Sun-Spots-Request@RICE.EDU (William LeFebvre) (01/26/88)
SUN-SPOTS DIGEST Monday, 25 January 1988 Volume 6 : Issue 7
Today's Topics:
Editor's "misteak" in Volume 6 issue 5
Re: Sendmail problems (where does the Sun.COM come from)
Re: Small SCSI Disks -- Early Report On Availability
Re: Removable storage for Suns
Floating point arithmetic investigations
dumpregion tips
Bug in getpwent() causes core dump in sa.
Interphase 3200 disk controller on a Sun-4
Licensing? NO question!
Help needed with RPC, XDR, UDP Routines
Routing Sun I/O over Ethernet?
Salvaging a dump tape - HOW?
TN3279?
Using sendmail to strip local names?
Mapping bad blocks to file names?
Integrating Sun mail with VMS mail?
MacDraw and/or MacPaint for Sun
Mac in a window
vt100tool
----------------------------------------------------------------------
Date: Mon, 25 Jan 88 15:42:26 CST
From: William LeFebvre <phil@Rice.edu>
Subject: Editor's "misteak" in Volume 6 issue 5
My apologies to one and all. I somehow got my names mixed up in v6n5 and
inadvertently said that the HP LaserJet Series II filters were stored in
"sun-source/hp-filters.shar". In fact, they were stored in
"sun-source/lj-filters.shar". I have rectified this problem by placing
the shar file under both names (linked to the same file). Those who
unsuccessfully tried to retrieve it, please try again. Sorry.
William LeFebvre
------------------------------
Date: Sat, 16 Jan 88 11:09:21 -0500
From: randy@ncifcrf.gov
Subject: Re: Sendmail problems (where does the Sun.COM come from)
The default top level domain in the files sun provides is 'com', and the
default second level is 'sun'. These are not at the same place in the
file, so a grep 'sun.com' will not catch them. Look for 'DDsun' and
'CDsun' and 'DUcom' and 'CUuucp com', which are the four lines in the
original sun distributed sendmail.cf that are relavant. I'm not sure how
this could have been left in, given that you are using a hand-written
sendmail.cf file, but your problem sounds *very* much like one of your
machines wasn't updated from this original. Hope this helps.
-- Randy Smith
Randy Smith @ NCI Supercomputer Facility
c/o PRI, Inc. Phone: (301) 698-5660
PO Box B, Bldng. 430 Uucp: ...!uunet!ncifcrf.gov!randy
Frederick, MD 21701 Arpa: randy@ncifcrf.gov
------------------------------
Date: Fri, 15 Jan 88 11:35:50 EST
From: bzs%bu-cs.bu.edu@bu-it.bu.edu (Barry Shein)
Subject: Re: Small SCSI Disks -- Early Report On Availability
Aviv is now selling disks and tapes of several useful flavors specifically
into the Sun market at what I would call very reasonable prices. This
means you can get correct cables, support and service also, I presume
(haven't bought any yet, but their brochures and salesperson are all
geared towards Suns.) They can send you a package of brochures and prices
(no phone number handy, they're in Woburn, MA, 617 area code.)
-Barry Shein, Boston University
P.S. I have no commercial interest and all that.
------------------------------
Date: Sat, 16 Jan 88 14:25:37 +0200
From: leonid@TAURUS.BITNET
Subject: Re: Removable storage for Suns
kem@emory suggested using old DEC RM05 to connect to SMD controller.
Well, although this is uneconomic I have tried it and discovered that the
DEC RM05/RM03 disks are custom-made by CDC, and the interface was
MODIFIED! In a word, it will not work unless you replace some 3 boards
and rewire the backplane of the RM05/RM03 board set. Fortunately I have
the diagrams for this driver and if you are interested I can tell you
exactly what these differences are.
Leonid
------------------------------
Date: Sun, 17 Jan 88 00:19:40 PST
From: quintus!ok@sun.com (Richard A. O'Keefe)
Subject: Floating point arithmetic investigations
I just acquired the Cody & Waite Fortran Elementary Function test package
(ELEFUNT) and paranoia.c from anl-mcs.arpa!netlib . With
f77 -fsoft # SunOS 3.2, SUN-3/50 or SUN-3/160
the double precision tests reported a couple of glitches:
o DX**DY is supposed to signal an error or return NaN when
DX is negative, but if DY happens to equal an integer IY it
returns the value that DX**IY would have returned
o ELEFUNT calls the builtin DCOTAN(), which doesn't exist
o datan2(0.0d0, 0.0d0) is undefined, so should return NaN,
but actually returns 0.785+
I hacked dmachar.f to return dflmin() and dflmax() rather than relying on
its own calculations, which returned a denormalised number and Infinity
respectively.
I was considerably surprised, however, to find that
f77 -f68881 # SunOS3.2, Sun-3/160
yielded quite different results. (machar thought there were 64 bits in
the significand, not 53.) So I decided to run paranoia in both modes. I
tried
cc -fsoft paranoia.c # with and without -O
and cc -f68881 paranoia.c # with and without -O
Under the -fsoft condition, paranoia was entirely happy. But under the
-f68881 condition, paranoia reported
Some subexpressions appear to be calculated extra # does not
precisely with about 11 extra B-digits, i.e. # appear if
roughly 3.31133 extra significant decimals. # -fswitch
* is neither chopped nor correctly rounded. # these do
/ is neither chopped nor correctly rounded.
Addition/Subtraction neither rounds nor chops.
Sticky bit used incorrectly or not at all.
FLAW: lack(s) of guard digits or failure(s) to correctly round or chop
(noted above) count as one flaw in the final tally below.
Square root is neither chopped nor correctly rounded.
Observed errors run from -5.0000000e-01 to 5.0000000e-01 ulps.
Does anyone know whether the 'neither chopped nor correctly rounded'
reports are an artefact of the extra bits (but surely the results should
be rounded back correctly?) or reflect a real problem?
It is interesting that the 'extra B-digits' message does not appear if I
specify -fswitch on a machine with the 68881, but the 'neither chopped nor
correctly rounded' messages do. What's going on?
I can't honestly say that any differences between -fsoft, -fswitch, and
-f68881 other than speed would have pleased me (I thought that portability
was supposed to be one of the goals of IEEE P754), but this much of a
difference bothers me.
------------------------------
From: Eric Ole Barber <mcvax!nw.stl.stc.co.uk!sizex@uunet.uu.net>
Date: Thu, 14 Jan 88 12:05:50 GMT
Subject: dumpregion tips
Chris Uppal, a resident guru, points out the following -
i) plain vanilla terminal and software - no problems
ii) greyscale terminal with no options to suntools - uses 8-planes of
greyscale. All rasterfiles from screendumps will be 8 planes deep, and
PostScript software won't take it. Convert to 1-plane with "rasfilter8to1".
dumpregion | rsfilter8to1 | lpr -v
iii) run only the black and white overlay plane of the greyscale terminal by
starting suntools with the -overlay_only option. Then set the
environment variable FRAMEBUFFER to /dev/bwtwo0 so that dumpregion knows
to run in the (1-plane) overlay plane. /dev/bwtwo0 must be created by
running MAKEDEV bwtwo in /dev.
------------------------------
Date: Thu, 14 Jan 88 18:17:34 CST
From: "Matt Crawford" <matt@oddjob.uchicago.edu>
Subject: Bug in getpwent() causes core dump in sa.
Recently our sunos 3.4 "/usr/etc/sa -m" began core dumping just before
writing to stdout. I traced the problem to a bug in getpwent which hits
if yellow pages are used, and if malloc() is called between two runs
through the password file. The following short program will duplicate the
problem.
================
#include <pwd.h>
i;
main()
{
setpwent();
while (getpwent())
;
endpwent();
for ( i=2048; i>2; i>>=1 ) {
bzero(malloc(i), i);
}
setpwent();
for ( i=100; i>0; i-- ) {
printf("%s\n", getpwent()->pw_name);
}
endpwent();
}
================
Running this will cause a core dump before printing out the first
user from the yellow pages (if you have fewer than 100 users in
/etc/passwd itself). A dangling pointer to some freed memory is left
after the first loop of getpwent(), and is freed again during the
second loop.
Given the nature of the bug I don't see why "sa -m" ever worked in
the first place. I suppose our yp passwd file finally hit a critical
size (it has 256 lines). A source fix to lib/libc/gen-4.2/getpwent.c
follows.
RCS file: RCS/getpwent.c,v
retrieving revision 1.1
retrieving revision 1.2
diff -c -r1.1 -r1.2
*** /tmp/,RCSt1a11925 Thu Jan 14 18:02:24 1988
--- /tmp/,RCSt2a11925 Thu Jan 14 18:02:24 1988
***************
*** 400,407 ****
}
if (oldyp)
free(oldyp);
! oldyp = key;
! oldyplen = keylen;
}
static
--- 400,412 ----
}
if (oldyp)
free(oldyp);
! if (yp) {
! oldyp = key;
! oldyplen = keylen;
! } else {
! oldyp = NULL;
! oldyplen = 0;
! }
}
static
***************
*** 420,427 ****
}
if (oldyp)
free(oldyp);
! oldyp = key;
! oldyplen = keylen;
}
static struct passwd *
--- 425,437 ----
}
if (oldyp)
free(oldyp);
! if (yp) {
! oldyp = key;
! oldyplen = keylen;
! } else {
! oldyp = NULL;
! oldyplen = 0;
! }
}
static struct passwd *
________________________________________________________
Matt University matt@oddjob.uchicago.edu
Crawford of Chicago {astrovax,ihnp4}!oddjob!matt
[[ Our /etc/passwd file has much less than 100 lines, but your test
program ran fine on our SunOS 3.2 machine. --wnl ]]
------------------------------
Date: Fri, 15 Jan 88 15:12:14 EST
From: Jon Hull <hull@cs.buffalo.edu>
Subject: Interphase 3200 disk controller on a Sun-4
We have purchased three Sun-3/260 to Sun-4/260 upgrade kits and have
started to install one.
We discovered a problem regarding our Interphase controllers. Sun made
changes which require re-labelling of the disks under a new diag. We only
have a version of diag for the Sun 3, and need a new one to work under the
sun4.
Our local tech people are not sure if there might also be some changes to
the driver that we need too.
Interphase is working on this problem but they don't expect to have a
solution for a month or two (and you know how those things go!). However,
they said they do know that some other places have already made the
required mods.
So, if you have those mods and are willing to share them, please let me
know. Your assistance would be greatly appreciated.
Jon Hull
hull@cs.buffalo.edu
hull@sunybcs.BITNET
..!{ames,boulder,decvax,rutgers}!sunybcs!hull
------------------------------
Date: Fri, 15 Jan 88 18:48:58 EST
From: Root Boy Jim <rbj@icst-cmr.arpa>
Subject: Licensing? NO question!
Dick St.Peters (stpeters@ge-crd.arpa) says:
> If the Suns are not source-licensed, then you are violating your license
Breaks my heart.
> if you put any licensed source on them. Furthermore, you should read the
> fine print on your non-source licenses very carefully:
Then do what you want.
> I suspect things are not so simple.
Definitely not.
> First, what's an "update"? If I've
> got licensed source on my VAX for some utility not provided by Sun, it
> would be hard to argue that cross-compiling on the VAX and copying object
> code amounted to an "update". The original posting asked about source
> code "with modifications". When does "modified source code" become a "new
> program" incorporating parts of licensed source?
Yeah, and is adb'ing a "modification" as well?
> After all, having a
> reference on how to do things is part of why we pay money for source in
> the first place.
AND to fix bugs, AND to add features or tailor to our particular tastes.
> Second, when does licensed source become "on my Sun"? If my Sun NFS-
> mounts the partition where the VAX with the license keeps the source, does
> that count? If I rlogin to the VAX from a commandtool and cat the source,
> the source is then stored in my Sun's memory. Does that count? I wrote a
> remote-editing GNUemacs package based on using rsh buffer processes (there
> are others around based on ftp) - if I have the VAX-licensed source in an
> emacs buffer, does that count?
> How about the following:
> vax% /lib/cpp <source.c | rsh sun '/lib/ccom | /lib/c2 >source.s'
> Was the source ever "on the Sun"?
Yeah, and for a company with the motto, "The Network is the Computer",
that means that if ANY machine ANYWHERE on the ENTIRE INTERNET has
source, everywhere else "is licensed" for it. Too bad it's ultimately
licensed from TPC. Maybe they made "The Right Choice" after all.
> It's pretty evident there's a big gray area here - but there are also some
> clear rights and wrongs.
What is clear is that it is wrong to sell software without source. I find
it slightly ironic that Mark Weiser, the author of a paper so passionately
championing the virtues of source code should be quoting the party line.
(Root Boy) Jim Cottrell <rbj@icst-cmr.arpa>
National Bureau of Standards
Flamer's Hotline: (301) 975-5688
I have many CHARTS and DIAGRAMS..
P.S. Mark, aren't you glad it was my brother working for you and not me :-)
------------------------------
Date: 14 Jan 88 19:12:37 GMT
From: ames!lll-tis!dsacg1!dsacg3!nff1847@ucbvax.berkeley.edu (Jennifer M. Taulbee)
Subject: Help needed with RPC, XDR, UDP Routines
We are trying to use the Sun Systems, Inc. Client/Server model to
implement a database system. We are having problems sending a string
across to the Server from the Client. We can send a string of less then
800 characters. However, when we try to send one bigger than 800/900
characters we get an error message stating that RPC can not encode the
arguments. We are using the Gould UTX32 Unix operating system on a Gould
Powernode 9050 series minicomputer.
We would like to know if anyone else has experience in using the Sun
Systems, Inc. RPC mechanisms, and if so HELP!
Please, any replies would be greatly appreciated.
Thanks in advance.
****************************************************************************
* Jennifer Taulbee << DLA Systems Automation Center, Columbus, Ohio >> *
* PHONE: 614-238-1315 UUCP:{uunet!gould,cbosgd!osu-cis}!dsacg1!jtaulbee *
****************************************************************************
------------------------------
Date: 15 Jan 88 15:55:08 GMT
From: flan!jlc@mcnc.org (Josh L. Clevenger)
Subject: Routing Sun I/O over Ethernet?
I am writing network management application software for a Sun 360.
Within this application I need to access the following external hardware
types:
async, raw character level communications channels
modem dial-out connection
modem dial-in connection
printer
Due to limitations of the workstation's external ports, I wish to access
the aforementioned hardware via a port extender over an ethernet link.
What software will be necessary on the workstation and port extender to
route my high-level buffered I/O of my C-based code out to my remote
hardware ?
------------------------------
Date: Fri, 15 Jan 88 17:10:20 EST
From: csrobe@icase.arpa (Charles S. Roberson)
Subject: Salvaging a dump tape - HOW?
We have a 0-Level dump tape that has gone bad on us. It was written using
the standard dump command that comes with our Sun 3. I know that it was
successfully written because I have read it before. I also know that it
is not our drive since our drive can read other tapes and another Sun
cannot read this tape.
Here is what dump spits out:
"Tape read error while restoring <directory file - name unknown>"
and here is what is displayed on our console:
xt0: lost interrupt
xt: bad command synchronization
xt: bad command synchronization
xt0: spurious interrupt
When i try to read it on the other machine dump gives me the same message
but there was a different console message:
xt0: hard error bn=1820 er=6
Can somebody tell me how we can salvage what we can from the tape?
Perhaps using dd or even writing our own program (or using one somebody
has already written)? Any and all help concerning the dump tape format,
the xylogic error codes, block sizes and blocking factors, incantations
and spells will be appreciated.
PLEASE RESPOND DIRECTLY TO ME.
thanks,
-chip
Chip Roberson ARPANET: csrobe@icase.arpa
1105 London Company Way BITNET: $csrobe@wmmvs.bitnet
Williamsburg, VA 23185 UUCP: ...!uunet!pyrdc!gmu90x!wmcs!csrobe
------------------------------
Date: Sat, 16 Jan 88 10:10:25 EST
From: Ira Fuchs <FUCHS@pucc.bitnet>
Subject: TN3279?
Has anyone written an enhanced version of TN3270 that implements the
attribute bytes (including color) for 3279 emulation?
------------------------------
Date: 16 January 1988 1:10:38 pm
From: pplace!penn!ted@sun.com (Ted Goldstein)
Subject: Using sendmail to strip local names?
Our mail configuration doesn't strip out my local machine name (in my case
'penn') before sending a mail message off site (in my case 'pplace' or
'parcplace'). Is it possible to get sendmail to strip out the local
machine name, but to leave the site name intact? I notice this problem
coming from many Sun sites, but not all.
Sincerely
Ted Goldstein
------------------------------
Date: Mon, 18 Jan 88 0:28:23 MET
From: Karl Kleine -- FZI Karlsruhe <kleine@ira.uka.de>
Subject: Mapping bad blocks to file names?
I have problems with a disk (here: ECC on F-M2361). The driver reports
problems with block numbers on the console. Ok, that's what it should do
and does. I know how to handle darling `diag' :-( ...
Now my problem: these bad blocks are part of files in active file systems.
Is there a utility of the icheck/ncheck/fsck/... family which accepts a
list of defective disk block numbers and spits out the names of the files
of which these blocks are part of?
Current brute force method is reading off the whole filesystem, which may
also reveal more bad spots and can turn the console go gaga:
( find /BadFileSys -type f -a -exec cp {} /dev/null \; ) 2> BadList
File Badlist will contain lines like
cp: /Pathname: I/O error
That however takes quite a number of execs and time... Somebody must have
written such a utility for BSD systems (last time i went to internals like
that was under V6/V7 on the good ol' 11 :-), and now I just don't have the
time -- HELP!)
Karl Kleine
Forschungszentrum Informatik
Haid-und-Neu-Str. 10-14
D-7500 Karlsruhe, West Germany
phone: +49 - 721 - 69 06 39
networks: kleine@germany.csnet
kleine@uka.uucp
------------------------------
Date: Sat, 16 Jan 88 17:18 EST
From: <STELLABO@CSHLAB.BITNET>
Subject: Integrating Sun mail with VMS mail?
Does anyone know of a way to integrate the Sun mail system with the mail
system on a MicroVAX using TCP/IP. We are using a MicroVAX II as our
BITNET server machine. We purchased an Excelan Ethernet board (EXOS 203)
for this computer so we could transfer mail between the VAX and our Sun
computers. We would like the VAX to automatically forward the BITNET mail
to the appropriate user on the Sun. Also, the ability to reply to
forwarded mail without rlogging into the VAX. The Excelan board came with
FTP and TELNET software so file transfer is not a problem.
Has anyone developed shell scripts, DCL command procdures and/or software
that will handle the file transfer and address parsing with a friendly
front end.
Fred Stellabotte
Computer Systems Manager
Cold Spring Harbor Laboratory
Cold Spring Harbor, NY
(516) 367-8420
------------------------------
Date: 14 Jan 88 20:08:51 GMT
From: umix!oxtrap!rsj@rutgers.edu (Rick Jaffe)
Subject: MacDraw and/or MacPaint for Sun
I know this question has been asked before, but I would appreciate an
update: Are utilities with functionality similar to MacDraw and/or
MacPaint available for Suns?
[[ There is a program called "fig" that is similar to MacDraw. We got it
via anonymous FTP from "sally.utexas.edu" in "fig/fig.tar.Z". It comes
with a program that converts fig output files into pic. One of these
days I may get ambitious and put it in the Sun-Spots archives. --wnl ]]
------------------------------
Date: 14 Jan 88 20:16:36 GMT
From: umix!oxtrap!rsj@rutgers.edu (Rick Jaffe)
Subject: Mac in a window
Will I ever be able to get my hands on a "Macintosh" running in a Sun
window, or is it a hopeless "non-product" at Sun?
------------------------------
Date: Fri, 15 Jan 88 07:42:07 EST
From: chuck@wooglin.scc.com (Charles Williams)
Subject: vt100tool
A question of curiosity. Is vt100tool the suntools interface to your vt100
terminal emulator that is purchased as add-on software, or is it a vt100
emulator in it's own right?
If the answer is the latter (suntools), what do I have to do to get a 1/4"
tape cartridge of the tape??
Charles B. Williams
CONTEL Federal Systems
12015 Lee Jackson Highway
Fairfax, VA 22033-3346
[[ vt100tool isn't "ours". It is an emulator in its own right---it runs
as a tool under SunView. I am of the impression that it is available on
the Sun User Group (SUG) software tape---at least there is something
called "vttool" (claiming to be a VT100 emulation tool) on the tape. For
more information about SUG and about the tape, you can send mail to either
"users@sun.com" or "sun!users". --wnl ]]
------------------------------
End of SUN-Spots Digest
***********************