Sun-Spots-Request@RICE.EDU (William LeFebvre) (06/02/88)
SUN-SPOTS DIGEST Wednesday, 1 June 1988 Volume 6 : Issue 100 Today's Topics: In celebration of this year's 100th issue Re: Suns lose track of console (3) Re: Problem with multiple groups Re: Csh bug (2) silo overflow Phillips Monitors and Sun Workstations Phillips Monitors-JUST SAY NO!! if( vs. if ( in CSH persistent clock drift problem 68881 on a 3/50? 68020 upgrades for sun2s? 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: Wed, 1 Jun 88 14:57:14 CDT From: William LeFebvre <phil@Rice.edu> Subject: In celebration of this year's 100th issue If someone had told me late last year that I would produce well over 100 issues of the Sun-Spots digest in 1988, I would have either (1) laughed in their face or (2) politely declined the job. I am absolutely amazed that the readership has produced so much input to this forum. We aren't even half way through the year and we are already reading issue number 100. I must say that this digest has proven to be more work than I anticipated, but at the same time I am very pleased with what I am seeing. I think that the quality of messages is still high overall, and that the exchange of information brought about by Sun-Spots is phenomenal. With just a little bit of pride I'll step out on a limb and say that I think this list has one of the highest signal to noise ratios of any newsgroup or mailing list, especially considering its volume. I want to take the time to thank everyone who has sent me words of encouragement. It means quite a bit to me to know that there are people who appreciate my efforts and are happy with what they see. I am also thankful that everyone has been so patient with me while I struggle to get rid of the backlog of messages and while Rice copes with its networking problems. Finally, I want to say that this list would not be nearly as successful as it is without the readers who take time to contribute good messages, whether they are bug reports, answers to simple or difficult questions, parts of a discussion, programs, or even icons. A big thank you to everyone who has contributed to this forum. Without all of you, we would have nothing to read! Now I get to find out how much of my homebrew digestification software breaks with three digit issue numbers! William LeFebvre Department of Computer Science Rice University <phil@Rice.edu> ------------------------------ Date: Mon, 23 May 88 15:29:45 mdt From: era@scdpyr.ucar.edu (Ed Arnold) Subject: Re: Suns lose track of the console (1) We've noticed the same thing under 3.3, so it's not a bug peculiar to 3.5. We also have a Wyse terminal as a console, attached to a 3/280. The problem is *not* in the terminal; the console interface locks up. It's not a simple matter of someone having typed ctl-s on the console. ------------------------------ Date: Mon, 23 May 88 10:24:36 EDT From: mcb%jones@research.att.com Subject: Re: Suns lose track of console (2) My servers (OS3.2) also occasionally have console lock-up, which in this case is caused by a hung `getty' process. It is cured by logging in as root from another machine and doing: ps aux | grep getty to find a line like root 123 ... -2 (getty) and `kill -9' it. These sometimes appear shortly after reboot, to judge from the process numbers. BTW, this is no permanent cure. After a day or so, it often hangs again. No great solution, but it avoids a reboot. If someone happens to know the cause &/or cure for this problem with tty consoles, please post. Mark Beutnagel AT&T Bell Labs mcb@research.att.com ------------------------------ Date: Wed, 25 May 88 20:42:40 EDT From: Stephen J. Roznowski <sjr@mimsy.umd.edu> Subject: Re: Suns lose track of console (3) I had this same problem. The fix was to remove all the devices associated with the display hardware. [Since they are not being used.] Remove /dev/mouse /dev/kbd /dev/fb on the server. Since I have removed them I have not had any problems [In over a month] The reasoning I was given (which I don't really believe) was that a user logging into the server and trying to start up suntools would confuse the server as to where the console is. Stephen P.S. I'm assuming that your console is attached either to ttya or ttyb. ------------------------------ Date: Mon, 23 May 88 17:01:53 EDT From: libes@cme-durer.arpa (Don Libes) Subject: Re: Problem with multiple groups Summary: User logs in on yp server and is in 4 groups, but only 2 groups on client. What's going on is that the client is getting the yp group data, but its wrong. For example, if a line is missing a colon, all the groups after that are ignored. What is confusing is that yp reorders the data, so that you have to look for the error between the last group you make it into and the first group you don't in "ypcat group", not "cat /etc/group". Don Libes cme-durer.arpa ...!uunet!cme-durer!libes ------------------------------ Date: 23 May 88 15:41:47 PDT (Mon) From: mills%tessi@tektronix.tek.com Subject: Re: Csh bug (1) >psune% echo !#:h >Segmentation fault (core dumped) That bug was in the early releases of SUN's csh in 1983/84. I assume it was the classic dereferenced NULL pointer bug you see porting from VAX-based Berkeley Unix. I may even have reported it to SUN. I thought it had been fixed sometime between then and now. Has it been with us all along? -- Charlie Mills Test Systems Strategies, Inc. ...!ucbvax!tektronix!tessi!mills [[ I just checked and it still exists under version 3.2 and 3.5. Note that the "echo" is not necessary, only the argument "!#:h". It's a great way to log off (provided you have set "limit core 0")! --wnl ]] ------------------------------ Date: Tue, 24 May 88 09:53:07 N From: Sven-Ove Westberg <sow@cad.luth.se> Subject: Re: Csh bug (2) A typical nil pointer!! The monster is still lurking around :-) (sigh) By the way, is the nil pointer in "sed -e" still alive?? This is the fix. Check for nil pointer in any() in sh.misc.c. any(c, s) register int c; register char *s; { if(!s) return(0); /* Check for nil pointer */ while (*s) if (*s++ == c) return(1); return(0); } ------------------------------ Date: Mon, 23 May 88 19:12:47 EDT From: smb@research.att.com Subject: silo overflow When a character arrives on a serial port (or any sort of port, actually, but the message usually refers to a serial port), it's put in the top of the ``input silo'', and an interrupt is generated. When the CPU fields the interrupt, it pulls the character out of the bottom of the silo. In the mean-time, though, more characters may have arrived; they're put in the top also, and the CPU can retrieve them all at the expense of one interrupt. Another way to look at the silo is that it's simply a multi-character buffer, though just how large depends on the implementation. Silo overflow occurs when characters are arriving faster than the CPU can field them. It may be a genuine problem (if, say, you've got a rackful of Trailblazers all talking at 19.2Kb/sec at the same time); more often, though, it's a symptom of something else -- the CPU is too busy fielding higher-priority interrupts during a burst of heavy input. I've often seen silo overflows during panics, when the kernel is syncing the disks with interrupts masked; the silo overflow messages are gratuitous, and should be ignored. Do not confuse this condition with a user process running too slowly; the characters are pulled out of the silo immediately, and either queued for the process or discarded if the input queue is too long. ------------------------------ Date: Mon, 23 May 88 23:04:16 EDT From: hull@cs.buffalo.edu (Jon Hull) Subject: Phillips Monitors and Sun Workstations We are upgrading a Sun-3/260M to a Sun-3/260G by adding a Sun CG3 color board as well as a Phillips 2064M monitor. Before going with this monitor, I was wondering if anyone has experience with it or with other Phillips monitors when used as grey scale workstation monitors. How well can you discern differences in grey level? Is the resolution adequate for good text display? Any other comments you might have about this or other Phillips monitors would be greatly appreciated. Thanks in advance, Jon Hull hull@cs.buffalo.edu ..!{ames,boulder,decvax,rutgers}!sunybcs!hull ------------------------------ Date: Tue, 24 May 88 08:23:10 PDT From: Steve Blair <spar!ascway!scb@decwrl.dec.com> Subject: Phillips Monitors-JUST SAY NO!! You may be surprised to find this out, but a fellow I work with has found a way to rid himself(and his group) of those @#$% Phillips Monitors that Sun so thoughtfully sells. We were on our 3/rd failure of the usual Phillips on the same machine. Expense of repair was driving this guy nuts. Since Sun has now decided that WE THE CUSTOMERS SHOULD PAY MORE FOR BAD CHOICES IN MONITORS (i,e,: 30 day color repairs now=$1600.00, monochrome=$1100.00). (old cost" " was=$800.00, monochrome=$600.00) I find it very hard to swallow that Sun is passing this additional cost of poor monitors onto the customers. I have told my local Sun representative that I was unwilling to get "ripped-off"(exact quote from conversation) for monitors that I didn't pick. Why, oh why Sun did you double the prices INSTEAD of getting NEW MONITOR Suppliers?? Oh, well, here's some sources of new monochrome and color monitors: **(They are 100% SUN compatable(and the users have not had ANY COMPLAINTS)*** Moniterm -------- P/N vy1962 19" Monochrome monitor cost=$1525.00 Sony Trinitron(16") ------------------- P/N gdm-1604-15 Color Monitor cost=$~2200.00 In each of these cases, I know that they cost a little more than sending the @#$% things back to Sun for repair. But please read a little further, there's some good reasons here: Moniterm Monitor is ECL/TTL. It has a jumper inside that changes the interface. This means that you can use it on Sun 3/50's & 3/160's. It has a 1yr. warranty vs. Sun's 90day warranty. It will pay for itself the second(and perhaps 1/st) time that you have to send one back to Sun. I got mine 4 days after ordering. I can now use the parts from the Phillips(groan) to repair the other Phillips until I get rid of them. You also get the luck of it being field repairable and parts are common for it. Sony Trinitron(16") is an INCREDIBLY sharp picture reminiscant of Sony's usual quality. If one measures screen real estate of the presentation surface, you'll discover what I did: The Sony's presentation surface is only 3/4" less than that of the 19" Hitachi/Ikegami that Sun sells. My user who has one has said the Sony will only leave when I pry his mutated dead fingers off of his mouse. Sure, it costs a little more, but what with the track record that Sony has in the world, cone can expect that the thing will be more reliable. I can get the Moniterm from almost any vendor. Sun uses these also. If you have access to a GOLD BOOK, they're in it. I can get the Sony from my local Electronics stores. I know not where from Sun is getting them. In either case, I highly urge other Sun administrators and users to consider these sources. It can be a good budget win for YOU the user!!!!! Steve Blair Schlumberger Technology Corporation Austin, Texas uucp: {backbone}!decwrl!spar!ascway!scb DISCLAIMER: The info given is from my research and doesn't reflect any policy OR opinions of Schlumberger or associated companies! ------------------------------ Date: Tue, 24 May 88 10:08:18 EDT From: csrobe@icase.arpa (Charles S. Roberson) Subject: if( vs. if ( in CSH Can someone please explain why the addition of a space (' ') between the "if" and the '(' in line 7 below causes the script to take two different execution paths? [41] work18: > cat csh1 [43] work18: > cat csh2 #!/bin/csh #!/bin/csh echo start echo start if(0) then if(0) then echo 1 echo 1 if (0) then if(0) then # line 7 echo 2 echo 2 endif endif echo 3 echo 3 else else echo 4 echo 4 endif endif echo end echo end [42] work18: > csh1 [44] work18: > csh2 start start 4 3 end end It might be easier for me to understand if the space only affected the second "if" statement but it affects the flow of the first as well! thanks, -c Chip Roberson ARPANET: csrobe@icase.arpa 1105 London Company Way BITNET: $csrobe@wmmvs.bitnet Williamsburg, VA 23185 UUCP: ...!uunet!pyrdc!gmu90x!wmcs!csrobe [[ This is a long-standing csh problem. The command line "parser" was pretty much thrown together and it causes quite a few little inconsistencies like this. In general it is a good idea to always put spaces around keywords in a cshell script. If you want more fun, check out the difference between "set foo=a", "set foo= a" and "set foo =a". --wnl ]] ------------------------------ Date: Tue, 24 May 88 00:00:43 PDT From: George Rinker <gcr@aristotle-gw.jpl.nasa.gov> Subject: persistent clock drift problem I have a persistent clock drift problem (on a Sun-3/50 with Release 3.4) even after the clock patch has been installed correctly (and verified by chuq@sun.com). Does anyone know the solution? George Rinker Jet Propulsion Laboratory gcr@aristotle.jpl.nasa.gov (818)354-6711 [[ Dying battery? Dying crystal? --wnl ]] ------------------------------ Date: Tue, 24 May 88 18:26:36 CDT From: wucs1!wuccrc!jst@uunet.uu.net (Jon Turner) Subject: 68881 on a 3/50? I need to add a 68881 math co-processor to my Sun 3/50. Sun has an upgrade available for $700 if you first ship your system back to them. On the other hand, the 68881 is available for about $200. Is there any reason that I can't simply buy the chip and plug it into the socket? Are there any jumpers I need to worry about? If you've done this before, please let me know. Thanks. Jon Turner jst@wuccrc.wustl.edu ------------------------------ Date: Wed, 25 May 88 09:22:18 EDT From: ted@braggvax.arpa Subject: 68020 upgrades for sun2s? We have a large collection of sun2 hardware and were wondering if there are any third party cpu boards that can upgrade a multibus sun2 to a 68020. I know this wouldn't give us sun3s, but it seems as though it would be worthwhile if the price were reasonable. I'd appreciate hearing about any such thing, and your experiences with it. Thanks, Ted Nolan ted@braggvax.arpa ------------------------------ End of SUN-Spots Digest ***********************