Sun-Spots-Request@Rice.edu (William LeFebvre) (10/13/88)
SUN-SPOTS DIGEST Wednesday, 12 October 1988 Volume 6 : Issue 257
Today's Topics:
Re: fpa board loops on SIGFPE
Re: Are Fig 1.4.FS and transfig in the archive
Re: pix
ftpd under SunOS 4.0
anonymous ftp and SunOS 4.0 ls
4.0/nfs bug followup
on rlogin, csh interprets all characters as ^D
Problems with MX sendmail for OS 3.5
problem with Sunview consoles under 4.0
Help needed with DATACUBE vision hardware
lpr filter for QMS PS 1500?
How many clients on a 3/60?
Word Processors on Sun/Unix?
Address for The Answer Center?
Some Benchmark Results (LONG)
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: Fri, 7 Oct 88 16:39:41 EDT
From: attcan!utzoo!henry@uunet.uu.net
Subject: Re: fpa board loops on SIGFPE
>Any attempt to use or even IGNORE the floating point exception on a Sun 3
>with the fpa board results in an infinite loop...
What you want to do is dig through the crate of documentation Sun ships
and find a document titled "Floating-Point Programmer's Guide for the Sun
Workstation", and read it. Down in the fine print you will find a
discussion of the use of floating-point signals with the FPA. The FPA
generates SIGFPE as a normal part of execution, and any signal handler for
SIGFPE must be prepared to cooperate with the FPA properly (-ffpa supplies
a default one that does this right).
Actually, the rest of that manual makes interesting reading too...
Definitely required reading if you're trying to do serious floating-point
work on a Sun.
Henry Spencer at U of Toronto Zoology
uunet!attcan!utzoo!henry henry@zoo.toronto.edu
------------------------------
Date: Fri, 07 Oct 88 17:45:46 +0000
From: Gordon Joly Statistics UCL <gordon%stats.ucl.ac.uk@nss.cs.ucl.ac.uk>
Subject: Re: Are Fig 1.4.FS and transfig in the archive
There are now at least three "info-server" machines in the UK and the list
sun-request@uk.ac.cardiff can help with info. The archives are quite
exensive and include X as well as SUN code.
Gordon Joly.
------------------------------
Date: Fri, 07 Oct 88 13:54:13 EDT
From: Daniel R. Ehrlich <ehrlich@blitz.cs.psu.edu>
Subject: Re: pix
(I got this from someone in Sun in response to my posting about the
appearent oversight on Sun's part in omitting pix from the Pascal
distribution. It has been sanitized to eliminate any reference to the
source.)
>>(Sun are you listening? This is a customer speaking.)
>...
>and "yes! I'm listening!" We didn't really "omit" pix; we discontinued
>it as a product. So it's not like we're trying to get extra money--we
>simply don't have such a product anymore; it is totally unavailable under
>4.0.
*** FLAME ON ***
Sun has missed the point. PIX is a useful and used part of the Pascal
environment in BSD UNIX. It appears to me that Sun has made 'yet another
arbitrary decision' to eliminate a 'useless feature' without checking with
any users first. Sun doesn't have a lock on the knowledge about what
tools are useful and which aren't. PLEASE ASK FIRST.
*** FLAME OFF ***
Dan Ehrlich <ehrlich@blitz.cs.psu.edu>
The Pennsylvania State University
Department of Computer Science
University Park, PA 16802
------------------------------
Date: Mon, 10 Oct 88 14:21:40 CDT
From: david@wubios.wustl.edu (David J. Camp)
Subject: ftpd under SunOS 4.0
Never since we have acquired our Sun 3/260 and SunOS 4.0 have we been able
to ftp in from the outside. We can do so with our 4/260 running 3.2. We
have called Sun, and they have been trying to help, but have not
reproduced the problem. The problem manifests itself in several ways.
First, root can login and transfer files with no problem. A normal user
(who can login) gets 'User <name> access denied.' when they try to ftp
login. The anonymous account can login, but gets 'crt0: no /dev/zero'
when I type 'dir'. Note we have no problem running ftp un the Sun3 and
accessing remote hosts. It is only a problem when we try to access the
Sun3 from the outside. Has anyone else had this problem? Can anyone
suggest what we need to do to set up ftp properly?
[[ The problem has to do with the chrooted environment that an anonymous
FTP session is placed in, and the fact that that environment no longer has
any sharable libraries (in fact it usually no longer has a "/usr/lib").
See the next message for more information. --wnl ]]
(314) 362-3635 Mr. David J. Camp
Room 1108D ^ Box 8067, Biostatistics
706 South Euclid < * > Washington University Medical School
v 660 South Euclid
Bitnet: david@wubios.wustl Saint Louis, MO 63110
Internet: david%wubios@wucs1.wustl.edu uucp: uunet!wucs1!wubios!david
------------------------------
Date: Fri, 7 Oct 88 10:56:42 EDT
From: Peter Marshall <peter%hadrian.uwo.ca@cornellc.ccs.cornell.edu>
Subject: anonymous ftp and SunOS 4.0 ls
I was following William LeFebvre's note from earlier in the summer about
how to set up an anonymous ftp area and ran into a problem with v4.0 of
SunOS. A colleague later showed me that LeFebvre's notes correspond
almost exactly to the notes in the ftpd(8c) manual entry from SUN, so
perhaps they haven't upgraded their description for v4.
Basically the idea is to define a private root for anonymous ftp access
and then put lots of restrictions on what the user can get at. (ls and
true are the only things in ftp's /bin for example.)
The problem shows itself when you make a request through ftp for /bin/ls
--a dir or ls command to the client. You get back the error message
crt0: no /usr/lib/ld.so
and if you make a private version of that file, you get the message
crt0: no /dev/zero
At that point it looked like I was going to have to define too many
private directories, so I quit.
ls seems to dynamically load various parts of itself. I am using a v3.5
compiled copy of "lc" named ls to get some functionality out of the
anonymous ftp, but it doesn't understand all the switches being sent to
it.
Without sources or object files, is there a way to make a dynamic
program static?
Peter Marshall, Data Com. Manager, NSC 220, CCS, Univ. of Western Ontario
519-661-2151 peter@hadrian.uwo.ca peter@julian.uucp pm@uwovax (BITNET)
[[ We looked into this problem as well when we set up our first (and still
only) 4.0 system. It turns out that you only need three extra files:
/usr/lib/ld.so, /usr/lib/libc.so.0.10, and /dev/zero. But there are two
problems. First, since it is very likely that FTP's home directory will
not be on the same partition as /usr/lib, they must be *copies*. This
wastes disk space (and before you say "use symbolic links" you should read
about "chroot"). Second, FTP's usr/lib directory *must* be readable in
order to do the dynamic linking (I also seem to recall that libc.so.0.10
must also be readable). This is not restrictive enough for my tastes.
The only solution that we have is to copy "ls" from a version 3 machine
and place it in FTP's bin directory. Version 3 executables still run
under version 4, and they are (of course) statically linked. If there is
a public domain version of "ls", one could use that. HEY SUN! We really
*need* a way to make a dynamically linked executable static. Okay? --wnl ]]
------------------------------
Date: Fri, 7 Oct 88 10:47:17 EDT
From: trudel@caip.rutgers.edu (Jonathan D. Trudel)
Subject: 4.0/nfs bug followup
A while back, I mentioned to sun-spots that I was having problems with
setting up a standalone sun 4/110 as a yp client of a sun 3/160 running
3.2. The problem I was having was that I could never mount NFS
filesystems when in this configuration. Every other yp function worked
except for this. I received several suggestions on how to fix this, but I
was told to run the machine completely independent of the yp master. SO,
I never followed up with testing all of the suggestions. The suggestions
included:
1) turn off the yp domain briefly just before and just after mounting, ie
domainname ""; mount -vat nfs; domainname originaldomain - This is what I
used as a bypass.
2) upgrade the yp server to software from OS 3.5 or higher (untested).
Some of the yp function calls changed somewhere around that version
3) set up the sun 4 as a yp slave server. I tried doing this, but the
same thing happened. Also suggested with this was to run ypset to another
server (in this case, the sun 4), and I STILL had the problem.
I suspect that I would have had to do both #2 and #3 to get it to work,
but I don't have to since the machine's gone completely standalone :-)
Jon
------------------------------
Date: Fri, 7 Oct 88 10:22:39 EDT
From: dickson@shire.cs.psu.edu (Scott Dickson)
Subject: on rlogin, csh interprets all characters as ^D
Machine Type: Sun 3/160C
O/S Version: SunOS 4.0
Organization: Center for Academic Computing, The Pennsylvania State Univ.
214D Computer Building, University Park, PA 16802
Phone Number: +1 814 865 0829
Description:
Intermittently, when using rlogin to connect from one Sun to
another, the shell (/bin/csh) that is opened on the destination
machine will interepret any character that is typed as a
control-D. This leads to the problem of not being able to do
anything in that shell. This occurs when rlogin-ing from a
machine running SunOS 4.0 to another machine running 4.0. It
occurs only on our Sun 3/160's, but not on 3/50's, 3/60's, all of
which are clients of the 3/160. It also does not happen on a Sun
4/260. This problem is intermittent at best, but when it occurs,
the only solution found so far is to reboot.
------------------------------
Date: Fri, 7 Oct 88 14:34:46 EDT
From: Lee A. Butler <butler@stsci.edu>
Subject: Problems with MX sendmail for OS 3.5
I'm having trouble getting MX sendmail running under OS 3.5. I've gotten
Bind 4.8 from Berkeley running, but the sendmail that came in the Sun
nameserver kit for OS 3.5 doesn't seem to query the nameserver EVER! As
long as the destination hostname is in the YP database, it's fine (it just
delivers to that host, regardless of MX records for the host). Mailing to
"hosts" which only exist as MX records just doesn't work (host unknown).
I am NOT running with YP/named (that was a whole different horror story).
Should I just chuck the nameserver kit and get sendmail 5.59 from
somewhere? Where can I get it (anon. ftp)?
Lee A. Butler Internet: butler@stsci.edu
Space Telescope Science Institute BITNET: BUTLER@STSCI
3700 San Martin Drive SPAN: SCIVAX::BUTLER (6549::BUTLER)
Baltimore, MD 21218 Phone: (301) 338-4531
------------------------------
Date: Fri, 07 Oct 88 08:44:49 -0700
From: Support Group (agent: Gary Scott Erickson) <support%humboldt.ics.uci.edu@paris.ics.uci.edu>
Subject: problem with Sunview consoles under 4.0
Problem:
all input on pty0 is converted to ^D's after a suntools has exited.
This happens on our 3/50's and 3/60's. I don't really want to try it
on one of our larger file servers.
Repeat by:
I can't get this to repeat all of the time.
Boot 3/50 or 3/60
enter suntools so that a console window appears. The
/usr/lib/.suntools file is good enough. No customization
is needed.
exit suntools
From another machine, or even the same machine, start up a
remote shell or an X window on pty0.
I thought that I had narrowed it down to cmdtool, but the same thing seems
to happen with shelltool. The only common factor that I can see is that
pty0 is used as a console (with the -C switch).
The only fix I've been able to use reliably is to reboot the machine.
Has anyone else seen this, and if so, is there a fix or an official Sun
bug number?
University of California Irvine | (714) 856-4222
Information & Computer Science Dept. | erickson@ics.uci.edu (Internet)
Support Group | ucbvax!ucivax!erickson (UUCP)
Agent: Gary Erickson | saericks@uci (Bitnet)
------------------------------
Date: 4 Oct 88 18:26:14 GMT
From: roston@robotics.jpl.nasa.gov (Gerry Roston)
Subject: Help needed with DATACUBE vision hardware
Apologies upfront since this may not be applicable to all of the
newsgroups, but...
I have a bunch of Datacube vision hardware. Although the hardware is
quite good, the software and documentation are terrible. (I am not
flaming them, I am simply stating a fact which the folks at Datacube are
aware of.) What I am looking for are folks who have been programming the
ROISTORE card, and are willing to send me sample programs, or speak with
me on the phone.
If you, the reader, have a ROISTORE or you know someone who does, please
get in touch with me via email. Thanks in advance for your help.
gerry roston, robotic systems research group
jet propulsion laboratory, 4800 oak grove drive, m/s 23
pasadena, california, 91109, (818) 354-9124 (818) 354-6508
roston@robotics.jpl.nasa.gov
------------------------------
Date: Fri, 7 Oct 88 09:31:05 EDT
From: spock%bnrmtl.UUCP@larry.mcrcim.mcgill.edu (Richard Young)
Subject: lpr filter for QMS PS 1500?
We have a QMS PS 1500 printer supporting a network of SUN 3/60
workstations. Has anyone implemented (any of) the filters used by lpr for
this printer? We are using version 4.0 of SunOS.
------------------------------
Date: Fri, 7 Oct 88 10:26:52 EDT
From: david@pyr.gatech.edu (David Brown)
Subject: How many clients on a 3/60?
How many diskless 3/60's can be effectively served by another 3/60? What
kind of server would you recommend for a lab of 6 to 8 3/60's (most or all
of which are diskless)?
Thanks,
David Brown
David Brown
Armstrong State College, Savannah, Georgia
uucp: ...!{akgua,allegra,amd,hplabs,ihnp4,seismo,ut-ngp}!gatech!gitpyr!david
ARPA: david@pyr.gatech.edu
------------------------------
Date: 7 Oct 88 10:00:48 PDT (Friday)
From: "Steven_S._Kang.ESAE"@xerox.com
Subject: Word Processors on Sun/Unix?
Where can I get info on popular word processing packages on the Sun? I am
specifically interested in the printers they support and how they interact
with the line printing system to print.
Thank You.
Steve at Xerox/El Segundo
------------------------------
Date: Fri, 7 Oct 88 11:53:20 EDT
From: albert@mssun7.msi.cornell.edu (Jay Albert)
Subject: Address for The Answer Center?
In Sun-Spots v6n225, there is an item dated 8 Sep 88 from
step!number1!perl@philabs.philips.com (robert Perlberg) about "The Answer
Center" run by Sun as a substitute or supplement for USA-4SUN. He gives
the e-mail address as only sun!hotline. I don't recognize this form of
address, or his either. The only addresses I know about are of the form
albert@mssun7.msi.cornell.edu, which is mine; no ! marks.
Would you please explain to me how I can get to these guys via e-mail?
Thanks,
Jay Albert
(novice) Sys.Adm., MSI, Cornell Univ.
[[ "hotline@sun.com". The machine named "sun" has both a uucp connection
and an Internet connection. Therefore: "sun!X" <=> "X@sun.com". --wnl ]]
------------------------------
Date: 7 Oct 88 15:28:40 GMT
From: ehrlich@blitz.cs.psu.edu (Dan Ehrlich)
Subject: Some Benchmark Results (LONG)
Here are the results of running 10000 passes of the Dhrystone V2 benchmark
on a number of machines in our lab. All were compiled with the vendor
supplied version of cc. There were two versions of the benchmark, dry2 and
dry2reg, with and without explicit register declarations. The level of
optimization is indicated in parentheses after the machine name. While
these benchmarks were not run standalone, all were run niced -20 which is
almost the same.
Machine Hwd OS Reg ms/dhry dhry/sec
Blitz(-04) 4/260 SunOS3.2 N 55.0 18181.8
Y 56.7 17647.1
Blitz(-O) 4/260 SunOS3.2 N 58.3 17142.9
Y 56.7 17647.1
Shire(-O4) 4/260 SunOS4.0 N 60.0 16666.7
Y 61.7 16216.2
Shire(-O) 4/260 SunOS4.0 N 61.7 16216.2
Y 61.7 16216.2
Curt(-O) RT/6150 AOS 4.3 N 213.3 4687.5
Y 206.7 4838.7
Sunws2(-O) 3/160 SunOS3.5 N 305.0 3278.7
Y 273.5 3658.5
Sunws2(-O4) 3/160 SunOS3.5 N 306.7 3260.9
Y 286.7 3488.4
Pele(-O4) 3/50 SunOS4.0 N 325.0 3076.9
Y 326.7 3061.2
Pele(-O) 3/50 SunOS4.0 N 396.7 2521.0
Y 350.0 2857.1
Gondor(-O) 11/780 4.3 BSD N 631.7 1583.1
Y 626.7 1595.7
Using only the results obtained with default optimization (-O) and no
explicit register declarations we find that a:
Sun 3/50 is 159% of a VAX 11/780
Sun 3/160 is 207% of a VAX 11/780
IBM RT 6150 is 296% of a VAX 11/780
Sun 4/260+ is 1024% of a VAX 11/780
Sun 4/260* is 1083% of a VAX 11/780
* Running Sun OS 3.2L-Gamma
+ Running Sun OS 4.0
What is interesting to note is that a Sun 4/260 running 3.2 is
significantly faster that the same hardware running 4.0. Anyone have any
thoughts as to why this should be so?
Dan Ehrlich <ehrlich@blitz.cs.psu.edu>
The Pennsylvania State University
Department of Computer Science
University Park, PA 16802
------------------------------
End of SUN-Spots Digest
***********************