info-kermit@ucbvax.ARPA (12/05/84)
From: Frank da Cruz <SY.FDC%CU20B@COLUMBIA.ARPA>
Info-Kermit Digest Tue, 4 Dec 1984 Volume 1 : Number 41
Departments:
ANNOUNCEMENTS -
Kermit for the ICL/Three Rivers PERQ
Kermit for the Pascal Microengine
CP/M-86 Kermit Version 2.9
CP/M-86 Kermit for Tektronix 4170
MISCELLANY -
How to Bootstrap Kermit for the Victor 9000
MS-DOS Kermit RUN Command
----------------------------------------------------------------------
Date: Tue 4 Dec 84 12:44:00-EST
From: Peter Thew, Computer Centre, RMC Duntroon, Australia
Subject: Kermit for the ICL/Three Rivers PERQ
To: Info-Kermit
The Pascal version of Kermit for RT-11 systems, written at the University of
Toronto, has been heavily modified for the ICL/Three Rivers PERQ. The
command parsing has been improved by using the PERQ's parsing routines which
allow pop-up menus and command files. Binary file transfer has been
included (but it is rather simple and needs to be improved).
Some features that are to be added in the future are:
. Server commands
. VT100 emulation during CONNECTS.
. Pop-up menus for all commands
(eg. SET --> SPEED --> baud-rates)
. General code clean up, including faster disk I/O using FileSystem
routines.
Peter Thew
Computer Centre
Australian Defence Force Academy
ACT 2600 Australia
[Ed. - The files are in KER:PQ*.* on CU20B, available via anonymous FTP.]
------------------------------
Date: 27 Nov 84 21:22:31 PST (Tue)
To: SY.FDC@CU20B
Subject: Pascal Microengine kermit implementation
From: "Tim Shimeall" <tim%uci-icsd@columbia.arpa>
I have adapted the UCTERAK version of kermit to run on a Western Digital Pascal
microengine. It was not a difficult translation (your protocol designers are
to be complemented on its clarity) and the Microengine is not a common system.
To get everything working properly I had to make rather widespread (but not
extensive) changes to the Cornell Terak version, such that it would be
difficult to tie it down to just a few files (every file in the program has
been changed at least slightly).
In the process of converting over, I spotted a few bugs in Terak Kermit. The
most serious is that it does *no* timed waiting for packets; it just checks
10000 times to see if SOH has arrived. From a colleague here, I understand
that UCIBM-PC kermit has problems as well. I have corrected this bug in the
transported version (can I suggest calling it UCMICRO?) The following is the
list of changes I've made in UCTERAK kermit to make UCMICRO kermit:
- Added device declarations copied from Microengine hardware documentation
- Replaced external assembly language routines with Pascal versions
- Modified debug messages to be label values printed
- Changed format of packetwrite display to show header fields
- Implemented machine-dependent packet timeout
- Added debug packetwrites in recsw
- Added wrap-around debug info region
- Added legality check in showparms
- Removed lf elimination check in echo procedure
- Unitwrite calls replaced by calls to device driving routines
- Most uses of char_int_rec replaced by ord and chr
- Removed queue (no interrupts)
- Used sets for integer ops to getaround Microengine bug
- Changed parser from a unit to a segment procedure to allow swapping
- Eliminated "sendbrk" procedure (couldn't determine its use)
Tim
[Ed. - The program is in KER:UCMICRO.PAS and KER:UCMICRO.DOC on CU20B.]
------------------------------
Date: Sun 2 Dec 84 15:49:29-PST
From: Ronald Blanford <CONTEXT@WASHINGTON.ARPA>
Subject: CP/M-86 Kermit Version 2.9
To: cc.fdc@CU20B.ARPA
I've been making changes off and on to 86kermit, but the next month or two
looks slack so I'll give you the current version for testing and possible
release. The files are in my account as usual with the source in
86KER*.* and the hex and binary in APCKERMIT.*.
The specific features that have been implemented in version 2.9 are:
o LOCAL DIRECTORY command now computes file sizes correctly
for all files. The size given is the actual allocation on
disk, and not the logical size (which might differ for
non-sequential files).
o LOCAL TYPE command has been implemented to display (text) files
on the screen. A wildcard filespec is accepted and files displayed
alphabetically. The display is paged in Unix fashion with --more--
displayed on the last line. Typein options at that point can be
obtained by hitting a '?'.
o Wildcard SENDs now send files in alphabetical order by name, and
accept an optional initial filename in the command line to allow
transmission of partial groups in the manner of TOPS-20 Kermit.
o Problems with use under Concurrent CP/M on the APC have been fixed.
In particular, a KERMIT.INI file is no longer required, the
SET DEFAULT-DISK command works correctly, and a process dispatch
is performed each time a call to the serial port status routine
returns negative, vastly improving the response of other jobs.
There is still no provision for mutual exclusion on the serial
port.
[Ed. - The files are in:
KER:86*.* source, documentation.
KER:APCKERMIT.H86 hex for NEC APC
KB:APCKERMIT.CMD 8-bit binary for APC
KER:RBKERMIT.H86 hex for DEC Rainbow
KB:RBKERMIT.CMD 8-bit binary for Rainbow
The old files will be set aside for a while in case problems appear.
Also see next message...]
------------------------------
Date: Tue 4 Dec 84 18:00:00
From: Frank da Cruz <SY.FDC@CU20B>
Subject: CP/M-86 Kermit for Tektronix 4170
To: Info-Kermit
CP/M-86 Kermit 2.9 also includes support for the Tektronix 4170. The
support comes from a system-dependent module, like the ones for the Rainbow
and the APC, contributed by Robert Raymond, TransEra Corporation, Provo, Utah.
He says it works just like the APC and Rainbow versions. His module compiled
and linked with the other modules with no apparent problems.
The Tektronix 4170 i/o module is in KER:86KERIO.TX4. The hex is in
KER:TX4KERMIT.H86 and the binary in KB:TX4KERMIT.CMD.
------------------------------
Date: Sat 1 Dec 84 12:51:06-EST
From: Peter D. Junger <JUNGER@CWRU20>
Subject: How to Bootstrap Kermit for the Victor 9000
To: info-kermit@CU20B
In response to the recent request for a way to get the
Victor 9000 Kermit onto the Victor without having a Kermit
already on that machine, I can explain what we did. We used
Kermit on another machine (I forget whether it was a North Star
Horizon or an IBM PC) which did run Kermit to download the Victor
Kermit source code. We then used Crosstalk to transfer the code
to the Victor and then assembled it on the Victor. As I recall
it took more than 128 K of memory to get the program to assemble.
For the assembler we used the one--I assume that its
Microsoft--that comes with the Victor Programmer's Toolkit (which
Victor didn't supply without cost).
Crosstalk isn't magic, as long as there is some file
transfer program which exists on the two micros. If there is no
communications program on the Victor in question--as I suspect
may be the case--I can send a floppy disk with the version which
works for us to the person who needs it. (I can't look at his
message while I am typing this.) I am very disorganized, so the
best way to reach me is over CCNet: JUNGER@CWRU20. I do not
believe that I can be reached directly or indirectly from
ARPAnet, so the sender may have to request you to relay it.
I hope that this is of some help. I will not quarantee
that our copy works very well, we hardly ever use it, since I am
the only one here that makes much use of Kermit and I do not use
the Victor as my main machine.
Peter Junger
------------------------------
Date: Mon, 3 Dec 84 15:04:08 cst
From: allegra!noao!utastro!nather@Berkeley (Ed Nather)
To: carina!allegra!ucbvax!info-kermit@Berkeley
Subject: MS-DOS Kermit RUN Command
The "run" command in mskermit is very useful in feeling around for a file
under MS-DOS 2.0, and for doing a few other things. It has a couple of
peculiarities, though: it requires the full name, with extension, for the
executable file; usual ms-dos procedure executes a .com, .exe or .bat file
if the extension is left off, and most users are used to it. It took me a
long time to guess what Kermit wanted.
By the way, it may not be obvious to some users that Kermit can't handle
a .bat file at all; the usual symptom is a hang that requires a power-off
reset to get the computer's attention again. A warning to this effect in
the next version of the manual might save someone a headache.
------------------------------
End of Info-Kermit Digest
*************************
-------