Sun-Spots-Request@Rice.edu (William LeFebvre) (09/13/88)
SUN-SPOTS DIGEST Sunday, 11 September 1988 Volume 6 : Issue 225
Today's Topics:
Re: old csh source (tcsh for SunOS 3.5 csh?)
Re: Hanging serial multiplexer
Re: Answer center
Re: Problems with atrun under 3.5
SunOS 4.0 Sources arrived! (2)
Non-Sun SCSI peripherals & Sun-4/110
PostScript GNUware
File descriptor shortage
problem with sunPS on sun 4/110
Problems with domain name and mail
uucp via ethernet?
lots of users on a Sun?
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: Thu, 08 Sep 88 12:55:50 PDT
From: (P. Graham) <pjg@solstice.unr.edu>
Subject: Re: old csh source (tcsh for SunOS 3.5 csh?)
Things are not as simple as they seem. I don't know about other people
but after I constructed tcsh using the posted diffs on my sun I found that
^C would terminate the shell. So I stopped typing ^C. Then I heard that
there was a new (beta) version so I got:
version tcsh 5.6 beta (Ohio State) 02/29/88 Patch level 0
from osu and tried to apply the diffs. Gack. No luck. Then I remebered
that I was using the 4.3tahoe csh. After rummaging around a bit I
sucessfully applied the diffs to the original 4.3 csh. Using patch of
course. I suppose I should mention the new csh to the good people at OSU
but I keep thinking they must know by now. How does this apply to suns?
The beta version no longer terminates the shell on ^C. Of course it has
lost it's place, but only once.
--
Paul Graham -- University of Nevada-Reno
uucp - uunet!unrvax!pjg
internet: pjg@solstice.unr.edu (MX mailing only)
------------------------------
Date: Thu, 8 Sep 88 15:18:48 BST
From: Chris Brown <mcvax!aivru.sheffield.ac.uk!chris@uunet.uu.net>
Subject: Re: Hanging serial multiplexer
Reference: v6n204
In v6n204 (sorry but I'm getting a bit behind reading these things!)
Anthony Datri asks about ALM2's and a Systech MTI board which hangs.
We have a System MTI (Multibus) board which is very susceptible to power
dips and regularly hangs when (for example) I switch on a nearby overhead
projector. The MTI has a row of lights on the back edge of the board.
Normally they move in a 'chaser' pattern. If the board is hung the lights
flash in a particular pattern and the Systech manual tells you what the
pattern means. Ours is invariably 'unexpected interrupt'.
Chris Brown
chris@aivru.sheffield.ac.uk
------------------------------
Date: 8 Sep 88 17:38:02 GMT
From: step!number1!perl@philabs.philips.com (Robert Perlberg)
Subject: Re: Answer center
The Answer Center is the place you get responses from when you send a
support problem to sun!hotline. In case you're not aware of this service,
you can send mail through the net to sun!hotline with your support
problems. It's the email equivalent of USA-4SUN. In your message, you
must specify the model of your machine, serial number, SunOS version, your
name, address, phone number (voice), and email address. They reply via
email through the net. Their reply always includes a Service Order
number, the name of the support person replying, his/her net address and
voice phone number. I have had lots of success with this service. I
almost never call USA-4SUN any more. You can even order patch tapes and
such via email.
One of the great things about it is that you can include the exact text of
error messages and such. It saves a lot of time, and you avoid having to
play telephone tag, which is as much of an advantage for them as for us,
which also assures that we get better service.
The next time you have to justify the net to your boss, tell him/her about
sun!hotline.
Robert Perlberg
Dean Witter Reynolds Inc., New York
phri!{dasys1 | philabs | manhat}!step!perl
------------------------------
Date: 9 Sep 88 15:49:12 GMT
From: battan@tc.fluke.com (Jim Battan)
Subject: Re: Problems with atrun under 3.5
Reference: v6n220
atrun under 3.5 tries to use an undocumented printf conversion character
(that Sun removed in the 3.5 release). Only atrun needs to be replaced
with the 3.4 version; the 3.5 at still creates correct scripts. This is,
of course, fixed (actually, rewritten) in 4.0.
--
Jim Battan
Voice: +1 206 356 6469
Email: battan@tc.fluke.COM || {uw-beaver,decvax!microsoft,sun}!fluke!battan
------------------------------
Date: Thu, 8 Sep 88 15:32:55 PDT
From: jonab@cam.unisys.com (Jonathan P. Biggar)
Subject: SunOS 4.0 Sources arrived! (1)
We just received our SunOS 4.0 Source tapes on September 8!
Jon Biggar
jonab@cam.unisys.com
[[ Michael Cooper was cc-ed on this note. Here is his response: --wnl ]]
------------------------------
Date: Thu, 8 Sep 88 19:24:01 PDT
From: mcooper@oberon.usc.edu (Michael A. Cooper)
Subject: SunOS 4.0 Sources arrived! (2)
One of the CS research groups received their copy of the 4.0 sources on
Sep 2. We (Computing Services) are still officially waiting for our copy
to show up.
mike
------------------------------
Date: Fri, 9 Sep 88 10:54:14 +0200
From: mcvax!cgch!wwtz@uunet.uu.net (Wolfgang Wetz)
Subject: Non-Sun SCSI peripherals & Sun-4/110
The contribution to Sunspots of Mr. Roberts issued an alert on our side,
because we are running EXABYTE backup units from Delta Microsystems and
we are planning to migrade to the Sun 4s. I have received a letter from
the manufacturer of this unit, which I include below:
__________inluded letter from Delta Microsystems__________
Sun Microsystems is ending a warning regarding third party SCSI peripherals
with each SUN 4/110 Workstation. The warning states that the Sun SCSI host
adapter is nonstandard and has pin 26 tied to ground. The warning further
states that connecting third party peripherals to the Sun 4/110 SCSI bus
may damage the Workstation. The implication is that only Sun peripherals
should be used.
The SCSI bus specification calls for the last peripheral on the bus to be
terminated. Bus termination consists of connecting each data and control
line to +5 VDC trough a 220 ohm resistor and to ground through a 330 ohm
resistor. The SCSI specification calls for bus termination power (the +5
VDC) to be provided on line 26. It is usually the function of the host
adapter (the SCSI port on the workstation) to provide termination power.
Therefore, almost all SCSI peripheral manufacturers provide peripherals
which can be jumpered to use internal power for the terminators.
I have experience with about 30 different SCSI peripherals from about
10 different manufacturers and I have never seen a SCSI peripheral which
puts termination power on the SCSI bus. There are undoubtedly some SCSI
peripheral and/or subsystem manufacturer who do put termination power on
the SCSI bus. If one of these devices were connected to a Sun 4/110 work-
station, damage would probably result. When in doubt, a subsystem or
device should be tested for +5 VDC on pin 26 before it is connected to
a Sun workstation. If a voltage is present on pin 26, then the peripheral
should NOT be used.
Delta Microsystems subsystems neither require power from nor supply
power to line 26. Line 26 is passed trough Delta Microsystems, but it
is completly isolated from the drive, power supply, etc. Delta Micro-
systems uses special circuitry to implement bus termination. Power
for bus termination is drawn directly from an internal power supply.
Since termination is done external to the drive and the drive is
completly isolated from line 26, Delat Microsystems peripherals will
not damage Sun workstations. In fact, Delta Microsystems uses Sun 4/110
workstations for burning in subsystems prior to shipping.
Sincerely,
Donald A. Trimmer, President
__________end of included mail__________
I have two comments to add:
- Sun should have cautioned users about this potential problem rather than
attempt to scare users away from third party peripherals, even on
peripherals which Sun does not produce (yet) as the video-8 backup unit.
- In view of Sun's claims, that they are providing an *open* system, they
provide a surprising number of obstacles to third party hardware vendors
(remember the disk controler story).
Wolfgang Wetz, Systems Administrator, Scientific Computing Centre
c/o CIBA-GEIGY AG, R-1045.330, CH-4002 Basel, Switzerland
Internet: wwtz%cgch.uucp@uunet.uu.net
UUCP: wwtz@cgch.uucp Phone: (+41) 61 697 54 25
BITNET: wwtz%cgch.uucp@cernvax.bitnet Fax: (+41) 61 697 32 88
------------------------------
Date: Thu, 8 Sep 88 14:22:59 PDT
From: parcplace!deutsch@sun.com (Peter Deutsch)
Subject: PostScript GNUware
GhostScript, a GNU project implementation of a language and graphics
library with a remarkable similarity to PostScript(tm), will be released
by the Free Software Foundation as soon as one or two pieces of paper get
signed.
GhostScript consists of a language interpreter for a language that is
essentially identical to PostScript, and a graphics library that provides
C-callable routines for all the graphics capabilities of PostScript, so
you can have the graphics without the language if you prefer. The
interpreter is implemented on top of the graphics library.
GhostScript currently has two device drivers: one for the EGA, which works
quite well, and one for X11, which is seriously brain-damaged because the
machine on which it was written had a badly buggy X server. (I hope
someone will write a good X11 driver.)
It should be trivial to write a PostScript previewer using GhostScript by
changing the top level of the interpreter a little.
GhostScript is covered by the same license as GNU Emacs, gcc, etc.
Roughly speaking, anyone can have it as long as they agree not to charge
(beyond reproduction costs) for redistributing it, and agree to always
redistribute it in source form. Read the GNU Standard License for the
full and accurate story.
L. Peter Deutsch
dba Aladdin Enterprises
P. O. box 60264
Palo Alto, CA 94306
(no e-mail address for GhostScript-related correspondence yet,
please don't use deutsch@parcplace.com -- I'll probably get an
account on portal)
------------------------------
Date: Fri, 9 Sep 88 06:24:48 +0200
From: unido!focus!bigmac!jum@uunet.uu.net (Jens-Uwe Mager)
Subject: File descriptor shortage
While playing with touchup I discovered a Bug in shelltool and cmdtool
(SUNOS 3.5). Both programs leave several file descriptors open for the
child process, causing touchup to fail while displaying the info popup. I
suppose most of you are using the c-shell, where this bug does not show up
because of the c-shell closing everything above file descriptor two. I am
using the korn shell, which does not close any files but sets the close on
exec flag for its local files.
Here the output of an forktest for a cmdtool:
argv="src/forktest"
SIGTSTP: SIG_IGN
SIGTTIN: SIG_IGN
SIGTTOU: SIG_IGN
file 0, flags: O_RDWR
owner -10563
inode 926, links 1, uid 0, gid 0, mode: crw-rw-rw-
rdev major 20, minor 3
file 1, flags: O_RDWR
owner -10563
inode 926, links 1, uid 0, gid 0, mode: crw-rw-rw-
rdev major 20, minor 3
file 2, flags: O_RDWR
owner -10563
inode 926, links 1, uid 0, gid 0, mode: crw-rw-rw-
rdev major 20, minor 3
file 4, flags: O_RDWR
owner -1
inode 1034, links 1, uid 0, gid 0, mode: crw-rw-rw-
rdev major 15, minor 18
file 5, flags: O_RDWR
owner -1
inode 917, links 1, uid 0, gid 0, mode: crw-rw-rw-
rdev major 22, minor 0
file 9, flags: O_RDONLY
owner -1
inode 2719, links 1, uid 100, gid 0, mode: -rw-------
size 0
file 10, flags: O_RDWR
owner -1
inode 2720, links 1, uid 100, gid 0, mode: -rw-------
size 0
file 11, flags: O_RDWR O_NDELAY
owner 0
inode 262838284, links 1, uid 100, gid 10, mode: prw-rw-rw-
file 12, flags: O_RDWR
owner 0
inode 262832140, links 1, uid 100, gid 10, mode: prw-rw-rw-
file 13, flags: O_RDWR
owner 0
inode 262779276, links 1, uid 100, gid 10, mode: prw-rw-rw-
Is there an easy way to fix this inside suntools?
In looking at files 11 through 13 the inode field for the pipes appear
somewhat fancy to me. Does anybody know what is stored there?
Jens-Uwe Mager
jum@focus.UUCP
------------------------------
Date: Thu, 8 Sep 88 16:56:05 CDT
From: deleone@cs.wisc.edu (Renato De Leone)
Subject: problem with sunPS on sun 4/110
I copied and compiled the PS preview and everything was fine. When I try
to run it on some demo file I receive the following error
typecheck in operator scalefont
What am I doing wrong? Any idea or suggestion? Thanks very much
Renato De Leone
arpa : deleone@cs.wisc.edu
UUCP : {allegra,seismo,ihnp4}!uwvax!deleone or deleone@uwvax.UUCP
na : na.deleone@na-net.stanford.edu
Telephone: (608) 263-2677
Telephone: (608) 262-5083
Telephone: (608) 233-4120 (home)
------------------------------
Date: 9 Sep 88 12:51:31 GMT
From: "Electrohome Ltd." <electrohome@watcgl.waterloo.edu>
Subject: Problems with domain name and mail
I am running across problems in getting UUCP and mail working on our home
system (I am posting from an account on our main feed) regarding domain
name.
At Electrohome, we have an Ethernet network of four Sun workstations. Our
domainname is "electro" (which it was back when we were on the net using
an Onyx and System III). Our main machine is called "scrooge" (with the
others being called "huey", "dewey", and "louie".
In reading the Sun documentation (SunOS 3.5, System Administration,
Chapter 4 - Communications), I get what I feel is conflicting
information as it says:
- "If [the name of your uucp connection to the outside
world] is your main machine, then your uucp
hostname should be the same as your domain
name." (p 119, Feb. 17, 1986)
- "In a network of Suns, the name of the main machine
becomes the name of your subdomain ..."
(p 125, Feb. 17, 1986)
Does this mean that we have to change the name of our main
machine (scrooge), our domain name (electro), neither, or both?
How would we go about making sure that all is OK if we have to
change something?
- Carlo Sgro
------------------------------
Date: Fri, 9 Sep 88 16:25:11 -0200
From: has@ztivax.siemens.com (Hans-Albert Schneider)
Subject: uucp via ethernet?
We have here two SUN 3/60 (running SunOS 3.5) and some APOLLO DN4000
(running AEGIS sr9.7 and DOMAIN/IX 9.5 (both BSD4.2 and SystemV)). The
APOLLOs are connected via token ring, and one of them functions as a
gateway to the local ethernet to which the SUNs are connected, too. None
of the SUNs has a printer, but one of the APOLLOs has.
Now we would like to use this printer with the SUNs, too. The simplest
solution would be to configure the default printer of the SUNs to be a
remote printer on the APOLLO. Unfortunately this requires that a user who
wants to print something has an account on the APOLLOs, too, because lpd
authenticates its clients the same way rshd does. Thus this approach is
out of question.
[[ Actually, a user can print something on a remote printer without an
account on the remote machine. That's the way lpd was set up. This
capability can be restricted, however, with the use of the "rs" printcap
capability. Also, in order for host A to queue any jobs for a printer on
host B, A must be in B's /etc/hosts.equiv (or, under 4.3BSD in
/etc/hosts.lpd). --wnl ]]
The next idea is to have the machines run uucp via the ethernet (which
would be System V uucp on the APOLLO), but I do not know how to configure
this.
Can anybody please give me some advice? Are there any other solutions?
Hans-Albert Schneider
------------------------------
Date: 9 Sep 88 13:36:03 EDT (Fri)
From: mark@cblpf.att.com (Mark Horton)
Subject: lots of users on a Sun?
I have an application that needs lots of light users (reading netnews) on
a single time sharing machine. I'm interested in a high end Sun 4, or a
few of them, in this configuration. However, I just heard a rumor that
Suns have problems with lots of users. I also recalled a rumor that
there's a 3 bit register somewhere in the hardware, and that if the load
average goes above 8 the performance suffers. (This was in the Sun 2
days, and I don't know if it applies to the 3 or 4.) Does anyone have
further information? Is anyone running with lots of users on a big Sun
that can give me feedback? What if the users were coming in with NNTP or
NFS or RFS?
Please reply by mail to mark@cblpf.att.com (horton@att.arpa if that fails.)
Mark
------------------------------
End of SUN-Spots Digest
***********************