Date: Thu, 21 Jan 93 04:30:09 PST
From: Advanced Amateur Radio Networking Group <tcp-group@ucsd.edu>
Errors-To: TCP-Group-Errors@UCSD.Edu
Reply-To: TCP-Group@UCSD.Edu
Precedence: Bulk
Subject: TCP-Group Digest V93 #21
To: tcp-group-digest


TCP-Group Digest            Thu, 21 Jan 93       Volume 93 : Issue   21

Today's Topics:
                           10 GHz link help
                           Bug in ftpserv.c
                           Clarkson Drivers
                DK5SG-bbs for wampes please?? (2 msgs)
                        Future of NOS (2 msgs)
                   ka9q e/w 4 serial ports @ 57.6K?
                         Ka9q with AST 4 port
                      Microsoft 7.0 & WG7J JNOS
             N6GN Microwave Link Construction Info Sought
                             RCS (4 msgs)
                        Several minor things 

Send Replies or notes for publication to: <TCP-Group@UCSD.Edu>.
Subscription requests to <TCP-Group-REQUEST@UCSD.Edu>.
Problems you can't solve otherwise to brian@ucsd.edu.

Archives of past issues of the TCP-Group Digest are available
(by FTP only) from UCSD.Edu in directory "mailarchives".

We trust that readers are intelligent enough to realize that all text
herein consists of personal comments and does not represent the official
policies or positions of any party.  Your mileage may vary.  So there.
----------------------------------------------------------------------

Date: Wed, 20 Jan 93 8:33:22 PST
From: Glenn Elmore <glenne@srlr12.sr.hp.com>
Subject: 10 GHz link help
To: tcp-group@ucsd.edu

From: bob@ke9yq.ampr.org (Bob Van Valzah, ke9yq)

> I'd like to build a few of the n6gn 2 Mbps 10 GHZ links described in the

> If anyone else has successfully built one of these links or knows an E-Mail
> address for Glenn Elmore (n6gn himself), I'd like to correspond.

  Bob, 
    I'm reading here and also reachable at glenne@sr.hp.com
I'd be happy to help answer questions.

Glenn

------------------------------

Date: Thu, 21 Jan 1993 00:31:42 -0600 (CST)
From: MSgt S. Sampson <ssampson@sabea-oc.af.mil>
Subject: Bug in ftpserv.c
To: TCP-Group@ucsd.edu

Joe/K5JB found an interesting feature in NET/NOS in that you could CD to a
file name in FTP.  So we changed access() to stat() and fixed that.  It's
pretty old code, so is probably in all versions.

In function ftpserv() of ftpserv.c the following bug exists in
the CWD_CMD switch:

     case CWD_CMD:
          file = pathname(ftp.cd,arg);
          if(!permcheck(ftp.path,ftp.perms,RETR_CMD,file)){
               usprintf(ftp.control,noperm);
               free(file);
#ifdef    MSDOS
          /* Don'tcha just LOVE %%$#@@!! MS-DOS? */
--------> } else if(strcmp(file,"/") == 0 || access(file,0) == 0){
#else
--------> } else if(access(file,0) == 0){    /* See if it exists */
#endif


change this too:

#ifdef MSDOS
          } else if ((strcmp(file,"/") == 0) ||
               ((stat(file, &cwdstat) == 0) &&
                    (cwdstat.st_mode & S_IFDIR)))  {
#else
          } else if (stat(file, &cwdstat) == 0 &&
               (cwdstat.st_mode & S_IFDIR)) {     /* See if it exists */
#endif

Also at the top of the function add:

static void
ftpserv(s,n,p)
int s;    /* Socket with user connection */
void *n;
void *p;
{
     struct ftpserv ftp;
     struct stat cwdstat;     <------------ or whatever you want to call it

And then add #include <sys/stat.h> at the top.

I also uploaded ftpserv.c%v.Z to incoming.

73, Steve/N5OWK

------------------------------

Date: Wed, 20 Jan 93 21:48:15 CET
From: BARRY TITMARSH <BTITMARS%ESOC.BITNET@vm.gmd.de>
Subject: Clarkson Drivers
To: TCP-GROUP <TCP-GROUP@ucsd.edu>

Can some one tell me where to find the latest driver package..
I was told by a local friend that there was a Version 11 driver pack out
cant find i on the normal ftp sites  still the old version 10
Barry.

------------------------------

Date: Wed, 20 Jan 93 6:55:28 MST
From: Dieter Deyke <deyke@mdddhd.fc.hp.com>
Subject: DK5SG-bbs for wampes please??
To: tcp-group@ucsd.edu

>   This may be the wrong group for this, but I think the ka9q-unix group
> has dried up? and there are probably more listeneres on here with Unix.
>   I am looking for the DK5SG bbs software that I belive works with Wampes??
> I have looked on Ucsd.edu, and cannot see it there. Also are there any usres
> with ISC Unix, that are also runing Wampes??
>   Thanks in advance, Mike

The BBS code is inside the wampes.tar file.

Dieter

------------------------------

Date: Wed, 20 Jan 93 9:40:16 CST
From: dave@holl.com (David Vrona)
Subject: Dk5Sg-bbs for wampes please??
To: TCP-Group@UCSD.Edu

> Date: Sun, 17 Jan 93 05:49:57 UTC
> From: g6phf@wg7j.ECE.ORST.EDU
> Subject: DK5SG-bbs for wampes please??
> To: tcp-group@ucsd.edu
> 
>   This may be the wrong group for this, but I think the ka9q-unix group
> has dried up? and there are probably more listeneres on here with Unix.
>   I am looking for the DK5SG bbs software that I belive works with Wampes??
> I have looked on Ucsd.edu, and cannot see it there. Also are there any usres
> with ISC Unix, that are also runing Wampes??
>   Thanks in advance, Mike
> 

Is the ka9q-unix group a separate mailing list from this one?  If so, how do I
sign up?  I'm very interested in packet on UNIX.

Also, I have the latest version of WAMPES from ucsd.edu.  I want to run this
on my 386 with Dell SVR4 v2.2.  However, I'm having trouble getting started
due to the lack of documentation in WAMPES.

So, could someone who is using WAMPES please send me a copy of their startup
files????  This would really help me get started.

Thanks.
--
 David Vrona N9QNZ               +1 708 680 2829 (voice)               
 Hollister Incorporated          +1 708 680 2123 (fax)                 
 2000 Hollister Drive            Internet: dave@hp1.holl.com           
 Libertyville, IL  60048-3781    UUCP: {well connected}!ddsw1!hp1!dave 
 Opinions expressed are my own and not those of Hollister Incorporated.

------------------------------

Date: Wed, 20 Jan 93 13:46:47 CST
From: jks@giskard.uthscsa.edu
Subject: Future of NOS
To: tcp-group@ucsd.edu

Jeff Comstock writes (about drive mounting//NFS):

> The stuff you guy's are talking about isn't futuristic, I did all this
> about two years ago.  This doesn't work too good at 1200/VHF, but one
> of the 9600/UHF weinersnitzels here in the Twin Cities told me they have
> stuff NFS mounted all over the place up there, and everybody is using
> NOS as a router.

Maybe you missed Walts point tho'. (What you say is True and Good!) He was 
addressing the concept from the standpoint of user interface.... While the 
"engine" is the same (tcp/ip) He plans to write to the SOM in OS/2. He thinks 
that (for example) treating the SMTP server as an object (rural route mailbox) 
on the desktop metaphors used in OS/2, Next, Sys7, Xwin... etc is more 
intuitive to the non-UNIX j.a.luser.

In the SMTP example, you create a message in "your favorite editor".  The 
message become an "object" which is dropped in the mailbox.  The mailbox 
queries for an address (if not included with the object--reply wise) and passes 
the mail to the SMTP server.  Incoming mail would be treated in the same 
manner, becoming an object in your mailbox.  When you open your mailbox there 
is a stack of mail waiting...  you can open, read, and reply to as you 
wish.....  

Already there is a "semi-object oriented" NFS client function in current IBM
TCP/IP suite for OS/2 and a more complete one in the Next implementation. The 
extension to FTP is natural....(and a little less bandwidth intense.)

Telnet and Talk objects could be created as well for your favorite login.... 
The XWindows implementation is already out there.... well enuf of the pipe 
dream......

73.

********************************************************************* 
* Dr. John Spitznagel                  *   Sancho Panza Institute   * 
* Internet: jks@giskard.uthscsa.edu    *    for Advanced Studies    * 
* AMPRNet:  kd4iz@kd4iz.ampr.org       *  Department of Bogometrics * 
* CIS:      76044,476                  *                            * 
* Tel:      (210) 567-6616             *  (C) JKS, 1992             * 
********************************************************************* 

------------------------------

Date: Wed, 20 Jan 93 09:44:34 EST
From: kz1f@RELAY.WESTBORO.LEGENT.COM
Subject: Future of NOS
To: jrc@brainiac.mn.org, tcp-group@ucsd.edu

Jeffrey Comstock writes:
>   B ---ether/slip----> A -----radio waves-------> D ----ether/slip----->C
>PC/PCNFS             PC/KA9Q                    PC/KA9Q               Sun/nfsd

>The stuff you guy's are talking about isn't futuristic, I did all this
>about two years ago.  This doesn't work too good at 1200/VHF, but one
>of the 9600/UHF weinersnitzels here in the Twin Cities told me they have
>stuff NFS mounted all over the place up there, and everybody is using
>NOS as a router.

Jeff, Since you looked like you were responding to my comments, not Steve's
I question what you mean by the above quote. I dont think what you described 
having done 2 yrs ago is what I had in mind. Neither KA9Q nor Unix derivitives
treat nodes/hosts/remotely mounted drives/mail converse services etc as 
objects. In order to use/abuse any of the above mentioned services one needs to
interact with the remote node thru some sort of cmd line interface. What I was 
refering to is, my D: drive is an icon that can be manipulated icononically or 
exploded and its contents manpulated iconically. There is no reason why I 
shouldnt be able to access UCSD\hamradio... with equal ease. Granted, in a 
strict sense this is not a new form of transport access method, but rather a 
human interface paradigm shift. Look how long it took Unix to get a graphical 
interface. Even with that it misses the boat. I dont think the future is having
multiple views of your desktop but rather viewing and interacting with real 
world objects as such, graphical representations of what we really are dealling
with. Consequently, taking an OBR view, one doesnt ftp to some symbolic name 
then signin and get/put symbolic names. They drag a object (file 
representation) from one object (disk drive) to another(remote node's disk 
drive). All the vulgarities of accomplishing that are abstracted away from the 
user, who really doesnt care how to do it just that it get done.
I think this relates to what has affectionately been referred to as the OS holy
wars in that if "NOS" is to grow, it really cant be done in the confines of 
640K. The issue is not whether PMNOS works better than JNOS because of platform
issues its whether its more "user friendly"/intuitive/easier. Frankly, thats 
almost a moot point, Its still a cmd line interface with different wrapping 
paper on it. This somewhat touches on another issue, What constitutes "advanced
Networking" ? Is it the differences between JNOS 107b and 108a, a different way
to ftp? Is it PMNOS? Or perhaps is it providing the user a straight forward 
intuitive way to accomplish what he/she wants to do? We, techno-weenies, take 
great pride in mastering the really arcane syntax of Unix or obscure 
definitional syntax of establishing forwarding routes or encapsulated 
protocols. This is all very nice but at best represents a small piece of 
"Advanced Networking". Even among the readers of tcp-group, nos-bbs and of 
course all those NOS users that dont have i-net access there is a large number 
that frankly have no clue how this stuff should be set up, much less, how it 
works. I dont feel its important they know how it works, unless they really 
want to, but certainly if they cant figure out how to use it, we have missed 
the point and are only existing to be self serving techno-weenies. Look at why 
NOS-BBS group came to exist. I, for one, think the REALLY "Advanced Networking"
goal is to make this science as easy and straight-forward as possible, not to 
see how quickly we can turn out really archane services it takes a PhD to use.
Well, thats my 20 cents (inflation, ya know). As always, comments are welcomed,
critizisms can be kept to one's self:-).

Walt Corey - kz1f@kz1f.legent.com
--------------------------------
|                                |
| Space for Rent apply within    |
|                                |
----------------------------------

------------------------------

Date: Wed, 20 Jan 1993 15:26:03 -0500 (EST)
From: Steven E Frazier <sfrazier@norstan.com>
Subject: ka9q e/w 4 serial ports @ 57.6K?
To: tcp-group@ucsd.edu

Does anyone have ka9q compiled to support 4 ports, that will run 57.6K
on each port, has PPP, SLIP, telnet, ftp compiled in?  Telnet and ftp
aren't really needed but it would be nice.  If so, could you email me,
please?

Thanks.

Steve

-- 
Norstan Communications, Inc.        | Steven E. Frazier
8101 N. High Street Suite 100       |---------------------------------------
Columbus, OH   43235                | Local : sfrazier
Voice 614-785-7335 Fax 614-785-7337 | Remote: sfrazier@norstan.com

------------------------------

Date: Wed, 20 Jan 93 20:09 EST
From: gws@n8emr.cmhnet.org (Gary Sanders)
Subject: Ka9q with AST 4 port
To: tcp-group@ucsd.edu

Anyone have a version of Ka9q compiled to support
an AST 4 port card?

Gary W. Sanders gws@n8emr.cmhnet.org, 72277,1325
N8EMR @ N8JYV (ip addr) 44.70.0.1 [Ohio AMPR address coordinator]
HAM BBS 614-895-2553 (1200/2400/V.32/PEP) Voice: 614-895-2552 (eves/weekends)

------------------------------

Date: Wed, 20 Jan 93 11:02:07 EST
From: kz1f@RELAY.WESTBORO.LEGENT.COM
Subject: Microsoft 7.0 & WG7J JNOS
To: JT@zfja-gate.fuw.edu.pl, tcp-group@ucsd.edu

>C standard tells the setjmp/longjmp were designed for non-local
>jumps between functions rather than between tasks.
>Does anyone know other reasons for MicroSoft cannot be used for NOS?

Agn, with 6.0, my only problem was with the setjmp/longjmp. I came up with no 
convenient way to implement Phils "multitasker" using uSoft c60. Since I was 
building an OS/2 protected mode version, I simply used real multitasking. That 
aside the C6.0 compiler generated really good code and all OS/2 1.x versions of
NOS I wrote used it. The only solution I could come up with, that I chose not 
to implement is to supply a user version of setjmp/longjmp that would carry 
the stack selector with it. That would be a pretty gutsy move though.

 Walt Corey -
kz1f@kz1f.legent.com

----------------------------------
|                                |
| Space for Rent apply within    |
|                                |
----------------------------------

------------------------------

Date: Wed, 20 Jan 93 10:51:52 EST
From: crompton@NADC.NAVY.MIL (D. Crompton)
Subject: N6GN Microwave Link Construction Info Sought
To: bob@ke9yq.ampr.org, tcp-group@ucsd.edu

Our local (EPA) ARRL technical Coordinator is a professor at a 
local school. His students built a 10GHZ data link which he 
demonstrated at our club meeting a few months ago. As part of
the project a detailed document with schematics was written. If
this would be of help I could get you in touch with him. He does
have an internet address - I don not have it handy at the moment.

Doug

------------------------------

Date: Wed, 20 Jan 93 2:29:09 HST
From: Antonio Querubin <tony@mpg.phys.hawaii.edu>
Subject: RCS
To: luru@stekt.oulu.fi

> Hi!
> 
> I recently downloaded the latest NOS source (now in RCS files),
> grabbed the latest RCS for DOS I could find and tried to make.
> 
> No such luck. Apparently there is a conflict between RCS convention of
> using filenames like VERSION.C%V and the command processor's desire to
> expand any %-sign-something to an environment variable? Or something,
> in any case, make results in:
> 
>          co  mktl.c
> co error: RCS\mktl.c: No such file or directory
> 
> I have an (empty besides makefile) source directory and below that an
> RCS directory containing all the source files. As far as I could tell
> from the RCS manual page, this is as it is supposed to be, no?
> 
> Now, I am new to RCS and I am sure there is something I am missing
> here. Can someone help me out, please?

In the makefile, add a '-x%v' switch to all of the 'co' lines.  

You may also find that a few files will generate compile errors.  Most
of these files will have a '\' at the end of one or more macro
definition lines.  Get rid of the '\' and join the macro lines
together into a single line.  I wonder if there's an obscure compiler
switch that would allow leaving the '\' in?

Finally, if you see a whole bunch of errors at the final link step the
first time you run make in a clean directory, delete the *.tl files
and do the make again.  For some reason, the *.tl files aren't
generated properly when starting clean.

Tony

------------------------------

Date: Wed, 20 Jan 93 11:57:14 EST
From: jbloom@arrl.org (Jon Bloom)
Subject: RCS
To: luru@stekt.oulu.fi

>I recently downloaded the latest NOS source (now in RCS files),
>grabbed the latest RCS for DOS I could find and tried to make.
>
>No such luck. Apparently there is a conflict between RCS convention of
>using filenames like VERSION.C%V and the command processor's desire to
>expand any %-sign-something to an environment variable? Or something,
>in any case, make results in:
>
>         co  mktl.c
>co error: RCS\mktl.c: No such file or directory

Try changing the rules in the makefile to generate:

     co -x%v mktl.c

The GNU RCS ported to DOS--I assume that's what you are using--seems to
have no default filename pairing.  Also, you'll notice that all of the
files checked out from the RCS directory will have two carriage returns
at the end of each line.  That's because the GNU RCS keeps the RCS
workfiles in unix format, while whatever Phil is using apparently keeps
them in DOS format (cr/lf).  Damn DOS eol convention!
-------
Jon Bloom, KE3Z              |  jbloom@arrl.org
American Radio Relay League  |     Justice is being allowed to do whatever
225 Main St.                 |     I like.  Injustice is whatever prevents
Newington, CT 06111          |     my doing so. -- Samuel Johnson

------------------------------

Date: Wed, 20 Jan 93 16:10:26 EDT
From: "William Allen Simpson" <bill.simpson@um.cc.umich.edu>
Subject: RCS
To: tcp-group@ucsd.edu

Well, I just spent 20 fruitless hours trying to make this work.

First there was the L-O-N-G upload.  The RCS.tar.Z file is 4 times
as long as the old .zip files.

I'm using GNU RCS from simtel20 pgmutl:rcs56dos.zip, and filutl:tar4dos.

I had to put the RCS directory just above the directory I wanted to get
the files (nos\rcs).

Even then, RCS wouldn't recognize the files.  So, I tried making a new
file, and discovered that GNU RCS didn't put a %v on the end.

I had to remove all the %v.  Fortunately, I have 4dos, or I don't think
that I could have done it.  It took quite a bit of experimentation.

        Note: I since learned that "co -x%%v <filename>" should work on
        the %v files.  But according to the .man pages, the default of
        no suffix will work on all systems, even Unix.

The makefile didn't work as is, and I couldn't edit it because the long
lines are folded by my editor, qedit.  So, I had to check the files out
by hand.

Finally, I got the files all checked out.  The compile wouldn't work,
because the various #defined macros don't work.  This turns out to be
because the CRLF is already in the RCS file, but GNU RCS assumes that
the file is in Unix format, and expands the LF into CRLF for DOS,
resulting in CRCRLF.

Borland C thinks this is a bogus line without a continuation, and gets
completely unhappy with the __ARGS macros.  (At least, that's what I
think is happening.)

        Late breaking news.  I used Desi's "flip", from
        txtutl:flip1exe.zip, to first flip to unix (converted CRCRLF to
        LF) and back to MSDOS (converted LF to CRLF).  Now it compiles.

Bill.Simpson@um.cc.umich.edu

------------------------------

Date: Thu, 21 Jan 93 10:15:25 PST
From: "Jerzy Tarasiuk" <JT@zfja-gate.fuw.edu.pl>
Subject: RCS
To: jbloom@arrl.org (Jon Bloom)

> Date: Wed, 20 Jan 93 11:57:14 EST
> Message-Id: <38985@jrb>
> The GNU RCS ported to DOS--I assume that's what you are using--seems to
> have no default filename pairing.  Also, you'll notice that all of the
> files checked out from the RCS directory will have two carriage returns
> at the end of each line.  That's because the GNU RCS keeps the RCS
> workfiles in unix format, while whatever Phil is using apparently keeps
> them in DOS format (cr/lf).  Damn DOS eol convention!

Use RCS 5.5 for DOS:  oak.oakland.edu: (sth)/pgmutl/rcs55.zip  or
wuarchive.wustl,edu: (sth)/gnuish/rcs55ax.zoo,(sth)/gnuish/rcs55as.zoo
- will not get the double CR-s. It is also here on zfja-gate
(148.81.6.100), file utils/rcs55.zip (taken from oak).
BTW, the EOL convention (CRLF) isn't DOS only, it's ANSI.
73's, JT

PS: I rearranged directory structure on zfja-gate; CPU checking programs
are now in directory ./cpu, all nos-source-related files in ./nos and
subdirectories (./nos/ka9q, ./nos/pa0gri, ./nos/n1bee).
For Polish subscribers: I have Mike's NOS here in directory ./nos/n1bee,
FTP it from me if need, to avoid use of slow Warszawa-Stockholm link.

------------------------------

Date: Wed, 20 Jan 1993 11:47:14 EST
From: "Russell Nelson" <nelson@crynwr.com>
Subject: Several minor things 
To: tcp-group@ucsd.edu

In article <1367.9301200839@pyr.swan.ac.uk> you write:

  > 1.) There have been several messages about things like 2MBps
  > 10GHz links. I was I have to admit looking more in the 56KBit
  > range since I can't see a PC coping with a 2MBps link at any
  > speed. I'm looking for any recommendations  > on books/papers on
  > the subject that people have found useful. ESPECIALLY working
  > with NOS/NET since it's a bit hard to run a full service Linux
  > system at 1200 baud

Well, NOS/NET is not the fastest TCP stack in the world.  I have full
faith in Phil's ability to make it the fastest, but that was not one
of his design goals.  But a PC can cope with 2MBps, so long as the
receiving device is buffering the bits.  Heck, it can cope with
10MBps (Ethernet) just fine...

-- 
-russ <nelson@crynwr.com> What canst *thou* say?
Crynwr Software           Crynwr Software sells packet driver support.
11 Grant St.              315-268-1925 Voice  |  LPF member - ask me about
Potsdam, NY 13676         315-268-9201 FAX    |  the harm software patents do.

------------------------------

End of TCP-Group Digest V93 #21
******************************
******************************