Received: from PACIFIC-CARRIER-ANNEX.MIT.EDU by po7.MIT.EDU (5.61/4.7) id AA21892; Fri, 1 Dec 95 19:43:46 EST Received: from SENATOR-BEDFELLOW.MIT.EDU by MIT.EDU with SMTP id AA26845; Fri, 1 Dec 95 17:24:39 EST Received: (from root@localhost) by senator-bedfellow.MIT.EDU (8.6.12/2.3JIK) id RAA02481 for Linux-Development-System-Dist@senator-bedfellow.mit.edu; Fri, 1 Dec 1995 17:14:33 -0500 Message-Id: <199512012214.RAA02481@senator-bedfellow.MIT.EDU> From: Digestifier To: Linux-Development-System@senator-bedfellow.MIT.EDU Reply-To: Linux-Development-System@senator-bedfellow.MIT.EDU Date: Fri, 1 Dec 95 17:14:28 EST Subject: Linux-Development-System Digest #71 Linux-Development-System Digest #71, Volume #2 Fri, 1 Dec 95 17:14:28 EST Contents: Re: Netscape 2.0b3 problems, anyone?/Poss Bug in XFree86 S3 server (Brendan Bartlett) Re: [HELP] "Virtual memory exceeded" compiling the kernel (Juergen Exner) Re: dump,restore,rdump (Lars Wirzenius) Re: How to make HD spin down? (Christoph Lameter) Re: Linux for DEC/MIPS (Ralf Baechle) Possible bug in slackware 3.0's new libraries (Sam Trenholme) Re: Socket listening (Andreas Hammarstedt) Re: dump,restore,rdump (Arie Maaskant) Re: Anyone working on a tape file system ? (BThompson1) Re: Netscape 2.0 breaks RedHat 2.0 Linux (Michael Nelson) Re: Has any one heard of VAX? (Jordi Guillaumes i Pons) Netscape/Mosaic for ELF (Darren J Moffat) Re: 8088 port (Dan Newcombe) [Q] sendmsg(), recvmsg() ? (Klaus-Peter Gulden) Re: Netscape 2.0b3 problems, anyone? (Clive Messer) ---------------------------------------------------------------------------- From: Brendan Bartlett Crossposted-To: comp.os.linux.misc,comp.os.linux.x Subject: Re: Netscape 2.0b3 problems, anyone?/Poss Bug in XFree86 S3 server Date: 29 Nov 1995 17:27:14 -0500 Reply-To: bartlett@io.org Alex.Butcher@bris.ac.uk (Alex Butcher) writes: >davis@evergreen.cc.usm.edu (Jonathan A. Davis) wrote: >>mbison (raiden@shellx.best.com) wrote: >>: >where it can be found. After you do that, you get the crashing >>: >behavior. >>: >>: If you are getting "Signal 7" or "Bus Error" when you >>: attempt to load a java applet, check your display setting. >>: It appears that Netscape doesn't like anything above >>: an 8 bit (256 color) display when it comes to Java. >>: >>: -MIC >>It "may" be related to the X server itself. I'm running 2.0B3 on my >>office workstation in Highcolor mode (16bpp S3) and have had no major >>problems with beta Java applets. With the exception of the odd applet >>that crashes Netscape entirely. Incidently, I've verified (so far) >>that the same applets crash Netscape on the following machines: >I'd actually second this. I've found something similar with my setup >(iDX4-100, 8MB RAM, 2-1-1-1 cache, 0 DRAM read ws, 2 DRAM write ws, >Kernel 1.2.13+KSwap+Voxware 3.x/Kernel 1.3.43, XFree86 3.1.1 S3 server >on Diamond Stealth 64 Vision864). To check it out, try this: [crashing X text deleted] Well, I have a pretty plain ELF Slackware linux distribution with XFree 3.1.2 for the Diamond Viper PCI. Netscape 2.0b3 will crash with a bus error if I try to run java apps with 16 bpp, but it runs fine at 8 bpp. The system is a Pentium 66, IDE disks and 40 megs of memory + 20 megs swap. -Brendan ------------------------------ From: jue@cs.tu-berlin.de (Juergen Exner) Subject: Re: [HELP] "Virtual memory exceeded" compiling the kernel Date: 01 Dec 1995 08:39:19 GMT In article <30BCB4CC.32B2@ipvvis.unipv.it> Fabrizio Capobianco writes: -> I've a Toshiba 400cs (notebook) with 8M RAM, 25 MEG swap. -> I'm trying to compile the kernel I've an error (after a while): -> "Virtual memory exceeded". -> I cannot compile the kernel because I've only 8M RAM ? 8 MB of TOTAL memory (RAM + swap) is definitely to low. Seems, as if your swap isn't working. What does "free" tells you? Greetings jue -- -- Juergen Exner: jue@cs.tu-berlin.de FAX: 030/314-73623 snail-mail: TU Berlin; Sekr. FR 5-13; Franklinstr. 28-29; 10587 Berlin =-=-=-=-= "I'm sorry Dave, but I can't do that" HAL 2000 =-=-=-=-=-= ------------------------------ From: wirzeniu@cc.Helsinki.FI (Lars Wirzenius) Subject: Re: dump,restore,rdump Date: 1 Dec 1995 11:54:24 +0200 koenig@tat.physik.uni-tuebingen.de (Harald Koenig): > any tar version changes the access date of any backuped file Not necessarily, since it is possible to remount the filesystem read-only while doing the backup, then remount it read-write again after the backup is done. If you're using dump, you need to unmount the filesystem anyway, so there is no problem doing it for tar. FWIW, I use dump myself, and use a shell script like the one below to do the backup (I'm not at my own system at the moment, so I can't give the exact script): umount /home dump 0uf /dev/ftape /home mount /home (Doing backups of live filesystems is usually not a good idea, since the file may be modified while it is being backed up. But in some environments it is not possible to have the system off-line even during backups.) -- Lars Wirzenius MIME and PGP mail welcome. Key: finger @kruuna, or check keyservers. Publib 0.5: ftp://ftp.cs.helsinki.fi/pub/Software/Local/Publib/ ------------------------------ From: clameter@fuller.edu (Christoph Lameter) Subject: Re: How to make HD spin down? Date: 25 Nov 1995 18:32:11 GMT : I got a tool from a friend, thats makes my hard disc spin down after a : given time (it's for D*S). So I watched out for a similar one for Linux, : cause i didn't know before, how silent my comp can be if it wants to. :-) : I have D*S and Linux on different drives, so I always do not need both of : them. : But I only was able to find hdparm. I know, this is a very powerful tool, : but I do not need that much power. I only need to spin down one of the : disks. Also I do not want to patch my kernel every time I install a new one. hdparm can spin down your harddrive. You can select which. You do not need a less powerful tool if the standard one does the job. : Does anybody know a tool that will work for me? Read the manpage for hdparm. -- ============================================================================= Christoph Lameter FTS Box 466, Pasadena, CA 91182 Internet Administrator Who is like Jesus... who is like God...? ------------------------------ From: ralf@informatik.uni-koblenz.de (Ralf Baechle) Crossposted-To: alt.os.linux,comp.os.linux.misc Subject: Re: Linux for DEC/MIPS Date: 1 Dec 1995 11:58:18 GMT Reply-To: linux@uni-koblenz.de In article <49h51q$ee8@news.belwue.de>, 681329@rhds02.rz.fht-esslingen.de (Nils Philippsen) writes: |> Does anyone have an idea how to install Linux on a DEC/MIPS based System |> (DEC 5000)? We've got the sources, the required Assemblercode, but we need |> a hint on compiling and installing it (how to make the system boot up with |> the new OS). |> |> Any idea would be appreciated. Paul Antoine is currently porting my R4400 Linux/MIPS port for R3000 based DECstations. He just started with his work so the code isn't usable yet. For more infortmation about Linux/MIPS reply to me or write to the Linux/MIPS mailing list at linux-mips@fnet.fr. Ralf ------------------------------ From: set@oryx.llnl.gov (Sam Trenholme) Subject: Possible bug in slackware 3.0's new libraries Date: 30 Nov 1995 12:17:39 GMT I have found a possible bug in the new libraries in Slackware 3.0. Namely, the following program, without WORK_AROUND defined, works improperly. It should exibit the same behavior irregardless of the definition of WORK_AROUND, and infact did act this way until I upgraded to Slackware 3.0: #include #include #include #include #include main() { FILE *tty; char ljb; if((tty = fopen(ctermid(0),"rb+")) == NULL) return -1; fprintf(tty,"Hit control-L to exit.\n"); while(ljb != EOF && ljb != '\f') { ljb = getc(tty); #ifndef WORK_AROUND fprintf(tty,"%d\n",ljb); #else printf("%d\n",ljb); #endif } close(tty); } -- - Any opinions expressed in this message are strictly my own opinions - Sam Trenholme - set@oryx.llnl.gov - http://ucsee.eecs.berkeley.edu/~set I use Linux at home. For more information, look at http://www.linux.org ------------------------------ From: Andreas Hammarstedt Subject: Re: Socket listening Date: Thu, 30 Nov 1995 14:36:38 +0100 Two processes communicating with sockets have one socket each and therefore have separate uid/gid, right? And if I understand it correctly, even if you change uid/gid on a process holding a socket, it doesn't change what that socket is connected to (as logically, it wouldn't be any of the peer process' business.) Hope you didn't mind me slipping in on the proverbial banana shell there. I might have missed a post or two on the subject. :) -- ___________________ ._____/Andreas Hammarstedt\________. |aha@solace.mh.se | haman@nts.mh.se| `. http://www.solace.mh.se/~aha .' `---------.aha@vbb.se.----------' `----------' ------------------------------ From: ari@fwi.uva.nl (Arie Maaskant) Subject: Re: dump,restore,rdump Date: 1 Dec 1995 15:50:01 +0100 wirzeniu@cc.Helsinki.FI (Lars Wirzenius) writes: >FWIW, I use dump myself, and use a shell script like the one below >to do the backup (I'm not at my own system at the moment, so I can't >give the exact script): > umount /home > dump 0uf /dev/ftape /home > mount /home >(Doing backups of live filesystems is usually not a good idea, since the file >may be modified while it is being backed up. But in some environments it >is not possible to have the system off-line even during backups.) In practice this is partly true. In my 13 year experience with dump and restore ( from Unix6, through BSD-unix until Solaris2.4 ) on systems with hundreds of ( potential users ) I am always dumping from livefilesystems. I have never had any bad practices because of doing it this way. If Linux is using the BSD variant of dumping and restoring ( I don't know, I have not used that version yet ) I can't hardly believe that dumping livefilesystems can give errors. We dump however in quiet periods ( 2 o'clock in the morning ) Arie -- Arie Maaskant University of Amsterdam Email: ari@fwi.uva.nl Faculty of Mathmatics, Computer Science, Physics and Astronomy Phone: +31 20 525 6438 Plantage Muidergracht 24 ------------------------------ From: bthompson1@eworld.com (BThompson1) Crossposted-To: comp.os.linux.development.apps,comp.os.linux.misc Subject: Re: Anyone working on a tape file system ? Date: 1 Dec 1995 10:06:04 -0800 Reply-To: bthompson1@eworld.com (BThompson1) ===== Begin Original Message ===== Subject: Re: Anyone working on a tape file system ? From: pjb27@thor.cam.ac.uk (Philip Blundell) Date: 25 Nov 1995 16:38:44 GMT Message-ID: <497gqk$mku@lyra.csx.cam.ac.uk> In article <19571@817158898>, Frank van de Pol wrote: >I disagree! A tape fs is *very* usefull. ok, you can't (practically) use it >like a disk, but for archiving files, it's very good. VMS (yes, the >operation system running on DEC VAX and alpha processors) offers such a tape >fs. > >A tape fs is much the same as a worm, or CD writable device: you can add >files to it, but not delete them, as this would mean relocating lots of >blocks, or creating fragmented files. A deleted file would be just *marked* Indeed. Even a read-only tape FS would be useful. I would quite like to be able to dump an entire filesystem to tape, then mount it as normal and read files off. Yes, it would be slow, but I wouldn't expect to use it every day and (IMHO) it would be far more convenient than tar for looking at archived files. It also has the advantage that permissions and ownership would be preserved, so I could let other people restore their own files from tape with relative impunity. ===== End Original Message ===== If we ignore the concerns about wear-and-tear to the tape drive mechanism, and we ignore concerns about wear-and-tear to the tape itself, and if we further ignore concerns about speed of access then I think I have a suggestion that could make everyone happy. Enter the TARFS, or perhaps even better the AFIOFS. Using the userfs kernel module surely someone could write a file system that reads a tape archive to find files. The filesystem would be read-only, and the directory stsructure could be read of at mount time to determine inode numbers and the like. Cheers, Bruce. -- Bruce Thompson Newton Systems Group, Developer Technical Support Group BThompson1 (eWorld) BThompson (AppleLink) bruce@newton.apple.com ------------------------------ From: nelson@seahunt.imat.com (Michael Nelson) Subject: Re: Netscape 2.0 breaks RedHat 2.0 Linux Date: Fri, 01 Dec 1995 09:53:38 PST Reply-To: nelson@seahunt.imat.com In article <30BC9351.71B@siac.com>, Michael Breuer writes: -> It would seem that netscape takes whatever memory it can get. If you -> run as root, this is ALL your memory. -> -> Solution: Don't run netscape as root. Jeeze... where do you people come UP with this amazing stuff? It uses lots of RAM, but it certainly does NOT use "ALL your memory" unless you don't have much free before you start it. I have 32MB, and here's what "top" shows when netscape 2.0b3 is running: 9:39am up 5:21, 2 users, load average: 0.26, 0.18, 0.33 36 processes: 35 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 1.3% user, 0.8% system, 2.1% nice, 97.9% idle Mem: 31368K av, 30424K used, 944K free, 15020K shrd, 18196K buff Swap: 33788K av, 1724K used, 32064K free 10747 nelson 30 15 5924 3548 2284 S N 0.0 11.3 0:02 netscape If it's using ALL my memory, why do I have 18196K available in buffers? Running it as root makes no difference (a slight difference because I started a shell as root so I could run netscape as root) either: 9:42am up 5:25, 2 users, load average: 0.45, 0.20, 0.29 36 processes: 35 sleeping, 1 running, 0 zombie, 0 stopped CPU states: 1.6% user, 2.5% system, 4.0% nice, 96.0% idle Mem: 31368K av, 30740K used, 628K free, 15436K shrd, 18376K buff Swap: 33788K av, 1688K used, 32100K free 10814 root 19 15 5908 3648 2356 S N 0.8 11.6 0:02 netscape Do _you_ have any numbers to back up your bullshit? Michael -- Michael Nelson nelson@seahunt.imat.com San Francisco, CA ------------------------------ From: Jordi Guillaumes i Pons Subject: Re: Has any one heard of VAX? Date: Thu, 30 Nov 1995 02:41:51 -0100 Grant Edwards wrote: > : - The CLI (DCL) lacks some features. I _really_ miss pipes and > : redirections. The /OUTPUT=XXX or DEFINE/USER SYS$INPUT... thing is a > : poor substitute for the real thing... And I think it _can_ be done using > : mailboxes. > > Yes, it can. The DECShell package included a v7 Bourne shell that > impelmented pipes using mailboxes. It also included a surprisingly complete > set of v7 utilities including awk, sed, nroff, lex, yacc, etc. It worked, > but it was slow. Creating processes (which happens a _lot_ if you are in the > habit of stringing together a handful of commands to do something useful) > involves a lot more overhead on VMS than on Unix. As I remember it, running > a program from DCL didn't create a new process (the program ran in the same > process as DCL somehow). Running a program from the Bourne shell has to > create a new process. Yup!. I guess it does a LIB$SPAWN for each image, and this only after creating the mailboxes, assigning them the correct logical names, setting up the redirections and so on... Excessive image activation is a resource-eater in VMS, but process creation is definitely a system killer... When the ol'swapper awakes, all VMS users cry... :-) Of course, as a VMS System Manager you can tailor and fine-tune the resource share of every user, to the level of how many subprocesses can he/she have running at once, how much core RAM can they spare and how much virtual memory can they allocate; this is one of the things I miss in linux (yes, I know Linus has always seen linux as a 'desktop' OS, so this fine-tuning is not a must-have). Are the rest of unixes the same, or do they have that sort of system management tools? -- /---------------------------------------------------------/ / Jordi Guillaumes i Pons - Barcelona, Catalonia (Europe) / / jguilla@ibm.net - 100325.734@compuserve.com / / http://ourworld.compuserve.com/homepages/jguilla/ / /---------------------------------------------------------/ ------------------------------ From: Darren J Moffat Subject: Netscape/Mosaic for ELF Date: Fri, 1 Dec 1995 18:54:39 GMT Where can I get either Netscape or Mosaic (or Hot Java) compiled for ELF, preferably with only the Motif (libXm) linked statically and the rest dynamically linked. I've got a competely ELF system and I've trashed my a.out libs (and I don't want to install them for just one program). Any chance some kind person with an ELF Motif license could make a WWW browser available Ta. (Email replies prefered). -- Darren J Moffat | 128 Chapel Lane | High Wycombe, Buckinghamshire, HP12 4BY | Tel/Fax: +44 (0)1494-449-537 ------------------------------ From: newcombe@aa.csc.peachnet.edu (Dan Newcombe) Subject: Re: 8088 port Date: Fri, 01 Dec 95 16:03:17 GMT In article , Greg Myers wrote: >On Mon, 27 Nov 1995, Larry Gray wrote: > Hi, >> >> I've heard a rumor that somewhere someone is working on a Linux >> port for the old 8088 chip systems. Please point me to them. >> >> Larry Gray >Linux is a 32-bit operating system that supports true multi-tasking and >true multi-user operations. The 8088 chip is a 16-bit internal/8-bit >external chip. Meaning it could never happen. At least not without >SEVERE problems/consequences. Yup. that's the chip specs, but nonetheless it is happening. http://www.linux.org.uk/Linux8086.html >With used 386 prices in the dirt, why would anyone even consider running >it on a 8088??? Because 8088's are cheaper than dirt :) There are a lot of them out there, and they are powerful enuff to do some basic functions. Also, it's a lot easier to put 8088's in an embedded system (ie. Linux for toasters.) -Dan -- Dan Newcombe newcombe@aa.csc.peachnet.edu -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= "And the man in the mirror has sad eyes." -Marillion ------------------------------ From: orch021@wrzx13.rz.uni-wuerzburg.de (Klaus-Peter Gulden) Subject: [Q] sendmsg(), recvmsg() ? Date: 1 Dec 1995 13:13:20 GMT Hi folks! In BSD and other UNIXES, there are two functions messages from socket to socket can transmitted with. Normally, sendmsg() and recvmsg() are defined in sys/socket.h. Looking into the corresponding header file under linux revealed that the either aren't implemented yet or they are present with another name. The manpages to theses functions exist since kernel version 0.99.11 -- indeed with a warning statement: WARNING: This is a BSD man page. As of Linux 0.99.11, recvmsg was not implemented I need these two functions to port the DTM stuff from NCSA. Does anybody know how to manage this problem under linux kernel 1.2.8. thanx for any help Klaus-Peter ------------------------------ From: clive@epos.demon.co.uk (Clive Messer) Subject: Re: Netscape 2.0b3 problems, anyone? Date: 28 Nov 1995 20:46:45 -0000 Brian M Grunkemeyer (bg2k+@andrew.cmu.edu) wrote: >Excerpts from netnews.comp.os.linux.development.system: 27-Nov-95 Re: >Netscape 2.0b3 problems.. by Sam Trenholme@oryx.llnl. >> On libc: OK, I just looked at both Sunsite and Tsx, and the latest >> version of Libc is 5.0.9. I looked at the incoming directory, just to be >> sure. >> >> If anyone claims the existance of libc-5.2.14, please back up this claim >> with a full URL (ftp://foo.bar.com/pub/really_hidden/^L^P^R/libc-5.2.14 >> for example). > >I believe it's in >ftp://tsx-11.mit.edu/pub/linux/packages/GCC/private/tofu/. That's where >most of the latest stuff has been hiding for the last few months. It is *deliberately* hidden to ensure it is not used by people not able to deal with the consequences of running *ALPHA* code !!! Clive. -- C Messer. Epos Systems. UK. | | "I pressed her thigh and death smiled." | Jim Morrison. ------------------------------ ** FOR YOUR REFERENCE ** The service address, to which questions about the list itself and requests to be added to or deleted from it should be directed, is: Internet: Linux-Development-System-Request@NEWS-DIGESTS.MIT.EDU You can send mail to the entire list (and comp.os.linux.development.system) via: Internet: Linux-Development-System@NEWS-DIGESTS.MIT.EDU Linux may be obtained via one of these FTP sites: nic.funet.fi pub/OS/Linux tsx-11.mit.edu pub/linux sunsite.unc.edu pub/Linux End of Linux-Development-System Digest ******************************