Sun-Spots Digest, v6n78
William LeFebvre
Sun-Spots-Request at RICE.EDU
Mon May 9 10:55:41 AEST 1988
SUN-SPOTS DIGEST Sunday, 8 May 1988 Volume 6 : Issue 78
Today's Topics:
Re: csh file completion bug
Re: Looking for ypchfn
Re: A Question about Shared memory
Connecting SCSI disks to a 3/50?
Big partitions don't work in 3.4 !!!!
More fun for color fans, yet another colormap editor
More about 8mm videotape for backups
problem with VMS Fortran Resolved
Problem with making selections of emphasized text in tty windows
Looking for APL for Sun
SIMULA on SUN 4?
Appletalk on Sun MCP board ?
Using icons?
Send contributions to: sun-spots at rice.edu
Send subscription add/delete requests to: sun-spots-request at 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 at rice.edu" or mail the word "help" to the same address
for more information.
----------------------------------------------------------------------
Date: Mon, 2 May 88 18:08:34 edt
From: umix!oxtrap!rich at rutgers.edu (K. Richard Magill)
To: mrd at sun.mcs.clarkson.edu
Subject: Re: csh file completion bug
If you get escape completion in your csh you aren't running csh at all.
You are really running tcsh which is a set of modifications to the csh
source.
csh, as distributed does not do escape completion.
Not having sources, MY csh does not expand. (thus we use ksh, but that's
another story :-).
[[ You are either running a old release of SunOS or you don't know how to
enable the feature. Even 3.2 is *distributed* with a version of the csh
that does file name completion when the csh variable "filec" is set. Try
"set filec" and then "ls ^D" or "ls a-ESC". Or just look carefully at the
manual page for csh(1) (search for "filec"). --wnl ]]
------------------------------
Date: Tue, 03 May 88 11:41:34 PDT
From: Alan Stebbens <aks%mondas at hub.ucsb.edu>
Subject: Re: Looking for ypchfn
We have modified the "yppasswdd" server to add two additional "proc"
entries: one for updating the GECOS field, and one to update the "shell"
field. We have also modified the user programs "chfn" and "chsh" to use
RPC to invoke the new "procs" in the server. Our modified "chfn" also
uses an external description of the GECOS fields, allowing the format and
number of subfields to change without requiring recompilation of the
"chfn" program. Our "finger" uses this GECOS description file also.
We made this change realizing (and hoping) that Sun OS 4.0 might do it
differently (and correctly), using a better authentication mechanism (and
perhaps, a more general dbm-map updating tool). It is an interim
solution, as Sun had inadvertantly (at best) forgot to provide it.
Since OS 4.0 is out (any day now), we are waiting to see what's been done
in this area. If you would like more details on what we've done, please
send mail.
A related item is how Sun OS 4.0 handles "rwho" and "ruptime": the fact
that the report packets are UDP-broadcast limits its utility to a cluster
of machines between gateways. Usually, though, it is desired that "rwho"
and "ruptime" report on larger domains of machines, spanning several
gateways; the problem here is determining the boundary of interest. Using
YP domains is not satisfactory, since it cannot be guaranteed (nor is it
desireable) that all machines in a given administrative domain be running
YP. We are currently investigating the feasability of using the Domain
Name Server with "rwho" and "ruptime", with subdomains defining the
boundary of interest. It would also be interesting to see if it could be
made a general feature of all "rwho"'s, to respond to domain-based "rwho"
queries. That is: to generate a query for users logged on, and machines'
uptimes, within a given subdomain. If any others have done anything in
this area, or are interested, let's compare notes.
Alan Stebbens (aks at hub.ucsb.edu)
------------------------------
Date: Tue, 3 May 88 08:42:36 CDT
From: vuse!ip1!hsc at uunet.uu.net (Hsuan Chang)
Subject: Re: A Question about Shared memory
We've augmented our shared memory to 4 MB for some purpose and the system
works fine. I guess you know where to go in to make the change and how to
build the kernel. Good luck.
hsc%vuse.vanderbilt.edu
..!uunet!vuse!hsc
------------------------------
Date: Mon, 2 May 88 15:40:02 EDT
From: peter jensen <jarsun1!jensen at relay.cs.net>
Subject: Connecting SCSI disks to a 3/50?
We are looking at connecting inexpensive SCSI drives such as the MAXTOR
4380S or CDC Wren IV to a SUN 3/50 Workstation. Most people I've talked
to are selling an Emulex MD21 (SCSI to ESDI) controller with a comparable
ESDI drive. However, I know of several people who have hooked up one of
these SCSI drives directly to the SCSI port on the 3/50. I'm looking for
more information on this:
1) General drive information and SCSI vs. ESDI:
Who's using these drives? How do they perform? Has anyone had
any problems with them? What are the issues between ESDI and SCSI?
Any other drives anyone would recommend? Any to avoid? Why?
2) Hooking it up to the SUN 3/50:
I'm really interested in hearing from people who have done this.
Have you run into any compatibility problems? What's your configuration
like? Has anyone daisy-chained two or more drives? Is there anything to
look out for? Anyone not recommend this? We are currently running 3.5.
Will 4.0 change anything?
Thanks in advance! I will summarize any replies.
Peter C. Jensen ...!brunix!jarsun1!jensen pcj at cs.brown.edu
Jordan Apostal Ritter Associates ANSYS, ARIES, and Engineering Analysis
Admin Bldg 7, Davisville RI 02854 (401) 884-3014 or (401) 294-4589
------------------------------
Date: Tue, 3 May 88 14:56:02 EDT
From: ajb%cornea.mitre.org at gateway.mitre.org (Alan J. Broder)
Subject: Big partitions don't work in 3.4 !!!!
We have solved the problem since posting the message excerpted below. The
solution (and problem) turn out to be quite interesting and probably
relevant to many other Sun sites.
> We've just installed 4 Fuji-2372 disks on a Xylogics 451 controller
> with an AVIV VME-bus adapter on our Sun-3/280. After formatting up the
> disks, installing UNIX and reloading our dump tapes we discovered a
> very nasty problem. The problem is that some programs will not run
> anymore. Interestingly, if we make multiple copies of the same
> executable file, some of them will run just fine. Others however
> never even make it to main(), instead they report an EMT trap and dump
> core. Yet cmp reports that there are no differences between the
> executables that work and the executables that dump core. So,
> apparently, the executable's position on the disk affects whether it
> will run or not ! This problem can be reproduced consistently on each
> of the 4 2372's. By the way, no disk errors are being reported.
Well, we finally decided to go back to ground zero; we removed all the new
hardware and reinstalled the original Super-Eagles (2361's) and the
Xylogics 451, all of which we had obtained from Sun. When we made multiple
copies of the same executable and ran them from the 2361's, we discovered
the same problem !!
A quick call to USA-4SUN, revealed that this is a known bug in 3.4 .
Apparently we had partitioned our second 2361 drive (and our second,
third, and fourth 2372's) with a single large partition to cover the whole
disk. According to Sun, in 3.4 and before, any partition larger than
about 512 Mbytes, may result in unpredictable behaviour when exec'ing
programs from it. Apparently when a large program executes from disk,
Unix sets up demand paging to the disk where the program resides. And in
3.4, the field which defines the offset into the disk partition for demand
paging will only accomodate offsets up to 512 meg. This problem has been
fixed (according to Sun) in 3.5 .
So, in summary, if you must run large programs which are stored on
partitions which are larger than 512 Meg, you MUST run 3.5 or suffer the
consequences.
Alan Broder
ajb at mitre.arpa
The MITRE Corporation
(703)883-5614
------------------------------
Date: Tue, 3 May 88 12:36:00 CDT
From: vuse!ip1!hsc at uunet.uu.net (Hsuan Chang)
Subject: More fun for color fans, yet another colormap editor
As a color fan myself, I've built a window-mouse based color editor which
was based on a half cooked version called "palettetool" provided to me
from Sun (I still call it the same name). My version simulates an
artist's palette where the mouse serves as a brush pan. (...a word from a
fine arts drop-out: I know how a palette functions...) A few sliders are
used to adjust the saturations for rgb's. The "artist" may assign a
desired color to a region of entries of the colormap by brushing over
those entries using the mouse, or s/he may dip the color and put it to a
few selected colormap entries...
Well, so much for the advertising. I'd be glad to share this product with
everybody (with an on-line manual) if somebody would tell me how to make
this contribution. I don't mean to start a competition with other guys.
It is just that I am excited to know such a little thing also counts.
H. Chang
Image Processing Lab
Computer Science Department, Vanderbilt University
Box 1679, Station B, Nashville, Tn 37235
hsc%vuse.vanderbilt.edu
..!uunet!vuse!hsc
[[ To submit source to the sun-spots archives, merely send it to the
submission address, "sun-spots at rice.edu". If it is too large to include
directly in a digest, I place it in the archives. Please include a brief
description of what the software does, and package it in a shar file (or a
series of shar files). --wnl ]]
------------------------------
Date: 3 May 88 23:54:42 GMT
From: milano!sandel at tuvalu.sw.mcc.com (Charles Sandel)
Subject: More about 8mm videotape for backups
As a followup to my recent posting about using 8mm video-tape drives for
backing up Suns, I have more information. In my original posting, I said
that I liked the drive and it worked very well, but I was concerned about
the tape medium itself and how well it would stand the test of time.
Since then, I have talked with a representative from Exabyte about these
concerns. The information he gave me was very encouraging. The
advertised shelf life of 8mm video tape is 4 years, because that is how
long that particular medium has existed. There is no way to perform
accelerated tests on magnetic tape of any kind because you would have to
vary temperature and humidity and when you do that, the backing tends to
fall apart. The National Bureau of Standards recommends that accelerated
tests on mag tape are useless. So, all that any manufacturer is willing
to commit to is what their experience is, which in this case is 4 years
maximum.
The characteristics of mag tape that are measurable without accelerated
tests indicate that it should be at least as good as regular 9-track mag
tape, and probably much better. The coercivity ratings of metal particle
tape are about 5x that of 9-track tape, which means that it is harder to
erase, hence should also have longer shelf life.
But what it boils down to is: no one knows for sure. The indications are
that metal particle 8mm tape should be at least as reliable as ordinary
9-track magtapes, but the only real test is to put something on a shelf
and wait.
But I am encouraged. The technology looks promising and a 2 Gb tape drive
can solve a lot of problems. The future also holds some interesting
developments.
On another front, R-DAT is an up-and-coming format. It looks like an
R-DAT cassette will hold about 1 Gb of data. I can't see why anyone would
bother with a 1 Gb medium when a 2 Gb (and maybe more?) is available and
working.
Sun is definitely taking a good look at 8mm videotape systems, as are
other manufacturers. If you are interested in having 8mm videotape be a
standard for tape backup, I suggest you let Sun know of your desires.
Charles
Charles Sandel arpa: sandel at mcc.com
uucp: *!ut-sally!im4u!milano!sandel
------------------------------
Date: Tue, 3 May 88 08:58:29 PDT
From: danq%sag4.ssl.Berkeley.EDU at jade.berkeley.edu (Daniel Quinlan)
Subject: problem with VMS Fortran Resolved
I posted a complaint about obtaining VMS Fortran for Sun 2 recently.
Sun-Spots was very effective in clearing up the problem. I got mail from
Dave Hough at Sun, who personally hand delivered a copy of the release.
I'm happy with the resolution of the problem, and I appreciate the
personal attention from people at Sun.
Daniel Quinlan
Space Sciences Laboratory
UC Berkeley
[[ The power of the network. Amazing, isn't it? --wnl ]]
------------------------------
Date: 3 May 88 15:19:13 GMT
From: roy%phri at uunet.uu.net (Roy Smith)
Subject: Problem with making selections of emphasized text in tty windows
I'm running suntools on a 3/50 under SunOS-3.2. If I double-click on some
emphasized text in a tty window, I can't select a word. To be more
specific, I was just running emacs in a shelltool (i.e. ttysw) window.
After quitting emacs, I was left with the emphasized mode line near the
bottom of the window:
EMACS (Fundamental Fill) HirshonFsmPreprint: HirshonFsmPreprint --100%-- *
The next thing I wanted to do was "ptroff -me HirshonFsmPreprint" and to
get the filename, I wanted to just double-left-click to select
HirshonFsmPreprint from the left-over mode line then do "Put, then Get".
When I do the double-click on, for example, the "H" in the first
"HirshonFsmPreprint", I select "ill) Hi". Double-click on any of those
characters gets me the same selection. Double-clicking on either the "P"
or the "r" after it gets me just that one letter, either the "e" or "p"
following that gets "ep", the next "r" get just that one letter, and
anywhere in the following "int: Hi" gets all that.
It would appear that there are some magic cookies burried in the text
which screws up the parsing of the line into words. Tripple-clicking
anywhere in the line does get me the whole line and left-click on "H"
followed by middle-click on the "t" before the ":" does get me
"HirshonFsmPreprint" as it should. For what it's worth, I use:
/Tty/Bold_style "Offset_X_and_Y_and_XY"
in my .defaults file.
Presumably this is a bug in the suntools library?
Roy Smith, System Administrator
Public Health Research Institute
455 First Avenue, New York, NY 10016
{allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy at uunet.uu.net
------------------------------
Date: 3 May 88 19:31:41 GMT
From: quirk at hubcap.clemson.edu (Christine Reynolds,2639,G09J)
Subject: Looking for APL for Sun
I'm looking for an APL (preferably public domain) that will run on a Sun
3. Any help would be appreciated.
Chris Reynolds uucp: ... !gatech!hubcap!quirk
CS Dept, Clemson University inet: quirk at hubcap.clemson.edu
phone: (803)656-{2639,3444} ARPA: QUIRK at TECNET-CLEMSON.ARPA
------------------------------
Date: Tue, 3 May 88 09:18 N
From: <PETIT at HGRRUG5.BITNET>
Subject: SIMULA on SUN 4?
Does anyone kwon if there is a SIMULA compiler available for the SUN 4 ?
We have plan to move from VAX to SUN for our students. Thanks for any
informations,
Marc (EARN/BITNET address : PETIT at HGRRUG5)
------------------------------
Date: Mon, 2 May 88 11:27:30 MET DST
From: Toerless Eckert <faui10!eckert at sun.com>
Subject: Appletalk on Sun MCP board ?
Is it possible to run Appletalk directly on a Sun 3/160 with Sun's
Multiprotocol Communication Processor?
>From the Specification of the board it seems possible. The board supports
the speed ( 230.4kbps) and the link layer protocoll ( SLDC).
Are there any Products available, which implement Appleshare or TOPS on
the Sun, using this hardware configuration ?
Toerless Eckert
------------------------------
Date: 3 May 88 09:27:47 EDT (Tue)
From: dck at mtunn.att.com (David Kwong)
Subject: Using icons?
Pardon my elementary question ... How do I use the icons in that icon
issue? I am running SunOS 3.2 on SUN-4's. Thanks.
D. C. Kwong
AT&T
mtune!mtunn!dck
[[ Since there might be others out there with the same question, I will
answer it here. Actually, reading the "suntools" manual page will reveal
the answer, but you must be patient and go through most of the entry
before finding the section entitled "Generic Tool Arguments". There we
find out that most tools should recognize the argument "-WI icon-image".
A text file, such as those distributed through sun-spots, is accceptable
files for the "-WI" argument. --wnl ]]
------------------------------
End of SUN-Spots Digest
***********************
More information about the Comp.sys.sun
mailing list