[18630] in Athena Bugs
Fwd: screen
daemon@ATHENA.MIT.EDU (Bill Cattey)
Thu Dec 21 01:34:37 2000
Message-ID: <QuEOFt5z0001BEl6B0@mit.edu>
Date: Thu, 21 Dec 2000 06:34:33 +0000 ()
From: Bill Cattey <wdc@MIT.EDU>
To: bugs@MIT.EDU
CC: rtansley@MIT.EDU, dspace-local@MIT.EDU
(I originally sent this to bug-gnu@mit.edu, but that list does not exist.
Where DO bugs in the gnu locker go?)
A collaborator from HP, (athena user rtansley) is using a SOCKSified ssh
to connect to Athena from HP in Bristol England. The system he's logged
into is a Linux Athena 8.4.17 system with the gnu locker attached.
screen is not working for him.
Could we sort out why?
-wdc
Details:
---------- Forwarded message begins here ----------
From: "Tansley, Robert" <Robert_Tansley@hplb.hpl.hp.com>
To: "'Bill Cattey'" <wdc@MIT.EDU>
Cc: dspace-local@MIT.EDU
Subject: RE: screen
Date: Wed, 20 Dec 2000 09:34:13 -0000
> Are you still having trouble getting 'screen' to work?
> Compiling it for yourself is probably not a good long term solution.
>
> Would you repeat for me the exact command line and symptoms you see.
> I'll get it into the bug reporting pipe here.
>
> -wdc
OK:
I type "screen" (after "add gnu"). This should give me a new shell, and let
me create new shells with ctrl-a, c, but instead just hangs. "screen -ls"
in another login, shows that a screen has been created but is dead:
There is a screen on:
13825.pts-2.m10-423-3 (Dead ???)
Remove dead screens with 'screen -wipe'.
1 Socket in /tmp/uscreens/S-rtansley.
The hung screen doesn't respond to SIGTERM (ctrl-c). Killing it means the
dead screen is left there and needs to be wiped. I also tried creating a
detached screen using "screen -m -d". This appears to create a working
screen, screen -ls then gives:
There is a screen on:
13894..m10-423-3 (Detached)
1 Socket in /tmp/uscreens/S-rtansley.
When I try to attach to the screen with "screen -r" I get:
Suddenly the Dungeon collapses!! - You die...
and "screen -ls" lists the screen as "Dead" again.
"screen" is quite vital for me due to the flakeyness of the HP connection,
so I did compile up my own screen, which works perfectly. As long as my
compiled version of screen creates the screen session, the athena-supplied
screen can attach to it fine; it's something about the creation of the
screen session that's going wrong with the athena version.
Rob.
