SY.FDC@CU20B.COLUMBIA.EDU (Frank da Cruz) (09/14/85)
Info-Kermit Digest Fri, 13 Sep 1985 Volume 3 : Number 19
Departments:
MACINTOSH KERMIT -
New Macintosh Kermit Bug List
Macintosh Kermit VT100 Emulation Bottom-of-Screen Bug
Garbage in Upper Left Corner on MacKermit
MacKermit Comments
MacKermit/ScreenSaver Incompatibility
MacKermit Transfer Command
Macintosh Kermit Key Configurator Limitation
Mac Kermit Show-Response Window Too Narrow
Bug and Nit in Macintosh Kermit
Terminal Emulation Bug In Mac Kermit
Double Keying Option-Keys in Mac Kermit
Mac Kermit vs Yale ASCII
IBM MAINFRAME KERMIT -
Kermit for UTS with Series/1
CMS Kermit with 7171's (two messages)
Kermit-11 vs IBM VM/CMS
MISCELLANY -
Kermit Distribution in the UK
----------------------------------------------------------------------
Date: Fri 13 Sep 85 17:24:56 EDT
From: Frank da Cruz <SY.FDC@CU20B>
Subject: New Macintosh Kermit Bug List
The following messages report bugs in the current release of Macintosh
Kermit, or suggest improvements. These have been noted in the new "beware"
file, KER:CKMKER.BWR, which is available via anonymous FTP from CU20B. For
now, no changes are being made to the program, mainly because Mac Kermit is
still "between maintainers".
Although not mentioned specifically in conjunction with Mac Kermit, the
C-Kermit bug reported by Icarus Sperry (Bath Univ, UK) in Info-Kermit V3
#17, namely that checksums are computed incorrectly if padding is selected,
applies also to Macintosh Kermit.
------------------------------
Date: Thu 15 Aug 85 11:12:12-CDT
From: Don Nash <CC.DLNASH@UT-A20.ARPA>
Subject: Macintosh Kermit VT100 Emulation Bottom-of-Screen Bug
I just found found this simply wonderful bug in Macintosh Kermit. Here's how
to do it:
1) Set MacKermit 0.8(33) for 9600 baud, no parity, remote echo.
2) Connect to a VAX 11/780 running VMS 4.1 and Eunice and log in.
3) Set your terminal type by having the following lines in your
LOGIN.COM file:
$define TERM "vt100"
$set term /dev = vt100 /line
4) Type out a file which is long enough to fill the screen more
once using the following command:
type /page filename.ext
5) When you are prompted to hit return, type ^C. The word "Cancel"
should appear, but it should be partially below the screen.
6) Type <CR> several times. If you get the same results I did, then
the screen of the Mac should go crazy. It looks like snow on a
TV set, except that it is black snow on a white background (which
seems logical, since the Mac's screen is white). The Mac is now
completely wedged, the cursor does not respond to mouse movements.
The only way out is to reset the Mac and restart MacKermit.
Also, the number zero and the capital letter oh were identical on MacKermit,
and the same applies to the capital letter "I" (the letter after "H"), and
the lowercase letter "l" (the letter after "k"). They are also *exactly*
the same.
Don Nash
[Ed. - Thanks, noted in the beware file.]
------------------------------
Date: Thu 15 Aug 85 12:25:26-CDT
From: Don Nash <CC.DLNASH@UT-A20.ARPA>
Subject: Garbage in Upper Left Corner on MacKermit
I read the file CKMKER.BWR.8. In the section UNDIGESTED, item #3 said that
Dan Tappan from BBN reported garbage in the upper left corner of the terminal
emulator window which eventually went away. I have the same problem, but
it won't go away. It is nothing serious, but it does seem to be permanent on
my copy of MacKermit 0.8(33). It does not interfere with the display in any
way. Hope this aids in digestion.
Don Nash
------------------------------
Date: Mon, 19 Aug 85 11:45:56 PDT
From: msev%Phobos@CIT-Hamlet.ARPA
Subject: MacKermit Comments
I'm very pleased with the new MacKermit release. I use it now in
preference to MacTerminal. I like the speed: both initialization and
operation are much better than MacT. I have set up the key map to send
the right characters for EDT. File transfer with VMS is very smooth.
- Martin Ewing
Comments on 0.8(33) Kermit for Macintosh:
SERIOUS
1. The VMS TYPE/PAGE command types a page of output from a file and
types "Type enter to continue" (or some such) in reverse video on bottom
line of screen. If you ^Y at that point to terminate, Kermit writes the
next line ("Interrupt" in reverse video) on what appears to be line 25.
Kermit often dies shortly thereafter. I don't have the exact character
sequence that VMS is putting out.
[Ed. - Same as Don Nash's bug. Temporary workaround: don't use VMS
(just kidding...)]
NUISANCE
2. Screen is not thoroughly wiped by erase screen command. Garbage
characters left at right of screen (sometimes). This may have been
documented before.
3. A blinking cursor, or, better, a blinking solid cursor would be a
great help. It's nearly impossible to find the underscore in a page
full of text. A user-selectable cursor format would be preferable.
[Ed. - Good idea.]
SUGGESTIONS
4. It would be handy to distribute a standard VT100/numeric keypad
definitions file. This is important for running the VMS EDT editor.
I'll donate such a file if requested.
[Ed. - Please do.]
5. Apparently the VT10x graphics characters are not emulated. At least
the MONITOR command on VMS does not produce the desired line and block
characters. This would be a useful addition for those of us who like to
stare at these status displays.
6. Emulate the VT102 printer port commands.
------------------------------
Date: Thu, 8 Aug 85 10:14:02 mdt
From: jad%b@LANL.ARPA (John De Vries)
Subject: MacKermit/ScreenSaver Incompatibility
I have received a report that while making long uploads from MacKermit while
ScreenSaver (a desk accessory) is active will cause the upload to bomb and
will indeed screw up the Mac's system, necessitating a reboot. It happens
when ScreenSaver goes and blanks the screen...
Zoz
------------------------------
Date: Fri 9 Aug 85 16:55:15-EDT
From: Don Lanini <Us.Dcl@CU20D>
Subject: MacKermit Transfer Command
Whenever I try to use the transfer command under the transfer menu to launch
an application from another disk (a disk other than the one Kermit is on), it
blows up the machine with an ID = 26 error and hangs.
The version is 0.8(32)
------------------------------
Date: Wed 19 Jun 85 16:30:52-EDT
From: Bill Schilit <BILL@COLUMBIA-20.ARPA>
Subject: Macintosh Kermit Key Configurator Limitation
I just read a message on the net... you might want to configure a system
disk with the Dutch or German, or French international resources and check
out the results of this on MacKey...
I think the result will be an incorrect keyboard (check this out in
relation to the KEYCAPS display). Since the key names are hard
coded in an array in the program...
------------------------------
Date: Tue, 9 Jul 85 18:43:58 edt
From: Mike Ciaraldi <ciaraldi@rochester.arpa>
Subject: Mac Kermit Show-Response Window Too Narrow
When you give a remote command (e.g. Directory) to a server, the results
come up in a window. You can scroll the window from side to side, but even
if you move it all the way to the left (i.e. the little box in the
horizontal scroll bar is all the way to the right), you still can't see the
right end of the line of text. You have to stretch the window to see it.
[Ed. - I agree this might not be intuitively obvious to the user...]
When you do a "Get file from server", after the transfer is complete the box
that keeps you informed of the progress of the transfer still stays on the
middle of the screen until you click on it. This is not really obvious as
the way to continue. What makes it bad is that the mouse pointer still
shows as the little clock face indefinitely, implying that you ought to wait
before proceeding (maybe while it closes the file or something).
[Ed. - Right, this has been noted in the .BWR file all along.]
------------------------------
Date: 23 Jul 1985 02:04 PST
From: Charles Carvalho <CHARLES@ACC>
Subject: Bug and Nit in Macintosh Kermit
A friend pointed out a bug and a nit in the Macintosh Kermit:
bug: if you call SFGetFile with the numTypes argument equal to zero,
it apparently doesn't find all files; numTypes should be -1 instead.
[Ed. - From one of the authors: "I haven't noticed any problems, and I
believe the code reflects the documentation." Has anyone actually observed
files being skipped over?]
nit: the Close box at the left of the MacKermit window doesn't do anything;
it ought to be left out (define the window with the NoGoAway attribute).
/Charles
------------------------------
Date: Wed 31 Jul 85 15:18:44-EDT
From: Ben Fried <UI.Ben@CU20C>
Subject: Terminal Emulation Bug In Mac Kermit
I don't know if i can reproduce this, but i definitely found a bug in mac
kermit 0.8(33)
I was in emacs, typing a command into the echo area, when a send came in. the
first few scan lines of the text showed up at the bottom of the screen, then
it froze--i had mouse control, but the event handler must have frozen,
because i got no response to any sort of update event -- keydowns, clicking
the mouse, whatever.
it looks like it didn't handle the scrolling of the text i was sent correctly,
and died somewhere in there. but i'm not sure.
[Ed. - "Doubt it was the event handler, more likely it is a scroll problem with
the characters being drawn off the screen." Noted in .BWR file.]
------------------------------
Date: Sat, 7 Sep 85 19:10 MST
From: Barry Margolin <Margolin@HIS-PHOENIX-MULTICS.ARPA>
Subject: Double Keying Option-Keys in Mac Kermit
I think I've figured out why certain characters have to be doubled in
order to send them in MacKermit. These characters are normally set to
nonspacing diacritical marks. For instance, Option-e is normally a
nonspacing accent grave. In the normal Macintosh text input
environment, nothing gets sent until you type the next character after a
diacritical, so that it can send the appropriate accented character. If
you type the diacritical twice, it sends just the diacritial character.
If you type a character that cannot be used with that diacritical, it
sends the diacritical followed by the character. This is all consistent
with the behavior I have seen when I use the Option key as a Control
key. I suspect that the same problems would exist if I used the Option
key as a Meta key, which I believe is the default.
[Ed. - This was also noted in the .BWR file that originally went out with
0.8(33).]
------------------------------
Date: Thu, 22 Aug 85 19:47:45 edt
From: Mike Ciaraldi <ciaraldi@rochester.arpa>
Subject: Mac Kermit vs Yale ASCII
We just started using MacKermit to talk to our IBM 4381, which uses a Series
1 front end with the Yale Ascii package. On the other end we installed the
version of TSO Kermit from U of Toronto.
File transfers work OK, it seems.
Big problem with terminal emulation. The packages we use on MVS use lots of
boldface. The CKMKER.BWR file says that boldface characters are the wrong
size, but for us it's worse than that. The cursor proceeds along, putting
stuff on the screen, then skips back and erases several characters, leaving
a series of gaps in the line of text. If you open up a window over the
affected area, then close it again, the text is OK. It's not in boldface,
but it is perfectly readable and in the right place.
[Ed. - The problem is still probably due to the characters being a different
size from what they program thinks they are. Some IBM/Yale-ASCII sites have
found the problem so annoying that they define a new terminal type for the
Macs, which is basically a VT100 without boldface.]
We tried opening up the SHOW RESPONSE window, stretching it to almost fill
the screen, then switching back and forth between it and the main window to
force refreshing of the screen image in the main window. This works, but
the lower right corner of the SHOW RESPONSE window is "dead", i.e. clicking
on it will not bring the window to the front. The dead area is the part
where the size-changing icon usually appears. When the main window is in
the background, there doesn't seem to be this problem getting it back.
Mike Ciaraldi
ciaraldi@rochester
[Ed. - Mike mentioned in a subsequent message that they would try to come up
with a fix for the boldface problem.]
------------------------------
Date: 12 September 1985, 15:31:45 EDT
From: Philip Murton <MURTON@UTORONTO.BITNET>
Subject: Kermit for UTS with Series/1
We are no longer running UTS under VM here so that only version
of Kermit for UTS I have is based on the old UNIX Kermit.
[Ed. - This is the KERMIT.C from the Protocol Manual, with modifications
to support the Series/1, presumably it will also work with the 7171, 4994,
and other systems that run the Yale ASCII package. I've put it in
K2:UTSS1.C on CU20B.]
------------------------------
Date: Fri, 13 Sep 85 02:07:01 EDT
From: Peter DiCamillo <CMSMAINT@BROWNVM>
Subject: CMS Kermit with 7171's
I can confirm Steve Ostrove's experience that CMS Kermit through the 7171
does not log itself out, and only responds to FINISH. Even with FINISH, CMS
Kermit doesn't respond with "R;" until after I enter a carriage return back
in the terminal session.
From what I have heard, the packet size problem with 7171s is a performance
problem which only shows up when a 7171 is loaded. If Steve was testing
with only one or two users on the 7171, then a larger packet size would
work. I haven't had a chance to confirm this myself yet.
Kermit should be able to distinguish a Series/1 from a 7171 because a
Series/1 emulates a 3277 terminal, but a 7171 emulates a 3278 terminal. This
device type information is available to a program by using DIAG X'24' to
obtain information about the virtual console.
Peter DiCamillo
[Ed. - Can anyone provide a clue as to why CMS Kermit might refuse to log
itself out when run from a 7171, but log out OK when run from a 3705???]
------------------------------
Date: 13 September 85 14:35 EDT
From: NJG@CORNELLA.BITNET
Subject: CMS Kermit and 7171s
The problem with Kermit (CMS at least, and I expect the same with TSO) and
7171s should only show up if the 7171 can not field the incoming characters
as fast as they are sent. This forces the 7171 to use a 64 character long
circular type ahead buffer. Once this buffer fills up the 7171 will not
accept further input until the buffer starts to empty. As long as a set of
8 terminal ports is not 'over worked' it should be able to handle input from
a terminal at 9600 baud. It is only when more than one of the terminals on a
set of 8 ports is sending large data buffers at high data rates that there
will be overrun problems. I do not know at what data rates or how many
active terminals it takes to actually cause the problems, but 1 terminal at
9600 baud (plus the rest doing normal terminal type output traffic, not file
transfer) should not exhibit the problem. This problem has been discussed
with the IBM development team for the 7171. Personally I doubt that the
problem will be corrected in future versions of the device as it appears
that IBM does not intend to have the device support file transfer, but
rather to support attachment of dumb-terminals only.
":-)" Nick Gimbrone <NJG@CORNELLA.BITNET> (607)256-3747
------------------------------
Date: 13 SEP 85 07:39-EST
From: BRIAN%UOFT02.BITNET@WISCVM.ARPA
Subject: Kermit-11 vs IBM VM/CMS
The most recent Kermit digest had a comment regarding Kermit-11 and CMS.
Here is an explanation (?)
I should point out that Kermit-11 has never worked well with CMS Kermit
(before the S/1 support) as the U of Toledo does not use standard IBM
half duplex lines, thus making testing impossible. The controller here
is a CCI which fakes a FDX link. Kermit-11 does, however, work fine
with the S/1 support in the newest CMS kermit, as long as one remembers
to set parity to anything other than the default, none.
brian nelson
------------------------------
Date: 10-SEP-1985 10:39:01
From: SYSKERMIT%vax1.central.lancaster.ac.uk@ucl-cs.arpa
Subject: Kermit Distribution in the UK
Lancaster University (UK) have just installed Columbia tapes written there
on August 3. The files are all online on a VAX 11/780 and can be accessed
freely by anyone who wishes using NIFTP or KERMIT. We can also send out
files on 9 track tape, and on floppy for some versions. The collection is
regularly updated with new tapes and files pulled from cu20b direct. For
details mail to Alan Phillips, SYSKERMIT @ LANCS.VAX1 (JANET address
000010404000.FTP.MAIL, or PSS address 23425240101.000010404000.FTP.MAIL), or
phone 0524-65201 x 4881. Distribution is free to all educational
establishments: a small handling charge is made to commercial users.
------------------------------
End of Info-Kermit Digest
*************************
-------