Re: downloads...burst, then nothing
Lev Lvovsky wrote: I'm having a problem with downloads from my machine (slink). On the network I have at home, I can upload at the regular high speeds, but downloading form the machine (to a Win98 machine), whether it be by ftp, or by http gives me a little bit of the file, and then slows down to a total crawl. Obviously if it owrks up, I'm assuming all ofthe hardware is fineis there some sort of buffer for this sort of thing that might be set too low? Make sure that the network interface on both machines are set up for half duplex operation (unless you are using a switch). Having one or both interface at full duplex might cause problems like these. Also, check if you are getting an abnormal amount of collision on the network during download (ifconfig eth0 on the Linux box). HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Cubic mouse cursor
[x] Emailed to author [x] Sent to mailing list David R. Kohel wrote: Anyone know how to get rid of this 2x2 cm mouse cursor under X? I just re-installed X (3.3.3.1) from potato. You REALLY need to give some more information. At the very least, we need to know which graphics card and which X server you are using. Also, you might want to go to a virtual console and run $ X :1 -probeonly probe.log 21 This will create a file named probe.log with the startup messages from the X server. If there are any anomalous messages, please post them as well. You night also want to check http://bugs.debian.org/xserver-svga (modify for the X server you use) to see if anyone else has reported similar problems. HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Cubic mouse cursor
[x] Emailed to author [x] Sent to mailing list David R. Kohel wrote: The x probe output is below. I'm using the SVGA server with a Neomagic card. The cursor appears as a large square with upper left hand corner at the mouse point, replacing the normal arrow symbol. I've got no other problems with the xserver. If anyone has seen this and has a fix I'd appreciate it. It is probably due to a bug in the new X server. Try doing a search on http://www.deja.com/ to see if anyone else has the same problem. As a workaround, you can add a line saying Option sw_cursor to the Device section of your /etx/X11/XF86Config as a workaround. You should probably submit a bug report on xserver-svga (man bug) if it hasn't been done already. Include all the detail you included in your second posting here. HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Sound Blaster 128 Detection
Thomas MANGIN wrote: I am now facing another nice problem : my SB 128 isn t detected. I was befdore using a Crystal card which was perfecltly working but now my Standard Sound Blaster is not working. The SoundBlaster PCI cards are not hardware compatible with the ISA based SoundBlaster cards. When configuring the kernel you need to include support for the Ensoniq AudioPCI (ES1370) card to support the PCI 128. Last time I installed a SB16, I solved the problem removing the Os is PNP aware in my Bios but this time this didn t changed anything ;( I tryed both Native and Legacy dma but it didn t help too !! I suspect the PNP or Legacy stuff like they call it to be somewhere guilty !! For information my kernel is the 2.1.130 I don't think this is the problem. This PNP/legacy stuff should only matter for ISA Pnp devices. You may want to upgrade your kernel since you have to recompile anyway. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: boot floppies -- thanks!!
Ossama Othman wrote: I wanted to thank Robert Woodcock and everyone who helped me (sorry, I don't recall their names :( ) for helping me figure out how to boot my Dell PowerEdge Server 6300. I am now the proud user (not owner, I wish! :) of a Quad-CPU Xeon Dell system with a gig (960MB with linux lim) of memory running Debian GNU/Linux 2.1 (potato). Now if I could just get the rest of my colleagues to switch to Debian. :p Just out of curiosity, how long does it take to compile a kernel on this sucker? (For maximum performance, use the -j compiler option.) The system is _smooth_ and stable! Thanks again for the help and for a great OS! Btw., are you running a 2.0 or a 2.2 kernel? The 2.2 kernel should give a substantial speed increase, esp. on multiprocessor system. In case you haven't noticed, kernel 2.2.0 was released less than 12 hours ago. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: X manager installation changes /dev/mouse from psaux to ttyS0
Oliver Elphick wrote: BOHICA wrote: I tried re-symlinking /dev/mouse to /dev/psaux BOHICA:/dev$ rm mouse BOHICA:/dev$ ln -s /dev/mouse /dev/psaux but for some reason this causes /dev/psaux to link to /dev/ttyS0. You got the command back-to-front. So now I have: /dev/mouse = /dev/ttyS0 and /dev/psaux = /dev/ttyS0 Can anyone help me resolve this? I am quite out of my depth here (not that I was too deep to begin with). As superuser, do: ln -sf /dev/psaux /dev/mouse It seems that the device node has been replaced by a symlink. Your suggestion will not help much: mouse - psaux - ttyS0 To fix this, do (# means do as root): # cd /dev # rm psaux # ./MAKEDEV misc # ln -sf psaux mouse The MAKEDEV script will recreate the /dev/psaux device node for you. ln -sf /dev/ttyS0 /dev/modem Actually, # ln -sf ttyS0 /dev/modem will suffice. Just remember that if you have multiple programs that need to access the modem (e.g. pppd, mgetty, minicom, xringd) they must all use the same device name, ie. they all use /dev/modem or they all use /dev/ttyS0. Otherwise, locking of the device will not work properly. HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: rawrite for linux
[EMAIL PROTECTED] wrote: Jonathan Sprague dixit: dd if=FILE.EXT of=/dev/fd0 (for floppy a:) say I want to make an image of Slackware's boot root disks from my cdrom to the floppy, would it be as follows?: mount -t iso9660 /dev/hdb /cdrom mount -t ext2 /dev/fd0 /floppy ^^ You should not mount the floppy if you are going to write to the floppy device (/dev/fd0) directly. Besides, these floppy images do not contain ext2 file systems. dd if =/dev/hdb/bare.i of=/dev/fd0 dd if=/dev/hdb/color.gz of=/dev/fd0 You have mounted the CDROM on /cdrom, so these commands should read dd if=/cdrom/bare.i of=/dev/fd0 dd if=/cdrom/color.gz of=/dev/fd0 Alternatively, you could use cp to do this: cp /cdrom/bare.i /dev/fd0 ... BTW, I believe obs=18k deals with disk space/capacity somehow. This shouldn't matter as long as the floppies are low-level formatted in advance (use superformat if they aren't). HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: using xfig + latex
LOPARIC Marko wrote: Do you know how to export xfig figures with latex commands and insert them in latex files? I was unlucky with pictex, eepic and pstex_t. [snip] 3. pstex_t: That is the format suggested in the documentation of tranfig package. But I was unnable to find the file psfig.tex, which is needed. I export my figures as combined PS/LaTeX (pstex_t) and put \usepackage{epsfig} in the preamble of my documents. The figures are then included in my LaTeX document with sth. like \begin{figure}[htb] \begin{center} \input{foo.pstex_t} \caption{Bar} \label{fig:foo} \end{center} \end{figure} HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: upgrade from 14.4 to 56.6 modem - what settings to change?
Martin Waller wrote: I've been a lucky boy and have got a new 56.6k modem for xams to replace my old 14.4. An old linux book I have (from kernel 1.2.13, and for slackware...) says to get fast modem speeds I have to run setserial ttys1 115200 to get higher modem speeds. The modem works OK, but I'm only getting a 28.8k connection (could be my ISP). Regradless, do I need to change the 38400 bit in the connect string thing to 115200? Or change some script somehwre to gurantee fastest possible connection speeds? You'll want to either (a) Call setserial with the spd_vhi option (in /etc/rc.boot/0setserial) for the port that your modem sits on. You can then use the 38400 speed setting in your ppp script which the kernel will interpret as 115200 bps. or (b) Leave /etc/rc.boot/setserial alone and use 115200 in your connect script. Option (a) is a relic from the pre-kernel-2.0 days and is deprecated today (but will still work). -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Samba
Jiri Baum wrote: Hello, I've just installed the Samba-server (and read the man-pages). I might be missing something but I can't figure out how to get a list of what Wimpdos calls the entire network. Can this be done without knowing any server-names? Hmm, can't find that! The best I can come up with is: smbclient -L `hostname` which has the info under browse list. Anyone have a better way of getting it? The smb2www package gives you a nice web based overview of a smb network, similar to the Network Neighborhood in windoze. You'll probably want to get the version in potato (980804-3) as the version in slink is slightly broken. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: HELP slink upgrade: netscape: locale `C' not supported
Jeff Noxon wrote: Re-install xlib6g and it should work. Now if only I could get Acrobat working again... Removing /usr/i486-linuxlibc1/lib/X11/locale (empty dir) and linking it to /usr/lib/X11/locale instead worked for me. A small glitch in the upgrade process, I presume. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: DNS woes
Gregory Dickinson wrote: Here's my (somewhat strange) situation, I hope someone can help. I have a 2.0 box configured as a DNS for our intranet. Said Debian box has 2 NIC's in it, one for the public side (internet) and one for our provate side (intranet.) I was wondering (mainly) how to make the DNS only answer queries from the private side, not the public side also. Just put a listen-on directive in your /etc/named.conf (assuming you are running bind 8). I use listen-on { 192.168.0/24; 127/8; }; You will of course need to modify this for your setup. Check the documentation in the bind-doc package for details. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: OFF TOPIC: help with SMB wierdness?
David Karlin wrote: Where can I find smb2www? (I couldn't find it in .../stable/main/binary-i386/packages or .../stable/contrib/binary-i386/packages.) It's in slink. The current version (980804-2) is somewhat broken wrt. never versions of samba, but the maintainer told me that a new version will be released shortly. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: timezones
David S. Zelinsky wrote: Since I dual boot with Win95, I have my hardware clock set to local time. I don't use Win95 much any more (read: my wife now uses Linux :), so I figure I might as well change over to UTC, so Linux will handle daylight/standard time correctly (I hope). How do I make that change? I can't find any documentation on this. You'll need to edit /etc/default/rcS. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: lprng and disabling network printer
Pere Camps wrote: How do you disable the 'printer' port for remote connections if you're using lprng? Edit /etc/lpd.perms so it has a couple of lines saying sth. like this # reject all connections from remote machines ACCEPT SERVICE=X HOST=localhost REJECT SERVICE=X and make sure that you have no other `SERVICE=X' lines in that file. Also, these lines should appear before the `DEFAULT' line. Now you can do a `/etc/init.d/lprng reload', and remote connections will be blocked. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: **IMPORTANT** Slink sendmail and libdb2
John Goerzen wrote: Remco van de Meent [EMAIL PROTECTED] writes: They default to /etc/mail, except for aliases.db, which should be in /etc (Debian Policy, 4.5). I know what policy says, but this is not how it works. mailertable, etc. are in /etc and not /etc/mail. You can put them wherever you want. Just use something like this in your /etc/mail/sendmail.mc: FEATURE(mailertable, `hash -o /etc/mail/mailertable') It's a pain to manually have to run a long makemap command each time something is updated. There has GOT to be a better way... You can put a make file in /etc/mail to do the dirty work. Then all you need to do to update the databases is to go to /etc/mail and type make. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Notification of incoming telnet sessions..
Michael Beattie wrote: On Tue, 27 Oct 1998, Petter Adsen wrote: Michael Beattie writes: | I want to be able to have some form of notification when a telnet | connection is made to my linux box, most likely in the form of playing an [snip] You can probably do this with tcpwrappers. See the manpages for hosts_access(5) and tcpd(8). Hmmm.. thanks... I'll look into that one too. Actually, it should be pretty easy. You can just use the spawn option on the in.telnetd line in /etc/hosts.allow (see hosts_access(5) and hosts_options(5)). Something like this will work: in.telnetd: ALL: spawn /usr/bin/play /your/favourite/aufile : ALLOW -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Why can't I execute a script??
Evan Van Dyke wrote: Peter S Galbraith wrote: If some user is capable of putting a fake `ls' in a random directory where you might trip on it, that user is far more likely to put it in your ~/bin directory! (Same privileges are required) Just a thought. Just make the . directory the _last_ part of your path, that way it will search /bin /usr/bin /usr/local/bin and all the rest of your path first. This discussion crops up on this list once a week or so. The bottom line is that users (root especially) should not have a . anywhere in their path. Lets assume that root has a . as the last element of his path. He then goes to the home directory of a malignant user, intending to do an ls on his dir. Even root is not perfect, so he makes a typo and actually types sl instead. The malignant user has a script called sl in his home dir: #!/bin/sh # Do bad stuff as root... rm -f sl echo bash: sl: command not found 2 And root never knew what hit him... -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: mySQL
D'jinnie wrote: I recently visited http://mysql.org (was trying to figure out whether I should migrate to it). I haven't read their license - but there are a few links there that suggested to me that it's not very free - because they wanted money for it. Is it GPLed or how does it work? MySQL is under an almost free license, meaning that you get the sources, you hvae the right to make modification and redistribute these, and you can use it for anything you wan't (including commercial use) without having to pay any licensing fees. The only thing you aren't allowed to do, IIRC, is to sell MySQL or a derived product for profit. This is why the MySQL packages are in non-free and not in main. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: HELP: Repair of tar.gz files??
Rick Younie wrote: If the file was corrupted by being transferred in ascii mode instead of binary, you can fix it. There's a program in the window's world called NOCRLF that will do it. Or if you have a language or a good hex editor, replace the carriage return/line feed pairs with a linefeed and that will fix. No, generally you can't fix binary files downloaded in ASCII mode. The CR - CR/LF transformation is irreversible for binary files. That is, there may have been some (valid) CR/LF's in the binary file at first, which will then be converted into CR's by the reverse transformation. If you're lucky it might work, though. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: PPP logging?
Ed Cogburn wrote: W. Paul Mills wrote: 1.3-27 is faulty. -26 or -28 should work. Uh, I *am* running sysklogd 1.3-28 and my ppp logging is still broken. Make sure that you have a line in /etc/syslog.conf that reads local2.*-/var/log/ppp.log and make sure that you have tabs separating the fields and not just spaces (a small syslogd quirk). -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Telnet with slink
Phillip Neumann wrote: [telnet to localhost not working] I realize that after upgrading telnet program were not avaible so i try to install telnet package apt installed it ok. You need to install the telnetd package as well. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: 2.1.x series with IP-Masqing
Steve Lamb wrote: Just a warning to anyone who is considering using the 2.1.x series since 2.2 is rumoured (I stress that word highly) to be out soon. If you use IP-Masqing it appears that the 2.1.x kernels break ipfwadm. With a broken ipfwadm it isn't possible to add/remove IP-Masqing rules like normal. There may have been another way, but I did this on my production machine (silly, I know) and just simply reverted back to my 2.0.x series kernel. IP Masquerading with 2.1 kernels is no problem. Just install the ipmasq package from slink (currently 3.2.3). If you have a standard IP Masquerade setup (one default gw, one or more internal networks to be masq'ed) it will set up the rules automatically. No need to write those pesky ipfwadm rules by hand anymore :-). It will even re-run the setup scripts every time you bring up/down a PPP link. Did I mention that it is fully configurable? No, I didn't write it, I just use it and like it very much. Clever stuff and a bit of shell magic. It supports both 2.0 kernels (ipfwadm) and 2.1 kernels (ipchains). -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: WARNING! Do NOT upgrade sysklogd to 1.3-29
Bob Nielsen wrote: The latest slink update to sysklogd 1.3-29 is SERIOUSLY broken. I installed it and could no longer become root. If you have installed it, you will need to boot with your rescue disk (you do have one, don't you?) and either disable sysklogd (chmod -x /etc/init.d/sysklogd) or install the hamm version. It does indeed seem to be broken. I don't have the problems you seem to be having (missing ability to become root), though. My problem is that nothing is logged. Syslogd seems to run just fine--the marks are written to the syslog every 20 minutes as usual, but is seems as if other programs can't communicate with it. All the same, I concur with Bob: *don't* upgrade to sysklogd_1.3-29. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: PPP logging?
Michael Beattie wrote: For some reason, ppp is not being logged anymore on my hamm/slink box.. anyone else having the same results? If not, how do I fix it? I believe I upgraded ppp yesterday, as the last logged time was 2:55pm... does not even have disconnect logged. AFAICT, this is caused by a faulty sysklogd. Upgrading to 1.3-28 should fix the problem (it did for me). -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: New Install... pon not working
Bob Eidt wrote: I Have installed Linux 2.0 from floppy disk to a unit without any non- linux partitions and have tried both using pppconfig and manually editing the provider files as indicated in install.txt. FTP is the only method to my disposal of enhancing my installation. It would seem the best I can do is with the configuration files I had after running pppconfig. Using pon and plog I get an OK in response to ATZ, I get a CONNECT in response to the ATDT1231234. But then a string of some sort of PPP negotiation talk is sent and after a while I get a message it has been sent 9 times without a response and I disconnect. You should post some more information, such as your chat script (/etc/chatscripts/provider), your ppp config file (/etc/ppp/peers/provider) and the exact error message logged by pppd and chat (output from plog). If plog doesn't produce the complete output (only the last ten lines are shown), copy the output directly from /var/log/ppp.log. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Format of .deb files
Stephen J. Carpenter wrote: I guess your home PC can't be acessed from work (ie having it online) too bad...I need to set mine up (one of these days) to let me tell it to conenct remotely. xringd is your friend. Just remember that if you have a dynamic IP you need a way to locate your machine on the Internet. You can either set up a script to post the IP to a web site or mail you the IP, or you can use a dynamic DNS service such as the one at www.ml.org. Check out the mlddc package for a client for the latter. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: netscape4.x cannot read debian-user list archives
Jan Krupa wrote: I have installed netscape4.5b2 under debian2.0. When I try to read debian-user mailing list archives (e.g. from September, august) first time it's O.K. but next times on the same month archives netscape just hangs I have to kill it (the archives are very big). When I try to read much smaller archives (e.g. debian-changes, September) it's all right. I had the same behavior using netscape4.x, but netscape3.x reads the archives well. Does anybody have any idea what causes such behavior of netscape ? May be I have configured netscape wrong. My be I some hint what change in preferences? I think this is a cache related bug, as turning off the (disk) cache in Netscape will prevent this from happening (at least for me). One workaround is to turn off the disk cache (size = 0) and use a caching proxy server (such as squid) instead. This will of course use a fair amount of RAM. Another workaround is to follow links from the main article index using the middle mouse button. This will tell Netscape to open the link in a new window so that you never have to reload the index from cache in the first window. Oh well, I really ought to file a bug report with Netscape :-) YMMV. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: I can't compile anymore
Pierfrancesco Caci wrote: E.L. Meijer \(Eric\) [EMAIL PROTECTED] writes: linux/errno.h is in libc6-dev. You also should not try to compile 2.0 Ok, reinstalled libc6-dev, I get these messages further on in the compilation: [EMAIL PROTECTED]:/usr/src/linux # make bzImage make modules make modules_install egcc -D__KERNEL__ -I/scratch/usr/src/linux/include -Wall ^^^ is this *really* where your kernel sources are installed? It seems like egcc is picking up the headers that come with libc6-dev instead of the headers that come with the kernel source. Did you remember to run `make dep; make clean' before running `make bzimage'? Besides, you should take a look at the kernel-package package which can automate the entire kernel build process and produce a nice .deb package which can be installed with dpkg. Works like a charm. -Wstrict-prototypes -O2 -fomit-frame-pointer -pipe -fno-strength-reduce -m486 -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=586 -c -o init/main.o init/main.c /usr/include/asm/string.h: In function `__constant_memcpy': In file included from /usr/include/linux/string.h:39, from /usr/include/asm/termios.h:58, from /usr/include/linux/termios.h:5, from /usr/include/linux/tty.h:20, from /usr/include/linux/sched.h:26, from init/main.c:17: /usr/include/asm/string.h:443: warning: control reaches end of non-void function [snip] HTH -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: help w/ kerneld loading of modules
Shaleh wrote: Yes I did compile it w/ module support. I want to make it work, or find out why it does not work. No hacks please (-: Module support and module auto loading (kerneld support) are two different settings that must both be set to Y in the kernel config for kerneld to work. Assuming that kerneld support is compiled into your kernel, the next thing to check is that kerneld is actually running. Try running $ ps ax | grep '[k]erneld' and you should get an output like this 181 ? S0:01 /sbin/kerneld If kerneld is not running, something strange is going on. In that case you should try starting it by hand (as root) and report back to the list with the results. If all of this seems in order, check /var/log/daemon.log to see if modprobe is complaining about modules that it cannot locate. Perhaps you just have to add an alias to /etc/conf.modules (hamm version of modutils) or /etc/modutils/aliases (slink version of modutils). HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Help! FAT magic number corrupted!
me wrote: and exits. fdisk /dev/hda1 won't give me a partition table because it's a dos disk. Yes, you use fdisk on the whole-disk devices such as /dev/hda and /dev/hdb while you put file systems on the partition devices such as /dev/hda1 and /dev/hda3. thanks, yes, this works, and i can run cfdisk fine. cfdisk works fine. I didn't change anything, but i tried to write the partition table, and here's the error cfdisk gives: Wrote partition table, but re-read table failed. Reboot to update table. This was what it gave me originally when this whole mess started. Well, did you reboot? There's a very good reason fdisk emits that message. If you don't reboot before making any file system modifications (mkfs et.al.) there is a good chance that you hose the wrong part of the disk, since the kernel's idea of the partition layout is then different from the actual layout. Btw.: does anyone know why the re-read fails? AFAIR, this used to work fine on my SlackWare system some eons ago, but invariably fails on my Debian system--seemingly independant of the kernel version used. e2fsck on my ext2 drives (/dev/hda3 et al) works fine. e2fsck on /dev/hda1 of course doesn't work, because it's a non-ext2 partition. e2fsck on /dev/hda returns Couldn't find ext2 superblock, trying backup blocks... e2fsck: Bad magic number in super-block while trying to open /dev/hda. what's this superblock thing? I'm not up on the technical details of FATs. The ext2fs uses the so-called superblock for storing important data about the file system, such as size, volume label, and reserved blocks percentage. The ext2fs superblock is identified by a magic number which ensures that other file systems don't get misidentified as ext2 file systems. Since the superblock is absolutely essential for the ext2 file system, it is duplicated a number of times throughout the file system; that is what the trying backup blocks... above refers to. Since you are trying to e2fsck a FAT file system, e2fsck doesn't find a valid ext2 superblock and aborts. fsck.msdos on /dev/hda returns: Currently, only 2 FATs are supported, not 243. This won't work, since there is no file system on /dev/hda. fsck.msdos on /dev/hda1 returns: Currently, only 2 FATs are supported, not 81. Hmm. Why does it think I have 81 FATs on one partition? It seems that the file system on /dev/hda1 is hosed :-( As for the Win95 system... you will probablky need to re-install yep, i figured so. but right now I can't even get a Win95 boot floppy to recognize my C drive (aka /dev/hda1). sigh Could fips help me with that problem? Try removing the broken FAT partition with Linux fdisk. Then boot with the win95 boot floppy and run DOS fdisk to recreate the FAT partition--make sure that you have a copy of fdisk.exe on the boot floppy! It is also a good idea to have a Linux boot floppy handy, just in case... HTH. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Help! FAT magic number corrupted!
me wrote: On 16 Oct 1998, Rene Hojbjerg Larsen wrote: scsi: 0 hosts. scsi: detected total.[this is usual when it's working.] Partition check: hda: hda1 hda2 hda3 hda4 hda5 hda6 [MS-DOS FS Ren 12, FAT 0, check=n, conv=b, uid=0, gid=0, umask=022, bmap] [me=0x52, cs=0, #f=81, fs=0, fl=48384, ds=52400, de=55699, data=55972, se=28213, ts=68224, ls=9471, rc=0, fc=4294967295] Transaction block size= 512 Kernel panic: VFS: Unable to mount root fs on 03:01 [HANG] Something is definately wrong here. The kernel tries to use /dev/hda1 as the root partition, which clearly won't work as it holds a FAT file system. Do you use LILO to boot the kernel? In that case it is not set up correctly. Try holding down SHIFT when the system boots--this should produce a boot: prompt at which you can enter the label of the Linux boot image you are using (press TAB to see the list of available images to boot). Enter at the prompt: label_of_image root=/dev/hda3 (without the quotes). This should tell the kernel where to find the root partition. Assuming the system boots up correctly, edit /etc/lilo.conf to reflect your setup (root=/dev/hda3, boot=/dev/hda), and re-run lilo. Now your system should boot up normally the next time. Hmm. Why does it think I have 81 FATs on one partition? It seems that the file system on /dev/hda1 is hosed :-( though i note that 81 appears in the output above (#f=81). wonder if there's a way to fix this at a low level. no, scratch that, it's probably a Bad Idea. yep, i figured so. but right now I can't even get a Win95 boot floppy to recognize my C drive (aka /dev/hda1). sigh Could fips help me with that problem? Try removing the broken FAT partition with Linux fdisk. Then boot with the win95 boot floppy and run DOS fdisk to recreate the FAT partition--make sure that you have a copy of fdisk.exe on the boot floppy! It is also a good idea to have a Linux boot floppy handy, just in case... sounds drastic. how likely is it that i'll lose the data on /dev/hda1? (did i mention that during an ftp process, my backup got corrupted, and /dev/hda1 (used to?) contain my only other copy of those files? sigh) Talk about learning the hard way. I really need to corrupt that partition as little as possible. Tough luck. At this point all bets are off. Try posting the output of fdisk -l /dev/hda. Perhaps something obvious will show up. I just got a crazy idea. Check if you have a file named /boot/boot.0301. If that is the case there may still be hope. Keep us posted. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: LaTeX book
M.C. Vernon wrote: Can anyone recommend a good one of these, please? Mainly for scientific writing... I can personally recommend Helmut Kopka and Patrick W. Daly: A Guide to LaTeX 2e, Addison-Wesley. It gives a thorough introduction to LaTeX and includes chapters on more advanced topics as well. I have found it far superior to the Lamport book (LaTeX, A Document Preparation System), which I also own. For advanced graphics, The LaTeX Graphics Companion is supposed to be the bible, but I don't know much about this one (these books are rather expensive). -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Debian's recommendation for the size of the swap.
[EMAIL PROTECTED] wrote: On Wed, Oct 14, 1998 at 11:34:16PM +0300, shaul wrote: (1) Linux accept up to 128MB for a single swap partition (2) There is (was) a rule of thumb to have a swap size as twice as the RAM the machine has (3) Having more RAM reduces the needs for swap. ya konw, i'm a little confused. people say more RAM reduce the needs for swap while swap is recommended for double size of RAM ? yeah, newbia i am :-P That swap must be = 2*RAM is a common misconception in the Linux world. On BSD's this is a good rule of thumb, due to a different VM[1] subsystem design but on Linux only the RAM+swap figure matters. The only general rule for swap size is that you should have enough. That means that RAM+swap should be large enough to run the programs you are likely to run at any given time. Note, however, that you should always have *some* swap, even if all your programs will fit in RAM, to allow the kernel to swap out unused pages and make room for more buffers. Similarly, you should always have some unused VM to allow for extraordinary memory requirements and to make room for buffers. Did I mention that buffers are a good thing? :-) It is, of course, better to have too much swap space than too little. Horrible things will happen if you run out of VM: processes will be randomly killed off, the system may crash, etc. i'm currently use 12M swap with 64M RAM, is it too few ? i feel ugly when open *guash*, and xemacs and netscape opened very slowly. i only have 2G hd, and 750M among it is spared for win98 'cause i need its support for chinese stuff. I would probably add some some swap space, but as I outlined above, this is a very individual thing. My box has 64MB RAM + 96MB swap and it seems quite happy with that configuration. However, I probably use more VM than most people... [1] VM = Virtual Memory -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: slink package upgrades fail
M.C. Vernon wrote: I also get similar messages for tetex-base and tetex-bin and can neither install nor remove these packages. Yes, I have problems like this as well... :( tetex-bin included as an example. Shall I file a bug? and how do I fix this? Matthew [EMAIL PROTECTED]:~# dpkg -i /mnt/public/pub/unix/Linux/debian/dists/slink/main/binary-i386/tex/tetex-bin_0.9.981008-1.deb Selecting previously deselected package tetex-bin. (Reading database ... 42207 files and directories currently installed.) Preparing to replace tetex-bin 0.9.981008-1 (using .../tex/tetex-bin_0.9.981008-1.deb) ... Unpacking replacement tetex-bin ... rm: /usr/lib/texmf/web2c: is a directory dpkg: warning - old post-removal script returned error exit status 1 dpkg - trying script from the new package instead ... rm: /usr/lib/texmf/web2c: is a directory dpkg: error processing /mnt/public/pub/unix/Linux/debian/dists/slink/main/binary-i386/tex/tetex-bin_0.9.981008-1.deb (--install): subprocess new post-removal script returned error exit status 1 rm: /usr/lib/texmf/web2c: is a directory dpkg: error while cleaning up: subprocess post-removal script returned error exit status 1 Errors were encountered while processing: /mnt/public/pub/unix/Linux/debian/dists/slink/main/binary-i386/tex/tetex-bin_0.9.981008-1.deb I had similar problems when I upgraded some TeTeX packages today. I then took a look in the bug reporting system to see if a bug report had been filed, and sure enough, one possible solution is to run this snippet of code (taken from the postinst script for tetex-bin): if [ ! -L /usr/lib/texmf/web2c ] then mv /usr/lib/texmf/web2c/* /var/lib/texmf/web2c rmdir /usr/lib/texmf/web2c ln -s /var/lib/texmf/web2c /usr/lib/texmf/ fi This ensures that /usr/lib/texmf/web2c is a link to /var/lib/texmf/web2c, as it is supposed to be. After I'd done this, the packages installed without a problem. YMMV. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Slow ifconfig
Mirek Kwasniak wrote: On Tue, Oct 13, 1998 at 06:27:30PM +, Rene Hojbjerg Larsen wrote: Somehow my ifconfig has become extremely slow. I'm not sure exactly when this happened, but it probably happened while upgrading to slink the other day. To give an example: $ time /sbin/ifconfig /dev/null real0m6.788s user0m4.450s sys 0m1.590s This used to take far less than a second... [..] Unless one of you has a solution, I will post a bug report. I checked kerneld tries load modules or reread conf file many times executing modprobe and command from config like this: path[*]=/lib/modules/`uname -r` path[*]=/lib/modules/`kernelversion` On my station invocation of ifconfig execute about 160 processes: 22:06:59 3:2'[EMAIL PROTECTED] ifconfig /dev/null;ps tp3 PID TTY STAT TIME COMMAND 15329 p3 S0:00 /bin/bash 22506 p3 R0:00 ps tp3 22:09:58 3:2'[EMAIL PROTECTED] ifconfig /dev/null;ps tp3 PID TTY STAT TIME COMMAND 15329 p3 S0:00 /bin/bash 22672 p3 R0:00 ps tp3 22672-22506 = 166 Bug should be posted against modutils: $ time ifconfig real0m4.923s :( user0m0.020s sys 0m0.000s $ killall kerneld $ time ifconfig real0m0.034s :) user0m0.010s sys 0m0.000s You're absolutely right, the problem is with modutils. My kmod executes, among others, the following command while running ifconfig: /sbin/modprobe -s -k net-pf-3 The execution of this command is very slow: # time /sbin/modprobe -s -k net-pf-3 real0m2.362s user0m1.520s sys 0m0.620s A full strace of this modprobe command (strace -f) comes in at about 13300 lines, so it is obviously doing a lot of (unnecessary) work. I will try to investigate further. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: Slow ifconfig
Torsten Hilbrich wrote: On: 13 Oct 1998 18:27:30 GMT Rene Hojbjerg Larsen writes: Somehow my ifconfig has become extremely slow. I'm not sure exactly when this happened, but it probably happened while upgrading to slink the other day. To give an example: $ time /sbin/ifconfig /dev/null real0m6.788s user0m4.450s sys 0m1.590s This used to take far less than a second... Already reported as bug #27776. As workaround reinstall modutils version 2.1.121-4. OK, thanks. For others having the same problem, you can (for the time being) find the older version of modutils at ftp://ftp.biostat.washington.edu/mirrors/debian/dists/slink/main/binary-i386/base/modutils_2.1.121-4.deb (Sorry about the long line) -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Slow ifconfig
Somehow my ifconfig has become extremely slow. I'm not sure exactly when this happened, but it probably happened while upgrading to slink the other day. To give an example: $ time /sbin/ifconfig /dev/null real0m6.788s user0m4.450s sys 0m1.590s This used to take far less than a second... If I strace the ifconfig process, it pauses around these lines: socket(PF_INET, SOCK_DGRAM, IPPROTO_IP) = 3 socket(PF_IPX, SOCK_DGRAM, [PF_UNSPEC]) = -1 ENOSYS (Function not implemented) socket(PF_IPX, SOCK_DGRAM, [PF_UNSPEC]) = -1 ENOSYS (Function not implemented) socket(PF_IPX, SOCK_DGRAM, [PF_UNSPEC]) = -1 EINVAL (Invalid argument) socket(PF_AX25, SOCK_DGRAM, 0) = -1 ENOSYS (Function not implemented) socket(PF_AX25, SOCK_DGRAM, 0) = -1 ENOSYS (Function not implemented) socket(PF_AX25, SOCK_DGRAM, 0) = -1 EINVAL (Invalid argument) socket(PF_APPLETALK, SOCK_DGRAM, 0) = -1 ENOSYS (Function not implemented) socket(PF_APPLETALK, SOCK_DGRAM, 0) = -1 ENOSYS (Function not implemented) socket(PF_APPLETALK, SOCK_DGRAM, 0) = -1 EINVAL (Invalid argument) brk(0) = 0x804fe0c Specifically, it pauses for about two seconds before (in?) the third socket call for each protocol (IPX, AX25, APPLETALK). My kernel is compiled without support for any of these, and I have aliasesed them to off in /etc/conf.modules. This happens with both kernels 2.0.36pre14 and 2.1.125. The version of netbase installed is 3.11-1. Unless one of you has a solution, I will post a bug report. TIA -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Printing with gimp 1.0.1
I installed gimp 1.0.1 from slink and can't make it print directly to the printer (through lpr). In the Print dialog, I can only choose to print to a file--no printers are accessible. I suppose I need to tell gimp which printer(s) I have but I can't find anything in the docs. I have noticed, however, that the print filter created an (almost) empty file called printrc in my .gimp directory but, again, I can't seem to find any useful docs on this. Hope you can help me out. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: xdm replacement?
Person, Rod wrote: Anyway, is there a replacement for xdm. I know about kdm but I want to use windowmaker so I don't want to take up all the space with kdebase and libs and support... Can I use kdm without all that or is there something else I can use instead of xdm? Perhaps you should take a look at wdm, which is an xdm replacement with a NeXT-alike look and feel. This should go perfectly with Window Maker. I haven't tried it out myself yet, though. You can find the Debian package in slink (and perhaps in hamm?). -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin
Re: completely strange system time....
Bruno Boettcher wrote: i recently installed a debian system, and since then have a completely strange system time: my clock reports: 09:13:43 yoda:~$ date Thu Oct 8 09:14:23 MST 1998 and a correctly working host: 16:17:21 erm6:~$ date Thu Oct 8 16:17:29 CEST 1998 that's a bit strange how do i get it corrected? theoretically i thought that the xntpd should take car of this, or am i wrong? Your time zone is wrong. Run tzconfig (as root) on the misbehaveing host to specify the correct time zone. -- /'`\ zzzZ | My PGP Public Key is available at: ( - - ) | http://home1.inet.tele.dk/renehl/ --oooO--(_)--Oooo-- Don't ya just hate it when there's not enough room to fin