Sun-Spots-Request@RICE.EDU (William LeFebvre) (08/05/88)
SUN-SPOTS DIGEST Thursday, 5 August 1988 Volume 6 : Issue 167 Today's Topics: Re: Multiple Swapping: three partitions? Re: eyestrain using domain name resolver Whither 68030? What (Sun) OS are you running? Linked or LCD Display of Sun 3/150's? make has a problem with "make makefile" mail question: Sun mail aliases? sendmail & PMDF problems Problem with fcntl 4GL on Sun? help with PostScript 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: Mon, 25 Jul 88 11:21:10 CDT From: "Matt Crawford" <matt@oddjob.uchicago.edu> Subject: Re: Multiple Swapping: three partitions? Reference: v6n148 Phone: +1 312 702 8207 This host, oddjob, has 96MB of swap space on three drives. The lines from fstab are: /dev/xy0b /xxx swap xx 0 0 /dev/xy1b /yyy swap xx 0 0 /dev/xy2b /zzz swap xx 0 0 And the configuration file includes: config vmunix root on xy0a swap on xy0b and xy1b and xy2b config vmunixtwo root on xy2a swap on xy2b and xy1b and xy0b (I use xy2a as a backup root partition updated nightly with rdist. Write for details.) Matt Crawford ------------------------------ Date: Tue, 26 Jul 88 09:54:42 BST From: Steve Platt <steve%mrc-applied-psychology.cambridge.ac.uk@nss.cs.ucl.ac.uk> Subject: Re: eyestrain I have been passing the various comments in recent sun-spots concerning eyestrain and workstations onto our group who work in the area; and the result is the following note(s), which I am passing on for one of them. Any errors are mine, the wisdom is his! Steve Platt, digital "minder" MRC A.P.U., Cambridge England. [[ I have only included excerpts in this digest. Anyone interested in the entire text of the original message can retrieve it from the archives. It has been stored under "sun-spots" as "eyestrain" and it is 8617 bytes in length. --wnl ]] Eye-strain Given the increasing collection of messages concerning eye-strain, I thought the readership might be interested in the following publications from our laboratory. [[ List deleted, although I highly recommend it to those who are interested. --wnl ]] Two of the above studies implicate the role of intermittent light in the causation of eye-strain, headaches and disturbances of ocular motor control, the third implicates the spatial characteristics of text. What can be done ? I would suggest the following: 1. Avoid working under fluorescent lighting. Easier said than done. One possibility is to ... use tungsten-halogen... 2. Use displays with a high refresh rate (the higher the better) and long persistence phosphor (eg P39), where possible. 3. Minimize the area of the retina exposed to intermittent light. This would indicate that it might be preferable, other things being equal, to use a black background with white letters... 4. Increase text size, as suggested by Tony Movshon... 5. Cover the screen with a sheet of dark plastic or fine nylon mesh... 6. Try tinted glasses. Many people find tinted glasses helpful, particularly those that absorb the blue end of the spectrum. One possible reason might be that most fluorescent lighting has a greater modulation at this end of the spectrum. Some people prefer other tints, and we do not know why. One common preference is for those tints that absorb green. You can obtain glasses that absorb the blue end of the spectrum from Bolle, 3890 Elm St, Denver, Colorado 80207: they are marketed as COMPUT-IREX. Or you can obtain tints cheaply from theatre lighting stores, and I suggest you look through a sampler, trying Lee filter number 109 to start with.... Perhaps it might be worth correcting the following statements from previous correspondence. The relaxed focussing distance is not 20 feet. The frequency at which epileptic seizures are most readily provoked by intermittent light is 16-20Hz, although patients may be sensitive to isolated flashes and to frequencies higher than 60Hz. The sensitive range is not related to the alpha rhythm, and varies from one individual to another slightly. Arnold Wilkins Research Psychologist Applied Psychology Unit 15 Chaucer Road Cambridge England CB2 2EF Tel 0 223 355294 ------------------------------ Date: Mon, 25 Jul 88 11:30:16 PDT From: jqj@hogg.cc.uoregon.edu Subject: using domain name resolver In v6n149, the editor notes: > [[ There is a way to configure the yellow pages hostname server so that it > will ask the name resolver if it cannot find the name in its own database. > But I do not know offhand how to do that. -- wnl ]] How to do it depends on whether you are running 3.x or 4.0. In either case, you should ONLY do it if you are connected to the Internet. You will need the ddn patch tape (available for anonymous ftp from various places, e.g. hogg.cc.uoregon.edu [128.223.20.5] pub/SunOS4.0_ddn.tar.Z). Then the procedure for running the resolver (assuming you already have a domain name server) is: 0/ get the patch tape, and read the README file on it. 1/ make sure you have the yellow pages working. 2/ on your YP server(s) create /etc/resolv.conf pointing to the domain name server you want to use. Note that if you want to use a Sun as the domain name server that's step 1a. Test resolv.conf by running nslookup (from the patch tape) to make sure that you can look up names not in your yp hosts.byname map. 3/ For 3.x: run ypserv from the patch tape, but with the flag "-i". For 4.0: build the hosts.* YP maps on the master server using the "-b" flag to the makedbm executions that build hosts.byname and hosts.byaddr (I changed my /var/yp/Makefile). This adds an entry, YP_INTERDOMAIN, in the map (visible only using makedbm -u, not using ypcat). Run ypserv.shared from the patch tape instead of ypserv. Note that Charles Hedrick at Rutgers has published a rather messy technique for building a version of SunOS 4.0 libc.so that includes resolver routines. Using his instructions, tcp/ip applications such as rlogin directly access the resolver even though they haven't been linked with -lresolv. This is clearly the right way to go -- Sun should NOT use the YP for host name translations in an Internet environment! ------------------------------ Date: Mon, 25 Jul 88 23:15:56 PDT From: mangler@csvax.caltech.edu (Don Speck) Subject: Whither 68030? Both HP and Apollo have recently announced 7-MIP 33 MHz 68030 workstations. I've heard not a single rumor about any possible 68030 Sun workstation. Has Sun decided to skip the 68030 to protect their current Sun-4 line? Or postponing it until they have faster Sun-4's, e.g. based on the just-announced Cypress 33 MHz SPARC chipset? Or saving it up for SigGraph, a popular time to announce new workstations? Don Speck speck@vlsi.caltech.edu ------------------------------ Date: 25 Jul 88 20:28:15 GMT From: randy@umn-cs.cs.umn.edu (Randy Orrison) Subject: What (Sun) OS are you running? Just a simple little question for you Sun users (I'm mainly interested in sites): What version of SunOS are you using? 3.x, 4.0, or what? Thanks! -randy [[ I imagine that quite a few are currently in the process of switching over to 4.0, making it hard to answer your question. --wnl ]] ------------------------------ Date: Mon, 25 Jul 88 15:40:59 PDT From: howardg@sun.com (Howard Greenfield) Subject: Linked or LCD Display of Sun 3/150's? Does anyone know where to purchase or acquire either: 1) a plasma display that will show bit-mapped screen displays (Interleaf) through an over-head projector for Sun 3/150's or 2) technology that will send the display from one screen (eg instructor's monitor) to 5 or 10 other screens (eg students' monitors). I know that choice 2 can be done because I've seen in kludged together at another training facility. Also, there is an LCD for overhead projectors out for the Mac and ASCII from PC's. Any leads on Sun-3 display products or solutions would be appreciated Sincerely, --Howard G. ------------------------------ Date: Mon, 25 Jul 88 12:18:33 BST From: Dr R M Damerell (RHBNC) <damerell@nss.cs.ucl.ac.uk> Subject: make has a problem with "make makefile" make does something I do not understand: This is on a SUN 3/60, SUN-OS 3.5, with sunpro installed. In an empty directory, create a file called Makefile containing this text: Makefile: Imakefile\ echo "HELLO hairyface" Then do as follows: mathsun1: touch Imakefile mathsun1: ls -l -r--r--r-- 1 sysdiag 1071 Jul 14 11:28 Imakefile -rw-r--r-- 1 sysdiag 5745 Jul 14 10:52 Makefile mathsun1: make Makefile `Makefile' is up to date. mathsun1: ls -l *file -r--r--r-- 1 sysdiag 1071 Jul 14 11:28 Imakefile -rw-r--r-- 1 sysdiag 5745 Jul 14 10:52 Makefile Please can anybody explain this? Let me anticipate the natural question " why do you want to make Makefile?" I have been trying for some weeks to get X windows running on our SUNs -- and not winning. This is an extremely complicated bit of software, and I assume that the authors found that their Makefiles were becoming unbearably long. For whatever reason, they provided a supplementary program called imake which generates Makefiles . On the version of X that I got from a colleague, I found that several of these files refer to Vax macros; so I did make Makefiles in the top level directory. The newly-made Makefiles still refer to Vax macros; so eventually I found the behaviour described above. Mark ------------------------------ Date: Mon, 25 Jul 88 18:52:10 PDT From: acad!robert@uunet.uu.net (Robert Wenig ext 609) Subject: mail question: Sun mail aliases? I notice that when I get a mail reply from somebody inside of sun, very often the return path is sun!personame, and all of the internal machine name business is stripped out. I would like to know how to set this up on my own machine. Robert Wenig Autodesk {sun,decwrl,uunet}!acad!robert ------------------------------ Date: Mon, 25 Jul 88 17:01 EDT From: DAVIS@blue.sdr.slb.com Subject: sendmail & PMDF problems This is entirely the wrong forum for this, but I'm not gatewayed to comp.mail.misc - maybe someone could forward it. Anyway - I'm trying to get sendmail on our Suns to communicate with a PMDF SMTP server on our VAX Cluster, in particular, to allow "reply" to function properly on messages that originated in a VAX network using PSImail. The heart of thr problem seems to be this - somewhere between sendmail's address rewrite and the actual connection to the PMDF server, some *crucial* double-quotes are dropped. The addresses have the following form: "PSI%SDRRTR::SCRVX1::DAVIS"@scr-gateway and without those quotes, PMDF chokes and spits it back. I know this from long and extensive chat sessions with it using mconnect (8). Now, I checked sendmail using the flags -bt (to check on the rewrite), and -d13.1 (to look at the user name used by deliver.c, and both look OK in that the double quotes remain. SO where are they being dropped, and how do I stop it ? We're using SunOS3.2 (4.0 anyday now...) and the TCP/IP connection is via Excelan. Any help you can give would be much appreciated, but please reply direct - I've had a lot of problems getting Sun-Spots over the last few months. Paul Davis Schlumberger Cambridge Research Cambridge, UK davis%blue@sdr.slb.com +44 223 325282 ------------------------------ Date: Mon, 25 Jul 88 11:31:57 +0200 From: mcvax!cgch!wbwa@uunet.uu.net (Bernhard Wagner) Subject: Problem with fcntl Here is a problem I run into when using "fcntl" in order to lock a file via NFS. Given a program (you find the complete program in the appendix), which looks like: fopen(fn, "r+") loop_forever Lock_file() Write_file() Unlock_file() Short_delay() end_loop The Lock_file procedure is implemented as follows: static void Lock_file() { struct flock fl; fl.l_type = F_WRLCK; /* exclusive lock */ fl.l_whence = 0; fl.l_start = 0; /* lock the whole file */ fl.l_len = 0; /* to EOF */ if (fcntl(fileno(fp), F_SETLKW, (int)&fl) == -1) { perror("locking"); (void)fflush(stderr); } if (fseek(fp, 0L, 2) == -1) { /* seek to end of the file */ perror("fseek"); (void)fflush(stderr); } } /* end Lock_file */ If I have only one process executing this program, everything works well. If there are several processes however, (I tried up to seven) running concurrently and if I try to kill them by kill -9 <pid1> <pid2> ... after a some time, usually some of the processes stay alive. They can be removed from the system by booting only. On the other hand, when such zombies hang around, every process which tries to lock any file hangs in the fcntl system call and and cannot be killed too. Does anybody has comments on this problem? My opinion is that the protocol which is used for networkwide file-locking is implemented too simple. Bernhard Wagner Ciba-Geigy AG Scientific Computing Center Postfach CH-4002 Basel Switzerland Appendix: /*******/ #include <errno.h> #include <fcntl.h> #include <stdio.h> extern void perror(); #define fn "mist4" #define content "blabla" FILE *fp; static void Lock_file() { struct flock fl; fl.l_type = F_WRLCK; /* exclusive lock */ fl.l_whence = 0; fl.l_start = 0; /* lock the whole file */ fl.l_len = 0; /* to EOF */ if (fcntl(fileno(fp), F_SETLKW, (int)&fl) == -1) { perror("locking"); (void)fflush(stderr); } if (fseek(fp, 0L, 2) == -1) { /* seek to the end */ perror("fseek"); (void)fflush(stderr); } } /* end Lock_file */ static void Unlock_file() { struct flock fl; fl.l_type = F_UNLCK; fl.l_whence = 0; fl.l_start = 0; /* unlock the whole file */ fl.l_len = 0; /* to EOF */ (void)fcntl(fileno(fp), F_SETLKW, (int)&fl); } /* end Unlock_file */ static void Write_file() { rewind(fp); (void)fwrite(content, sizeof(content), 1, fp); (void)fflush(fp); } /* end Write_file */ main() { int i = 0; int j, k; (void)fprintf(stdout, "pid: %d\n", getpid()); fp = fopen(fn, "r+"); if (fp != (FILE *)NULL) for (;;) { Lock_file(); Write_file(); Unlock_file(); (void)fprintf(stdout, "%d", i); ++i; k = 0; for (j=0; j < 10000; ++j) k = k + j; } /* for */ else /* file does not exist */ (void)fprintf(stdout, "file %s did not exist\n", fn); return 0; } /* end main */ ------------------------------ Date: Tue, 26 Jul 88 11:12:41 BST From: mcvax!ritd.co.uk!mr@uunet.uu.net Subject: 4GL on Sun? I have a potential need to understand and run 4GL packages on Sun hardware. Could *any* of you who have information, suggestions, experience or war stories please post me. Even a good working definition of what a 4GL is/should be would be a start. I will re-post answers if there is interest. TIA, Martin Reed, Racal Imaging Systems Ltd uucp: mr@ritd.co.uk,{mcvax!ukc!ritd,sun!sunuk!brains}!mr Global String: +44 252 622144 Paper: 309 Fleet Road, Fleet, Hants, England, GU13 8BU ------------------------------ Date: Mon, 25 Jul 88 12:50:03 +0000 From: jock%CCL.UMIST.AC.UK@cunyvm.cuny.edu Subject: help with PostScript The Chinese Information Centre Cooperative in Manchester, UK, is currently developing a set of PostScript Chinese character fonts. They are interested to hear from other concerns involved in that area, with a view to exchanging information. Unfortunately they do not have access to e-mail, so please reply to: Simon Jones Chinese Information Centre 1st Floor 16 Nicholas Street Manchester UK M1 4EJ ------------------------------ End of SUN-Spots Digest ***********************