Sun-Spots-Request@Rice.edu (William LeFebvre) (10/10/88)
SUN-SPOTS DIGEST Friday, 7 October 1988 Volume 6 : Issue 252
Today's Topics:
Re: The Horrible 4/110 killing program! (2)
Re: interactive spelling checker (3)
Re: bad user stack error message
Re: kermit for sun 4
Sun-Spots index
Sunview + Graphics questions
nd question
NFS mount: hard vs. soft?
Nfs mounted /usr/spool/mail?
Has Van Jacobsen TCP/IP been successfully installed in 4.0?
GKS to X11R2 conversion?
sun to Datakit?
PASCAL compiler for 3.5?
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: Tue, 04 Oct 88 10:01:56 +1000
From: Craig Bishop <munnari!lupus.cc.deakin.oz.au!craig@uunet.uu.net>
Subject: Re: The Horrible 4/110 killing program! (1)
> If we run it we get a "watchdog reset" and the system is down.
This is caused by exactly the same hardware problem meantioned in an
article about "spurious watchdog resets". I tried the program and it
causes a segmentation violation on our 4/110.
The fix is to get the modified "trap.o" from Sun which gets around the
hardware fault in software.
Craig Bishop ARPA: craig%lupus.cc.deakin.oz.au@uunet.uu.net
UUCP: ...!uunet!munnari!lupus.cc.deakin.oz!craig
------------------------------
Date: 4 Oct 88 01:56:46 GMT
From: munnari!natmlab.dms.oz.au!neilc@uunet.uu.net (Neil Crellin)
Subject: Re: The Horrible 4/110 killing program! (2)
In <Sun-Spots Digest, v6n245>, darrell@midgard writes:
>We have found that mixing library types is a bad thing to do on a 4/110.
>If we run it we get a "watchdog reset" and the system is down.
We've had essentially the same problem with the following code. Is it the
type mismatching in the executable that's causing the watchdogs? Will the
fixes mentioned by craig@lupus.cc.deakin.oz.au solve both these
problems? The problem here is obviously(?) the incorrect declaration of
stack in thingy.c
File 1: fart.c
===================
double stack[20];
main()
{
stack[0] = 42.0;
thingy();
printf("stack[0] is %g\n");
}
File 2: thingy.c
===================
extern double *stack;
thingy()
{
stack[0] = -stack[0];
}
Repeat-by:
===================
Wheresya-sensayuma% cc -c fart.c
Wheresya-sensayuma% cc -c thingy.c
Wheresya-sensayuma% cc -o fart main.o thingy.o
Wheresya-sensayuma% ./fart
------------------------------
Date: Mon, 3 Oct 88 22:08:03 EDT
From: seth@violin.ctr.columbia.edu
Subject: Re: interactive spelling checker (1)
Reference: v6n247
This sounds like a job for ispell
Ispell is an interactive speller available at:
celray.cs.yale.edu
with anonymous login. It is under the directory /u/public/ispell
you need the files: dict.shar.Z (137,622)
ispell.shar.Z (108,435)
ispell.el ( 22,712)
You can use the ispell.el to hook ispell into emacs so that emacs, when
you check spelling, will give you alternate spellings.
Seth Robertson Systems Manager
seth@ctr.columbia.edu Telecommunications Network Laboratory
uunet!columbia!ctr!seth Center for Telecommunications Research
#include<std/disclaimer.h> Columbia University
[[ Thanks also to the many other people who sent in similar information.
--wnl ]]
------------------------------
Date: Wed, 5 Oct 88 22:11:34 CDT
From: schultz@mmm.3m.com (John C Schultz)
Subject: Re: interactive spelling checker (2)
This may be more than you are looking for but I recommend GNU emacs with
the ispell.el package. GNU emacs runs under Sunwindows (though its X
support is arguably better). I byte-compiled and ran ISPELL with no
difficult and found it immediately useful, particularly the ability to
spell the entire buffer, only regions, or only the current word. It knows
about TeX and LaTeX commands and automatically maintains you own
dictionary of previously unknown words (which you presumably spelled
correctly).
The most serious drawbacks that I have found are that it checks names and
abreviations.
GNU emacs and ispell.el and its dictionaries are available via UUCP or FTP
from many sources. Refer to comp.emacs, gnu.emacs, or watch
comp.sources.d for archive sites which allow UUCP access.
--
john c. schultz schultz@mmm.3m.UUCP (612) 733-4047
3M Center, Bldg 518-1-1, St. Paul, MN 55144-1000
------------------------------
Date: Tue, 04 Oct 88 09:48:25 PDT
From: grand!day@uunet.uu.net
Subject: Re: interactive spelling checker (3)
Reference: v6n247
We sell a screen editor for plain ASCII text that incorporates an
industrial-strength typo and spelling checker and corrector. Words are
checked against the 118,533-word Proximity/Merriam-Webster Linguibase
(TM). Suggested corrections are chosen by a sophisticated phonetic lookup
algorithm. Included are features to add words to a user dictionary,
detect repeated words, understand troff sequences, etc.
The spelling checker is so sophisticated that it can even detect "usenet
hacker fads" such as the non-word "alot". Maybe we should bundle it with
netnews software! }8')
[[ The use of the non-word "alot" is hardly restricted to the Usenet. How
about "foo", "fubar", "IMHO", "FYI"? Besides, if I was using the checker
on a real paper, I would *want* it to flag those words as incorrect. Is
the acceptance of "slang" optional? --wnl ]]
This is all working now and will be in the next release. Send email for
info.
Dave Yost
Grand Software, Inc.
day@grand.COM
uunet!grand!day
------------------------------
Date: Tue, 4 Oct 88 11:38:56 EDT
From: libes@cme-durer.arpa (Don Libes)
Subject: Re: bad user stack error message
This diagnostic appears when the user hands the kernel a bogus stack
pointer. For example, if you casually reassign the third arg of a signal
handler this will happen upon return. The reason it is reported this
strange way is because your process can never regain control from the
kernel since you've munged its stack.
Take a close look at the sig-11 handler in the deceased process.
Don Libes cme-durer.arpa ...!uunet!cme-durer!libes
------------------------------
Date: Mon, 3 Oct 88 19:10:13 MDT
From: Dick Wiley <wiley%unm-la@lanl.gov>
Subject: Re: kermit for sun 4
I recently brought up Kermit 4E (069) on a Sun 4/110 running SunOS4 r4.0.
I use it to talk to a MicroVAX over a direct hookup (we haven't gotten our
DECNET setup, license, etc. yet). It seems to work ok, at least at 1200
baud. I haven't had the time to try it at higher speeds yet, but will
within the next week or so.
Thinking back, about the only thing I had to do this version of C Kermit
was change the declaration for signals (SunOS4 uses the System V
declaration). I did this by adding a new define (SUN4) to ckcdeb.h and an
additional target to the make file. I can provide additional details if
you need them. A more recent version of C Kermit should be available from
Columbia (cu20b.columbia.edu). The one I have is from January 1988.
Dick Wiley
Flow Science, Inc. Internet: <wiley%unm-la@lanl.gov>
Los Alamos, NM Phone: (505) 662-2636
------------------------------
Date: Tue, 4 Oct 88 09:38:08 EDT
From: srs!matt@uhura.cc.rochester.edu
Subject: Sun-Spots index
(possibly for putting into the digest, but more a question that I don't
know who to ask/mail)
I requested the Sun-Spots index awhile back and I have a feeling that the
file is too big to get by many uucp mailers (many have real 64Kb limits,
others have artificial 64Kb limits set in their sendmail.cf files). Would
it be possible to break up the index into smaller chunks?
uucp: {rutgers,ames}!rochester!srs!matt Matt Goheen
internet: matt@srs.uucp OR matt%srs.uucp@harvard.harvard.edu
[[ Sigh. I may have to do that. It is currently over 100K. "Stay tuned
for more details." --wnl ]]
------------------------------
Date: Tue, 04 Oct 88 10:39:52 SST
From: TNG TH <ISSTTH@NUSVM.BITNET>
Subject: Sunview + Graphics questions
I am new to Sun-Spots, but have a few questions which I hope any kind soul
can help.
I am developing a graphical interactive application on a SUN 3/160 and
have chosen Sunview. But I can't seem to get any graphics packages
(circles, ellipse, fill areas etc) such as CGI and CORE to work with
Sunview's windows and buttons. I was forced to write the graphics routines
using pixrects - and it is real slow. Can someone offer me solutions? If
Sunview isn't the correct way, then what are the alternatives? How come
programs like Framemaker works on Sunview with a superfast set of graphics
routines? What did they use? Or do I have to resort to writing my own User
Interface packages on top of CGI?
Requirements
____________
1. User interface tools - buttons, icons, windows, frames, much like Sunview
2. Graphics routines - fills, patterns, shapes, to enable me to write a
Macdraw-like program complete with patterned outlines for shapes etc.
3. Animation package (much like Apollo's GMR-3D)
4. Supports multifont text environment.
Am I asking too much? Or am I ignorant?
Please mail your solutions to me directly at ISSTTH@NUSVM.
[[ If you are using pixrects or pixwins, the first thing you should check
is screen locking. If you don't have a lock on the area of the screen you
are dealing with, pixrect calls will attempt to place a lock on it for
you. This can take a long time. If you are going to do many pixrect
operations (say, in a loop) then explicitly lock and unlock around the
loop. This will speed up screen access immensely. --wnl ]]
------------------------------
Date: Mon, 3 Oct 88 16:39:46 PDT
From: garlick%ucsco.UCSC.EDU@ucscc.ucsc.edu (Tim Garlick -- UCSC Computer Center)
Subject: nd question
We have a Sun 3/280 (OS 3.5) with two Super Eagle disks. Currently, the
machine is stand-alone, but I want to add a 3/50 or two via nd. I'm
hoping to avoid repartitioning the disk, but I'm not sure if that's
possible. I'd like to use the beginning of xy0d (our user file system)
for the nd partitions, and the rest of the disk for local user's disk
space.
So let's say I set up my user file system, xy0d with 4 soft partitions for
the root and swap of two workstations (after dumping, of course!). Then I
run /etc/nd, make the file systems, etc. But how does the system know
that xy0d, mounted on /a really starts several Mbytes into the partition?
And when I do a newfs on xy0d after doing the mkfs for the soft
partitions, how does the system know not to do the newfs right on top of
the soft partitions? Or do I really have to have a dedicated disk
partition on which to put my nd soft partitions?
I'm sure this is obvious some place, but I can't find it in the manuals.
I'm afraid I'll leave out some important step and trash the disk...
Thanks,
Tim Garlick
garlick@ucsco.ucsc.edu
garlick@ucscc.bitnet
...ucbvax!ucscc!garlick
------------------------------
Date: 3 Oct 88 23:46:30 GMT
From: etnibsd!vsh@uunet.uu.net (Steve Harris)
Subject: NFS mount: hard vs. soft?
Can anybody give me a good explanation of the difference between hard and
soft when mounting an NFS file system? Or a pointer to such an
explanation?
The Sun documentation explicitly states that if a NFS file system is
mounted read-write, it should be mounted hard. Why? What will (or won't)
happen if I mount it soft? What does it mean to be mounted hard or soft?
Currently, we mount everything hard. When a NFS server goes down,
processes (e.g., shells) which try to access the NFS partition become
hung. This is especially noticable when the partiton is /usr/spool/mail.
(In this case, I unset my csh mail variable IMMEDIATELY!) But there are
other ways that processes become hung when NFS fails, and, despite
mounting things "intr", I cannot interrupt out of these hung processes.
(I suppose my real question is: what can I do to prevent a crashed NFS
client from hanging my system?)
I recall being told that if a soft mounted NFS server goes down, it can
crash its clients. Except for possible file system corruption, this would
almost be better than being hung with no response until the server comes
back up -- at least you wouldn't have to wonder whether or not to reboot
the hung system.
--
Steve Harris -- Eaton Corp. -- Beverly, MA -- uunet!etnibsd!vsh
------------------------------
Date: 4 Oct 88 13:27:57 GMT
From: david@pyr.gatech.edu (David Brown)
Subject: Nfs mounted /usr/spool/mail?
Hardware: Vax 11/750 & several Sun 3/60's
OS: 4.3BSD+NFS and SunOS 3.5
The Vax is the 'mail server'. It's /usr/spool/mail directory is remotely
mounted on all of the Suns, so that regardless of where a user logs in,
they can read their mail. I have the 'nobody' feature turned on, because
most of the Suns are publicly accessable. The problem is that whenever I
send mail from my workstation, its owned by 'nobody' when it finally gets
to the spool directory (because sendmail runs suid-root, and root is
mapped to nobody on the remote machine).
Is anyone else trying this type of approach? Can sendmail run suid to
someone else? Are you using a different mailer? Should I just tweek the
sendmail.cf on the Suns to route everything through the Vax (Bleah!)?
Thanks in advance.
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: 4 Oct 88 13:27:11 GMT
From: ehrlich@blitz.cs.psu.edu (Dan Ehrlich)
Subject: Has Van Jacobsen TCP/IP been successfully installed in 4.0?
Has anyone tried to install the Van Jacobsen TCP/IP that Berkeley released
a while back into SunOS 4.0? I would appreciate hearing about any
experiences that people have had trying (successfully or not) to do this.
Thanks in advance.
Dan Ehrlich <ehrlich@blitz.cs.psu.edu>
The Pennsylvania State University
Department of Computer Science
University Park, PA 16802
------------------------------
Date: Tue, 4 Oct 88 11:21:08 CDT
From: shell!jody@cs.utexas.edu (Jody Winston)
Subject: GKS to X11R2 conversion?
Is there a public domain or commercially available product which allows a
program to make GKS calls in X-Windows version 11?
Please reply directly to me and I will summarize for posting to the net...
Thanks
Jody Winston jody@shell.uucp
...!{sun,psuvax1,bcm,rice,decwrl,cs.utexas.edu}!shell!jody
Shell Development Company, Bellaire Research Center
P.O. Box 481, Room 2202
Houston, TX 77001
(Voice: 713 663-2050)
------------------------------
Date: 4 Oct 88 16:49:44 GMT
From: pgf2@scobee.att.com (59457[djm]-p.fisher)
Subject: sun to Datakit?
Has anyone connected an outgoing tty port on a sun workstation to Datakit
and gotten uucp to work?
I have a sun 3/260 running SunOS 3.5 and access to a DK port which I can
configure any way that I want to. I can see how to set up to run tip, but
not how to get uucp going since I have neither a true direct connection or
a hayes/ventel modem on the port. Does this require that I write a device
driver that knows how to respond to the DESTINATION: prompt or is there
some way to fool uucp into doing the dialing correctly? Does sun provide
a basic tty driver that can be modified?
Responses via e-mail would be appreciated.
Thanks.
Paul Fisher
LC 2ND06
(201) 580-4157
8-243-4157
[mtunk | att ]!scobee!pgf2
------------------------------
Date: Mon, 3 Oct 88 16:42:16 PDT
From: liou@parns.nsc.com
Subject: PASCAL compiler for 3.5?
Does anyone out there know how to get hold of a copy of an excellent
"PASCAL" compiler for SUN 3.5?
email: liou@nsc.com phone: (408) 721-3683
------------------------------
End of SUN-Spots Digest
***********************