SY.FDC@CU20B.COLUMBIA.EDU (Frank da Cruz) (10/18/85)
Info-Kermit Digest Thu, 17 Oct 1985 Volume 3 : Number 24
Today's Topics:
MS-DOS Kermit for the DECmate II and III
Crosstalk XVI 3.6 Kermit Implementation
Kermit Throughput Problems with "Smart" Multiplexer
MacKermit and TACs
VMS Kermit 3.1.066 Bug
Re: VMS C Kermit problem...
Commodore-64 Kermit 1.7 1200 Baud Fix
CP/M Kermit on a Remote Machine
File Transfer Between VAX (Unix 4.2) and IBM 3081 (VM/370) ?
File Transfer Between VAX And IBM PC Through LAN ?
Victor 9000 CP/M-86 Kermit Binary Wanted (On Floppy)
MS DOS Kermit vs DTR
Kermit for 1750A, MOS, or Jovial?
----------------------------------------------------------------------
Date: Thu, 17 Oct 85 10:02:47 EDT
From: Frank da Cruz <SY.FDC@CU20B>
Subject: MS-DOS Kermit for the DECmate II and III
An implementation of MS-DOS Kermit 2.28 (the current release) is now
available for the DECmate-II and III with the XPU (MS-DOS) option, from an
anonymous donor at DEC. The files are in KER:MSDM2.BOO (printably encoded
.EXE file, use KER:MSPCTRAN.BAS or KER:MSPCBOO.BAS to decode it, if indeed
DECmate MS-DOS has a Microsoft-like Basic), KER:MSDMII.BAT (for building from
source), KER:MSXDM2.ASM (system-dependent source module), and KB:MSDM2.EXE
(8-bit binary). No documentation is available, but it is said to work, and
to use the DM-II/III's built-in VT100/200 firmware for terminal emulation.
------------------------------
Date: Wed, 9 Oct 85 08:57:06 cdt
From: blake@astro.utexas.edu (R. Blake Farenthold)
Kermit: Kermit on The Source
The Source (a commercial time sharing service) has just set up their
SIGS (Special interest groups) containing discussions & downloads for
people with various interests. Aside from straight ASCII up and down
loads they offer Kermit transfers.
A couple of days ago I sent a 110K file to the Source using Kermit. At
1200 baud this should have taken about 15 minutes. IT TOOK OVER AN
HOUR! Is Kermit that inefficent, is it horribly hampered when using
a packet switching network (I used Telenet), or has The Source slowed
things way down so they can collect more per-minute connect time!
In otherwords who do I yell at?
Blake Farenthold | CIS: 70070,521 | UUCP: {ut-sally,ut-ngp,noao}
P.O. Box 3027 | Source: TCX023 | !utastro!blake
Austin, TX 78764-3027 | Delphi: blake | Intr: blake@astro.UTEXAS.EDU
BBS: 512-442-1116 | MCI: BFARENTHOLD | ESL: 62806548
[Ed. - The people at The Source are well aware of the problem, which is
twofold: (a) TELENET, and the Prime computer itself (which is what you are
talking to) provide only 7-bit channels, so you incur the overhead of
8th-bit prefixing for binary files, and (b) any packet-switched wide-area
public network like TELENET has its own built-in delays, which, due to the
stop-and-wait nature of the Kermit protocol in its present form, will tend
to dominate any file transfer over TELENET. To cope with these problems,
The Source has proposed (in Info-Kermit v3 #7, with discussion in following
issues) a sliding window protocol to allow multiple packets to be sent back
to back, which can result in a dramatic performance improvement under these
conditions. Prototype programs are running now, and should be announced
before the end of the year. By the way, don't complain too much -- most
other protocols don't work in this environment at all!]
------------------------------
Date: Wed, 16 Oct 85 20:02:47 EDT
From: RAF@UMDC
Subject: Crosstalk XVI 3.6 Kermit Implementation
I just received a copy of Crosstalk XVI 3.6, which includes KERMIT support.
I gave it a quick try and encountered two problems. One is that when I send
a text file from the PC to my TSO KERMIT, Crosstalk does not stop at the
Control-Z. Thus I get a Control-Z plus some additional garbage at the end
of the file. Microstuf customer support confirms that there is no way to
get Crosstalk to stop at the Control-Z. Since Control-Z for end of file is
a PC convention, it seems to me that Crosstalk should support it. Customer
support said they would pass the suggestion on for consideration.
The other problem is much more minor: after a Crosstalk KERMIT LIST command
is done to list KERMIT protocol options, everything put on the screen after
that is in high intensity.
Roger Fajman <RAF@UMDC>
National Institutes of Health
------------------------------
DATE: 16-OCT-1985
FROM: BRIAN@UOFT02.BITNET
SUBJECT: Kermit Throughput Problems with "Smart" Multiplexer
A note to those of you who support a Kermit implementation on odd modems and
muxes:
I recently had a call from a Kermit-11 user who found that a download from
the host to the RT11 system for a given file took 2 minutes, whereas the
upload of the same file to the host took 14 minutes. Solution: the mux they
were using severely downgrades the uplink bandwidth in an ill conceived
attempt to improve the host to local link bandwidth. From what I have seen
recently in the various trade rags, this seems to be an approach that some
vendors are trying out. Beware...
brian
[Ed. - One of the hardest problems Kermit or any similar protocol has to
cope with is that so much communication equipment is designed under the
assumption that input from a "terminal" will only be human keystrokes.
Another example follows...]
------------------------------
Date: Thu, 10 Oct 85 21:35:27 EDT
From: Dave Swindell <dswindel@BBN-LABS-B.ARPA>
Subject: MacKermit and TACs
I've found that you have to explicitly set the send-packet length to
something less than 64 when uploading data from ANY PC over a TAC. The TAC
input buffer just isn't big enough to handle the 90 (or is it 94?) character
default send-packet size used by MacKermit. As was mentioned in the digest,
you also have to use the TAC commands @ B O S and @ B I S (in that order) to
allow the Kermit protocol to function correctly over a TAC.
Dave Swindell
BBN Laboratories
------------------------------
Date: Tue 8 Oct 85 16:15:54-EDT
From: Michael Fuchs <EXT1.FUCHS@CU20B.ARPA>
Subject: VMS Kermit 3.1.066 Bug
I don't know if 3.1.066 is state-of-the-art, but it can't handle CRCs and
8bit data simultaneously in server mode. One can't set the micro to CRC and
send a binary file to the VAX unless one explicitly makes the micro
*request* 8-bit-prefixing. CRC works fine with non-8bit-files, and 8bit
files can be sent micro to VAX with Checksum error detection.
The VAX puts a "3" in the appropriate place in the init packet. Then, if
the data packet has bytes with the eighth bit high, it sends NAK packets
back to the micro. (Indeed, the NAK packets correctly have CRCs.) The only
way to use CRC error detection is to also have the micro request
8-bit-quoting (if one is sending 8bit data to the VAX).
------------------------------
From: Ivan Auger <augeri%rpics.csnet@CSNET-RELAY.ARPA>
To: info-kermit <@csnet-relay.arpa:info-kermit@cu20b.columbia.edu>
Subject: Re: VMS C Kermit problem...
There is a default mailbox buffer size on any VMS system (mailboxes are used
for process communication). You can have kermit set the size of mailboxes,
change the defaulf mailbox size, or you can indeed increase BYTLM quotas.
(By theway on our system you can do a SET HOST with a BYTLMC quota of 4096).
Ivan Auger, NYS Dept. of Health.
------------------------------
Date: Tue 8 Oct 85 17:20:07-EDT
From: Robert Lenoil <LENOIL%MIT-EECS@MIT-MC.ARPA>
Subject: Commodore-64 Kermit 1.7 1200 Baud Fix
You must download the .INI file as a SEQ file. The proper parameters for the
kludge baud rates are contained in that file. 1200 baud will not work without
it. That was the problem. There is no way (from within Kermit) to set the
optional baud rate parameters in the .INI file, so you better download
KERMIT.INI properly, or 1200 baud won't work. As as alternative, you might
wish to create KERMIT.INI with this small program:
10 open 8,8,8,"0:kermit.ini,s,w"
20 for i=1 to 45 : read b : print#1,chr$(b); : next
30 close 8
40 data 25,1,0,0,0,0,1,0,5,0,1,0,1,1,0,0,0,60,1,56,0,38,38,0,0,0,0,13,10,8,10
50 data 93,93,4,0,35,0,0,0,0,0,0,0,0,0
------------------------------
Date: Thu, 10 Oct 85 17:37:27 edt
From: Mike Ciaraldi <ciaraldi@rochester.arpa>
Subject: CP/M Kermit on a Remote Machine
I sent this before, but I think it got lost:
Is there any way to run CP/M Kermit as the "remote" kermit rather than the
"local" kermit? e.g. Suppose you had a computer bulletin board system
running under CP/M, and you wanted to use Kermit to upload and download
files. You would use the Kermit on your own micro to call the BBS, and
could then run Kermit on the remote machine. But when you told the remote
Kermit to, say, SEND, it would try to send the data out some communication
port, and would send only the status reports about the transfer to the
console (which is REALLY the modem ultimately connected to your local
machine). Even if you use Generic Kermit and have it communicate with the
TTY or CRT device, it will still be sending both file data and the status
reports to the same device. I know the IBM PC version has a flag you can
set to disable displaying the status, and the mainframe versions of Kermit
must be able to tell when to send status reports to the console and when not
to (by looking at the SET LINE perhaps?), but I couldn't find anything like
this in the CP/M Kermit 4.05 manual.
Thanks for your help.
Mike Ciaraldi
ciaraldi@rochester
seismo!rochester!ciaraldi
[Ed. - For now, there's no way to do this. I've been forwarding messages
regarding CP/M-80 Kermit to the current maintainer, Charles Carvalho
(CHARLES@ACC), but haven't had a response in months. Charles, are you
there???]
------------------------------
Date: 11 Oct 85 02:14:56 GMT
From: wei@princeton.UUCP (P Wei)
Subject: File Transfer Between VAX (Unix 4.2) and IBM 3081 (VM/370) ?
We have vax running UNIX 4.2BSD , ckermit program, modem (/dev/tty18), vcu
(to use modem). The ibm3081 also has cmskermit program. the question is:
is it possible to transfer ascii file between this two system via that
modem? I tried to use vcu to connect to ibm3081, but when I type 'kermit'
in CMS , I got 'an ascii terminal must be used' error message. Even if i
didn't get that message, I couldn't get (escape) back to vax to invoke
ckermit because vcu doesn't allow me to 'escape'. Then I tried to use
'kermit -l /dev/tty18 -b 1200'. The phone line was connected. However, when
ibm ask my terminal type, whatever I answer it warned me to check parity
etc... I don't have time to go through further. I would like to ask if any
one knows my intention (file transfer) is possible. If anyone has
experience, would you please mail me some sample session (as detail as
possible)? (e.g. the parameter settings...) thanks a lot !
HP Wei
[Ed. - It is entirely possible to use Kermit to transfer ASCII text as well as
binary files between a 4.2BSD VAX and an IBM 370-Series mainframe running
VM/CMS. Very briefly, here's how:
If you have a Series/1 style front end (7171, 4994, etc) running the Yale
ASCII package, you must also have version 2.00 or later of CMS Kermit on the
IBM mainframe (if you don't, you'll get the "An ASCII terminal must be used"
message. If you're going through some other kind of 3270 protocol emulator,
then you can't use Kermit. If you're going through a 3705-style front end
as an ordinary line mode TTY, you can use any version of CMS Kermit.
You should have version 4C(057) of C-Kermit on the Unix system -- certain
earlier versions had bugs that might have prevented them from working with
IBM mainframes. To C-Kermit you should give the following commands if you're
coming in as a line-mode TTY:
set duplex half (half duplex = local echo)
set flow none (no full duplex xon/xoff flow control)
set handshake xon (use half duplex line turnaround handshake)
set parity mark (use whatever parity your IBM system expects)
or else the following commands if you're coming in through a Series/1 style
front end:
set parity even (or whatever)
And of course you must also include whatever "set" commands are necessary
to set up the communication line ("set line", "set modem", "set baud", etc).
If all this seems a little strange, you can blame the IBM style of data
communications, which is different from everyone else's. ]
-----------------------------
Date: 11 Oct 85 04:49:02 GMT
From: wei@princeton.UUCP (P Wei)
Subject: File Transfer Between VAX And IBM PC Through LAN ?
We have several ibmpc's connected to ETHERNET LAN with the communication
program 'yterm'. I can connect the ibmpc to the vax (running UNIX 4.2) using
yterm. Therefore , the ibmpc serve as a terminal for the vax. Now , I escape
the yterm program to DOS and invoke MSKERMIT and then type 'connect'. I get
back to the unix shell. CKERMIT is then invoked. Then type 'send filename
<CR>'; escape (^]C) back to MSKERMIT ; type 'receive <CR>'. However, there is
no packet coming in. The screen just display 'transfer in progress......'
message forever!
(1) Am I totally wrong ? The kermit doesn't work on LAN ??? (only on tel
line ? )
(2) Or I forgot to set some parameters ??? Can someone mail me sample session
if same attempt has been done?
Thanks in advance.
HP Wei
[Ed. - Kermit can indeed be used over LANs, so long as the PC's connection to
the LAN looks like a serial communication port to the PC. This would seem to
the case in HP Wei's query, since a terminal connection can be made.
In general, when the Kermit CONNECT command works but file transfer does not,
the culprit is usually (a) parity, (b) flow control, (c) packet size, or
(d) interference:
(a) Check to see if your LAN terminal interfaces are using some kind of parity
and if so, tell BOTH Kermit programs about it, using the SET PARITY command.
(b) It is command for LAN terminal interfaces to want to do xon/xoff flow
control with the PC. Make sure you PC is set up for this (MS-DOS Kermit
does this by default in most cases). If some other kind of flow control is
required (like RTS/CTS) you could be in for trouble.
(c) Some LAN terminal interfaces have small buffers and can't accept a normal
Kermit packet (90-95 characters) at 9600 baud. Try using SET SEND/RECEIVE
PACKET-LENGTH for smaller packets, or else reducing the baud rate.
(c) Some LAN terminal interfaces intercept a certain character for control
purposes. If this is a printable character, a Control-A, or a carriage return,
then this will prevent Kermit file transfers from taking place. In this case,
try to find out how to change the box's intercept character to a control
character other than ^A or CR. If the ^A or the CR are at fault, you can use
Kermit's SET START and/or SET END to change these. If it's a printable
character, and it can't be changed in the LAN box, you're out of luck.]
------------------------------
Date: Sunday, 13 October 1985 06:56-MDT
From: Gijs Mos <gijs%ark.uucp@BRL.ARPA>
Subject: Victor 9000 CP/M-86 Kermit Binary Wanted (On Floppy)
I want to get some files from our Victor 9000 computers to various other
machines. The best program to use seems to be kermit. The problem is how to get
Kermit running on the Victor. I have a recent kermit distribution with a
V9000-CP/M 86 kermit on it. But I don't have an assembler, nor a way to get the
sources on the Victor 9000's disks. So I guess I'm looking for a kermit CP/M
86 executable in Victor 9000 format. Can somebody provide such a beasty at
media/handling costs?
Gijs Mos
Dept. of Biology
Free University
De Boelelaan 1087
1081 HV Amsterdam
The Netherlands
{seismo,philabs,decvax}!mcvax!vu44!gijs
------------------------------
Date: 14-OCT-1985 11:41
From: Heather Holmback <IWTS@HNYKUN52.BITNET>
To: Problems with MS-DOS Victor Kermit
We have been unsuccessful in using Kermit for a connection between a VICTOR
PC with operating system MS-DOS and the VAX, at the Max-Planck-Institut for
Psycholinguistics in Nijmegen, The Netherlands. We are using Sirius version
1.20 by Barry Devlin, University College Dublin, April 1984, translated by
SIREXE.BAS (by Daphne Tzoar, CUCCA). Could you please send information on
any later versions of Kermit or any recent solutions to problems. Our main
problem is that only one file can be successfully uploaded without exiting
and reloading Kermit. Also MSKERMIT.INI does not work as indicated in the
documentation. Thanks.
[Ed. - Unfortunately, this is still the latest version of MS-DOS Victor/Sirius
Kermit. No one has ever taken the trouble to write an MSXSIR.ASM module so
that Victor support would be included automatically in new MS-DOS Kermit
releases. Any volunteers?]
------------------------------
Date: Wed, 16 Oct 85 19:11:05 CDT
From: <uwmacc!uwmcsd1!u1100u!os1100!Mark=Zinzow@rsch.wisc.edu>
Subject: MS DOS Kermit vs DTR
MS DOS Kermit leaves the modem signal DTR high upon exit. This is useful
if a person wishes to exit kermit and then resume an online session, however
there are times when it is useful to drop DTR when running Kermit or when
finished with Kermit. For instance some modems will answer the phone only
when DTR is high, so you might want to drop DTR after a Kermit session so
that YOU can answer the phone rather than your modem.
On our campus a great many of our computers are linked together by a Gandalf
PACX communications switch. Our switch polls all ports not in use that have
DTR high. Since we have hundreds of micro's hard-wired to the switch, if
they keep DTR high when the switch has no active connection on that port to
a given system, they bog down the polling and after a short period produce
copious timeout error messages on the switch console. For this reason it is
important that our users drop DTR when finished with a session.
Furthermore, on some lines it is impossible to reconect or change systems
without dropping DTR first.
Ideally, it would be nice to have another SET parameter called DTR or DTR-
EXIT or DTR-QUIT or DROP-DTR that could be set ON or OFF so that when Kermit
is terminated it would dictate what state DTR would be left in. It would
also be nice to just have a command like RESET COMx or DROP-DTR COMx.
In lieu of this facility in Kermit I have written a short COM program
which I call OFFCOM1.COM or OFF.COM for short. The commands to create this
8 byte COM file follow. I suggest typing them in without the comments
(things beggining with a semicolon).
DEBUG OFFCOM1.COM
A
; PROGRAM OFFCOM1
; by Mark Zinzow
;
; Provides an external DOS command to clear
; the serial port.
; This has been tested on a Zenith Z150 running MS Dos 2.11 and an
; IBM PC AT running PC Dos 3.1.
; Note that this program is hardcoded for the address of the RS232 port.
; A better way would be to use an offset on the base address normally
; found in memory at location 40H:0.
;
MOV DX,3FC ; Port address for serial modem register.
; (Use 2FC for com2 rather than com1)
MOV AL,0 ; Store a zero as the value to output.
OUT DX,AL ; SEND the zero to the port.
;
INT 20 ;return to DOS
RCX
8
W
Q
After typing the Q you should get the DOS prompt back. A DIR should
show the 8 byte com file OFFCOM1.COM. This program can be run after
exiting kermit or from kermit using the run command. I use a macro to
drop DTR with the command do off. The macro definition looks like this:
DEFINE OFF RUN A:OFF.COM
The first program of this kind at our site was written by Mitch Blank for
the Zenith Z100. Here's the MASM source:
[Ed. - Source omitted, stored in KER:MSXZ100.DTR. Adding DTR control for
every system that MS-DOS Kermit runs on (not to mention all the different
internal modems in use on them) would be a very big job, so little programs
like this are probably the best approach.]
------------------------------
Date: Thu, 10 Oct 85 13:47 CDT
From: "David S. Cargo" <Cargo@HI-MULTICS.ARPA>
Subject: Kermit for 1750A, MOS, or Jovial?
Has anybody seen or heard of a version of Kermit written in Jovial, or
the assembly language of the 1750A for MOS (the MATE Operating System)?
We have some parts of the company that want such a thing.
[From "Burton J. Ewing" <Ewing@HI-MULTICS.ARPA>: By the way, MOS is a
derivative of Unix (by way of IDRIS?). This might be more useful info to the
Kermit community than MOS. Our people who have been peering about in MOS's
innards tell me that it is largely identical to Unix in most respects. Same
structure, same subroutine names, arguments, etc. Therefore, if we could
find another Unix hosted Kermit that was written in something we can compile
to 1750A code we are in business. The last time I checked, that meant JOVIAL
- but, who knows what will happen in the future?]
------------------------------
End of Info-Kermit Digest
*************************
-------