Sun-Spots-Request@Rice.edu (William LeFebvre) (11/15/88)
SUN-SPOTS DIGEST Sunday, 13 November 1988 Volume 7 : Issue 13 Today's Topics: Re: SUN 386i dos windows problem, and fortran Re: single user while dumping Re: Memory for a Sun 3/50 Re: helios memory for SUN-3/50 Re: Postscript Interpreter (2) Patch to fstat.c sgi <-> sun 1/4" tape transfers Appendix Examples from "Writing Device Drivers" Sun Education January - June 1989 Schedule Lurnix announces two Sun related classes 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: Thu, 3 Nov 88 14:35:29 EST From: shenkin@cubsun.bio.columbia.edu (Peter Shenkin) Subject: Re: SUN 386i dos windows problem, and fortran Reference: v7n2 >Does anybody know what is the status of SUN fortran for the 386i (as well >as the status of decnet) ? I just got a 386i with Sun Fortran. My number-crunching programs run just twice as fast (actual measurements) on it as on a Sun 3-180 also running SunOS 4.0. On the latter machine, there was no improvement over f77 running under SunOS 3.4. On the Sun 3, the times were VAX 11/780-like. Peter S. Shenkin, Department of Chemistry, Barnard College, Columbia University New York, NY 10027 NEW TEL !!: (212) 854-1418 (work); (212) 829-5363 (home) shenkin@cubsun.bio.columbia.edu shenkin%cubsun.bio.columbia.edu@cuvmb.BITNET ------------------------------ Date: Wed, 2 Nov 88 10:46:12 GMT From: mcvax!ritd.co.uk!mr@uunet.uu.net Subject: Re: single user while dumping albert@mssun7.msi.cornell.edu (Jay Albert) asks whether it is safe to make level 0 (or any for that matter any other) dumps whilst multi-user. I am no expert in this area, other than 4 years practice. We *always* perform large dumps on live servers. A level 0 dump takes us 1-2 working days; no way am I taking the net off-line 2 days a month (plus 1 day a week for level 0's). We have *never* had any problem restoring files. I believe that in theory active files (i.e. being written to whilst being dumped) may be munged. We don't care, as such files are generally not important to us. The manual is "correct" in that Sun cannot recommend a practice that could theoretically result in (partially) defective dump sets. After many hundreds of dumps, I conclude that we haven't just been running on luck and the probability of problem is acceptably small. When I *know* that I might have to restore a disk (e.g. moving a server which might get damaged) then I become paranoid and dump single user. Martin Reed, Racal Imaging Systems Ltd uucp: mr@ritd.co.uk,uunet!ukc!ritd!mr,sunuk!brains!mr Global String: +44 252 622144 Paper: 309 Fleet Road, Fleet, Hants, England, GU13 8BU ------------------------------ Date: Wed, 2 Nov 88 11:29:20 +0100 From: Jeremy Cook <mcvax!heops.cmi.no!jeremy@uunet.uu.net> Subject: Re: Memory for a Sun 3/50 I would like to know whether Sun gives its blessing on expansion memory for 3/50s. Most importanly would Sun refuse to service a machine if such a memory card was installed? -- Jeremy Cook Chr. Michelsen Institite, Bergen, Norway jeremy@kheops.cmi.no ------------------------------ Date: Thu, 3 Nov 88 09:21:35 CST From: kusalik%SKLPL.USask.CA@cornellc.ccs.cornell.edu Subject: Re: helios memory for SUN-3/50 >Each card is a daughter board that sits on top of the 3/50 mother board. >The connecting cable must be SOLDERED (!) to the mother board. What's effect will this have on SUN maintenance of that 3/50? ------------------------------ Date: Fri, Nov 88 12:02:14 EST From: John.Myers@pie8.pie.cs.cmu.edu Subject: Re: Postscript Interpreter (1) (I would like to remind the moderator that ghostscript is not in the public domain.) [[ Yes, okay, it's Gnu-ware. So shoot me. --wnl ]] Crispin Goswell wrote a PostScript interpreter which is freely distributable (though not public domain) It is available from the various comp.sources.unix archives. The ones I know of are: j.cc.purdue.edu, news/comp/sources/unix/volume12, v12i050.Z through v12i067.Z uunet.uu.net, comp.sources.unix/volume12/postscript, part01.Z through part18.Z There are a number of bug fixes and improvements available via anonymous ftp to z.andrew.cmu.edu (128.2.30.8) in the subdirectory psdiffs. The patches should be applied to an unpacked virgin distribution. READ THE INSTRUCTIONS AT THE BEGINNING OF EACH FILE BEFORE APPLYING THEM WITH PATCH. _.John G. Myers Internet: John.Myers@cs.cmu.edu LoseNet: ...!seismo!inhp4!wiscvm.wisc.edu!give!up [[ Ted Nolan sent me the README from the package. Here it is: This is a virtually complete implementation of PostScript. Just type make and it should produce a version for suns. Since the only complete driver is for suns and since some driver writing will be necessary on any other machine, I have not packaged everything up in an idiot-proof box. I am not aware of any Gotchas in building this source. The X driver used to work on X10 - It will need work for X11. The pixrect driver is very new, and should be regarded as didactic, as it scribbles straight onto the display. Persuading suntools to look after the bits would have taken too long. Writing your own driver involves merely hacking pixrect.c until it will drive your display. There are only 11 routines to write, and they're all quite short.... --wnl ]] ------------------------------ Date: Sat, 5 Nov 88 02:07:50 EST From: Mark Moraes <moraes@csri.toronto.edu> Subject: Re: Postscript Interpreter (2) In Sun-Spots Digest, v7n3, you ask: > Does anyone know if there is a public domain Postscript Interpreter > running on SUN with full source? > > [[ There's "ghostscript" from the gnu project. Sorry, I don't know any > more about it. --wnl ]] Ghostscript doesn't run too well on our Suns. It dumped core on Sun4s, and sort of worked on Sun3s, dumping core after a while. It only runs reasonably for us on uVaxen. Apparently, it only runs well on IBM-PCs with EGA. But it is still young... An alternative is the PostScript interpreter posted by Crispin Goswell to comp.sources.unix a while back - check your nearest archive site. You can also get it from expo.lcs.mit.edu by anonymous ftp - See oldcontrib/xps.tar.Z. That works somewhat better on Suns. Compile it with -Dfloat=double if you want it to work on Sun4s. It has drivers for the X Windows system (a fairly good X11 driver, a not-very-complete X10 driver) and bare pixrect. It saves some paper for debugging PostScript code, and is interactive, so you can do fun things like print your stack without needing to go look in obscure log files for the output. Neither of these will preview the output from dvi2ps, or psroff, and both are very slow. (Not that we really want them to - the dvi/ditroff previewers are much better, and faster) xps has a slightly better collection of fonts - it uses the Hershey fonts for imaging. Still, it is a far cry from a proper PostScript interpreter. ------------------------------ Date: Thu, 3 Nov 88 15:12:40 GMT From: Nick Holloway <alfie%CS.WARWICK.AC.UK@cunyvm.cuny.edu> Subject: Patch to fstat.c I hope this subject has not been mentioned before, which is possible considering age of fstat. I recently pulled fstat.c out of the rice archives, and compiled it up. I did, however, find a couple of problems, but have managed to sort them out for my needs. I did try to send mail to the author, but it didn't make it, so I am sending the patches here, to sun-spots. The first problem was the test for whether the filesystem was nfs didn't work for me. The test was that the major part of device number was 0xFF. For me it is 0x82. Can anybody supply a better test? It is now possible to specify major device number to look for with -DNFS_DEV=0x82 (or whatever). (Maybe the change is a side effect of 4.0?) I also changed the printing of the permissions. If a file has mode 02700 (g+s,u+rwx), then it should be shown as rwx--S---, but fstat used to show it as rwx------. (see manual entry for ls). From the entry for ls, I would expect mode 01000 to be shown as --------T, and mode 01001 as --------t, and this is the way I have modified fstat to work. However, ls in practice seems to fail. The final change I made was to teach it about fifo pipes. try $ mknod fred p; fstat fred Bye for now, Nick Holloway JANET : alfie@uk.ac.warwick.cs 16 Queens Rd UUCP : ..!mcvax!ukc!warwick!alfie Hertford BITNET: alfie%uk.ac.warwick.cs@ukacrl Herts ARPA : alfie%cs.warwick.ac.uk@cunyvm.cuny.edu England Here are the diffs for fstat (in a shar of course :-). -----------------------C U T H E R E------------------------------- #! /bin/sh # This is a shell archive, meaning: # 1. Remove everything above the #! /bin/sh line. # 2. Save the resulting text in a file. # 3. Execute the file with /bin/sh (not csh) to create the files: # fstat.c.diff # This archive created: Thu Nov 3 15:01:09 1988 export PATH; PATH=/bin:$PATH if test -f 'fstat.c.diff' then echo shar: will not over-write existing file "'fstat.c.diff'" else cat << \SHAR_EOF > 'fstat.c.diff' *** fstat.c Thu Nov 3 10:59:59 1988 --- fstat.c.old Thu Nov 3 15:01:38 1988 *************** *** 36,49 **** ** Daniel Trinkle - really want to use major(st_dev), this seems ** to be 0xff on all systems we have. minor(st_dev) is NFS type ** mntent index. - ** - ** Modified - November 3, 1988 - ** Nicholas Holloway - found that major device may vary - made it - ** possible to specify as -DNFS_DEV=?? [ I found it to be 0x82 -- - ** really need a 'isnfs(major(pst->st_dev))' macro/fn ]. - ** - Also added "fifo pipe" to mpfmtsb[]. - ** - Modified printing of mode to show set[ug]id bits correctly. - ** - To reverse these changes compile with -DORIG ******/ #include <stdio.h> --- 36,41 ---- *************** *** 59,69 **** /* file type messages */ char *mpfmtsb[] = $ - #ifdef ORIG "unknown", - #else ORIG - "fifo pipe", - #endif ORIG "character special", "directory", "block special", --- 51,57 ---- *************** *** 75,89 **** #define FMTTOSB(fmt) mpfmtsb[(fmt) >> 13] /* BEWARE hwc! */ #define CCHMAX 128 #ifdef NFS - #ifdef ORIG #define NFS_DEV 0xFF /* may be bogus, but it works for Mt. Xinu, Sun and Sequent */ - #else ORIG - #ifndef NFS_DEV - #define NFS_DEV 0xFF /* may be bogus, but it works for Mt. Xinu, Sun and - Sequent */ - #endif !NFS_DEV - #endif ORIG #endif NFS time_t tNow; --- 63,70 ---- *************** *** 159,165 **** pue = getpwuid((int)pst->st_uid); pge = getgrgid((int)pst->st_gid); printf("Mode: %c%c%c%c%c%c%c%c%c %9s(%3d) %9s(%3d)\n", - #ifdef ORIG mode & 0400 ? 'r' : '-', mode & 0200 ? 'w' : '-', mode & 0100 ? (mode & S_ISUID ? 's' : 'x') : '-', mode & 0040 ? 'r' : '-', mode & 0020 ? 'w' : '-', --- 140,145 ---- *************** *** 166,185 **** mode & 0010 ? (mode & S_ISGID ? 's' : 'x') : '-', mode & 0004 ? 'r' : '-', mode & 0002 ? 'w' : '-', mode & 0001 ? (mode & S_ISVTX ? 't' : 'x') : '-', - #else ORIG - mode & S_IREAD ? 'r' : '-', /* user */ - mode & S_IWRITE ? 'w' : '-', - mode & S_IEXEC ? mode & S_ISUID ? 's' : 'x' - : mode & S_ISUID ? 'S' : '-', - mode & S_IREAD >> 3 ? 'r' : '-', /* group */ - mode & S_IWRITE >> 3 ? 'w' : '-', - mode & S_IEXEC >> 3 ? mode & S_ISGID ? 's' : 'x' - : mode & S_ISGID ? 'S' : '-', - mode & S_IREAD >> 6 ? 'r' : '-', /* other */ - mode & S_IWRITE >> 6 ? 'w' : '-', - mode & S_IEXEC >> 6 ? mode & S_ISVTX ? 't' : 'x' - : mode & S_ISVTX ? 'T' : '-', - #endif ORIG pue ? pue->pw_name : "<unknown>", pst->st_uid, pge ? pge->gr_name : "<unknown>", pst->st_gid); --- 146,151 ---- SHAR_EOF fi # end of overwriting check # End of shell archive exit 0 ------------------------------ Date: Thu, 3 Nov 88 14:43:40 EST From: kc@rna.rockefeller.edu (Kaare Christian) Subject Problematic 3/50 This is just a brief note describing a long-term problem that I've had with a 3/50, and the solution. Over the past 18 months I have had repeated problems with one of our 3/50s. It has failed in numerous ways, and at various times the monitor has been replaced (3x or 4x), the main board (2x or 3x), and the power supply (2x). All "repairs" were by Sun maint. Recently it failed again, so I decided to play around. I discovered that the main (only) board fit very tight, and that when in the machine its center was bowed down. I surmised that it was touching the bottom of the chassis, inserted some cardboard to act as an insulator, and since then I haven't had a problem. Kaare Christian kc@rna.rockefeller.edu ------------------------------ Date: Thu, 3 Nov 88 09:53:26 PST From: Al Kossow <aek@apple.com> Subject: sgi <-> sun 1/4" tape transfers SGI tapes appear to be byte-swapped. Just do a "dd" with conv=swab on the input ------------------------------ Date: 12 Sep 88 22:37:15 GMT From: tekbspa!tss!joe@uunet.uu.net (Joe Angelo) Subject: Appendix Examples from "Writing Device Drivers" We made a service call requesting the electronic copy of the appendix/examples in Sun's "Writing Device Drivers" section of the SUNOS 3.X documentation distribution. The following shar file is what Sun mailed me. Please save in the comp.sys.sun archive *only* if you feel that Sun won't get upset... [[ The shar file have been placed in the archives under "sun-source" as "drivers.sun.shar". It is 32950 bytes long. It can be retrieved via anonymous FTP from the host "titan.rice.edu" or via the archive server. For more information about the archive server, send a mail message containing the word "help" to the address "archive-server@rice.edu". --wnl ]] Thanx much, & Thank Sun, Joe Angelo -- Senior Systems Engineer/Systems Manager at Teknekron Software Systems, Palo Alto 415-325-1025 **Temporarily in Boston 617-570-6953,570-2614 joe@tss.com - uunet!tekbspa!joe - tekbspa!joe@uunet.uu.net ------------------------------ Date: Thu, 3 Nov 88 17:58:20 PST From: susan@sun.com (Susan Morin) Subject: Sun Education January - June 1989 Schedule The course catalog and schedule for January through June 1989 has been completed and is being printed. Hard copy will be mailed to all customers and Sun sales offices within the next 6 weeks. We do have the ASCII version of the catalog available now. To assist you in scheduling your training for the next six months, we can e-mail you the online version immediately. If you would like to receive a copy, please send your request to customer-training@sun or sun\!customer-training. Here's what's new: -Expanded Calif. and Mass. training facilities -Clearer, easier to use course descriptions -Video-based training courses -Improved programmer's curriculum If you have any questions, please feel free to contact me. Sun Educational Services looks forward to continuing to serve you in 1989! Thank you. Susan Morin Course Development Manager Educational Services Sun Microsystems, Inc. (408) 276-3150 ------------------------------ Date: Fri, 4 Nov 88 17:42:52 PST From: shipley%WEB.Berkeley.EDU@lilac.berkeley.edu Subject: Lurnix announces two Sun related classes Lurnix, the Unix education company, announces two new classes: Sun Installation and Sun System Administration. Both classes are offered on-customer-site only, and can be taken together or separately. The Sun Installation class lasts 2 or more days, depending upon the number of workstations, students, and software packages involved. The client should provide the workstations to be installed; they should be unpacked but need not be connected or unconfigured. The course teaches up to 6 students how to install a Sun workstation. Installation of the workstation on a local area network is included if applicable. Significantly, at the conclusion of this course, the customer will have up to 6 Sun workstations fully installed and running. Student materials for up to six students are provided by Lurnix, the same company that created "Unix for People". The Sun System Administration course lasts three days and is designed to be taken in conjunction with the Sun Installation course or by itself. It is limited to 10 students and teaches the students how to back-up, maintain and restore the file system, add and remove users and peripherals, use system administration utilities, including NFS and Yellow Pages, and troubleshoot problems. Lurnix is a firm based in Berkeley, California that has been providing education and training for the Unix marketplace for the last five years. It was founded by the same people that wrote "Unix for People." Lurnix offers a wide spectrum of courses in both their Berkeley training center and on-site for larger customers. Other services they offer include curriculum and course design as well as documentation and project management. For more information please call Scott Anderson, Director of Marketing, at 800-433-9337. email: scott@lurnix.lurnix.com uunet!lurnix!scott lurnix!scott@uunet.uu.net ------------------------------ End of SUN-Spots Digest ***********************