Sun-Spots Digest, v6n275

William LeFebvre Sun-Spots-Request at Rice.edu
Thu Oct 27 02:01:43 AEST 1988


SUN-SPOTS DIGEST         Tuesday, 25 October 1988     Volume 6 : Issue 275

Today's Topics:
                   Re: ALM-2 and hardware flow control
                       Re: Window Lock Broken msgs 
                       Re: Is NFS mounted mail safe
                        Re: communication with IBM
                       Re: calentool - how to beep
                  Thanks for info about adding SMD disks
                         SCSI vs. SMD performance
                      Hardware Flow Control with MTI
                         Responses to tape survey
                       3/50 memory upgrade options
                         icon->cursor conversion
           Problems with oob data on sockets using the notifier
           How to dynamically change the font in a tty window?

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:    Fri, 21 Oct 88 14:48:48 PDT
From:    auspex!guy at uunet.uu.net (Guy Harris)
Subject: Re: ALM-2 and hardware flow control

>What you can do depends on what version of SunOs you are running:
>
>SunOs 4.0:
>	You can set flag CRTSCTS flag on some termio struct

"termios", not "termio".

>	to get hardware flow control on *outgoing* stream.

(The hardware doesn't support flow control on incoming data....)

>	If you want to use that on printer or call-out-modem lines, you
>	have to patch source of lpd, uucico, tip etc...  to set it with
>	ioctl. (What, no sources. Too bad).

Not true for "lpd".  Check out PRINTCAP(5), paying special attention to
the "ms" capability....

------------------------------

Date:    Fri, 21 Oct 88 11:43:02 MDT
From:    erik at cadnetix.com
Subject: Re: Window Lock Broken msgs 

Chuck Musicano wrote that you can basically ignore these "Window display
lock broken" which appear in the console window. This is true with the
important understanding that it will greatly slow down window operations.

1. You acquire a lock, do a bunch of related window operations, then 
   release the lock. This speeds up window access since  you acquire
   the lock once for many operations.

2. If the lock is broken (by a timeout of the lock mechanism), *each 
   and every* subsequent window operation acquires the lock, does 
   its thing and releases the lock.  Since acquiring a lock is time
   consuming, this is *really* expensive.

3. If however, you acquire the lock again before it has been released
   or broken, you mearly increment the lock count (cheap).  This means
   that you can avoid breaking the lock and it's disasterous side effects
   by incrementing the lock count every "n" of your window operations in
   your own code (i.e., reset the timer to prevents a timeout).

We have tried hosing with the actual timeout value using one of sun's
provided routines, but the value is, as Chuck writes, tied up with other
functions in interesting ways, so we were not successful with this method.

erik at cadnetix.COM  (Erik Hyypia)

------------------------------

Date:    Sat, 22 Oct 88 12:51:41 -0400
From:    Henry B.J. Krempel <krempel at pacrat.npac.syr.edu>
Subject: Re: Is NFS mounted mail safe

I don't have an authoritative answer to this, but we have been doing this
for some time without any problems like the concern mentioned (file
corruption if mail is delivered when it is being read).

One thing that might be helping this in our environment is that all the
users use MH to read their mail.  Since MH "inc's" the mail into the users
home directory, the amount of time changing the remote mail file during a
read is reduced.

Perhaps a safer method that doesn't involve NFS at all is MH "POP"
service,  where the remote system fetches the mail from the mail server.
We haven't had the time to set this up,  but it has the advantage of
servicing that vanishing breed of Unix machines that don't have NFS.

Most of our staff is quite happy with mail reading using MH in GNU-emacs,
and with emacs under NeWS or emacstool, it's even somewhat workstation-y.

Henry B. J. Krempel	<krempel at pacrat.npac.syr.edu>
Northeast Parallel Architectures Center (NPAC)
Syracuse University
250 Machinery Hall
Syracuse,  N.Y. 13244

------------------------------

Date:    Sat, 22 Oct 88 19:29:00 CDT
From:    David Lippke <LIPPKE at UTDALVM1.BITNET>
Subject: Re: communication with IBM

Even though I understand that it's possible to connect the Sun directly to
an IBM channel, I'd vote for changing the angle of attack.  I would
connect the IBM to the Sun via ethernet.  IBM's own TCP/IP product,
5798-FAL, has been getting rave reviews ever since it was released a year
ago and would be a good option to take (especially with one of the
inexpensive and fast third party ethernet controllers).

We ran some benchmarks a few months ago (and will likely run more this
December) to compare some of the software and hardware that's available.
I'd be glad to forward the results or provide more details on this whole
issue to anyone with an interest.

In any case, there's more than one software/hardware combination that
would give you a solid program interface between a Sun and an IBM.  I
suspect that you'd be more pleased with one of them than a direct channel
interface (which is not as fast or clean as it may seem at first glance).

Kind Regards,
	David Lippke

p.s. There's a LISTSERV discussion list called IBMTCP-L at CUNYVM (.CUNY.EDU)
     devoted to the subject of IBM mainframes and TCP/IP.

------------------------------

Date:    Fri, 21 Oct 88 17:34:29 EDT
From:    "Bruce L. Rosen" <brosen at spcb.bbn.com>
Subject: Re: calentool - how to beep

There was a recent query about how to get calentool to beep at the same
time that it is turning the calentool icon into reverse video to inform
the user that an appointment is coming up.

The following change to calentool.c appears to work OK:

438a439,442
>   		/* Also ring bell */
>   		gtime.tv_sec = 1L;    /* Set beep time in seconds */
>   		gtime.tv_usec = 0L;
>   		win_bell(tool->tl_windowfd, gtime, 0);

Note that the program appears to go through its "there's an appointment"
logic each time it wakes up and notices that there is a current
appointment.  Since the icon is already in reverse video, you don't see it
change to reverse video again, but you will hear the beep again.  A little
more thought and logic would get the beep to happen only the first time.

Regards -

Bruce Rosen
BBN Software Products

------------------------------

Date:    Fri, 21 Oct 88 16:56:21 EDT
From:    betty symonds  <symonds at chagall.cad.bnl.gov>
Subject: Thanks for info about adding SMD disks
Reference: v6n264-entry

Reply of thanks to all who responded to my request for information on
adding SMD disks.  The information I received is very useful and I will be
contacting some people as soon as we discuss the various options offered.
THANKS ALL.

Betty Symonds

------------------------------

Date:    Fri, 21 Oct 88 10:42:54 EDT
From:    cucard!proexam!glen at nyu.edu (Glen Brydon)
Subject: SCSI vs. SMD performance

A company here in NYC just made an interesting presentation about disk
drives (and high volume tapes) which you ought to know about.

First, we have a 3/280 with an SMD/super-eagle drive and a 1/4" tape on a
SCSI interface.  An outfit called BoxHill Systems Corp at 1841 Broadway,
Suite 609/NYC 10023 (800) 727-3863 claims that they can offer up to 4
760MB disks in a single rack-mount cabinet at about half the price of
Sun's offerings and about half the access time.  All of this is on the
SCSI interface and the salesperson claims that they couldn't achieve such
performance on the SMD!

BTW, the high-end tape drive is rated at 2.3 GBytes on an 8mm cartridge.
In NYC they have a relationship with another outfit that offers service on
Sun equipment for about half the price again.

Sun 3/60, internal SCSI int., BoxHill 380 MB disk subsystem with embedded
controller reads at 655 kB/sec under Unix using 512 byte reads.

Just how good is this and what about the SCSI vs SMD issue???

Glen Brydon
PES (Professional Examination Service, Inc.)
475 Riverside Drive, Suite 740
New York, NY  10115
212-870-2262	glen at proexam

------------------------------

Date:    Sat, 22 Oct 88 10:06:59 PDT
From:    "John D. Polstra" <polstra!jdp at uunet.uu.net>
Subject: Hardware Flow Control with MTI

Many people have been asking how to get hardware flow control on Sun
serial ports.  Although it's not guaranteed anywhere in the Sun manuals, I
have found that hardware flow control works just fine for outgoing data on
the Systech Multi-Terminal Interface (sometimes called the ALM-1, I
think).  The following statement in MTI(4S) prompted me to try it: "The
Systech requires the CTS modem control signal to operate."  I found that
connecting the flow-control signal to the CTS pin of the MTI (pin 5) did
exactly what I had hoped it would do.  I've been using this with a daisy
wheel printer for over two years, and there has not been a single problem
with it.  I know that the flow control is working, because I've seen what
happens to the printer when there is no flow control.

Now for the disclaimer:  I've used this on a 3/180, under SunOS 3.0, 3.2,
3.4, and 3.5.  I don't know what will happen under 4.0.  Also, I've used
it only in connection with an output-only device.  Conceivably, dropping
CTS might cause bad things to happen to input on the port, rendering the
technique useless for modems.  (I don't know, because I've never tried it.
If anybody has experience with this, I'd appreciate hearing from them.)

Anyway, if you have an MTI and need hardware flow control for a printer,
try using the CTS pin.  It works for me.

John Polstra (jdp at polstra.uucp)

------------------------------

Date:    17 Oct 88 18:41:39 GMT
From:    phri!roy at philabs.philips.com
Subject: Responses to tape survey

A few weeks ago (in Sun-Spots Digest, v6n233) I asked people to let me
know what kind of tape drives they had.  Now that the flood of replies has
stopped, here's the results.  BTW, I got 79 replies, far more than I
expected; I must have hit a raw nerve.  To save space, I won't list
everyone by name, but thanks to everyone who took the time to send me
mail.

	Votes	Catagory

	31	1a) Both 1/4" and 1/2"
	19	1b) Local 1/4" with resasonable over-the-net access to 1/2"
	9	1c) Local 1/2" with resasonable over-the-net access to 1/4"
	2	2)  Only 1/4"
	12	3)  Only 1/2" (I'm one of the 12)

I'm not 100% sure why that doesn't add up to 79, but it's pretty close and
I'm certainly not going to slog through all 79+- letters all over again to
find the error.  Some people seemed to delight in chatting about their
situation but never actually coming out and stating which category they
were in; in those cases, I made the best guess I could.

Those people who only had one kind of tape were pretty loud in their anger
that companies just don't seem to listen when you tell them which kind of
tape to send.  From what I can tell (and this is my experience too), Sun
is just as bad as anybody else, if not worse (and all they have to do is
look up our sales and/or service records to see what kind of drive they
sold me).  Interestingly enough, some of the people from software
suppliers who wrote to me did a pretty good job of making fools of
themselves -- I mean, here I am trying to show the world that it really is
important to listen to your customers when they say they need a specific
kind of media, when along comes:

	we ship [product name deleted to protect the guilty] for the
	Sun3 and Sun4 on 1/4" only -- we won't even bother with 1/2",
	since nobody seems to need it.

BTW, the reason I ran this poll is because I'd just finished getting about
a half dozen 1/4" tapes from various suppliers when I specified 1/2".
Among the recently guilty companies were Sun (who gets the prize for
sending me 4 different pieces of software on 1/4" tape in the course of a
month), BBN/Prophet, and Sabre.  In the not-too-distant past, I've had the
same problem from the people who put out FrameMaker and Interleaf.  One
person said that he has the same problem with DEC, who seems to insist on
shipping software on TK-50's.

I'm actually surprised at how few people don't have 1/4"; only about 15%
said they couldn't read 1/4" if they had to.  Lots of people had bad
things to say about 1/4" tape, but had them and could use them in a pinch.
Meanwhile, the other 15% of us are left out in the cold with companies who
insist on blindly shipping 1/4", regardless of what the customer requests.

Roy Smith, System Administrator
Public Health Research Institute
{allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy at uunet.uu.net

------------------------------

Date:    Fri, 21 Oct 88 17:20:12 EDT
From:    "Tim G. Smith" (Mechanical) <tsmith at usna.mil>
Subject: 3/50 memory upgrade options

< I spoke with a company in Cambridge, MA recently (Cambridge Automation
< perhaps?)  recently, that claims to have upgrade memory for Sun-3/50s....

Adding memory to a 3/50 would not be the most difficult thing in the
world. Some of our local techies were going to come up with the hardware
themselves but after talking to third party memory vendors they decided to
save themselves some work and wait for the products that the vendors were
developing.

If anyone wants more info don't ask me- call your favorite third party
memory vendor and ask them about it. I got two of the major vendors to
admit to working on it at least 6 months ago. One of them is in beta test
and expects to have products RSN. The other is hoping for a December or
January date.

In any event there will be some buyable products on the street Real Soon
Now for all of the people suffering from sluggish 3/50s.

	Tim Smith		
US mail:ECSD/CADIG mailstop: 11G	E-mail:
	US Naval Academy		internet:tsmith at USNA.MIL
	Annapolis, MD 21402		uucp	:...!uunet!usna!tsmith
MaBell :(301)267-4413

------------------------------

Date:    21 Oct 88 13:17:13 GMT
From:    fedsys!scotty!jwr at kodak.com (Dier Retlaw Semaj)
Subject: icon->cursor conversion

Is there a way to convert an icon into a cursor?  I have a few icons I
like, & would like to use one of them for my cursor.  Please use E-Mail if
you know.  Thank you much.

Dier R. Semaj	{ames,cmcl2,rutgers}!rochester!kodak!fedsys!wally!jwr

------------------------------

Date:    18 Oct 88 19:18:18 GMT
From:    phri!roy at philabs.philips.com
Subject: Problems with oob data on sockets using the notifier

Either I'm going crazy, or notify_set_exception_func() doesn't work as
advertised (I'm currently using SunOS-3.5.2, but I've gotten similar
results on 3.2).  I'm trying to write an rlogintool which has essentially
the same functionality as running rlogin inside a shelltool window, but
all in one process.  The rlogin protocol uses out-of-band signals to
negotiate window size changes and stuff like that.

What I do is register an urgent data handler for the remote socket with
the notifier by calling notify_set_exception_func().  Sometimes, however,
it seems that my handler gets called when there really isn't any oob data
pending; select() says there is no pending exceptional condition, a recv()
returns 0, and ioctl (SIOCATMARK) says we're not at the mark, but if I do
a read(), I end up blocking.

Anybody have any idea what might be going on?

Roy Smith, System Administrator
Public Health Research Institute
{allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy at uunet.uu.net

------------------------------

Date:    21 Oct 88 18:48 +0100
From:    Dominique Petitpierre <petitp at cui.unige.ch>
Subject: How to dynamically change the font in a tty window?

Is it possible to change dynamically the font in a tty or text window
under sunview (SunOS 3.4)? Or is it fixed at creation time?  I know you
can change the default font with defaultsedit(1) or by setting the
environmment variable DEFAULT_FONT. But what I want is change my mind
after a window is opened (say a shelltool window).

I have RTFM, but you know what TFM is like sometimes when you ask the
wrong question! The example program /usr/src/sun/suntool/examples/font_menu.c
is a nice start, but, at least for me, it doesn't even change the font in
its panel (just the name changes)!.

At our site, we use the Unipress emacs editor, and it has this feature:
you open a window containing the file names of the fonts, click on a name
and here you are with courier bold 24 points (for example). But emacs
doesn't use a tty window.

My purpose is to adapt to various kinds of output produced by a program
run under shelltool or commandtool, while giving a demo. It is possible to
run it under emacs, but the load is then too big with almost 2 MB (1MB
suntools, 1MB emacs) on a 4MB SUN 3/60 => bad response time => bad demo!!

Thanks in advance.
--
Mr. Dominique Petitpierre   |EAN, BITNET, EARN, MHS, X.400: petitp at cui.unige.ch
ISSCO, University of Geneva |UUCP: mcvax!cernvax!cui!petitp , petitp at cui.uucp
54 route des Acacias	    |JANET: petitp%cui.unige.ch at ean-relay.ac.uk
CH-1227 GENEVA (Switzerland)|CSNET, ARPA: petitp%cui.unige.ch at csnet-relay.csnet
Tel: 0041/22/20 93 33 extension 2117  

------------------------------

End of SUN-Spots Digest
***********************



More information about the Comp.sys.sun mailing list