From EAKatMIT-MC  Thu Jan 29 00:00:00 1981
From: EAKatMIT-MC (Earl A. Killian)
Date: 29 January 1981, 00:00
Subject: Page ahead
Message-ID: <mailman.702.1479821117.4695.its@nocrew.org>

Actually the documentation only existed on AI, not on the other
machines.  I've installed it elsewhere.


From KLH  Thu Jan 29 00:00:00 1981
From: KLH (Ken Harrenstien)
Date: 29 JAN 1981, 00:00
Subject: page ahead
Message-ID: <mailman.703.1479821117.4695.its@nocrew.org>

Did I miss something?  I have no idea what people are talking
about with respect to "page ahead".  Is that something like
disk file read-ahead in that if you reference a page, the next
sequential page is also swapped in?

From EAKatMIT-MC  Thu Jan 29 00:00:00 1981
From: EAKatMIT-MC (Earl A. Killian)
Date: 29 January 1981, 00:00
Subject: No subject
Message-ID: <mailman.704.1479821117.4695.its@nocrew.org>

When you first started hacking page ahead code in TECO/ITS I did
a quick calculation that showed that it would be a noticable
improvement on AI, but not so noticable on MC, because MC can
search a page faster.  I was moderately surprized at that result.
Anyway, glad to hear it really wins.  Of course, HIC's new memory
may help a lot too...


From RMS  Wed Jan 28 00:00:00 1981
From: RMS (Richard M. Stallman)
Date: 28 JAN 1981, 00:00
Subject: No subject
Message-ID: <mailman.705.1479821117.4695.its@nocrew.org>

Page ahead is a big improvement!
On a lightely loaded system, moving the gap in TECO
is about twice as fast.
When there gets to be more load, it slows down, but the
TECO, which was 165k with one 100k file, never even had
50 pages in at any time during moving the buffer from one
end to the other.

From RMS  Wed Jan 28 00:00:00 1981
From: RMS (Richard M. Stallman)
Date: 28 JAN 1981, 00:00
Subject: No subject
Message-ID: <mailman.706.1479821117.4695.its@nocrew.org>

When SPC told me he was working on a :S program,
I told him that there was already a :S.
But then I looked for it and didn't find any.
So I figured that the link had got lost somehow
and nobody had missed it, so I told SPC it was ok
to put in a new one.

I don't know when or how the old :S went away.
CENT, do you remember when you last used it?
I used to use it but got out of the habit a long time ago.

From RWKatMIT-MC  Wed Jan 28 00:00:00 1981
From: RWKatMIT-MC (Robert W. Kerns)
Date: 28 January 1981, 00:00
Subject: No subject
Message-ID: <mailman.707.1479821117.4695.its@nocrew.org>

    Date: 27 January 1981 23:24-EST
    From: Pandora B. Berman <CENT at MIT-AI>
    To: RMS at MIT-AI
    cc: BUG-ITS at MIT-AI, BUG-DDT at MIT-AI
    
    what did you do to :s? when i tried to use it to send a msg to
    more than one person (using :s because :send only takes one
    recipient) it kept saying No message? :KILL

    This is most frustrating; a standard thing like this should not
    change so suddenly/incompatibly/without warning so that when a
    user expects it to act as it always has, it breaks incomprehensibly
    instead. please explain and fix.

This was not RMS, this was SPC.  I have deleted the offending program
and restored the link to QSEND.  *** SPC DO NOT DO THIS! ***.  Replacing
public programs with your own version is not at all polite.  If this
happened because you did not look before you put it on SYS3;, you should
be aware that you *MUST* check for the name being in use on any of
SYS, SYS1, SYS2, and SYS3.  Please excersize more restraint and caution
in the future.


From CENT  Tue Jan 27 00:00:00 1981
From: CENT (Pandora B. Berman)
Date: 27 JAN 1981, 00:00
Subject: No subject
Message-ID: <mailman.708.1479821117.4695.its@nocrew.org>

what did you do to :s? when i tried to use it to send a msg to more than one
person (using :s because :send only takes one recipient) it kept saying
No message? :KILL
this is most frustrating; a standard thing like this should not change so
suddenly/incompatibly/without warning so that when a user expects it to act
as it always has, it breaks incomprehensibly instead. please explain and fix.

From EB  Tue Jan 27 00:00:00 1981
From: EB (Edward Barton)
Date: 27 JAN 1981, 00:00
Subject: SAUTH bug
Message-ID: <mailman.709.1479821117.4695.its@nocrew.org>

I have been playing around with MLDEV for a while, trying to figure
out why it so often hangs up while the DDT :COPY command is trying to
do a SAUTH call.  If anyone thinks the following fix is wrong, please
explain and delete it.  If anyone knowledgeable thinks it is right,
please assemble and install it.  It seems not to hang, and to
otherwise work, in my tests.

AI   SYSENG
FREE BLOCKS #1=91 #2=53 #14=116 #4=78 #5=101 #13=355 #15=298 #3=95 
  2   MLDEV  61       6 +919    6/09/80 03:28:23  (1/12/81) RMS   
  5   MLDEV  62       7  +86    1/12/81 19:41:51  (1/22/81) USERS1
  4   MLDEV  63       7 +139    1/22/81 00:51:45  (1/27/81) MOON  
  4   MLDEV  64       7 +143 !  1/27/81 20:08:40  (1/27/81) EB    

;COMPARISON OF DSK:SYSENG;MLDEV 63 AND DSK:SYSENG;MLDEV 64
;OPTIONS ARE    /3

**** FILE DSK:SYSENG;MLDEV 63, 17-39 (30351)
	SKIPE CALING	;AND IF USER IS STILL WAITING FOR RETURN FROM SYSTEM CALL
	 SOSLE CALACK	;AND THIS REPLY IS TO HIS CURRENT ATTEMPT, NOT A PREVIOUS,
	  JRST GOLOOP
**** FILE DSK:SYSENG;MLDEV 64, 17-39 (30351)
	SOSG CALACK	; NOTE THE ACKNOWLEDGE AND SKIP IF NOT CURRENT YET
	 SKIPN CALING	 ; CURRENT; SKIP IF USER IS WAITING
	  JRST GOLOOP	; NOT CURRENT YET, KEEP WAITING; OR DON'T WANT IT
***************


From GJC at MIT-MC  Mon Jan 26 00:00:00 1981
From: GJC at MIT-MC (GJC at MIT-MC)
Date: 26 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.710.1479821117.4695.its@nocrew.org>

Does anybody know about this ARGUS frob, SYS2;TS ARGUS on MC?
Seems innocent enough, but it takes up incredible amounts
of resources for a trivial program, was getting like
33% to 44% of MC for quite some time before I killed it.
Sounds kind of buggy. I noticed it because it was being
run by a lot of turists, that kind of thing usually is.



From abrax  Sat Jan 24 00:00:00 1981
From: abrax (Jeff Shrager)
Date: 24 JAN 1981, 00:00
Subject: Rubout to first char fails in DDT mode
Message-ID: <mailman.711.1479821117.4695.its@nocrew.org>

If you forget the ":" to issue a real command and use rubout to get
back to the beginning of the line and insert it, ITS forgets that
there is a ":".  Looks like someone's crufty code looks for the ":"
ONLY on entry of the first character rather than at the beginning
of the line entered after all of it comes in.

From KMP at MIT-MC  Wed Jan 21 00:00:00 1981
From: KMP at MIT-MC (KMP at MIT-MC)
Date: 21 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.712.1479821117.4695.its@nocrew.org>

The system just hung for about a minute or so without accessing the disk.
Now I did a peek about 5 seconds after it started letting me get to disk
and there were only about 13 disk channels taken. Unless a lot were closed
in a very short time, I wonder if the problem is really disk channels all
used up or something else...


From GJCatMIT-MC  Tue Jan 20 00:00:00 1981
From: GJCatMIT-MC (George J. Carrette)
Date: 20 January 1981, 00:00
Subject: No subject
Message-ID: <mailman.713.1479821117.4695.its@nocrew.org>

I know you need specific info about the state the running jobs where
in. What I'll do is set up some example runs, and make sure the 
behavior is repeatable, and catch you sometime system conditions
are right for a demo. 

-gjc


From MOON at MIT-MC  Tue Jan 20 00:00:00 1981
From: MOON at MIT-MC (MOON at MIT-MC)
Date: 20 Jan 1981 00:00
Subject: Terminal in 812
Message-ID: <mailman.714.1479821117.4695.its@nocrew.org>

    ITS doesn't seem to be listening to the terminal in 812.
    The AI ITS nnnn CONSOLE 36 ... message is coming through
    alright but the system doesn't respond to ^Z.
The system believes that Teleray has a meta key.  Evidently the parity
control switches in the back are not set correctly, such that
c-Z sends c-m-Z.  You want 8-bit characters, no parity, 1 stop bit.

From MOON at MIT-MC  Tue Jan 20 00:00:00 1981
From: MOON at MIT-MC (MOON at MIT-MC)
Date: 20 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.715.1479821117.4695.its@nocrew.org>

Were these jobs running in a system call or in user mode?  Were they
taking page faults?  Were they getting some sort of interrupts such as
pdl overflow?  Were you able to reown them and stop them with the
DDT control-X command?  Without some specific information there is
nothing I can do.

From JPG at MIT-MC  Tue Jan 20 00:00:00 1981
From: JPG at MIT-MC (JPG at MIT-MC)
Date: 20 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.716.1479821117.4695.its@nocrew.org>

Forwarded for GJC:
   GJC at MIT-MC 01/19/81 20:09:53
   To: JPG at MIT-MC, JM at MIT-MC
   foo, its worse now than ever. BALCH's two dissowned macsyma
   jobs are still running hours later, its 8:00pm, and
   the fair-share is 3%. But, somehow his jobs get
   a total of 60% share, plus all the physical memory they
   need, almost a Moby unshared. How come these CPU-intensive
   memory-intensive dissowned jobs don't get swapped out?
   It should be something like active-for 1 minute, inactive for
   4 minutes.


From jerryb  Mon Jan 19 00:00:00 1981
From: jerryb (Gerald R. Barber)
Date: 19 JAN 1981, 00:00
Subject: No subject
Message-ID: <mailman.717.1479821117.4695.its@nocrew.org>

ITS doesn't seem to be listening to the terminal in 812.
The AI ITS nnnn CONSOLE 36 ... message is coming through
alright but the system doesn't respond to ^Z.

From MOON  Mon Jan 19 00:00:00 1981
From: MOON (David A. Moon)
Date: 19 JAN 1981, 00:00
Subject: No subject
Message-ID: <mailman.718.1479821117.4695.its@nocrew.org>

CHANNA;RAKASH DMNSTR does not work in the new ITS,
I think because it tries to slave from a not logged in sty

From FREND at MIT-ML  Sun Jan 18 00:00:00 1981
From: FREND at MIT-ML (FREND at MIT-ML)
Date: 18 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.719.1479821117.4695.its@nocrew.org>

I just logged in from the net, and before my login init was thru,
I got a message from system overseer saying I did not look like an
authorized user. there are about 8 people on now, half are tourists.
I think this is a bug because if PWORD let me in at this time, the
overseer program should not say he doesn't recognize me. Anyway, i
will leave as soon as I ^C this message
					michael bloom


From KLH  Sun Jan 18 00:00:00 1981
From: KLH (Ken Harrenstien)
Date: 18 JAN 1981, 00:00
Subject: Spying on com-links
Message-ID: <mailman.720.1479821117.4695.its@nocrew.org>

Tell your friend to ask the UC Berkeley Comp Sci dept.  Their UNIX
systems have a talk program which, while a feeble imitation of ITS com
links (eg no multiple party links), is much easier to modify for
scripting, because it is a user program rather than part of the
operating system.  I suspect the patterns will be different (more
sophomoric dialogue, owing to thousands of freshlings) but why not
take advantage of a home-grown resource...


From ___022 at MIT-ML  Sun Jan 18 00:00:00 1981
From: ___022 at MIT-ML (___022 at MIT-ML)
Date: 18 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.721.1479821117.4695.its@nocrew.org>

A friend of mine would like to do a study of language pattern in com-links
as a project for her class in Spoken/Written language differences.  (She
views com-links as a cross between the two.)  Is there any easy way to have
ITS record com-links somewhere for a week or two?  The person doing the study
is Nancy Levidow, a grad student in linguistics at UC Berkeley.  (I'm assuming
there are no privacy problems with this, as any comlink could be spied upon;
just technical problems.)  Easiest way to respond is via GNU at MC, since she's
not on the net.



From EAKatMIT-MC  Wed Jan 14 00:00:00 1981
From: EAKatMIT-MC (Earl A. Killian)
Date: 14 January 1981, 00:00
Subject: system hanging on disk...
Message-ID: <mailman.722.1479821117.4695.its@nocrew.org>

Gee, you could implement real resource management, and require
that any program that uses more than one disk channel, first say
"reserve me N channels", avoiding the deadlock.  Just kidding.


From Moon at MIT-MC  Wed Jan 14 00:00:00 1981
From: Moon at MIT-MC (Moon at MIT-MC)
Date: 14 Jan 1981 00:00
Subject: system hanging on disk...
Message-ID: <mailman.723.1479821117.4695.its@nocrew.org>

Yes, exhausting the supply of disk channels is the usual cause of this.
There are a fairly humongous number I believe.  Maybe I should see if ITS
can be made to do something useful in this case.  (It used to return errors
instead of hanging, but someone decided it would be better to hang.)  Maybe
I can make it return an error after a while, which might encourage killing
of jobs.  TEX is a gross offender in this respect (so are PUB and COMPLR)
in that they open many channels, making a deadlock more likely.

From ELLEN at MIT-MC  Wed Jan 14 00:00:00 1981
From: ELLEN at MIT-MC (ELLEN at MIT-MC)
Date: 14 Jan 1981 00:00
Subject: system hanging on disk...
Message-ID: <mailman.724.1479821117.4695.its@nocrew.org>

I just happened to have a peek today when one of those hangs happened...
and by gosh and by golly, there were two full VT52 screens plus a few lines
of disk channels in use from a C in peek.  I dunno... lots of TeX's and various
INQUIRS (nasty Ellen here noting that some of the INQUIR channels were from
unlogged lines doing WHOIS or NAME...).  Anyway, it probably was running out
of disk channels.  How many are there?


From KMP at MIT-MC  Tue Jan 13 00:00:00 1981
From: KMP at MIT-MC (KMP at MIT-MC)
Date: 13 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.725.1479821117.4695.its@nocrew.org>

There has been a very high incidence rate of jobs hanging for a long time
(rlb: `several tens of seconds') waiting to do i/o to the disk. Is it possible
that this is due to an ITS bug or is it more likely to be people using up all
the disk channels/buffers/?? If it's possible that it's the former, could 
someone look into it? It's extremely irritating and has almost caused us to
reload the system unnecessarily a couple times because disk i/o was unavailable
for so long (one or two times ITS just started working again just as we were
about to reload it; these were several-minute pauses) ... 
-kmp


From EBM  Mon Jan 12 00:00:00 1981
From: EBM (EBM)
Date: 12 Jan 1981, 00:00
Subject: Slight protocol change, and new MLDEV/MLSLV installed
Message-ID: <mailman.726.1479821117.4695.its@nocrew.org>

The MLSLV protocol has been upgraded in response to some catastrophes
that occurred on DM about 6 weeks ago.  Specifically, a number of files
on SYS directories were deleted using MLSLV, and because of the way the
protocol is constructed, there was no way to track down the culprit,
even thought the usual console message were written.

MLSLV has been extended by the addition of a LOGIN command and reply
code.  Logging in is REQUIRED only for deletion of files, but for
simplicity and friendliness, MLDEV starts a login (but does not wait for
the response) almost immediately.  This should not particularly slow
down the protocol.

Instead of logging in as hhhJuu (host and user) MLSLV, it now logs in as
XUNAME (of the foreign user) MLSLV, with a "terminal name" of hhhJuu.
It should still be easy to match up jobs, connections, etc., via PEEK,
and much easier to identify who is doing what.

If it is deemed necessary to de-install the new MLSLV and MLDEV, it was
installed using links:
	On each site, SYS;ATSIGN MLDEV is a link to SYS;ATSIGN NMLDEV,
	and SYS;ATSIGN OMLDEV is the old MLDEV.

	MLSLV is organized the same way, but lives in DEVICE;
Bugs can be sent to me.

Although this change is slightly incompatible, it should not prove to be
a real problem for users, or for upgrading of other programs using the
protocol.

-------

From KMP at MIT-MC  Sun Jan 11 00:00:00 1981
From: KMP at MIT-MC (KMP at MIT-MC)
Date: 11 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.727.1479821117.4695.its@nocrew.org>

:COPY TTY:,DSK:KMP;> >
Hi.^M^J^C

created a file called _COPY_ OUTPUT on my dir with `hi' in it ... While this
*is* the greatest file of the last type in my dir, I should think there are
more aesthetic filenames it could have chosen.


From BARRYG at MIT-MC  Fri Jan  9 00:00:00 1981
From: BARRYG at MIT-MC (BARRYG at MIT-MC)
Date: 09 Jan 1981 00:00
Subject: strange interaction between RMAIL and TCTYP GLASS
Message-ID: <mailman.728.1479821117.4695.its@nocrew.org>

I normally run under TCTYP GLASS SCRL 22 (See MC: GUEST0; BARRYG LOGIN)
and most programs (INFO, PRINT, MSGS, MAIL) that I use get along well with
that.  When using RMAIL, however, when the screen  fills I get **MORE**
(instead of --MORE-- as with other programs).  No matter WHAT character is then
typed, RMAIL prints out the rest of the message (again printing out **MORE**
approximately every screenful but not pausing for input).

I haven't tested this in EMACS, so I can't tell if the problem is general with
EMACS or specific to ^R RMAIL.  If you need help with this, I'll try a few
things under EMACS to try isolating it.  

This is a minor bug (so far), but since we tourists are supposedly here
for the purpose of testing programs, I present this bug for your delectation.

	barry gold

From PDL  Thu Jan  8 00:00:00 1981
From: PDL (P. David Lebling)
Date: 8 Jan 1981, 00:00
Subject: No subject
In-Reply-To: Message of 08 Jan 81 at 1211 EST by BERN@MIT-ML
Message-ID: <[MIT-DMS].178600>

The author is stored as the directory number, which is why you see
USERS1.  Unfortunately, there isn't room in the directory entry for
the full name.




From BERN at MIT-ML  Thu Jan  8 00:00:00 1981
From: BERN at MIT-ML (BERN at MIT-ML)
Date: 08 Jan 1981 00:00
Subject: No subject
Message-ID: <mailman.729.1479821117.4695.its@nocrew.org>

Now that passwords are implemented on most machines, it is more
appropriate when a file is created, written, or linked, that the
author be the uname rather than the sname (or whatever it currently
is).  It is hard to find out who wrote in a file when the author 
name is USERS1 for example.