SY.CHRISTINE@CU20B.COLUMBIA.EDU (Christine M Gianone) (03/02/88)
Info-Kermit Digest Tue, 1 Mar 1988 Volume 7 : Number 7
Departments:
ANNOUNCEMENTS -
Recent FTP (and general TCP/IP) problems with CU20B
New BOOing and UnBOOing Programs, and German Kermit Documentation
Keasy - "Easy Kermit" documentation in TeX
Kermit-CMS Updates
MS-DOS KERMIT -
Minor problems in Kermit-MS V2.30
MSKermit on Zenith eaZy-PC Serial (Mouse) Port
Kermit & OS/2
Int 14h, Kermit and Networks
MS-Kermit Tek 4010 Emulation Typeover -vs- Overlay
IBM PC MS-Kermit vs Datatronics Internal Modem
MS-DOS Kermit Using EGA 43 Line Mode
Toshiba T3100 versus MS-Kermit Tek Emulation
MISCELLANY -
C-Kermit on sys-V Based PCs
Using TOPS-20 Kermit with 9-Bit Files
Re Mark Zinzow's request for MicroBee Kermit
----------------------------------------------------------------------
Date: Mon 29 Feb 88 18:13:55-EST
From: Ken Rossman <sy.Ken@CU20B.COLUMBIA.EDU>
Subject: Recent FTP (and general TCP/IP) problems with CU20B
Keywords: FTP
Recently, there has been an increasing frequency in the number of problems
experienced with FTP file transfers from the Kermit directories on CU20B.
While I am not sure of all of the different possible causes for this recent
set of problems, I do know that we have sorely needed some IP fixes in the
TOPS-20 monitor on CU20B.
Well, the fixes are finally installed!
I'm hoping that the IP free space fixes we've recently installed here will
help FTP performance to/from CU20B. I'd be interested in hearing from some
of the sites out there who were so recently having FTP performance problems
to CU20B. Do things appear to be working better now?
/Ken
------------------------------
Date: 12 February 1988, 16:21:39 SET
From: Gisbert W. Selke <RECK@DBNUAMA1.BITNET>
Subject: New BOOing and UnBOOing Programs, and German Kermit Documentation
Keywords: .BOO Files, German
I have converted MSB*.PAS to Turbo Pascal 4.0, and sent them as MSBMKB.PAS
(which makes .BOO files) and MSBPCT.PAS (which decodes them).
I've also sent updated copies of the corresponding Fortran versions,
MSB*.FOR. I have removed some more system dependencies which I hadn't even
been aware of before (thanks to Stefan Kaufmann for pointing them out), and
also fixed a minor bug in MSBMKB.FOR due to which garbage was added at the end
of a BOO file.
And I have also sent a new, corrected German translation of the MSKERM.HLP
file, under the name MSKGER.HLP.
I just noticed that there is no 63 second restriction on MS-Kermit 2.30's SET
INPUT DEFAULT-TIMEOUT and the like, as I seem to recall there used to be.
That's great, since our host is sometimes *quite* sluggish!
\Gisbert
[Ed. - Thanks Gisbert! Your contributions have all been installed with the
MS-DOS Kermit distribution.]
------------------------------
Date: Mon, 8 Feb 88 16:53 N
From: Barbara Rosi <ROSI%VAXPR.INFNET%IBOINFN.BITNET@CUVMA.COLUMBIA.EDU>
Subject: Keasy - "Easy Kermit" documentation in TeX
Keywords: TeX, LaTeX
Thank you very much for giving us Kermit, it has been a valuable resource and a
pleasure to participate in the community. Enclosed is a translation of part of
the Kermit User Guide from Scribe into LaTeX. Thank you very much for your
interest.
[Ed. - Thank you! It has been put into KER:KEASY.TEX.]
------------------------------
Date: Mon, 1988 Feb 29 18:34 EST
From: (John F. Chandler) PEPMNT@CFAAMP.BITNET
Subject: Kermit-CMS Updates
Keywords: CMS Kermit 4.0
I have forwarded new versions of IKCKER.BWR and IKCKER.UPD covering
all bug reports and complaints to date about release 4.0 of Kermit-CMS.
The updates are largely, but not entirely, to the generic part of the
code, and some have no effect on CMS operation (they were needed for
the TSO version coming out soon). Only three of the updates touch on
problems that had no simple workarounds in the original release:
a. Files of RECFM V being downloaded are no longer trimmed of the
trailing blanks on TEXT transfers,
b. the user may now explicitly upload a file to a specific filemode
(and filemode number) through a Kermit-CMS server, and
c. there is no longer a garbage message when WARNING is ON and the
automatic renaming facility runs out of names to try.
The details can be found in IKCKER.BWR for the other changes.
John
[Ed. - Thanks John!]
------------------------------
Date: Fri, 5 Feb 88 09:05 EST
From: Pete Kanaitis <X979PK0P@VB.CC.CMU.EDU>
Subject: Minor problems in Kermit-MS V2.30
Keywords: MS-DOS Kermit 2.30
I wanted to point out two minor problems with KERMIT-MS 2.30.
The first is about using flow of control. With flow of control
enabled, (which it is by default), some programs on your DEC hosts that use
^S and ^Q as special characters, rather than XOFF/XON, (such as EMACS) will
get an extra ^S or ^Q sent into the input buffer. In the case of EMACS, a
^S or ^Q gets sent into the editing buffer. Depending on your baud rate, I
have noticed that ^Q gets sent into the input buffer at faster baud rates,
(9600), and ^S gets sent at slower baud rates (1200).
This problem did not exist in the 2.29b versions. I noticed that this
started to occur when the Tektronix code was added to Kermit in the later
2.29c versions. It was mentioned to me that the serial port handling was
changed when the Tektronix code was added.
A work around for this problem is to set flow of control to none. But by
doing this, I have found that some of my host characters get garbled at 9600
baud, when long lists go scrolling by on my screen.
[Ed. - The real, though painful, workaround is to SET FLOW NONE before you
start EMACS, and SET FLOW XON/XOFF when you exit EMACS, see below.]
I would like to suggest that this be mentioned in the beware file, unless
that there is some solution for this problem.
The second problem is about starting up Kermit in Tektronix mode on a port
connected to a modem (in my case, a Hayes 1200 external).
First starting Kermit by:
C>KERMIT
IBM-PC Kermit-MS: V2.30 8 Jan 1988
Type ? for help
Kermit-MS>set term tek
Kermit-MS>c
Now try to type a few "AT" commands to the modem. You may find that you may
have to hit Control-Break once to get the "AT" to work. Even before that,
you are unable to use Alt-Minus to switch terminal types.
If you already have a host connected to the port, you will not experience
this problem. If you start up on a modem port (with no connection
established) in VT102 mode, hit Alt-Minus a few times to get into Tektronix
mode, the same problem occurs. You cannot type "AT" until you hit the
Control-Break once. Even after that, toggling back to VT102, then back to
Tektronix, and typing "AT" sometimes causes a bell to sound. (This is *NOT*
the keyboard lock high-pitched bell).
If you have an external modem, you can make this observation. Once
connected, look at the Transmit and Receive lights on your modem when you
switch into Tektronix mode by pressing Alt-Minus. You will see the Transmit
light flicker every time you enter Tektronix mode. I am guessing that this
is where the serial port handling is different. Is the serial port being
reinitialized at this time? Maybe this is why I cannot type the "AT" from
Tektronix mode.
I have tested these problems with my modem with both DTR turned on and off.
There was no other resident software on my PC at this time.
Thank you for your time.
Peter Kanaitis
Research Systems Analyst
Allegheny-Singer Research Institute
Pittsburgh, PA 15212
(412) 359-3180
X979PK0P@VB.CC.CMU.EDU
[From JRD -- Pete. Once again, thanks for the extended comments. Emacs and
flow control: what can I say? If Kermit is permitted to do flow control and
Emacs responds to xon/xoff as commands then one must turn off flow control,
or else move the Search and Quote EMACS commands to to some other keys
besides Ctrl-S and Ctrl-Q.
Speed: the speed/baud rate/code sequences are the same for XON and XOFF so I
don't quite understand your observations unless you are referring to lag
time of the VAX responding to our hand-typed XON/XOFFs. Please note that
changing screen modes to enter or exit Tektronix graphics emulation involves
a full Bios video Mode Set and that takes a long time with interrupts turned
off. To prevent overruns from the serial port I bracket the mode-set with
XOFF/XON to suspend the host. Some modems will echo these characters and
thus suspend Kermit for Set Receive Timeout (13) seconds unless SET TIMER
OFF is given. That is the cause of troubles starting the modem. Bell noises
mean characters were lost in an overrun. Try SET TIMER OFF before starting a
Tek session since the timer is mainly for (a) packet timeouts and (b) to break
XON/XOFF deadlocks of precisely the kind which you encountered (local and
remote sides tell each other XOFF simultaneously); expect lost characters
during text-graphics screen changes. It's a limitation of the machine's
architecture.]
------------------------------
Date: Fri, 5 Feb 88 19:25:22 PST
From: Samuel_Lam@mtsg.ubc.ca
Subject: MSKermit on Zenith eaZy-PC Serial (Mouse) Port
Keywords: MS-DOS Kermit 2.30, Zenith Kermit, eaZy-PC Mouse Port
On a Zenith eaZy-PC, there is a built-in serial *mouse* port. Experiments
had shown that ProComm can use it as a communications port at speed up to
19,200 baud without problem, but MS-Kermit will hang the machine at the
first character. Does anyone know why?
(To rephrase the question, if a serial port works for a serial mouse *and*
ProComm (at high baud rate), why wouldn't it work with MS-Kermit 2.30?)
BTW, the port address of this only serial port is 0x02F8, if that matters.
Any pointers would be appreciated.
..Sam
Internet: <Samuel_Lam%UBC@um.cc.umich.edu> or <Samuel_Lam@mtsg.ubc.ca>
UUCP: {ihnp4!alberta,watmath,uunet,uw-beaver}!ubc-vision!ubc-mts!skl
BITNET: <USERSKL@UBCMTSG>
[From JRD - My Microsoft serial mouse does the same thing when I talk to it
from Kermit while in TEKTRONIX mode (for testing, naturally, since the mouse
is a poor conversationalist). The cause for me is the mouse echoes the
XOFF/XON sent while the screen changes to graphics mode, which is a very slow
process partly done with interrupts turned off. But communications resume
about 13 seconds (SET RECEIVE TIMEOUT interval) later as Kermit breaks the
XON/XOFF deadlock. Another likely situation is the port does not use the same
Interrupt ReQuest line as IBM machines or even that the mouse driver grabs
material away from Kermit. Release 2.30/A will help resolve the IRQ business
automatically but can do nothing about aggressive drivers.]
------------------------------
Date: Mon, 8 Feb 88 14:59:03 est
From: snorthc@NSWC-OAS.ARPA
Subject: Kermit & OS/2
Keywords: MS-DOS Kermit, OS/2 Kermit
I tried to run MS-Kermit 2.29B and 2.30 under OS/2 in the DOS compatibility
box. It sort of worked. I think it might be OK at a slower speed, but this
was at 9600 baud. Problems include "bell sounds" (Ctrl-G's), and data that
should have been arranged in columns not lining up correctly. I think the
other side of the connection thought I was sending BREAKS as well.
I have seen similar results on an ATT 6300 running Kermit as a DOS task
under UNIX at 9600 baud and a 6mhz AT running Kermit under Wendin's PC UNIX.
I think the problem in each case is the PC is not fast enough to manage 9600
baud and a 'real' operating system.
There was a plan at one point to merge PC Kermit into C-Kermit. This might be
the best migration path towards OS/2.
Stephen Northcutt (snorthc@nswc-g.arpa)
[From JRD -- Slowness is the result of all the overhead of context switching
from protected to real mode which must be occuring once for each received
character. The Control-G Beeps mean the UART received a new character before
the last was extracted - an "overrun" situation. Microsoft warns of this
effect on 80286 based machines due to the design deficiency of that Intel
chip; 80386's are vastly better. The context switch to real mode involves a
full reset of the cpu chip (yikes!) and, clearly, interrupts are off;
selected 80286 chips have a faster reset pathway but they seem to be present
or used in fast reset mode only in the PS/2 machines. Conversion to a full
protected mode OS/2 version of C Kermit is indeed on my mind; that is a lot
of work and costs a few dollars. I have no direct information on Wendin's
O/S products beyond that stated in a few magazine articles.]
------------------------------
Date: Mon, 8 Feb 88 15:36:43 est
From: snorthc@NSWC-OAS.ARPA
Subject: Int 14h, Kermit and Networks
Keywords: MS-DOS Kermit 2.30
The documentation for 2.30 Kermit mentions communicating with the BIOS via
Int14h on some clone PCs. (much improved documentation by the way, many
thanks).
The documentation for my PC's Bridge Communications Ethernet adapter card
mentions communicating with terminal emulation sw via Int 14.
I read and re-read the Kermit documentation, but could not find out how to
tell Kermit to treat my PC as a clone and use the BIOS. Finally I modified
line 1548 of MSXIBM.ASM to read "jmp chkpor1". The resulting Kermit used
the BIOS. When I Connected the Bridge PCS1> prompt appeared. With Kermit I
was able to telnet to a host computer. I was even able to use a Kermit
server on the host computer to transfer a file to my PC. Kermit at 10 mbs
takes a little getting used to, but I am sure I will.
Have I missed something? Was there a Kermit command to force use of the
BIOS? The problem with my version of Kermit is that now it always finds the
Bridge board, not the serial port.
If I didn't miss anything is it an option worth having? I am under the
impression there are several networking boards and software that allow Int
14 interfacing. Perhaps an addition to the Set Port command to include BIOS
along with 1 - 4 and NET. I am even willing to attempt the coding, but be
warned, the MSXIBM.ASM change was my first foray into assembler.
Stephen Northcutt (snorthc@nswc-g.arpa)
[From JRD -- Stephen. The story here is Kermit uses the Bios only if a
suitable UART is not available, but not by command of the user. LAN vendors
have invented a variety of interesting extensions to the serial port Bios
interrupt vector but I do not have documentation from them to do anything
about those methods. The best suggestion currently is to do what you did:
get sources and make a local change; that's one of the reasons for sources
being available. At least Bridge chose to emulate regular Bios calls rather
than using proprietary settings of registers. Did Bridge really let you run
TCP/IP's Telenet? If so we'd really like to know!]
------------------------------
Date: Mon 08 Feb 1988 15:39:27 EST
From: <SS@LL.ARPA>
Subject: MS-Kermit Tek 4010 Emulation Typeover -vs- Overlay
Keywords: MS-DOS Kermit, Terminal Emulation
>>>>Date: Sun, 10 Jan 88 20:59:48 PST
>>>>From: Ya'akov_Miles%UBC.MAILNET@um.cc.umich.edu
>>>>Subject: MSKERMIT ver 2.29C tektronix 4010 won't overlay ALPHA and VECTOR
>>>>Keywords: MS-DOS Kermit 2.29C
>>>>
>>>>Possible bug with MSKERMIT version 2.29C:
>>>>
>>>>I have been using MSKERMIT vers 2.29C and have encountered a bug (feature?)
>>>>in the Tektronix 4010 emulation mode. Specifically, alpha mode characters
>>>>will erase and NOT OVERLAY vector mode data. This frequently causes plot
>>>>titles and labels to obliterate the graphed data.
>>>>P.S. I have an IBM-PC/XT clone with the CGA adaptor.
>>>>
>>>>[Ed. - This seems to be the behavior on the CGA, but not the EGA...]
>>
>>[From jrd - A complete 8 by 8 dot character cell is written for each
>>character, so that both foreground and background colors are maintained and
>>characters can be erased. Supression of the background is being considered
>>but may not be do-able and still maintain the above characteristics. If
>>annotation is done first then the problem does not occur. I'd like
>>overlaying too.]
I have already undone this 'improvement' to the Tek 4010 emulator in my copy
of kermit (2.30). It involved only about a half dozen lines in the
MSGIBM.ASM file.
Stuart Scharf (SS@LL)
[From JRD - Making the change in the code is easy. The reason for doing it
my way is explained above, mostly erasure (after all, we are a few years
beyond punched cards and paper tape). The Tek storage tube characteristic
of overwriting everything is messy in today's environment, especially with
my typing skills. If the Kermit community would rather have the overwriting
then I'll remove the full cell approach and hence erasures. What's the
consensus?]
------------------------------
Date: 10 Feb 88
From: Robert W. Lane, Iowa City, IA (via US mail)
Subject: IBM PC MS-Kermit vs Datatronics Internal Modem
Keywords: MS-DOS Kermit 2.30, Modems
I have a Datatronics internal modem, 1200 baud, and I'd like to configure
it as COM3 at 03E8h using the 8259 interrupt line IRN4 (also used by COM1).
MS-Kermit 2.30 will let me access this modem using SET PORT 3 after running a
short program called SETCOM3 in order to update memory at 40:0000h with the
address of the modem. Unfortunately, this modem is built around a 8031
microprocessor and when Kermit tries to ascertain whether or not the port found
in 40:0000h has an 8250 async controller, it fails and prints a message that
all calls will go through the Bios. Is there any way to make Kermit skip this
test and just use the interrupt vector at IRQ4 (at 0030h)?
[From JRD -- Selecting an IRQ (4 in this case) is only part of the story.
Kermit needs to talk with the UART chip to set/get the baud rate, send/receive
characters, and so forth, and thus Kermit needs to know a lot about the UART.
That is why the tests are done before activating a serial port and why the
Bios is used if an 8250 chip is not found. If the Datatronics modem has such a
specialized pseudo-UART then only software tailored to it can be expected to
function. You might wish to query the manufacturer about this situation.]
------------------------------
Date: Thu 18 Feb 88 11:43:24-PDT
From: PAWKA@NOSC-TECR.ARPA
Subject: MS-DOS Kermit Using EGA 43 Line Mode
Keywords: MS-DOS Kermit 2.30
I've found a couple of bugs in MS-Kermit 2.30 for the IBM when using EGA 43
line mode, I'm going to try to look at the code when I get a chance, but in
case someone else fixes it first (or jrd has already fixed them!):
1. The cursor is a dash after exiting connect (I saw the note in
MSKERM.BWR, it's kind of a pain to have to run a program to fix the
cursor every time you pop in and out of kermit.
2. The STAT command doesn't clear the screen, it starts from the top
but leaves stuff at the bottom.
3. Scroll up doesn't work, it only goes 3 lines and only once. I tried
to fix this in 2.29, but gave up when I heard 2.30 was imminent, hoping
it would be fixed.
Mike Pawka
PAWKA@NOSC-TECR.ARPA
[From jrd - Mike, EGA 43 line mode stuff. Thanks.
Cursor is a dash after exiting Connect mode. This appears to be an EGA board
problem. IBM had bugs in their original EGA boards and Kermit takes steps to
avoid it within Connect mode. However, testing here with a Video 7 EGA board
in all kinds of strange screen modes indicates no problems.
Similarly, the lack of screen clearing for the STATUS command does not
appear here. I just check the item by going into 43 line mode to write this
note. If you are letting ANSI.SYS be active then it knows only 25 lines
whereas in Connect mode Kermit takes charge.
Only 3 lines on Scroll Up commands. That means Kermit found insufficient
memory for screen buffers and Command.com together so sacrificed buffer
space to allow subsidary tasks to be run via Command.com. Use CHKDSK or
Kermit's SPACE command to see the total free memory. However, memory space
is easily fragmented by loading TSR's.
What brand of EGA board and machine are you using?]
------------------------------
Date: Monday, 22 February 1988 10:24am EST
From: Frank da Cruz <SY.FDC@CU20B>
Subject: Toshiba T3100 versus MS-Kermit Tek Emulation
Keywords: MS-DOS Kermit 2.30, Terminal EMulation, Toshiba Kermit
It has come to our attention that Tek Emulation doesn't work on the
Toshiba laptop. Not only doesn't it work, it hangs the machine so badly
that only powering it off and on will revive it. It seems that the Toshiba,
although apparently IBM compatible in other respects, requires a special
function call to put it in graphics mode. The question is, how can MS-Kermit
tell, at runtime, that it is running on a Toshiba 3100, rather than a real
IBM PC? Is there a ROM location that has a unique machine identifier?
------------------------------
Date: Fri, 19 Feb 88 00:56:53 EST
From: hedrick@aramis.rutgers.edu (Charles Hedrick)
Subject: C-Kermit on sys-V Based PCs
Keywords: C-Kermit 4C(058)
Re: Info-Kermit Digest V7 #6
Because of the question about Kermit on various sys-V based PC's, I
thought I'd confirm that version 4C(058) works without any
modifications on Microport System V/AT, which is Sys Vr2 for a PC/AT.
I compiled it as sys3nid. (There is no need to request separate I&D
space. You always get that effect.) Microport supplies 4D, which
also works fine, but I built 4C (the most recent source we had lying
around at Rutgers) because I like to have my major applications
programs built from source. (Of course I have no way to know what
Microport may have done to bring up 4D.)
------------------------------
Date: Wed, 17 Feb 1988 23:03 MST
From: "Frank J. Wancho" <WANCHO@SIMTEL20.ARPA>
Subject: Using TOPS-20 Kermit with 9-Bit Files
Keywords: TOPS-20 Kermit
The KCC DEC-20 C compiler defaults to 9-bit files when you create a file
with "wb" for its own reasons - the internal character types are 9-bit, and
four 9-bit quantities on a read will properly pick up all 36 bits in a
PDP-10 word. However, all you have to do is change the "wb" in the open to
"w8" and you will get the expected behavior, i.e. 8-bit files which DEC-20
Kermit can recognize automatically.
[Ed. - Thanks, Frank! Your message has been added to the DEC-20 Kermit
"beware file".]
------------------------------
Date: Fri, 19 Feb 88 14:13 +1000
From: Andrew Hunt <munnari!rpepping.oz.au!ANDREW@uunet.UU.NET>
Subject: Re Mark Zinzow's request for MicroBee Kermit
Keywords: MicroBee Kermit, CP/M Kermit
Yes CP/M Kermit does run on a MicroBee - a colleague of mine has one
and uses kermit betweem it and his PC. He warned that the only problem
he had found occurred when using the optional dual serial line (SCC)
ports on the memory card in place of the one on the mother board. In
this case there are 3 serial ports on the machine which confuses the
poor wee beasite - solution is to disable the single mother-board port
and then all performs well up to 38400 Baud.
There exists - somewhere in the vicinity of "Public Domain" - a program
called PC-Alien (incl PC-Alien Jnr and MSA) for reading foreign CP/M and
MS-DOS diskettes on ordinary PCs. This includes support for MicroBee
format 360KB diskettes and I have successfully used it for reading and
writing such.
Regards ...Andrew HUNT, CSIRO Radiophysics, Australia.
------------------------------
End of Info-Kermit Digest
*************************
-------