 GMT
Message-Id: <9045.9211241040@pyr.swan.ac.uk>
Sender: owner-linux-activists@joker.cs.hut.fi
X-Note1: Remember to put 'X-Mn-Key: normal' to your mail body or header
From: Alan Cox
To: linux-activists@joker.cs.hut.fi
Subject: Re: Pathworks Server for Unix available
Date: Tue, 24 Nov 1992 12:39:55 +0200
X-Mn-Key: NET
Well the NetBIOS over TCP is an RFC 10xx document somewhere so anyone
wanting to tidy that part of the code up just needs to go have an argument
with wais or gopher about it. I did look at the spec once and its not
too bad. The concepts it implement are quite nicely planned too.
Note that TCP netbios is different to novell netbios and various other
netbios systems - netbios doesnt have to interoperate , it just all lives
on interrupt blah and works like so...
Alan
From owner-linux-activists@joker.cs.hut.fi Wed Nov 25 01:49:05 1992
Status: RO
X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil]
["253" "Wed" "25" "November" "1992" "01:15:53" "+0200" "Dan Shearer"
"ccdps@lux.levels.unisa.edu.au " nil "8" "Pathworks server now accessible :-)"
"^From:" nil nil "11"])
Received: from joker.cs.hut.fi by hutcs.cs.hut.fi with SMTP id AA22668
(5.65c8/HUTCS-S 1.4 for ); Wed, 25 Nov 1992 01:49:02 +0200
Received: from joker.cs.hut.fi by niksula.hut.fi id <61854-4>; Wed, 25 Nov 1992
01:46:16 +0200
Received: from Levels.UniSA.Edu.Au ([130.220.16.10]) by niksula.hut.fi with SMTP id
<62219-3>; Wed, 25 Nov 1992 01:35:37 +0200
Received: from lux.levels.unisa.edu.au by Levels.UniSA.Edu.Au (PMDF #2428 ) id
<01GRKBQ1DXQO9S3T07@Levels.UniSA.Edu.Au>; Wed, 25 Nov 1992 09:46:48 +1030
Received: by lux.levels.unisa.edu.au (4.1/SMI-4.1) id AA29831 for
linux-activists@niksula.hut.fi
Sender: owner-linux-activists@joker.cs.hut.fi
X-Note1: Remember to put 'X-Mn-Key: normal' to your mail body or header
Illegal-Object: Syntax error in To: address found on niksula.hut.fi:
To: linux-activists@niksula.hut.fi(Linux (Remember X-Mn-Key !)))
^-missing closing ')' in token
Message-Id: <9211242316.AA29831@lux.levels.unisa.edu.au>
X-Envelope-To: linux-activists@niksula.hut.fi
Content-Type: text
Content-Transfer-Encoding: 7BIT
X-Mailer: ELM [version 2.4 PL6]
Content-Length: 252
From: ccdps@lux.levels.unisa.edu.au (Dan Shearer)
To: linux-activists@joker.cs.hut.fi
Subject: Pathworks server now accessible :-)
Date: Wed, 25 Nov 1992 01:15:53 +0200
X-Mn-Key: NET
Thanks to the persistence of Juergen Henke I have at last
got the permissions on the pathworks correct. As I said before, it is available
from lux.levels.unisa.edu.au:pub/pathworks. Sorry for any inconvenience.
Dan.
Path: sparky!uunet!dtix!darwin.sura.net!jvnc.net!news.edu.tw!twnmoe10!levels!ccdps
From: c...@levels.unisa.edu.au
Newsgroups: comp.unix.misc,comp.sources.wanted,comp.os.linux
Subject: Free Pathworks for Unix now available
Message-ID: <19405.2b15bc7a@levels.unisa.edu.au>
Date: 26 Nov 92 19:43:14 GMT
Organization: University of South Australia
Lines: 143
I thought there might be people in this newsgroup interested too.
Dan.
From ccdps Tue Nov 24 10:15:44 1992
Subject: Pathworks Server for Unix available
To: linux-activi...@niksula.hut.fi (Linux (Remember X-Mn-Key !))
Date: Tue, 24 Nov 1992 10:15:44 +1030 (GMT+1030)
X-Mailer: ELM [version 2.4 PL6]
Content-Type: text
Content-Length: 6091
X-Mn-Key: Net
I have had a lot of enquiries about the free pathworks server after I followed
up a comment on the list. Alan Cox has tracked
down the author and I have got the full story from him. He is Andrew Tridgell,
at the Australian National University.
Summary
- A free Pathworks server exists, and I have made it available for anon ftp
in the directory lux.levels.unisa.edu.au:pub/pathworks.
- It implements NetBIOS, on top of sockets.
- It only took about two weeks to write.
- He no longer has a need for it, and so has stopped development.
--
Dan Shearer email: Dan.Shea...@UniSA.edu.au
Information Technology Branch Phone: +61 8 302 3479
University of South Australia Fax : +61 8 302 3385
In Andrew's words (compiled from several messages to me) :
Yes, I did write a pathworks server for unix. I wrote it because
we were interested in the PC X server offered by Dec but it required
using Pathworks as the net. We used PC-NFS and couldn't get the same
functionality from pathworks because we couldn't mount from non Ultrix
boxes (such as suns). So I wrote a server for a general unix box. I
didn't have a spec so I looked at the packets to and from a dec server
and worked out the structures they used. It works but has a few minor
bugs. I have never fixed them because the dec X server now works with
PC-NFS so our local site is no longer interested in pathworks at all!
Anyway - you are very welcome to the code, I'm never going to
commercialise it so I don't mind who does what with it. I'm not going
to be doing anything with it but I would hate to see all the work go to
a total waste. It would be particularly interesting to write a unix
client (not very hard I think) so you have a complete replacement for
NFS. It should (in theory) work with DECNET as well and should talk to
some other clients that are netmanager compatible (again I haven't tried
as I don't have one).
[another message - Dan]
>From start to finish it took under 2 weeks. Most of that time was spent
staring at packets and trying to find out what the protocol was. I wrote
a short socket program that grabs all packets (to disk) then passes them
on to the destination. It's like a simple sniffer I suppose and it means
I only have to modify /etc/inetd.conf to look at any programs socket
communications. I got the idea from the dec gateway sample code.
As you have probably noticed the actual code is fairly simple. The hard
bit is knowing things like that when answering a dfree request the 43rd
and 44th bytes are the byte swapped result and the 38th byte has to be a
'%'. If you had a copy of the spec I'm sure my code would be hilarious.
That '%' has got to mean something - I'm just buggered if I can figure
out what. By the end of it I got the feeling that the spec was generated
with a random number generator.
[another message - Dan]
> 1. Did you not have a spec because no-one would give you one or because
> you couldn't find one easy enough? (ie, if someone wanted to make sure
> it really was legal to put on say, a Decnet network, would it be possible
> to say "yes, because it conforms to XYZ.")
The real problem was that I didn't know there was a spec! I had assumed
that it was a proprietry DEC protocol and only after I released it did
someone say "hey! you've implemented netbios!". I dealt only with the
bytes I saw in the dec server, not knowing it was copying a microsoft
spec.
I have checked with dec and they said what I did was legal - because
there was a published spec (even if I didn't have it).
>
> 2. What lots and *lots* of people want is a way of doing printer/file
> sharing for PCs, most commonly Macs and IBM clones. How do you see your
> efforts fitting in with this aim? In particular people/institutions
> would like to have royalty-free servers accessible with their existing
> workstations. We already have a free, reliable Unix working on cheap
> hardware (Linux) and its ethernet interface is coming together. The
> concept of putting Lan Manager (=Pathworks, almost) on top of this is
> very attractive indeed.
I've no idea about print serving, but it wouldn't be hard to write a dos
client for this for file sharing. All you need is a socket library and a
few dos type changes (\ for / etc). The NCSA socket lib would probably
do the trick. I have used my server for quite a while on a Sun, using
the DEC PC client and mounting serveral remote disks.
My server should certainly work with mininal changes (if any) on almost
any unix implementation that has sockets. The headache with writing a
dos client would be the memory resident stuff, as it needs to reliably
access the disk. Of course this wouldn't be a problem with Linux.
> 3. I understand (from Alan Cox in the UK, who has run your server) that
> you have implemented NetBIOS, what other protocols can/do you implement
> in your code? How do they interface with Unix?
>
If only I knew it was netbios when I wrote it! I might have been able to
get a spec. I suspect that I'm missing many calls of the spec - I only
implemented the ones I saw the dec server using. Adding new calls is
pretty easy though. I use sockets as the communication method. This
allows decnet to be used as it uses basically the same syntax (for
simple things like I do). I've no idea about other transports or
protocols. One problem is that I know next to nothing about networking,
I just know that a socket is like a file and write to it like a file.
> However I'd like to know which is the best ftp site. Once I understand
> a little bit more about it I will let other potential developers/users know.
> We'll see what happens. If it catches the imagination of enough people
> some development may well take place.
I just checked and the machine I was distributing it on has gone! I
attach a uuencoded version below.
[which I have made available for anonymous ftp - Dan]
Cheers,
Andrew
--
--
Dan Shearer email: Dan.Shea...@UniSA.edu.au
Information Technology Branch Phone: +61 8 302 3479
University of South Australia Fax : +61 8 302 3385
From owner-linux-activists@joker.cs.hut.fi Mon Nov 30 18:43:55 1992
Status: RO
X-VM-v5-Data: ([nil nil nil nil nil nil nil nil nil]
["371" "Mon" "30" "November" "1992" "18:42:42" "+0200" "David Barr"
"davidb@mcis.washington.edu" nil "11" "Re: Pathworks server now accessible :-)"
"^From:" nil nil "11"])
Received: from joker.cs.hut.fi by hutcs.cs.hut.fi with SMTP id AA26183
(5.65c8/HUTCS-S 1.4 for ); Mon, 30 Nov 1992 18:43:52 +0159
Received: from joker.cs.hut.fi by niksula.hut.fi id <62065-3>;
Mon, 30 Nov 1992 18:43:41 +0200
Received: from olympia.mcis.washington.edu ([128.95.150.66])
by niksula.hut.fi with SMTP id <61702-2>; Mon, 30 Nov 1992 18:43:23 +0200
Received: by olympia.mcis.washington.edu
(5.57/UW-NDC Revision: 2.21 ) id AA13083; Mon, 30 Nov 92 08:43:11 -0800
Message-Id: <9211301643.AA13083@olympia.mcis.washington.edu>
Sender: owner-linux-activists@joker.cs.hut.fi
X-Note1: Remember to put 'X-Mn-Key: normal' to your mail body or header
In-Reply-To: <9211242316.AA29831@lux.levels.unisa.edu.au>;
from "Dan Shearer" at Nov 25, 92 1:15 am
X-Mailer: ELM [version 2.3 PL11]
From: David Barr
To: linux-activists@joker.cs.hut.fi
Subject: Re: Pathworks server now accessible :-)
Date: Mon, 30 Nov 1992 18:42:42 +0200
X-Mn-Key: NET
> Thanks to the persistence of Juergen Henke I
> have at last got the permissions on the pathworks correct. As I said
> before, it is available from lux.levels.unisa.edu.au:pub/pathworks.
This sounds great. Is anyone working on a pathworks file system for
linux so that we can mount disks from other machines using pathworks?
David