Sun-Spots-Request@RICE.EDU (William LeFebvre) (01/16/88)
SUN-SPOTS DIGEST Friday, 15 January 1988 Volume 6 : Issue 4 Today's Topics: Re: Spreadsheets for the Sun in v5n71 Re: Text-retrieval program Re: Sun consoles and keyboards Re: Upgrade to SunOS 3.4 or not SunOS 3.4 f77 optimiser An irritating bug in f77 (3.4) Availability of 68881 chips Fixed problem with partial screendump program MacPaint to Sun Rasterfile converter Process limit under 3.3 Sun crashing due to interaction between RPC broadcast and IPC board Trouble with ditroff on Suns EPSON Printer Problems Sun vs. DELNI Tn3270 info and RSCS info? Changing UDP buffer space? Simulator for OS course? Color screen dump on a Tek 4696? Converting ST rasters to Sun? Sun-Spots archive? Networking info please ---------------------------------------------------------------------- Date: Fri, 18 Dec 87 15:52:36 CST From: steve@ncsa.uiuc.edu (Steve Christensen) Subject: Re: Spreadsheets for the Sun in v5n71 I have investigated several Sun spreadsheet programs in the past few months - C-Calc, Ultracalc and 20/20. If you are interested in a menu based system like 123 or Excel, then 20/20 is the most like those. Access Technology that sells 20/20 also sells a 123 to 20/20 import/export system that runs pretty well. It does not convert Excel spreadsheets, but there are ways to do that in a Mac or PC. The color graphics is also pretty good though not as easy to use as Excel or Cricket Graph on a color Mac II. Getting the files back and forth between the machines is the most complex part unless you have Kinetics boxes or ethernet boards in your micros. It can be done however. I have been told by most of the vendors that upgrades that use one of the X, Sunview or NeWS systems in a nice way are in the works. Steve Christensen NCSA, Illinois ------------------------------ Date: Mon, 21 Dec 87 09:53:24 +0100 From: mcvax!solaris!wyle@uunet.uu.net (Mitchell Wyle) Subject: Re: Text-retrieval program We have a large library of Modula-2 code on our sun-cluster which forms the core of three separate PhD dissertations in Information Retrieval. Before you look into re-inventing the wheel, however, RT*M: addbib(1) sortbib(1) roffbib(1) indxbib(1) refer(1) I use this family of programs instead of the large modula-2 modules. Indxbib is apparantly a shell script which calls word-inversion routinas called /usr/lib/refer/mkey and /usr/lib/refer/inv_[ac] which may be the building blocks you need for customization. These routines provide stemming (albeit primitive) stop-lists, and word-inversion. They were created by Michael Lesk, one of the lesser G-ds of Info Retrieval. The Modula-2 code here is more flexible, provides anti-stop lists, frequency loading, and associative arrays. It is much slower, however because it runs on Ceres PCs and was ported to suns. Please keep me informed on replys to your posting, and let me know if you want to look at the Modula-2 code. (I could send the definition modules which are well-documented.) wyle@solaris.uucp wyle@ethz.uucp wyle%ifi.ethz.ch@relay.cs.net +41 1 256 5237 ------------------------------ Date: Tue, 22 Dec 87 07:12:24 EST From: crepea@spam.istc.sri.com Subject: Re: Sun consoles and keyboards > From: John L. Shelton <jshelton@ads.arpa> > To: crepea@spam.istc.sri.com > Subject: Your request about Sun consoles and kbds > > I saw your note on Sun-spots. > > We regularly run Sun color consoles 100 feet away with ribbon cable > extensions for the keyboard, and normal RG-59U coax for the video. You > should be able to go to 1000 feet on the video, but I haven't tried. (If > necessary, you can get low-loss coax.) > > I suspect the keyboard can be extended further if you build some gauge 20 > cables to do the job. Make sure you get low capacitance cable. > > =John= John, We have successfully remoted the keyboard 500` with 22g shielded pair cable. The voltage drop was less than .4v for the 5v line. We used 500` rolls of Belden 8281 cable to remote the video. The picture was fair, but slightly blurred. We are going to look at better connectors for both the monitor/cpu and the cables. We are also looking at some video compensation amplifiers on the market. These are EXPENSIVE. Thanx for the help. Ken C. [[ Ken, please have the appropriate person at your site fix your mailer. It sent that message to me with a "From" address of "crepea@ccfs". This is clearly wrong. If it wasn't for your "Cc" line, I would have had no way of finding your return address. --wnl ]] ------------------------------ Date: Tue, 29 Dec 87 17:32:31 GMT From: James Davenport <jhd%maths.bath.ac.uk@nss.cs.ucl.ac.uk> Subject: Re: Upgrade to SunOS 3.4 or not 1) I believe that you should definitely install the compilers etc off the 3.4 tape. As I have said in this forum in the past, they fix many bugs (but there ARE still bugs in the FORTRAN optimiser). It's a but hard to say exactly what you need - I'll try to generate a list: /bin (or /pub/bin) -rwxr-xr-x 1 root 204800 Apr 22 1987 as /lib (or /pub/lib) -rwxr-xr-x 1 root 73728 Apr 22 1987 compile -rwxr-xr-x 1 root 196608 Apr 22 1987 c2 -r-xr-xr-x 1 root 229376 Apr 22 1987 ccom -rw-r--r-- 1 root 326548 Apr 29 1987 libc.a /usr/lib rwxr-xr-x 1 root 327680 Apr 22 1987 f77pass1 -rw-r--r-- 1 root 53752 Apr 22 1987 libF77.a -rw-r--r-- 1 root 62366 Apr 22 1987 libF77_p.a -rw-r--r-- 1 root 93520 Apr 22 1987 libI77.a -rw-r--r-- 1 root 99346 Apr 22 1987 libI77_p.a -rw-r--r-- 1 root 24066 Apr 22 1987 libU77.a -rw-r--r-- 1 root 26658 Apr 22 1987 libU77_p.a -rw-r--r-- 1 root 149562 Apr 22 1987 libm.a -rw-r--r-- 1 root 168250 Apr 22 1987 libm_p.a -rw-r--r-- 1 root 11934 Apr 22 1987 libmp.a -rw-r--r-- 1 root 55230 Apr 22 1987 libpc_p.a -rw-r--r-- 1 root 45584 Apr 22 1987 libpc.a -r-xr-xr-x 1 root 139264 Apr 22 1987 f1 -r-xr-xr-x 1 root 155648 Apr 22 1987 cg -rw-r--r-- 1 root 16664 Apr 22 1987 fswitch.il (or whatever) -rwxr-xr-x 1 root 212992 Apr 22 1987 pc0 (if you like pascal) 2) In fact, I upgraded completely to 3.4, and haven't noticed things being any worse (apart from some X25 problems, but I think I have them licked now). James Davenport ------------------------------ Date: Mon, 28 Dec 87 17:26:29 GMT From: James Davenport <jhd%maths.bath.ac.uk@nss.cs.ucl.ac.uk> Subject: SunOS 3.4 f77 optimiser I reported (Thu, 23 Apr 87) a bug with the beta-test version of this - I can now confirm that the 'real' version also has this bug. Basically, the optimiser screws up nested DO loops where the limits of the inner loop depend on the outer one. For those who missed my previous posting, here is the test program. Try it with and without -O. J.H. Davenport jhd@uk.ac.bath.maths REAL*8 A(10,10) INTEGER I,J DO 10 I=1,10 DO 11 J=1,10 A(I,J)=1000*I+J 11 CONTINUE 10 CONTINUE CALL JAMES(A,10,10) STOP END SUBROUTINE JAMES(B,IB,N) REAL*8 B(IB,N) DO 40 J=1,N C=B(J,J) JP1=J+1 IF (JP1.GT.N) GO TO 40 C = -1 DO 20 I=JP1,N C=-B(J,I) 20 CONTINUE WRITE(*,*) C 40 CONTINUE RETURN END ------------------------------ Date: Thu, 31 Dec 87 19:45:13 GMT From: James Davenport <jhd%maths.bath.ac.uk@nss.cs.ucl.ac.uk> Subject: An irritating bug in f77 (3.4) The following program (which is legitimate according to IBM compilers) causes the Sun f77 compiler (Sunos 3.4) to give: Error on line 5 of foo.f: Declaration error for MAIN: external name already used Error on line 5 of foo.f: syntax error f77: Fatal error in f77pass1: Segmentation fault (core dumped) Note that the original REAL*8 declarations are accepted fine. To make matters worse, the correction that makes it legitimate Sun FORTRAN (viz real*8 function ...) stops it from being legal IBM FORTRAN. The joys of non-standard extensions. James Davenport REAL*8 X0, X3, RES RES = SOLVE( X0, X3 ) STOP END REAL FUNCTION SOLVE*8( X0, X3 ) REAL*8 X0, X3, S1, S2 RETURN END ------------------------------ Date: 29 Dec 87 16:34:05 GMT From: roy%phri@uunet.uu.net (Roy Smith) Subject: Availability of 68881 chips Thought this might be of some interest to the more hardware oriented readers of this list. Not understanding why Sun charges $700 for the floating point option on their 3/50's (i.e. to plug in a 68881 chip) I did a bit of research a while ago (mostly by asking this list) and got several replies to the effect that indeed, all one has to do is get the chips and plug them in. I then discovered that the chips cost about $200 (for 16 MHz parts) and were essentially unavailable (12-16 week delievery times). Well, that's changed. Yesterday I called Schweber; they have them in stock (I think she said they have thousands of them in stock) and they now cost $145 in single-unit quantities. I ordered 9 of them and hope to have them in my hot little hands next week. I'll let you all know how it turns out. Roy Smith, {allegra,cmcl2,philabs}!phri!roy System Administrator, Public Health Research Institute 455 First Avenue, New York, NY 10016 ------------------------------ Date: Thu, 31 Dec 87 17:57:38 gmt From: Richard Tobin <richard%aiva.edinburgh.ac.uk@nss.cs.ucl.ac.uk> Subject: Fixed problem with partial screendump program Someone has reported a problem compiling the screendump program I sent a couple of weeks ago under release 3.0 (not surprising really, I wrote it under 3.2). I would reply directly to him, but the mail headers were sadly corrupted, and I suppose others might want the same information. Solution: change 'Pixrect' to 'struct pixrect', I think (or typedef it). -- Richard Richard Tobin, JANET: R.Tobin@uk.ac.ed AI Applications Institute, ARPA: R.Tobin%uk.ac.ed@nss.cs.ucl.ac.uk Edinburgh University. UUCP: ...!ukc!ed.ac.uk!R.Tobin ------------------------------ Date: Mon, 21 Dec 87 20:06:53 PST From: strout@lll-lcc.llnl.gov (Robert E Strout) Subject: MacPaint to Sun Rasterfile converter Well here is the program to convert MacPaint files to Sun Rasterfiles. I received many requests (>30) so I am going to send a copy off to the archives. I hope everyone that made the request directly to me received it. With mailers and pathnames, who knows? Apologies to those that did not get theirs (but now you can go to the archives). First, I must correct myself... I previously said that a fellow at Rice had developed this program. Getting back to the original, I find that actually someone (unknown) at Univ. of Texas did. Many thanks to him/her!!! Once you convert your file to raster format, you may screenload or pssun it. Have fun!!! Rob Strout II (resii) strout@lll-lcc.arpa Lawrence Livermore National Laboratory ...{ihnp4,pyramid}!lll-lcc!strout [[ It has been placed in the archives as "sun-source/paintraster.c". It can be retrieved via anonymous FTP from the host "titan.rice.edu" or via the archive server with "send sun-source paintraster.c". For more information about the archive server, send a mail message containing the word "help" to the address "archive-server@rice.edu". --wnl ]] ------------------------------ Date: Mon, 21 Dec 87 7:46:54 CST From: jpt@uf.msc.umn.edu (Joseph Thomas) Subject: Process limit under 3.3 We ran into the same problem here when several users created and executed recurssive scripts. Instead of killing off user processes, the system crashed with a panic "xfree rssize" ( found in vm_text.c ). We haven't located the exact cause of that error, but did a similar type fix. If you look at param.c, it seems to me that the real "error" is in why they chose to undefine/redefine MAXUPRC. The formula seems to work great if one uses Sun's assumptions of MAXUSERS = 4, but at 32, one get a user process limit of 517 !!!!!! We decided to add an "options" in the config file and then add some ifdef's to param.c, thus allowing different kernels to choose different MAXUPRC without having to edit param.c each time. Below is what we did: sys/conf: 'GENERIC' ( or other config file ) options MAXUPROC=50 'param.c' #ifdef sun #define NPROC (10 + 16 * MAXUSERS) #undef MAXUPRC > #ifdef MAXUPROC > #define MAXUPRC MAXUPROC > #else #define MAXUPRC (NPROC - 5) > #endif #else #define NPROC (10 + 8 * MAXUSERS) ; if not sun, why not redifine #endif ; MAXUPRC as above ???? ------------------------------ Date: Tue, 29 Dec 87 08:25:53 EST From: wpm@gateway.mitre.org (William P. Malloy) Subject: Sun crashing due to interaction between RPC broadcast and IPC board We recently installed a Sun IPC (integrated Personal Computer) board into our Sun 3/180. When we ifconfig-ed in the ``me0'' interface used to talk PCNFS between the server and the PC, we discovered it would respond to requests from ``rup'' and ``rusers'' giving two responses from this machine. That's a minor problem, the BIG problem is the machine started crashing 2 to 3 times a day immediately after people typed ``rup'' and ``rusers''. A message was blasted on the console identifying these programs as the ones responsible for the crash, with a ``panic: Bus error''. Doing a savecore and then an ``adb'' backtrace seems to confirm the culprit as being the ``meoutput()'' routine in "/sys/OBJ/if_me.o". The question is has anyone else with an IPC board seen anything like this ? And if you have, do they have some way of preventing it. I have called Sun's Answerline and so far as I can tell they've never heard of this problem before. My current solution is to turn off ``rup'' and ``rusers''. The machine is a Sun 3/180 with 16 Megabytes of memory, a Xylogics 450 disk controller, 3 Fujitsu Eagle disk drives, a Tapemaster 2000 tape controller with a CDC Keystone 25 1/2" 1600 BPI only tape drive, a 16-port Systech Multiplexor, an ACC M/1822 ARPANET IMP interface, and the SunIPC board. William P. Malloy/The MITRE Corp/Dept. W-31/McLean, VA/22102 {postmaster,root,wpm}@gateway.mitre.org ------------------------------ Date: 21 Dec 87 19:50:32 GMT From: dartvax!wbc@seismo.css.gov (Wayne Cripps) Subject: Trouble with ditroff on Suns I am porting ditroff from the 4.3 Berkeley distribution to a Sun-3 running sun 3.4. It works fine, except that headers don't format correctly when using me or ms. The me macros .eh and .oh take three fields seperated by single quotes, and on the vax they print to the left, center, and right of the top of the page. On the sun they all print to the left side, and reading the output file from ditroff, it appears that the input to the macro never gets expanded. (The page number does get expanded correctly). The printers I am using are postscript (ps) and interpress (ipress), but I don't think they are related to the problem. Has anyone had this problem? Wayne Cripps wbc@u2.DARTMOUTH.EDU wbc@dartvax.uucp Box 6188 Dartmouth College, Hanover N.H. 03755 (603) 646-3198 ------------------------------ Date: 22 Dec 87 18:07:46 GMT From: boulder!seri!arch@hao.ucar.edu Subject: EPSON Printer Problems I've got an Epson LQ-1500 I'm trying to interface to a Sun 3/50. We are taking the output from the ttyb cpu port, running through a Micro-Fazer buffer/serial<->parallel converter, and then into the Epson. As a print filter, I'm using a piece of software from the SUG donated software tape called sd.ep. The bug is this: a) Occasionally, the print is skewed, almost like italics, and seemingly at random. b) When I try to send a screendump to the printer, the first few 'lines' of output look O.K. (e.g Namestripes from windows come out alright,icons at the top of the screen, etc.) but then the printer mysteriously jumps into an odd mode and prints garbage ( e.g. strange looking escape sequences and other oddities). Admittedly, I'm ignorant of how the Epson distinguishes between text and raster data, but the problem does seem to manifest itself in either mode. We have tried different baud rates, different cables, etc, but I still don't know if this is a hardware problem or if there's something wrong with the filter. Any Suggestions???? Thanks in advance, Arch Arch Mott Solar Energy Research Institute Golden,Co 80401 (303) 231-7687 .....boulder!seri!arch ------------------------------ Date: Sun, 3 Jan 88 16:20:21 est From: cbmvax!grr@uunet.uu.net (George Robbins) Subject: Sun vs. DELNI Does anyone know how to get a Sun 2, with Sun Ethernet Board, running SunOS 2.0 or later to work with a DELNI? The theories my usual sources have come up with have been more interesting than effective, so I'd love to hear from someone with direct experience in the matter... ------------------------------ From: Joel Isaacson <joel@WISDOM.BITNET> Date: Fri, 25 Dec 87 09:27:40 -0200 To: sun-spots@rice.edu Subject: Tn3270 info and RSCS info? I have two requests. 1.) Has anyone integrated Berkeley's Tn3270 and Sun's 3270 SNA, DAI interface. We would for various reasons prefer to use tn3270 ove Sun's te3278. 2.) Has anyone gotten the Urep package (RSCS, Bitnet) to work on a Sun via the synchronous port. We would very much want to discontinue (pull the plug) on our Vax that serves as our Bitnet gateway. Thanks Joel Isaacson joel@wisdom.bitnet ------------------------------ Date: 30 Dec 87 22:41:56 GMT From: rfox@amelia.nas.nasa.gov (Richard Fox) Subject: Changing UDP buffer space? At the present we do not have a source lic. for our suns. What I need to know is how to change the amount of buffer space allocated for a udp socket. I notice that it is not done the same way as a 4.3bsd on a vax. Please send responses to: rfox@ames-amelia.ARPA or call: (415)694-4358. I will repost the results.... thanks, rich ------------------------------ Date: Mon, 28 Dec 87 10:20:34 EST From: <tom@williams.edu> Subject: Simulator for OS course? Does anyone know of a processor simulator suitable for use by students completing an operating systems course project that runs on Suns. The Chip simulator from Cornell (which simulates a PDP11-like machine) would be ideal, however, porting it from the Vax to the Suns appears to be a non-trivial undertaking. Thanks for your help. Tom Murtagh ------------------------------ Date: Fri, 1 Jan 88 11:22:56 PST From: ucrmath!dennis@sdcsvax.ucsd.edu (Dennis Michael) Subject: Color screen dump on a Tek 4696? Has anyone used a Tektronix 4696 color ink jet printer to print a color Sun screen dump? I would be very interested in how the connections were made, and what software was used. Dennis O. Michael Department of Mathematics and Computer Science University of California Riverside, CA 92521 (714) 787-3236 UUCP: ..!{ucbvax!ucdavis,sdcsvax,ucivax}!ucrmath!dennis INTERNET: ucrden@sdcsvax.ucsd.edu dennis%ucrmath.uucp@sdcsvax.ucsd.edu dennis%ucrmath.uucp@ics.uci.edu BITNET: dennis@ucrvms ------------------------------ Date: 4 Jan 88 19:31:02 GMT From: jjoshua@topaz.rutgers.edu (Jonathan Joshua) Subject: Converting ST rasters to Sun? I am new to using Sun terminals and I could use any help you can give me. If you use a Sun terminal, you probably are aware of the fact that you can change the background display to a pre-drawn design. Is there a way to convert an atari st picture file so that it could be used as a backdrop on a Sun terminal? If the displays on the Sun are bitmapped like on the atari st I don't think that there would be much of a problem. Any ideas? please send replies to me. Thanx! Jonathan Joshua jjoshua@topaz.rutgers.edu Rutgers - Since 1766 ------------------------------ Date: Tue, 1 Dec 87 20:51:31 EST From: Marty Hall <hall@alpha.ece.jhu.edu> Subject: Sun-Spots archive? Where is the Sun-Spots program archive? I tried anonymous ftp to rice.edu, but that was obviously the wrong place. Thanks. - Marty Hall [[ It's a lesson in reading ability, and it is amazing how many people fail. The correct host is "titan.rice.edu", as every editor's comment on the subject has clearly stated. --wnl ]] ------------------------------ Date: Mon, 4 Jan 88 15:21:48 EST From: clapper@nadc.arpa (Brian M. Clapper) Subject: Networking info please We're presently gathering information about workstation LANs, particularly those which support heterogeneous hardware configurations. For example, we forsee the need to hook up Apollos, Suns, MacroVAXes and the like. I am looking for *any* information I can get on: 1. Networking schemes (e.g., Ethernet, IEEE 802.2 standards, etc.) 2. Personal experiences with this sort of problem. 3. The degree to which various hardware vendors support the various services. E.g., when a vendor claims to support NFS, how easy is it *really* to drop one of their machines into a network of Suns? 4. Support for TCP/IP. References to useful publications or articles, informal personal anecdotes of experiences in this area, other electronic mail contacts -- any or all of these would be extremely useful. Please send replies directly to me, as I do not presently subscribe to this mailing list. Thanks in advance. Brian M. Clapper clapper@nadc.ARPA Naval Air Development Center Street and Jacksonville Roads Warminster, PA 18976 (215) 441-2118 ------------------------------ End of SUN-Spots Digest ***********************