SY.FDC@CU20B.COLUMBIA.EDU (Frank da Cruz) (10/08/85)
Info-Kermit Digest Tue, 8 Oct 1985 Volume 3 : Number 23
Departments:
New BOO-file Maker for MS-DOS
Further Problem with MS-DOS Kermit 2.28 GET Command
About MS Kermit and EXEPACK...
Communication Problems with MS-DOC Kermit 2.28 on Leading Edge PC
MS-DOS Kermit vs "Desk Accessory" Clocks
How to Build Z100 MS-DOS Kermit from Source?
Terminals for the Blind
The Claimed C-Kermit Bug for VMS
Four-Table ASCII/EBCDIC System for EBCDIC Kermits
Request for Osborne and Kaypro Kermit Diskettes
MacKermit TAC Problem
Re: Kermit 1.7 Loses at 1200 Baud
Kermit for Wang 2200?
Kermit for DG Machines?
Kermit for CDS 4000?
----------------------------------------------------------------------
Date: Thu 26 Sep 85 11:39:58-EDT
From: Frank da Cruz <SY.FDC@CU20B.ARPA>
Subject: New BOO-file Maker for MS-DOS
Alan Phillips at Lancaster University (UK) added Lattice C support to
MSMKBO.C, so that now MS-DOS Kermit .BOO files can be made directly on the
PC. The new version replaces the old one, as KER:MSMKBOO.C on CU20B.
If you don't know what a .BOO file is, it's yet-another-printable-encoding
for binary files, the format that we distribute MS-DOS Kermit binaries in.
It features 4-for-3 encoding and 0-compression so that the resulting .BOO
file is often shorter than the original .EXE.
------------------------------
Date: 24 Sep 1985 2221-EDT
From: Stephen H. Owades, c/o EIBEN@DEC-MARLBORO
Subject: Further Problem with MS-DOS Kermit 2.28 GET Command
I have had some problems with MS-DOS Kermit 2.28 on my IBM PC/AT, as I
described in a mail message some weeks ago. I downloaded the MSKERM.BWR
file, which offered a solution to the faulty file-name truncation (in
multi-file GETs) problem. I made the suggested changes, reassembled and
relinked, and tested the resulting program. Apparently no truncation
whatever is performed on a GET; DOS truncates the file name passed to it by
KERMIT. This does eliminate the problem of badly truncated file names
(wherein KERMIT was doing something bizarre to the name of the first
incoming file), but it has created a new, perhaps worse, fault. Now
collisions (incoming file name the same as a file name already presen are
only detected when the incoming file name is valid under DOS; if the
incoming file name is longer than XXXXXXXX.XXX, KERMIT doesn't know there is
a conflict and the system hangs. I had to reboot my AT! Collisions are
detected and avoided with the "0-1-2-..." naming method described in
MSKERM.DOC if the incoming file name is valid under DOS--evidently KERMIT
doesn't realize that there is a collision when the incoming file name was
truncated by DOS.
Stephen H. Owades
------------------------------
Date: 27 Sep 1985 0912-EDT
From: LCG.KERMIT
Subject: About MS Kermit and EXEPACK...
I have used EXEPACK on MS Kermit 2.26 through 2.28 and it produces workable
code (note I relink from source). In the 2.27 and older versions the .EXE
file went from 80+ K to around 36K. In V2.28 however the savings was minimal
(10% if I recall right). Thus EXEPACK is of marginal use with V2.28. Can't
say much about backward compatibility; it looks OK on PCDOS V2.0 on, but
could break some compatibles. Since MSDOS Link V3.XX is not commonly
available I believe its a mistake to use it in distribution; people should
be able to recreate the distributed .EXE via source with more common tools...
Glenn Everhart
------------------------------
Date: 2 Oct 1985 1152-PDT
From: Rob-Kling <Kling%UCI-20B@UCI-ICSA>
Subject: Communication Problems with MS-DOC Kermit 2.28 on Leading Edge PC
I normally use Kermit through COMM1 on my IBM-PC. I was trying it on a
Leading Edge PC last night which was configured to communicate through COM2.
It wouldn't send commands to the modem.
The Status command indicated that 1200 baud was illegal [even though we could
use PC-TalkIII to dial out on COM2].
I could not get Kermit to accept any baud rate (e.g., 110, 300, 1200) as
legal for COM2. That is, the Status command indicated an error at any baud
rate when I set the port to COM2, but it would accept 1200 baud on COM1.
When I came home, I tried setting Kermit to use COM2 on my IBMPC at home.
I have only one serial port on this machine, but tried to set COM2 as the
active port. I received the same error messages re inappropriate baud rate
["Baud rate not recognized" or equivalent message].
I couldn't find any information in the Kermit 2.28 manual to help me decide
where the problem might lie.
1. What may be the problem?
2. Do you know if anyone has tried to use MSKermit on one of the new leading
Edge D machines?
Thanks
Rob Kling
UC-Irvine
[Ed. - I'm pretty sure you can set the baud rate on COM2, as many people at
Columbia have two serial ports and use both. If the second serial board isn't
there, the status command would not be able to figure out the baud rate, since
reading the (non-existent) baud rate status port would return something
meaningless. I don't know anything about the Leading Edge machines, but it
sounds like they're another 'almost' compatible, at least in terms of the
second serial port. Does anyone out there know something about Leading Edge?]
------------------------------
Date: Mon, 7 Oct 85 10:21:49 PDT
From: walton%Deimos@CIT-Hamlet.ARPA
Subject: MS-DOS Kermit vs "Desk Accessory" Clocks
Any program which accesses the IBM PC timer interrupt to place a real-time
clock on the screen seems to put a real strain on Kermit. Continuous-update
programs basically trash Kermit. I have a shareware program called Deskmate
on my PC right now, which updates the time on the screen every 10 seconds or
so. It still badly interferes with Kermit. I don't want to have to reboot
my system in order to use Kermit effectively.
Does anyone have a solution to this problem, or is it inherent in the IBM PC
architecture?
Steve Walton
Caltech Solar Astronomy
walton@citdeimo.bitnet
walton%deimos@cit-hamlet.arpa
...!psuvax1!walton@citdeimo.bitnet
[Ed. - Thanks for the report; I've noted it in the "beware" file. There may
be a way to get two or more programs that use timers to coexist, we'll have
to look into it.]
------------------------------
Date: Mon 7 Oct 85 13:20:34-MDT
From: Dick Dysart <RDYSART@SIMTEL20.ARPA>
Subject: How to Build Z100 MS-DOS Kermit from Source?
I have downloaded the MS*.ASM files for Kermit for the Z-100, 13 of them..
I have run them thru MASM , each with no errors from MASM..(not sure if or
what I should modify for MY machine)..
Then when I LINK them, the linker responds -> I/O run error in EXE, and
aborts without creating the .EXE file...
I thought that I should try from the beginning as my Besterm still won't work
properly, and MAYBE a version of Kermit compiled on MY machine would work..
With the LINKER aborting this way, and I cant find that error message in the
MS-DOS manual as yet, I don't know what's wrong, nore do I have any idea
what to fix..
Any help would be appreciated......................Dick
[Ed. - Can anybody who has experience with MS-DOS Kermit on the Z100 offer
any help?]
------------------------------
Date: Mon, 7 Oct 85 20:31:19 EDT
From: Doug Gwyn (VLD/VMB) <gwyn@BRL.ARPA>
Subject: Terminals for the Blind
I don't know why nobody seems to be mentioning the VersaBraille (another
company makes a similar device). I used to have a blind programmer working
for me, and we tried various talking terminals, optical scanners, and so
forth. Her conclusion was that the VersaBraille (with communications
software cassette) was much easier and faster, although for graphics (yes!)
she resorted to an optical scanner (sorry, I forget the trade name).
This topic really seems orthogonal to KERMIT, other than to the extent to
which it points out the silliness of fancy user interfaces in what was
supposed to be a file transfer program.
[Ed. - So true. By the way, I can't find any other forum for discussion
of these issues in the "list of lists", so I don't mind if the topic
continues in Info-Kermit for a while.]
------------------------------
Date: Mon 7 Oct 85 21:47:01-EDT
From: Jin Au Kong <F.KONG%MIT-EECS@MIT-MC.ARPA>
Subject: The Claimed C-Kermit Bug for VMS
The problem with "! XXX" in VMS is related to the BYTLM in user authorization
file, as we discovered soon after we installed C-kermit on our system. For
VMS to create a subprocess, the BYTLM must exceed 4096 (which is our previous
setting), and of course, PRCLM must be at least 1. I don't know the exact
minimum for BYTLM, but after we set it to 6144, it worked.
But note that the created subprocess will not be stopped after you get out
from Kermit. Hope somebody can fix the problem.
[Ed. - Thanks, noted in the CKVKER.BWR file. Does this mean that the
previous report is wrong and that no change needs to be made to the code,
or that both are necessary? Anybody want to contribute "installation
instructions" for C-Kermit under VMS, and/or a review of its usefulness?]
------------------------------
Date: Sat, 27 Jul 85 15:55 PDT
From: IVA3GET@UCLAMVS.BITNET
Subject: Four-Table ASCII/EBCDIC System for EBCDIC Kermits
[Ed. - This message was recently excavated and is now belatedly published.
It was apparently provoked by the new ASCII/EBCDIC translation feature of
VM/CMS Kermit v2.]
The two ATOE's and two ETOA's would differ in the following way.
First let us denote the system tables ATOE0 and ETOA0. We will construct
ETOA1 to be the left inverse of ATOE0, i.e. for every printable ASCII
character x, ETOA1(ATOE0(x)) = x. Similarly, we will construct ATOE1 to be
the right inverse of ETOA0, i.e. for every printable ASCII character x,
ETOA0(ATOE1(x)) = x. The -1 tables are used to postmap incoming packets
back to ASCII and to premap outgoing packets out of ASCII. They form an
outer layer to Kermit so it can analyze and build packets in ASCII. If the
-0 system tables are nonstandard, then the -1 will be too. Note that the
right inverse may not be unique and that either inverse may fail to exist.
The other function of translation tables is to map text messages back and
forth between ASCII and EBCDIC as packets are analyzed and synthesized.
Call these ETOA2 and ATOE2. Ideally these should be based on the "standard"
in the IBM System 370 Reference Summary or the Appendix of the Kermit User's
Manual, and thus would differ from the -1 tables in the nonstandard
situation. Currently, -1 tables do double duty as they perform the text
message translation function as well. The result is various distortions in
text as it is transmitted to and from EBCDIC systems, including undesirable
substitutions and swallowing of characters. In a four-table system as I
have outlined, these distortions would not occur.
What is left is to state mathematically the conditions under which the
-1 tables can be constucted, and to present the appropriate algorithms.
. ETOA1 exists iff ATOE0 is 1-1 (unambiguous) on the printable set.
. ATOE1 exists iff the range of ETOA contains the printable set.
With standard tables, these questions do not arise since in this case
the system tables are both left and right inverses of each other.
This is all I have time for now. Hopefully, I can sketch the algorithms
later. I ran into the problem of distorted characters at UCLA where TSO
Kermit has recently been installed with modified translation tables. I
wanted to download Kermit-MS in the .BOO format as well as the Basic program
to decode it. I noticed that the tilde (or degree sign) was getting
swallowed and that the backslash was being blanked out. It would have been
easy enough to hand patch the Basic, but hardly the .BOO file. I hope I
have convinced you that there is a problem.
Ciao -
Glenn Thobe
iva3get@uclamvs and vss7853@uclavm (bitnet)
[Ed. - Given the number of calls I get every day from IBM mainframe sites
who have "customized" their translate tables, I don't need convincing that
there's a problem. But I'm not sure I understand how a 4-table system will
necessarily solve it. Your level 1 tables that invert the system's tables
will only work if the system's tables are already unique and invertible --
in many cases they are not. If 2 different printable ASCII characters are
mapped by the system to the same EBCDIC character, then the even the low
level stuff won't work -- some packets will be perceived has having wrong
checksums. In such cases, the only solution is to fix the system's tables.]
------------------------------
Date: Wed, 25 Sep 85 13:10:12 PDT
From: spacerad@JPL-VLSI.ARPA
Subject: Request for Osborne and Kaypro Kermit Diskettes
To: info-kermit@cu20b.arpa
I have been in touch with Frank da Cruz regarding our local (Los Angeles
area) Osborne club acting as a distribution house for Osborne and Kaypro
Kermit diskettes and documentation. the details are all worked out, but
I do require copies on disk of latest versions and doc or library files
for these programs. I would also like to obtain a copy of the Kermit
User Guide.
Anyone who can assist in this matter may reply directly to this message or
contact me also via:
1) dantas@jpl-vlsi.arpa
2) BOB DANTAS
% JET PROPULSION LAB
4800 OAK GROVE DRIVE
MAIL SLOT T-LL
MAIL SLOT T-1180 (CORRECT ONE)
PASADENA, CALIF. 91109
3) (818)354-4932
[Ed. - I'll send a User Guide. Could someone who has Kermit on Kaypro or
Osborne diskette please send in a copy? Thanks!]
------------------------------
Date: Tue 1 Oct 85 09:07:52-PDT
From: Steve Dennett <DENNETT@SRI-NIC.ARPA>
Subject: MacKermit TAC Problem
The NIC has gotten several calls lately from users having trouble getting
Macintosh Kermit to work through a TAC. I've tried doing file transfers
and have been equally unsuccessful. The odd thing is that it's a one-way
problem.
Going through a TAC, files can be downloaded from host to Mac without
difficulty. But when uploading, MacKermit sends three packets then
starts re-trying until it finally times out.
I've read the general information about using Kermit through a TAC and
have successfully moved files in both directions through a TAC using the
IBM PC version of Kermit, so the problem is something specific to MacKermit.
Also, I've tried all the different TAC twiddles (BIS/BOS, flow control,
varying packet sizes, etc.) but no combination seems to make a difference.
The version of MacKermit I'm using is .8(33) July 1985. Since you're listed
as one of the authors, I hope you can help. With the growing number of net
users and the popularity of the Mac, this question is certain to come up
with increasing frequency.
Thanks for your help.
Steve Dennett ( dennett@sri-nic.arpa )
DDN Network Information Center
[Ed. - Well... You've got the latest version of Mac Kermit, so that's not
the problem. I've never used a TAC personally, so all my information is
second hand. @B I S/@B O S makes the TAC transparent, so once you've done
that, you should be able to rule out any interference by XON/XOFF (which the
Mac doesn't do anyway), atsigns, etc. The fact that you can download files
to the Mac seems to confirm this. Therefore, I'd look again at the TAC's
buffers. If there's a way to make them bigger, do that. If not, you've got
to get the Mac to send shorter packets; to do this, tell BOTH Kermits to
reduce their packet sizes. What may be happening is that the effect of
commands like "set send packet-length" might be the opposite of what you
expect -- some Kermits take this to mean that you want to override whatever
the other Kermit asks for, and while others do the opposite. Let me know
how it works. If you still have problems, find out as much as you can from
the two Kermits involved -- note all the current communications and protocol
settings, get debug and/or packet logs if possible.]
------------------------------
Date: Wed 2 Oct 85 21:30:17-EDT
From: Robert S. Lenoil <LENOIL@MIT-XX.ARPA>
To: prindle@NADC.ARPA
cc: info-kermit@CU20B.COLUMBIA.EDU, lavitsky@RED.RUTGERS.EDU
Subject: Re: Kermit 1.7 Loses at 1200 Baud
I tried your suggested fix of setting the RS-232 registers to 8 so that my
modem could autobaud correctly, and then resetting them to zero. This worked
in that my modem did autobaud correctly and go into high-speed mode. However,
when I reset the rs-232 regs to zero, the host could no longer understand me.
Of course, if I left the registers at 8, I dropped characters. The symptoms
are this: my transmit data light goes on, but the host does not return any
character (I am in full duplex). After restarting with Kermit 1.5 (what I'm
using now), I saw that the DEC-20 was receiving nothing but back-quotes ("`").
I've rejected the possibility that my download went poorly, since I used
Kermit, and because the hex file has its own checksums. Again, my modem is a
ProModem 1200, by Prometheus. Has anyone else seen this behavior exhibited?
------------------------------
Date: Wednesday, 18 September 1985 17:43-MDT
From: pjohnson@sdcsvax.ARPA (Paul Johnson)
Subject: Kermit for Wang 2200?
Does anybody have the source for Kermit on a Wang 2200?
paul johnson
{akgua,allegra,dcdwest,decvax,ihnp4,helios,ucbvax}!sdcsvax!pjohnson
[Ed. - To my knowledge, no Kermit program exists for any Wang systems other
than the PC, and no one is working on any. Does anybody know something to
the contrary?]
------------------------------
Date: 26 Sep 1985 1331-EDT
From: LSM.DUNCAN at DEC-MARLBORO.ARPA
Subject: Kermit for DG Machines?
Is there ayone who could provide a copy of a Data General Kermit for an
MV4000 system in binary form? We have a system with no compilers and a
cartridge tape. Alternatively, is there a way to get a binary version
from a system so it could be downloaded with a 'crude' transfer program?
Thanks,
Jeff Duncan (lsm.duncan@dec-marlboro)
------------------------------
Date: 2 Oct 85 22:13:21 EDT
From: Steven Christensen <SC1K@CMU-CC-TE>
Subject: Kermit for CDS 4000?
Is anyone working on a Kermit for ComputerVision's CDS-4000 computer? It's
basically a FORTRAN generic machine, with some strange idiosyncronicities.
Steven Christensen
Phone: (513) 752-4595
Address: 728 Stuart Lane Cincinnati, Oh 45245
------------------------------
End of Info-Kermit Digest
*************************
-------