home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Columbia Kermit
/
kermit.zip
/
newsgroups
/
misc.19990422-19990725
/
000114_news@watsun.cc.columbia.edu _Sat May 29 18:23:21 1999.msg
< prev
next >
Wrap
Internet Message Format
|
2020-01-01
|
5KB
Return-Path: <news@watsun.cc.columbia.edu>
Received: from newsmaster.cc.columbia.edu (newsmaster.cc.columbia.edu [128.59.59.30])
by watsun.cc.columbia.edu (8.8.5/8.8.5) with ESMTP id SAA06968
for <kermit.misc@watsun.cc.columbia.edu>; Sat, 29 May 1999 18:23:21 -0400 (EDT)
Received: (from news@localhost)
by newsmaster.cc.columbia.edu (8.8.5/8.8.5) id SAA20974
for kermit.misc@watsun.cc.columbia.edu; Sat, 29 May 1999 18:05:44 -0400 (EDT)
X-Authentication-Warning: newsmaster.cc.columbia.edu: news set sender to <news> using -f
From: "Brian K. White" <linut@squonk.net>
Subject: cku195b07.sco234c crashing on reentry into server mode
Date: Sat, 29 May 1999 05:42:59 -0400
Organization: Deja News Posting Service
Message-ID: <374FB6A3.83E4DAF1@squonk.net>
To: kermit.misc@watsun.cc.columbia.edu
Anyone else seeing this?
hardware: 486-50, 16M ram, Digiboard PC/Xe (ISA, with 80186)
software: Xenix 386 GT 2.3.4, digiboard driver v5.3.0, cku195b07.sco234c
(also 195b05, will test old versions today)
Recipe to get symptom:
enter kermit
<do any command mode operations you like, or connect as a client, all
works ok>
enter server mode
<server works ok, even for long periods of uptime (tested overnight so
far)>
ctrl-c out of sever mode
<command-mode or connect as client works ok as before>
try to reenter server mode--->crash (detail of crash below, it doesn't
have to crash, but will if you let it.)
Symptoms: kermit hangs/crashes any time you try to reenter server mode
without exiting-restarting kermit too.
the first switch to server mode works fine, and the server works fine
for as long as you care to leave it up, tested overnight so far. but if
you exit server mode, and then reenter, it hang after printing two
messages:
"Now eneterin server mode on ttyi1h"
"Press ctrl-c to quit"
it gets one more linefeed after that, but doesnt clear the screen or
redraw the server file transfer display.
at this point, it is not necessarily crashed yet. it responds to a
ctrl-c and you get back a kermit prompt, and command-mode and connect as
client functions all work ok. if you try to go back into server mode the
same thing happens. you can try and abort and retry as many times as you
like.
if, when it is hung as I described, you don't ctrl-c out of it, the hard
drive becomes very active and the whole system (all tty's not just this
shell) becomes unresponsive or very slow to respond as though under
extreme load. this lasts about a minute or two untill kermit coredumps
with the message:
"Memory fault(coredump)" and the system returns to normal, and no harm
is done other than the console that was running kermit must be logged
out / back in to get the display working properly, but the shell itself
is still accepting commands even.
for the record: using the following scripts has been a perfectly
functioning work-around
the net effect i sthat you start the server with the shell script, and
any time you stop the server, it exits all the way out of kermit, so you
always restart with the shell script, and never enter server mode more
than the first time.
btw: this reminds me of a question: is #!/usr/local/bin/kermit supposed
to work on xenix? it's working on open server 5.05, but I had to use the
following two-file combo on xenix.
/usr/local/bin/ks:
----start----
kermit /usr/local/lib/kermit/server
----stop----
/usr/local/lib/kermit/server:
----start----
set port /dev/ttyi1h
!ditty fastbaud fastcook forcedcd -ixon -ixoff -ixany rtspace ctspace
ttyi1h
set flow keep
set speed 110
eightbit
enable all
set server display on
set server get-path /
set server login abc abc
server
----stop----
with the above setup, I can type "ks ctrl-c ks ctrl-c ks ctrl-c..."
starting/stopping the server as many times as I want with no problems.
notes on the above: ditty is a souped up stty that comes with the
digiboard, and is used to set communications parameters outside of the
stock xenix norm. the combination of setting 'fastbaud' on the port, and
then asking for speed '110' set the port to 115200. this version of
xenix does on it's own know about rts/cts, but this version of kermit
does not, so I set rts/cts with ditty, and then tell kermit not to look
at flow control at all. btw: the communication itself seems to be
working quite nice, even at the breakneck speed. downloading files from
the xenix box in server mode, to a Open Server box, running
cku195b07.sco32v505netgcc also via a digiboard, using a short,
hand-wired, sheilded, full null cable, I get about one retransmitted
packet per megabyte and the cps rate is between 10,000 and 17,000 ! the
files themselves arrive clean, accurate copies.
(binary mode, including executables which ran (are running) fine.)
also, setting the above ditty settings and just enabling the port for
normal getty login, and logging in from the open server box with kermit
works perfect, including lots of heavy ansi full color / upper ascii
screen drawing.
--
Brian~