From GLS  Fri Nov 30 00:00:00 1979
From: GLS (Guy L. Steele, Jr.)
Date: 30 NOV 1979, 00:00
Subject: No subject
Message-ID: <mailman.911.1479821121.4695.its@nocrew.org>

    MSG: LOSER  1     
    DLW at MIT-AI 11/29/79 19:23:09 Re: Don't delete other people's files!
Subject: Don't delete other people's files!
    Some very recently deleted a group of files on my directory.
    These were very important information, and some had been
    quite recently modified.  If you are trying to create disk
    space, ASK people to delete their own files--don't do it yourself!
    The ITS no-security system can only work in an environment of
    reasonable, responsible people; deleting other people's files
    like this is highly irresponsible.

I wonder how hard it would be to make the following modifications
to the disk routines:
[a] when a file is deleted, shuffle its five-word block to the
middle of the directory rather than just overwriting it.  Thus, a
rotation of a number of five-word blocks would occur, I think.
[b] install the uname of the deleting job as the file's "author".

This would tend (though not guarantee) to preserve information
as to who deleted a file.  It might make it easier to deal with this
problem, which has come up a lot recently.

From RWK at MIT-MC  Wed Nov 28 00:00:00 1979
From: RWK at MIT-MC (RWK at MIT-MC)
Date: 28 Nov 1979 00:00
Subject: No subject
Message-ID: <mailman.912.1479821121.4695.its@nocrew.org>

Translations should canonicallize DSK: to MC: etc, or the result is
unpredictible.  (Do YOU know if MM Load LibraryFOO uses MC: or DSK: ??)


From RMS  Wed Nov 28 00:00:00 1979
From: RMS (Richard M. Stallman)
Date: 28 NOV 1979, 00:00
Subject: No subject
Message-ID: <mailman.913.1479821121.4695.its@nocrew.org>

Control-CALL echoes as just a Z.  It should probably echo as ^Z.

From CFFK at MIT-MC  Tue Nov 20 00:00:00 1979
From: CFFK at MIT-MC (CFFK at MIT-MC)
Date: 20 Nov 1979 00:00
Subject: Continuation of previous message
Message-ID: <mailman.914.1479821121.4695.its@nocrew.org>

Thus MC is receiving burst of ~ 70 characters at a time.
I realize that the real culprit is the TELNET program
that runs on the MFE-NET.  However is there any fix I
can make at MC to circumvent this problem?


From CFFK at MIT-MC  Tue Nov 20 00:00:00 1979
From: CFFK at MIT-MC (CFFK at MIT-MC)
Date: 20 Nov 1979 00:00
Subject: No subject
Message-ID: <mailman.915.1479821121.4695.its@nocrew.org>

If I log into MC from LLL-MFE, MC occasionally drops the last
few characters of the lines I type in (and the <cr>).  The problem
arises because I'm logged into LLL-MFE via the MFE-NET which works
a-line-at-time.  Thus MC is receiving 

From MAP  Thu Nov 15 00:00:00 1979
From: MAP (Michael A. Patton)
Date: 15 NOV 1979, 00:00
Subject: No subject
Message-ID: <mailman.916.1479821121.4695.its@nocrew.org>


The node A/1/a in ITSTTY cannot be reached by any normal means (that I
could find).  The only way I could get there was with G.

From REM at MIT-MC  Sun Nov 11 00:00:00 1979
From: REM at MIT-MC (REM at MIT-MC)
Date: 11 Nov 1979 00:00
Subject: Not a bug, rather a query/gripe/request
Message-ID: <mailman.917.1479821121.4695.its@nocrew.org>

Given the SIXBIT name of a STY device, such as "S20", or just the number
such as 20, it should be easy to find out whether or not that STY is in use
by any job (your own job, or anybody else).  Anybody know how to do it?
Apparantly currently you can't do it unless you know the TTY that is
associated with it, thus making it impossible to write a subroutine that
gets you a new STY you don't already have unless you have  been keeping
a list of STYs-you-have all along since you started acquiring STYs.


From DWOODSatMIT-AI  Wed Nov  7 00:00:00 1979
From: DWOODSatMIT-AI (Donald R. Woods)
Date: 7 November 1979, 00:00
Subject: net ports
Message-ID: <mailman.918.1479821121.4695.its@nocrew.org>

Hmm, okay, I guess I hadn't been willing to believe that completely
unrestricted simulated ttys would number only 6, so I assumed that
only a subset of them were allowed as net ports.  Amazing that you
can survive with only 6!  Assuming most of them are normally in
use for net ports or lisp-machine users, I find it hard to believe
that you can get enough STYs for CRTSTY and other uses.  Maybe you
don't use them as much on ITS as we do at SAIL.


From CEH  Tue Nov  6 00:00:00 1979
From: CEH (Charles E. Haynes)
Date: 6 NOV 1979, 00:00
Subject: No subject
Message-ID: <mailman.919.1479821121.4695.its@nocrew.org>

Recently I have become involved in a discussion of the relative merits
of CRTSTY vs TCTYP support for C100's.  My personal feeling is that
the TCTYP support is inadequate in a number of areas, the largest being
its lack of support for the "Sail" character set.  Granted CRTSTY
takes an extra job slot, I feel that its support is enough better
that I prefer it (CRTSTY).

What do other people think?

-- Charles



From RWKatMIT-MC  Wed Nov  7 00:00:00 1979
From: RWKatMIT-MC (Robert W. Kerns)
Date: 7 November 1979, 00:00
Subject: net ports
Message-ID: <mailman.920.1479821121.4695.its@nocrew.org>

    Date: 5 Nov 1979 2308-PST
    From: CSD.DON at SU-SCORE
    To:   bug-crtsty at MIT-AI
    Re:   net ports

    Incidentally, I consider it a crock of the finest shit that, for whatever
    reason(s), CRTSTY uses up an extra network port (of which I'm told there
    are only 6) instead of a system-internal simulated tty.  Do all STYs on
    ITS use net ports, or just CRTSTY STYs?  With net ports so rare, such an
    implementation seems rather silly.
It does use a system-internal simulated TTY.  That's what a STY is.  That is
also what a "net port" is.  It's not a limitation on how many network
connections there are, but on the number of system-internal simulated tty's.
While I would agree that 6 may be a little too few in the case of AI, it IS
a very heavily loaded machine, and there is justification for wanting to limit
the number of users coming in via the network.  This has nothing to do with
whether CRTSTY is implemented in the reasonable way; it is.


From ED at MIT-MC  Sun Nov  4 00:00:00 1979
From: ED at MIT-MC (ED at MIT-MC)
Date: 04 Nov 1979 00:00
Subject: I don't believe this
Message-ID: <mailman.921.1479821121.4695.its@nocrew.org>

DDT version 1388 runs on AI and MC.  On MC, lower-case
characters typed as the first thing to DDT's command
loop (as in foo/ or foo, as opposed to :foo) generates
an OP? error.  On AI, this is not the case.  Further,
loading SYS;ATSIGN DDT into a job gets the same foul
behavior, but doing :copy sys;atsign ddt,common;random file
and then loading common;random file WINS!
It seems that the pure page that's always in core is broken.
Starting up several 256K jobs that referenced all their pages
managed to get the offending page swapped out which fixed things.
>From this I gather that when a pure page is swapped, it is replaced
from the original, rather than put a dupe in the swap area.
Wierd.