Sun-Spots-Request@RICE.EDU (William LeFebvre) (06/07/88)
SUN-SPOTS DIGEST Monday, 6 June 1988 Volume 6 : Issue 104
Today's Topics:
A few words about the archive server
Re: 875 Meg Disk yet
Re: Opinions about ArborText preview versus VorTeX dvitool
Re: Consult/Gateway
Re: "panic: iechkcca" from 2 machines in one night
[Unofficial] Patch to Fig 1.4
SUN delivery times
Extra security for logins?
AI/LISP development on SUN 4-series?
psvf with bounding box?
Searching for past messages?
Send contributions to: sun-spots@rice.edu
Send subscription add/delete requests to: sun-spots-request@rice.edu
Bitnet readers can subscribe directly with the CMS command:
TELL LISTSERV AT RICE SUBSCRIBE SUNSPOTS My Full Name
Recent backissues are available via anonymous FTP from "titan.rice.edu".
For volume X, issue Y, "get sun-spots/vXnY". They are also accessible
through the archive server: mail the request "send sun-spots vXnY" to
"archive-server@rice.edu" or mail the word "help" to the same address
for more information.
----------------------------------------------------------------------
Date: Mon, 6 Jun 88 14:28:12 CDT
From: William LeFebvre <phil@Rice.edu>
Subject: A few words about the archive server
Well, actually it is many, many words:
For those who do not know, the archive server is a means by which
non-ARPANet sites can access the Sun-Spots archives. It is an automatic
mail-response program originally written by Brian Reid of DECWRL. There
are a few things concerning the use of the server that I would like to
take the time to clarify. Primarily they have to do with failures.
The archive server only has one source for a reply address: the mail
header of the requesting message. It merrily mails its reply to the
address it finds in the mail header. If the header contains a "Reply-to:"
field, it will use the address specified there. Otherwise, it will (in
most circumstances) use the contents of the "From:" field. The most
common problem that the server has is that the address it mails to is
bogus. This could happen for any number of reasons, the two most common
being either the header was not constructed correctly at the originating
end or some site along the way "munged" it in an invalid way.
Unfortunately, when this happens, the mail message gets returned to the
server. The server detects the message as a failure and forwards it to
me. I do not have the time to track down the true address for every
failed mail message I get from the server. The end result from the
requestor's point of view is that the request fell into a black hole.
There is a way around this problem. If the request contains a line of the
form "path address", it will mail the response to "address". All requests
received by the server are either acted upon or acknowledged within an
hour of receipt. So you should receive some sort of reply (even if it is
just an acknowledgement) within a day of mailing your request. If you do
not, then add a "path" line with an address that would work *from Rice*
(not necessarily from your machine) to your request and try again.
Another common problem is "I got the acknowledgement but never saw the
actual file". This is almost certainly caused by the size of your
request. Large mail message have difficulty travelling through the mail
system. Sometimes they have difficulty getting out of Rice! Uucp sites
typically place an upper limit on the size of a mail message. If your
requested file exceeds that limit, the failed mail message gets sent
to---you guessed it---me.
The automatic detection of mail failure messages also gives rise to a few
problems. First, the userid from which the request is mailed cannot
contain certain substrings: daemon, mailer-daemon, system, root. This
means that "root" cannot request things from the server. Unfortunately,
it also means that "root2" cannot request things either (someone did try
that recently). Second, the server does not distinguish between comments
and the true address: it pays attention to text found in parentheses
(this is a bug I hope to fix soon). If your "From:" address looks like:
"john@golf.com (John Smith, systems programmer, golf clubs international)"
then the server will think it is a failed mail message. Why? Because the
string "system" appeared in the mail address. Using "path" in this case
won't fix the problem. When one of these two problems exist, the request
gets treated as a failed mail message and gets forwarded to---you guessed
it---me.
I will try to be more conscientious about turning around failures that are
not caused by bad paths, but please understand my severe time contraints.
The archive server is provided on an "as is" basis. I can only put in so
much of my time in making it work smoother. If you do not get a response
to a request within a few days, it is very likely that the server's
response failed in its delivery attempt. If you send mail to
"archive-management@rice.edu" saying "what happened to my request", I will
try to track it down as time allows.
People with direct ARPANet access shouldn't even bother---use anonymous
FTP instead.
William LeFebvre
<phil@Rice.edu>
------------------------------
Date: Tue, 31 May 88 09:04:18 EDT
From: nesheim@think.com
Subject: Re: 875 Meg Disk yet
We have received several systems with the Hitachi drives, (4/280 systems).
Our local sales rep says that there is some delay in getting them not
because of the drives, but because of a shortage of mounting brackets! In
any case we have recieved one system 3 drives short, and are eagerly
awaiting the rest of the system.
Bill Nesheim
Thinking Machines Corp. Cambridge MA
------------------------------
Date: Tue, 31 May 88 09:08:21 PDT
From: rusty@cartan.berkeley.edu
Subject: Re: Opinions about ArborText preview versus VorTeX dvitool
ArborText's Publisher is very good. It has WYSIWYG equation and table
editors and integrated graphics. If you have technical typists that
aren't using TeX or aren't using Unix it could well be the way to go. On
the other hand, if you are already using TeX or Unix you might not like
the Publisher. My big complaint about the Publisher is that you have to
use their editor in order to use the Publisher because they store the .tex
files in their own format. People that already use a powerful editor with
macros and whatnot may have a setup that makes editing TeX files fairly
easy and straightforward; for example emacs. Going from emacs to the
Publisher's editor is a big step down. I wish they would split out the
WYSIWYG equation and table editors into separate programs which output raw
TeX, and allow their previewer to accept plain ascii text files and allow
you to use whatever editor you want. (It would also be nice if they would
make it work under X windows so that it could be used on other machines.)
The current monolithic setup just isn't flexible.
As far as previewers go, dvitool is easily in the top 5, if not #1.
------------------------------
Date: Tue, 31 May 88 15:55:40 EDT
From: texsun!sunrock!ed@sun.com (Ed Pendzik {Consulting} Sun Rochester)
Subject: Re: Consult/Gateway
Reference: v6n97
>>Has anyone used a product called (I think) Consult/Gateway for the Sun 3?
>>It is supposed to allow using up to 4 ethernets under SunOS 3.x.
The standard SunOS release supports two ethernet connections per slotted
Sun. One on the CPU itself and one Sun ethernet board in a slot. The
consulting group within Sun sells a "special" called CONSULT-GATEWAY which
supports four ethernet boards. It is available for SunOS 3.x and Sys4-3.2.
>[Could this be true? Are third-party vendors selling SunOS bug-fixes?]
No.
>Doesn't the operating system (ab)normally support this?
Umm .., No. As I said above, some kernel sources need to be recompiled to
support more than two ethernet connections per slotted Sun.
Regards,
Edward F. Pendzik
Sun Consulting Services
Rochester, NY
------------------------------
Date: Tue, 31 May 88 17:41:30 EDT
From: hedrick@aramis.rutgers.edu (Charles Hedrick)
Subject: Re: "panic: iechkcca" from 2 machines in one night
Reference: v6n96
Re the iechkcca crashes reported by Roy Smith: iechkcca means that the
Intel Ethernet controller chip did not acknowledge a command after a
reasonable amount of time. Sun says that this is normally caused by
either broken hardware or some problem with Ethernet cabling (e.g. use of
a DEC DELNI). There are people in their support group (the guys you get
at 800-USA-4SUN) who are experts in dealing with this problem, and claim
to have had good success at fixing it. You should probably consult with
them first to make sure you don't have a hardware problem. However I
believe that there are situations when this crash will happen even on
systems whose hardware is working and that are correctly installed. When
working with a standalone gateway (i.e. not a Sun), I got far more
experience than I ever wanted with the blasted Intel chip. This chip is
slightly flaky. The condition that causes the ieckhcca crash on a Sun
occured now and then in our gateways even when the hardware is working
correctly. The chip simply hangs now and then, for reasons that we have
been unable to determine. It is possible to recover from these hangs by
completely reinitializing the chip. I tried applying the same approach in
the Sun device driver. In the last version of my Sun code, I got recovery
to work at least half the time. The other half of the time, the Ethernet
ended up being declared down (as opposed to panicing in the standard Sun
code). When recovery didn't work automatically, simply doing "ifconfig
ieX up" appears to bring the Ethernet back. Based on that fact, I have
just put up a new version of the code that does a bit more retrying.
However my code has certainly not received a lot of testing yet. If you
are seeing iechkcca a lot, and have verified that the problem is not
hardware or other things, I'd be willing to give you my changes. You will
however need source to be able to use it. I have tested it only on the
Sun 4 version of 3.2, but it should be applicable to Sun 3's as well.
A reliable source informs me that he thinks this problem has been fixed in
SunOS 4.0. However until I get source for it, I won't be able to be sure.
It is a matter of philopsophy as to whether problems with the Ethernet
chip should be able to cause crashes. In many cases, if recovery is not
possible, it is better to panic than to leave the system in a state where
it can't talk to the network.
------------------------------
Date: Tue, 31 May 88 13:48:08 -0400
From: mesard@bbn.com
Subject: [Unofficial] Patch to Fig 1.4
Phone: 617-873-1878
Enclosed is a patch to Fig 1.4 which gives the program the ability to save
files in SunView raster format (in addition to bitmap format). I find
this functionality invaluable, as I regularly use Fig to create graphs and
illustrations to be included in programs and online presentations.
I told the author about this and he said I was the first one to request
raster file capabilities, so it's unclear whether this enhancement will be
incorporated into future versions. But he did say it was okay to post the
patch to the net.
The modified Fig will have a new menu option called "Save as Raster..."
It will behave exactly like "Save as Bitmap..." except the saved file will
be a one pixel-deep standard format rasterfile.
There are two files enclosed. The first will patch a virgin copy of Fig
1.4. The second is for Fig 1.4 with the first official patch already
applied to it. (The official patch is available via anonymous ftp from
sally.utexas.edu.)
--
unsigned *Wayne_Mesard(); MESARD@BBN.COM BBN Labs, Cambridge, MA
[[ The patch has been placed in the archives under "sun-source" as
"fig.mesard.patch". It is 7989 bytes in length and it can be retrieved
via anonymous FTP from the host "titan.rice.edu" or via the archive server
with the request "send sun-source fig.mesard.patch". For more information
about the archive server, send a mail message containing the word "help"
to the address "archive-server@rice.edu". --wnl ]]
------------------------------
Date: Tue, 31 May 88 10:59:14 PDT
From: jqj@hogg.cc.uoregon.edu
Subject: SUN delivery times
Sun customers in our area received a letter from the district sales
manager last week quoting extended delays on numerous Sun products. For
example, it quotes:
Sun 3/50: 60 to 75 days ARO
Sun 3/60: 120 to 150 days; even longer with more than 4MB
Sun 3/100: 120 to 150 days
Of particular concern is the 3/60 lead time and the additional warning
that lead times on other products are likely to stretch equally
dramatically. The letter goes on to recommend ordering now (with the
intent of cancelling) even if your requirement is soft. Such a
recommendation of course works very much against the educational
marketplace since our grant cycles often make the strategy impossible.
Given these extremely long lead times, are any other longtime Sun
customers besides me seriously looking at alternatives? For example, I am
now for the first time seriously considering DEC VS 3000 and Tektronix
68020 boxes as direct alternatives to desktop Suns.
------------------------------
Date: Tue 31 May 88 15:22:45-PDT
From: HAC63CAD@ecla.usc.edu
Subject: Extra security for logins?
Is there anything available that I can add on top of my SUN workstation to
give extra login security features such as tighter login control more than
the sixty second timeout etc. Thanks ahead for any info.
Peter Ho
[[ This wouldn't be that hard. A sufficiently complex version of the
program "/bin/login" would do it. I don't know if anyone has done
something like this, but it is surely possible, and probably not that hard
for the average Unix wizard. Exactly what features would you like to see?
--wnl ]]
------------------------------
Date: Mon, 30 May 88 14:07:17 +0200
From: jan@eik.ii.uib.no
Subject: AI/LISP development on SUN 4-series?
I would very much like to hear from people who are AI/LISP development on
Sun 4/200's (or 4/110). Experiences, good or bad. Comparisons with
dedicated LISP-machines, TI Explorer & Symbolics? Have anybody used
expert systems (particularly KEE) on Sun-4's? Other Expert systems?
We have at present 1 Xerox and 1 TI Explorer, and we are running Sun
Common Lisp on a Sun 3/60. (The last we had to upgrade to 8MB of RAM. We
tested it with 12MB too, but 8MB seemed to be what was really needed.)
We are thinking about purchasing another 2 TI Explorer's, but they are
goddamn expensive. So if the Sun 4/200 is a good machine for this type of
research, we should be able to buy 1 4/200 and a couple of 3/60's for the
same amount of money as the 2 Explorer's.
Regards,
Jan Berger Henriksen email: jan@eik.ii.uib.no
Institute of Informatics, jan%eik.ii.uib.no@tor.nta.no
University of Bergen
Allegt. 55,
N - 5007 Bergen,
Norway.
------------------------------
Date: Tue, 31 May 88 06:21 EDT
From: DAVIS@blue.sdr.slb.com
Subject: psvf with bounding box?
Does anyone have, or know anything about getting, a raster->PostScript
filter that computes BoundingBox information ? I don't know where out
current psvf comes from (probably part of the Transcript package), but it
doesn't, and I want to include screendumps (or dumpregion selections) into
LaTeX documents using dvi2ps's understanding of \special's. At the moment,
the pssun PostScript prologue puts the raster in completely the wrong
place, *and* I have to guess its size...
Failing that, how about a utility for calculating the BB ? Our Uniras
graphics could then be painlessly sewn into LaTeX...
Paul Davis
Schlumberger Cambridge Research
Cambridge, UK
davis%blue.sdr.slb.com@relay.cs.net
------------------------------
Date: Tue, 31 May 88 08:18:46 PDT
From: texsun!sun!kyeoh@sun.com (kahkee yeoh)
Subject: Searching for past messages?
I am in the Manufacturing and scan the reports for failures that may be
interesting to me.
How do I search for reponses, for example to the question:
"panic: iechkcca" from 2 machines in one night
Thanks
[[ The sun-spots archives includes "master indexes" for volumes 5 and 6.
They are named "v5.index" and "v6.index" and are in the "sun-spots"
directory. You can retrieve the appropriate file from the archive server
by mailing the reqeust "send sun-spots v6.index" to
"archive-server@rice.edu". Hopefully you will get it back within a day or
two. This file is a list of subjects and issue numbers. You can use text
searching programs (such as editors or "grep") on this file to get issue
numbers. I'll save you the bother for this subject: there are two
messages about it, one in v6n96 and one in v6n97. --wnl ]]
------------------------------
End of SUN-Spots Digest
***********************