info-kermit@ucbvax.ARPA (05/15/85)
From: Frank da Cruz <SY.FDC@CU20B.ARPA>
Info-Kermit Digest Tue, 14 May 1985 Volume 2 : Number 27
Departments:
C-KERMIT 4C -
C-Kermit 4C Bugs (Still)
Missing File for C-Kermit VMS Make Procedure
VMS C-Kermit 4C Works
LCK files, parity, "backing out", aborting FINISH, make errors
MISCELLANY -
VM/CMS Kermit 2.00 vs IBM 7171
MSTIPRO Kermit with Internal Modem and TEK-4010
----------------------------------------------------------------------
Date: Tue, 14 May 85 16:15:21 pdt
From: tweten@AMES-NAS.ARPA (Dave Tweten)
Subject: C-Kermit 4C Bugs (Still)
I just finished getting the changed source files you recommended in the
latest digest, and rebuilding on both a 4.2 bsd Vax and a System V
Vax. I retested for the "rem dir"-hangs-on-NAKs problem. It's still
there. I used the bsd Vax as the local Kermit, and the System V Vax as
the remote.
[Ed. - Dave reports the other problems fixed for these systems -- binary
files transfer ok, etc. But sending any "remote" command to the C-Kermit
4C server on the System V VAX times out. The 4.2bsd VAX C-Kermit 4C server
handles "remote" requests just fine, no matter what system they come from,
even from itself. Anybody who can figure this one out will be a hero.]
------------------------------
Date: 13 May 85 14:05 EDT
From: (David H M Spector) <SPECTOR@NYU-CMCL1.ARPA>
Subject: Missing File for C-Kermit VMS Make Procedure
I found a problem with the distribution kit for C-Kermit. The VMS
com file called "ckvmak.com", which becomes "CCMAKE.COM" is a copy
of ckver.com, the 'makefile'. Attempting to run the make file will
cause lots of errors as the com file goes into an endless loop.
David Spector
NYU/acf Systems group
[Ed. - Sorry! The correct file is now installed as <CKERMIT>CKVMAK.COM,
available via anonymous FTP from CU20B.]
------------------------------
Date: Tue, 14 May 85 13:04:15 EDT
From: stew@harvard.ARPA (Stew Rubenstein)
Subject: VMS C-Kermit 4C Works
Seems to work OK; I have not tested it extensively, but things mostly
work. There are some problems with file name canonicalization (it doesn't
strip off version numbers before sending). This is with the new modules
announced Monday.
[Ed. - Thanks, Stew. I'm glad I didn't break your code too badly...]
------------------------------
Date: Monday, 13 May 1985 13:26:55-PDT
From: d_schullman%sarah.DEC@decwrl.ARPA (Dan Schullman)
Subject: LCK files, parity, "backing out", aborting FINISH, make errors
Is it possible to clear the LCK* files on an interrupt? When
I accidentally leave the line as a "modem" line and can't open
it, an interruption leaves the LCK* file for the line around.
Yes, and it is done in many cases -- e.g. when ^C is typed, when various
fatal errors occur -- the program goes through doexit(), which is supposed
to clear any lock file. But some other interrupts are not presently caught,
and some can never be (e.g. if the process is killed from another process,
or the system crashes). I agree this is an area that needs improvement.
The new manual (CKUKER.DOC) discusses the perils of lock files in greater
detail than the previous one did.
I had trouble copying a binary file (but not a text one), and
thought it was a KERMIT 4C/4.2 incompatability. Eventually
turned out to be a parity problem. I don't understand why
text files copy okay but binary ones don't. I thought KERMIT
"packaged" the transmission so that only ASCII graphics would
be used. Both "ends" were set for the same parity ("none"),
and as stated I WAS able to copy text files.
This should be fixed in the current (circa Monday, May 13) release of
version 4C of the program.
There seems to be some interactions with quoting from the command
line. For example, try:
! echo "this should \07 ring"
Even escaping the backslash doesn't help.
This will have to be looked at. Meanwhile, you can always use a real
control-G in the string -- it works both in shell commands and in Kermit's
own echo command.
Is there a way to "abort" questions if one has "gone down the
wrong path"? For example, if one has answered the "from file"
question and wants to get out of the "to file" question. I
tried EOF (Control-D) but that didn't work.
Currently no, short of typing Control-C, which gets you out of the program
entirely. This too will have to be looked at.
I often "hang" trying to do a FINISH, BYE, etc. Is there a way
to abort these operations without exiting KERMIT? As mentioned
in previous mail, this is sometimes because the remote server
has returned to command mode for some unknown reason.
Currently, there is no way to "abort" a protocol operation before data packets
start to flow, short of typing ^C. This is true of most Kermits. Macintosh
Kermit is an exception. Unix Kermit will have to have a more powerful
interrupt facility added to it to fix this and several of the other problems
you've mentioned.
In building KERMIT on a SYS-V VAX, I got the following in addition
to problems caused by multiple includes of types.h by ckxunx.h.
"ckuser.c", line 916: warning: illegal combination of pointer and
integer, op =
These problems should be fixed in the current release.
------------------------------
Date: 14 May 1985 0954-EDT
From: LCG.KERMIT@DEC-MARLBORO
Subject: VM/CMS KERMIT 2.00 vs IBM 7171
I installed CMS kermit on our 3032 system that uses a 7171 front-end.
I had to comment out two lines to get everything to work. At
OKDEV - 3, remove 2 lines:
* clc cons772,temp
* bne baddev
After that change, everthing worked.
Thanks to Columbia for another great version,
Wade Missimer
Abbott Labs
[Ed. - Thanks; this hint has been added to the CMSKERMIT.BWR file.]
------------------------------
Date: Thu 9 May 85 21:41:48-EDT
From: Joe Smith (415)794-2512 <LSM.SMITH@DEC-MARLBORO.ARPA>
Subject: MSTIPRO Kermit with Internal Modem and TEK-4010
In reference to Richard Berke's questions, 3-May-85.
1) The file KERMIT:MSTIPRO.HLP on MARKET tells how to use the internal modem
on a TI Professional. This functionality was added before version 2.27
was released, but appearantly did not make it to VERSIONS.DOC.
2) The Tektronix emulation is for line drawing only. Since the system I
developed it on drew labels using pen strokes, no attempt was made to
make the alpha cursor track the graphics cursor.
3) Full HEATH-19 emulation is being added. The new maintainer is:
Dan Smith (303)273-3396
Colorado School of Mines
Computing Center
1500 Illinois Street
Golden, CO 80401
------------------------------
End of Info-Kermit Digest
*************************
-------