locks and pppd
-BEGIN PGP SIGNED MESSAGE- I have a strange problem with my serial port /dev/ttyS0. /proc$ setserial /dev/ttyS0 /dev/ttyS0, UART: 16550A, Port: 0x03f8, IRQ: 4 /proc$ tail -f /dev/ttyS0 [1] 13643 /proc$ echo at /dev/ttyS0 This hangs infinitely, BUT only if I have used pppd. There are no lock files in /var/lock: /proc$ ls /var/lock iptraf.tag iptraf_detstat.tag wtmplock and there's no pppd running. ppp still runs fine if started up again, and minicom works too. File locking seems to work fine between minicom and pppd, but plain redirection with '' fails after having used pppd once. Rebooting helps :-) but that's definitely not what I want to do that often. You might want to know why I need that? I want to write a dialer script using the modem to dial and taking over with the normal phone. I'm sick of typing the numbers which are already on the screen into the phone when I have a modem. Please help me, I have spent hours trying to solve that problem today. (which is probably not even worth the effort, but that's the nerdy business, isn't it?) - -- Lukas Eppler iotalab gmbh, Limmatauweg 18a, 5408 Ennetbaden +41 76 373 07 87 | fax +49 89 244394808 -BEGIN PGP SIGNATURE- Version: 2.6.3ia Charset: noconv iQCVAwUBOWoC8UGzNEF6cQktAQGKNgP/SvyR+W57eA04FldOUJpsQCXie+zSB0qJ pyzXZeL9W5+LzWdc08yeJ0Qx18mcZrYaoutug8AMO66N7OD+B2PN4jl4J19N8wxo ntAXqefRFDx6ttBQmZBKakfzRJNhD8xTYGZo5yPpEvvUF6qC/WBkb/7dlLKCCL2z VxMNmMY9tj4= =bc1F -END PGP SIGNATURE-
Re: ssh won't forward x11 connections [solved]
-BEGIN PGP SIGNED MESSAGE- Hi there! Thank you very much for your help. I had to do something which I would consider a quick fix: placing my own hostname in /etc/hosts as localhost, i.e. added 'spy' in 127.0.0.1 localhost spy This is not a nice method, I suppose? But who else than me would use ssh to connect to my own computer? Thanks anyway, Lukas Am Mon, 27 Sep 1999 schrieb Martyn Pearce: Lukas Eppler writes: | Hi, | | when typing | | ~$ ssh localhost kvt | | I get | | _X11TransSocketINETConnect: Can't connect: errno = 101 | kvt: cannot connect to X server spy:10.0 Clearly you found the Debian defaults for switching off X11. An obscure hint, probably, but I got caught out because my laptop had a different (constant) hostname to the name I was connecting with, because of different environments. I.e., my laptop is called 'incubus', whilst I was connecting to it as baker-st because it was on a foreign network. This spooked ssh. I have to set the hostname a/c to the DNS lookup, it seems for this --- although everything else was happy... - -- http://www.fear.ch/godot Bülachstrasse 7a, 8057 Zürich +41 1 3130787 | fax +49 89 244394808 -BEGIN PGP SIGNATURE- Version: 2.6.3ia Charset: noconv iQCVAwUBN+9dMEGzNEF6cQktAQFsFQQAmMObQaSA9fjHynf9n1wbtUuIwQWqsPtF NWVy0+vEIUZ+4Lat2ZeuD2rjCZDDaeqINgulP0x++CmKKrtOQ/WsX2kby0QeqEcB Eb7WXXstMS+WfucZxHI9LQHyqDmaFusPWTR4neep1dqmBxo9pWfUFApe1ZT7LDpr bacLsml8lj8= =UElk -END PGP SIGNATURE-
ssh won't forward x11 connections
-BEGIN PGP SIGNED MESSAGE- Hi, when typing ~$ ssh localhost kvt I get _X11TransSocketINETConnect: Can't connect: errno = 101 kvt: cannot connect to X server spy:10.0 This happens also when I try to connect to my machine from the outside. Since it works on the other three machines I installed, I don't know what to do. More verbose output (-v): SSH Version 1.2.26 [i586-unknown-linux], protocol version 1.5. Standard version. Does not use RSAREF. spy: Reading configuration data /etc/ssh/ssh_config spy: Applying options for * spy: ssh_connect: getuid 501 geteuid 0 anon 0 spy: Allocated local port 1022. spy: Connecting to 127.0.0.1 port 22. spy: Connection established. spy: Remote protocol version 1.5, remote software version 1.2.26 spy: Waiting for server public key. spy: Received server public key (768 bits) and host key (1024 bits). spy: Forcing accepting of host key for localhost. spy: Host '127.0.0.1' is known and matches the host key. spy: Initializing random; seed file /home/godot/.ssh/random_seed spy: Encryption type: idea spy: Sent encrypted session key. spy: Installing crc compensation attack detector. spy: Received encrypted confirmation. spy: Trying rhosts or /etc/hosts.equiv with RSA host authentication. spy: Remote: Rhosts/hosts.equiv authentication refused: client user 'godot', server user 'godot', client host 'localhost'. spy: Server refused our rhosts authentication or host key. spy: Connection to authentication agent opened. spy: Trying RSA authentication via agent with '[EMAIL PROTECTED]' spy: Received RSA challenge from server. spy: Sending response to RSA challenge. spy: Remote: RSA authentication accepted. spy: RSA authentication accepted by server. spy: Requesting X11 forwarding with authentication spoofing. spy: Requesting authentication agent forwarding. spy: Sending command: kvt spy: Entering interactive session. _X11TransSocketINETConnect: Can't connect: errno = 101 kvt: cannot connect to X server spy:10.0 spy: Transferred: stdin 0, stdout 96, stderr 0 bytes in 0.2 seconds spy: Bytes per second: stdin 0.0, stdout 395.3, stderr 0.0 spy: Exit status 1 My /etc/ssh/ssh_config _ # This is ssh client systemwide configuration file. This file provides # defaults for users, and the values can be changed in per-user configuration # files or on the command line. # Configuration data is parsed as follows: # 1. command line options # 2. user-specific file # 3. system-wide file # Any configuration value is only changed the first time it is set. # Thus, host-specific definitions should be at the beginning of the # configuration file, and defaults at the end. # Site-wide defaults for various options Host * ForwardAgent yes ForwardX11 yes # RhostsAuthentication yes # RhostsRSAAuthentication yes # RSAAuthentication yes # TISAuthentication no # PasswordAuthentication yes # FallBackToRsh yes # UseRsh no # BatchMode no # StrictHostKeyChecking no # IdentityFile ~/.ssh/identity # Port 22 # Cipher idea # EscapeChar ~ ___ My /etc/sshd_config ___ # This is ssh server systemwide configuration file. Port 22 ListenAddress 0.0.0.0 HostKey /etc/ssh/ssh_host_key RandomSeed /etc/ssh/ssh_random_seed ServerKeyBits 768 LoginGraceTime 600 KeyRegenerationInterval 3600 PermitRootLogin yes IgnoreRhosts no StrictModes yes QuietMode no X11Forwarding yes X11DisplayOffset 10 FascistLogging no PrintMotd yes KeepAlive yes SyslogFacility DAEMON RhostsAuthentication no RhostsRSAAuthentication yes RSAAuthentication yes PasswordAuthentication yes PermitEmptyPasswords yes UseLogin no # CheckMail no # PidFile /u/zappa/.ssh/pid # AllowHosts *.our.com friend.other.com # DenyHosts lowsecurity.theirs.com *.evil.org evil.org # Umask 022 # SilentDeny yes If you have an idea what it could be, please let me know. Lukas - -- http://www.fear.ch/godot Bülachstrasse 7a, 8057 Zürich +41 1 3130787 | fax +49 89 244394808 -BEGIN PGP SIGNATURE- Version: 2.6.3ia Charset: noconv iQCVAwUBN+5eB0GzNEF6cQktAQFxewP6A9K4CztUEAcNiHLFCtBd1Xu/CwkFpa4t 6LS6iwjeVv3owfc7Xwht5v3+gmai+LAGm+CEf/Am2Doyxv6Wl1R1fDx18+DTuYG0 RH83hruwaFhNOsSj16BC1UVuUdvoot9AfNTcATVC+Z4x+7/FYQR/gnXydGttBxBK T+i75tddDoM= =wVJC -END PGP SIGNATURE-
strange modem behaviour
When using my modem by ppp, everything works fine, always. But when trying to dial with the following script (which I plan to use out of a database), it sometimes hangs at the stty command. I have a Laptop using a ClipperCom World V.34 PCMCIA Modem. Removing it and reinserting does not solve the problem. I am not sure if the ppp command shuts down the possibility to access it. With minicom I can talk to the modem even when the dial script here is not working anymore. It has nothing to do with blacklisting. Any hints? #!/bin/bash stty 115200 /dev/modem echo ATF0M3 /dev/modem echo ATDT $1 /dev/modem sleep 8 echo ATH /dev/modem in my /etc/chatscrips/provider: ABORTBUSY ABORTNO CARRIER ABORTVOICE ABORTNO DIALTONE ATF0%VM0 OK ATDT 0840 840 888 ogin T44 word \wonttell in my etc/ppp/peers/provider: noauth defaultroute /dev/modem 115200 -- http://www.fear.ch telnet://mud.fear.ch: finger://[EMAIL PROTECTED] Bülachstrasse 7a, 8057 Zürich, +41 1 313 07 87 (home)
Re: Oracle 8i has problems [solved]
Success! The Oracle 8i together with Oracle Webserver is running. We installed redhat and the same error occurred, so we (gladly) switched back, so it's no debian specific problem. The problem seems to be not a configuration error but a bug in oracle; a queue somewhere (we don't know) keeps one to two requests in the queue until it is flushed (about every 10 seconds). A workaround is to let a script start a request every second or so. Just wanted to let you know. On Thu, 26 Nov 1998, Lukas Eppler wrote: Hi, We managed to set up an Oracle Version 8i on our 2xPII/512MB debian box with the Oracle Webserver, but it answers webrequests only all 10.05 seconds. If I manage to wget the page right before the tick, I get it instantly, otherwise I have to wait up to 10 seconds. Logs seem to think the request is completed instantly, but it isn't... It seems the page is blocked afterwards, outside Oracle. Since no one out there seems to have that problem we consider to try out Redhat, which would be a pity. So, I would be glad if someone could give me a hint what to try. Thanks, -- Lukas Eppler (godot) -- Lukas Eppler (godot)
Oracle 8i has problems
Hi, We managed to set up an Oracle Version 8i on our 2xPII/512MB debian box with the Oracle Webserver, but it answers webrequests only all 10.05 seconds. If I manage to wget the page right before the tick, I get it instantly, otherwise I have to wait up to 10 seconds. Logs seem to think the request is completed instantly, but it isn't... It seems the page is blocked afterwards, outside Oracle. Since no one out there seems to have that problem we consider to try out Redhat, which would be a pity. So, I would be glad if someone could give me a hint what to try. Thanks, -- Lukas Eppler (godot)
Re: KDE Questions
On Sun, 25 Oct 1998, Cristov Russell wrote: 2 - When I start the editor application, it opens 3 times. The first 2 times it pops up an error message for a directory I have not loaded a file from previously but from a directory I used in bash prior to running KDE the first time!? Does anyone know how to fix this? This seems to be an error in the .deb - kde distribution. The startup line in /usr/share/applnk/Graphics/KFax.applnk is Exec=kfax %f -caption %c %i %m The options are explained in file:/usr/doc/kde/HTML/default/kfm/app.html : - Most old style apps may need the %f argument. This means that they can only accept files on the local hard disks. If you use URLs never the less, KFM will download the files first and start the application afterwards. If the files are modified by the app, the user is prompted wether he wants them to be uploaded. You may enter %u instead of %f. This will cause KFM to pass URLs to the application. External files are not preloaded. Some KDE apps support this cool feature. There are addtional tags you may use. Here is a short list * %k The path and name of the config file itself. This is useful if the app you want to start should read/modify the config file that started it. This way you can for example implement simple printer representations. The same file that repesents the printer on your desktop stores your prefered printer settings. Use %k to pass the config file to the printer GUI. * %n The filename of the documents without path. * %d The path of the documents without file name. The %d and the %n tag together are the same as %f. * %c Name of the application. This is the name of the config file itself without path and extension. This way you can set the title of the application. - Anyway, %i and %m are _not defined_ in the readme. kde seems to pass %i and directly to the program, so it tries to open the file you klicked on, the inexistent %i and %m (maybe even creating them on the hd... how many such files have you found lying around?) I solved the problem by taking these %i and %m out of the applnk files for the programs i need most. But I am very unhappy with this solution. I don't need KFax, that's why i have still the old config there. Is this problem solved with the slink version? -- Lukas Eppler (godot)
Re: Debian Menu Methods WMaker Menu Config
On Tue, 27 Oct 1998, jim r wrote: I've installed several non-debian packages (some RPMs and what not). How do I safely update my Window Maker menus? If I hand-jam the changes, will the results be wiped out the next time I install a .deb? orwill the changes cause a .deb install (specifically the menu update part) to screw up? Debian normally uses (if you installed the menu package) information provided in /usr/lib/menu, and normal .debs install their information there. After unpacking a new .deb, normally a program named 'update-menus' is called. So, the perfect way to solve your problem is to create a new file called /usr/lib/menu/rpms the same way like the others including the calls to your locally installed rpms, and then call update-menus as root. This way the changes would persist when a .deb calls update-menus after installation. And it should work for all window managers in one run. -- Lukas Eppler (godot)
Re: Crash
On Mon, 26 Oct 1998, Andreas Kahari wrote: At work, if I reboot my Solaris workstation, the last command reports something like rebootsystem boot Fri Oct 23 09:54 ... but at home, with Debian 2.0, it simply states that my system has crashed at a specific time. the 'halt' and 'reboot' commands are probably the hardcore method to bring the system down. On some distributions (maybe debian too, don't want to try out at the moment) 'sbin/shutdown -h now' goes into runlevel 0, goes trough /etc/init.d/* stop, halts then kills the remaining processes. this is the safe way, and would not 'crash'. I saw once that at the end of the shutdown, a 'halt' or 'reboot' is executed at the very end, so on these systems, typing 'halt' is like just pressing the power button. try if your problem goes away by using 'shutdown -r now'. -- Lukas Eppler (godot)
Re: Security problem
On Mon, 26 Oct 1998, Christian Hudon wrote: ... Well, you can also subscribe to debian-security-announce@lists.debian.org Information about every security fix released by Debian is posted there. (To subscribe, send an email to [EMAIL PROTECTED] with the single word 'subscribe' in the subject of the message. And if you're wondering, an announcement about the security-announce list was sent to debian-announce on its creation.) Sorry for having used such a rude tone. It is everything perfect if the stable release gets upgraded in a certain time. We were hacked here recently, on an old suse-linux-box through a pop3server. so I am pretty paranoid at the moment. I remember a discussion here long ago, when cdrom vendors were against changing the stable tree, only because they would end up in selling cds which would have been slightly out of date (if I understood the problem correctly). I remember a bo-updates directory lying around for a long time, without a Packages.gz file, so installations could only be made package by package with dpkg. I have [dists/hamm/main dists/hamm/contrib dists/hamm/non-free] in my selection in dselect. is there a directory to mention to have the security updates quicker than a week, without going slink/unstable? Thanks for your replies. -- Lukas Eppler (godot)
Re: Setting variables for menu access
On Sat, 24 Oct 1998, David Morris wrote: ...CVSROOT... Where do I need to set variables like that so my xdm-started session menus have access to them? I have it in .bash_profile. better than .xsession in my opinion. If more than one user uses it, you could also place it in /etc/profile, i think. -- Lukas Eppler (godot)
Re: mount NFS during boot problem
On Fri, 23 Oct 1998, Sergey V Kovalyov wrote: Hi, I have a wierd problem: during boot process mount fails with the following message: Starting portmapper... Mounting remote filesystems... mount: RPC: Program not registered I saw this once when rpc.mountd and/or rpc.nfsd on the server machine was not up (i.e. on the other computer). So, maybe it is really the missing local portmapper. Find out where it gets started, and take a look at the logs of the other computer, if your machine tries to connect or not. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED]
dinstall
No replies to my last post. Probably it was too long. Compressed: What does a kernel to be used with rescue disks need other than initrd/ramdisk? How can I find out what the problem is when dinstall exits with error code 100? How can I disable scsi bus scans at the boot prompt? Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED]
Re: Security problem
On Fri, 23 Oct 1998, King Lee wrote: The bug is real, and Debian has a fix. See security lists in Debian. If you are running Debian 2.0 you might have a security hole. There was also security problems with bind. The fixes appear in the current distributions (2.0.2 I think) not in package-updates. Why the bloody hell not? Sorry, this makes me angry. Debian does a whole lot on finding these holes, then spreading the information they are there, but then every one has to read at least debian-user or visit the security page on the web to find out. We have such a great distribution system and make no use of it. A pity, really. But thanks for the information. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED]
can't dinstall
Hi, Trying to set up a debian box. There is an odd device built in, it is a scanner card with a NCR scsi chip, but only to be used for the scanner. From the file:///usr/doc/HOWTO/BootPrompt-HOWTO.gz 4.2.16. Pro Audio Spectrum (`pas16=') The PAS16 uses a NCR5380 SCSI chip, and newer models support jumper- less configuration. The boot argument is of the form: pas16=iobase,irq I tried this. With all boot disks I found on the 2.0 Official CD, there is the same behaviour: - when I boot with pas16=0x388,255 syslinux fails to load after writing 'loading root.bin..' - when I boot with pas16=0x388 or without arguments the scsi part of the kernel tries to talk to that odd device and gets confused, scsi0: 1 hosts, but fails with the last message 'scsi command could have been completed before aborting'. - when I boot with pas16=0x220 (which is wrong, just something to confuse the driver) when looking for scsi 'waiting 5 seconds', then 'aborting', then it hangs again. I created a new kernel on my other computer, including initrd and ramdisk and iso9660 and cdrom support, and finally this way i could install up to a certain point up to where I would be able to install the base system. but then it gets very depressing: - I can mount the cd. I can select from a list of directories where the base files reside, but then dselect just exits (error code 100, as Alt-F3 tells me) and starts again black and white, letting me select color mode. - There is no tar program on the rescue disks. Have I forgotten to compile in something desperately needed to install the base system? Can I disable the kernel scsi detection on the boot prompt? Can you tell me what I should try? (I don't want to spend half an hour with the screwdriver when I want to use the scanner.) Helplessly yours -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED]
Re: ftp1.us.debian.org and FTP mirrors
On Thu, 22 Oct 1998, Frock wrote: I'm submitting this as a bug in hope that we might want apt to use a random ftp-site carrying debian when installing - instead of everyone using the same host (ftp1.us.debian.org) Once someone posted this script to find the fastest debian mirror, which could be used in some way in the postinst of apt: #!/bin/sh # Script to figure out the fastest Debian mirror sites zgrep \..*:/ /usr/doc/mirror/README.Debian-mirrors.gz | \ awk '{ print $1 };' | \ awk -F: '{ print $1 }' | fping -ae | \ awk '{ print substr($2,2),$1 }' | sort -n| head -10 I admit this produces a lot of traffic, but just once. Still better than ftp1. I don't know how it should be adapted for apt, tough. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED]
Re: just shoot me
On Mon, Oct 19, 1998 at 04:40:02PM -0700, Steve Lamb wrote: Personally, on production machines that you don't have access to easily enough (for me that is walking into the next room) I'd not even think of upgrading from a stable to a frozen. I'd just ride the stables out. There is one thing which should be possible: keeping a debian box at least secure against attacks. When a security problem is published, there must be a quick fix somewhere, and easily to download. It seems to me that stable is taken a bit too much like 'not changing' than 'not crashing' or 'not vulnerable'... I may be wrong and even paranoid, but I would feel more secure when using unstable when it is at least frozen. I did not do it, however, and with your sysklogd-problems i seem to be right. But, what is the policy with security holes? are they fixed immediately in stable, or will they come in an upgrade dir, like I saw with 'bo-upgrades'? I heard rumors that cd vendors are against every change in stable, which is understandable. Maybe it will get better with apt? Or can I feel secure by using stable? Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED]
Debian on Notebooks
Hi there, I am looking for a new laptop for debian. If you run debian nicely, please let me know. I had some problems installing and running debian on my current laptop, and if I can avoid these on my next buy, I would be very glad. So please drop a note if you are happy with your installation. I would like to buy one with TFT screen 1024x789 w 32k colors in X, if possible with a nice resolution with SVGATextMode. Most information I get is on 'how can I run X on my laptop', I look for information the other way 'round. On Thu, 15 Oct 1998, Infomed wrote: Acabo de comprar un notebook (ayer) y quisiera instalar Linux en él. Podriais aclararme cosas ???. Even tough I did not understand everything (Julián, please try to write english, even if it is difficult for you, people will understand better) it seems the same question. Thanks, -- Lukas Eppler (godot)
Re: BROWSER
On Thu, 15 Oct 1998, count zero wrote: On Wed, 14 Oct 1998, Phillip Neumann wrote: Im searching for graphicals internet browsers... What can you people recomend me?? (currently im using netscape) why u want to change? netscape it's not so ugly ! Hi, You might try kfm, the filemanager of the kde package. However, it's not totally free (like netscape up to now) and has some problems displaying forms. You might try lynx and configure it with zgv as graphical viewer (runs in the console, but is quick as hell). Both have no java, so keep netscape installed. Other browsers I only know the name: amaya, mozilla (from netscape), qtscape (mozilla port on qt-libraries). Gruss -- Lukas Eppler (godot)
No Deadkeys in some X apps
Hi, Everything works in xterm, but in kvt, netscape, lyx I have no deadkeys. Since I use a german keyboard layout, I have no chance to type a ~, for example. (ok, copy-paste from an xterm works, but...) Can someone give me a hint where to look at? Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
lpr to ethernet printer / apsfilter
Hi! When using apsfilter for printing to an ethernet printer the input filter 'if=..apsfilter...' is _not_ run. This has to do with the fact that normally the printer spooler does the filtering. But if I mention the printer itself in the printcap file, like :lp=:\ :rm=195.48.69.8:\ , the printer is introduced as if it was a computer, a printer server. So, the filter isn't run. Now:Someone said here, that one could use two entries in the printcap, so, that one printer filters, and prints it to the other one. I tried this: lp|lp2|PS_600dpi-a4-auto-mono|PS_600dpi auto mono:\ :lp=:\ :rm=localhost:\ :rp=raw:\ :sd=/var/spool/lpd/PS_600dpi-a4-auto-mono:\ :lf=/var/spool/lpd/PS_600dpi-a4-auto-mono/log:\ :af=/var/spool/lpd/PS_600dpi-a4-auto-mono/acct:\ :if=/usr/lib/apsfilter/filter/aps-PS_600dpi-a4-auto-mono:\ :mx#0:\ :sh: raw|lp3|PS_600dpi-a4-raw|PS_600dpi auto raw:\ :lp=:\ :rm=195.48.69.8:\ :sd=/var/spool/lpd/PS_600dpi-raw:\ :lf=/var/spool/lpd/PS_600dpi-raw/log:\ :af=/var/spool/lpd/PS_600dpi-raw/acct:\ :if=/usr/lib/apsfilter/filter/aps-PS_600dpi-a4-raw:\ :mx#0:\ :sh: but this produces an error: $ lpr embassy lpr: connect: No such file or directory jobs queued, but cannot start daemon. $ ps -ax | grep lpd 4117 ? S 0:00 /usr/sbin/lpd 4373 1 S 0:00 grep lpd $ lpq waiting for queue to be enabled on localhost Rank Owner Job Files Total Size 1stgodot 7embassy 649 bytes Or, the same question simple and easy: how can I print textfiles on a postscript-only printer with lpr (avoiding a2ps and such) ? Any help is appreciated. Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Bootdisk for PCMCIA?
Hi, Is there a boot floppy set which supports installation over a pcmcia network card? If not, how is the simplest way to do it? -- Lukas Eppler -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Removing obsolete packages
On Mon, 17 Nov 1997, Dirk Bonne wrote: Lukas == Lukas Eppler [EMAIL PROTECTED] writes: Lukas Hi, When I remove a package, which contains files also Lukas contained in other packages, is it removed? dpkg will refuse to remove a package if it removal would break another installed package (except if you force dpkg) Okk, I deinstalled it and everything still worked. I once removed the pcmcia-cs-binaries after having compiled the pcmcia-source, which is in the main tree (the binaries did not work) and it removed the newly compiled binaries. I assume this was a special case; I didn't want to risk anything, and perl is very important to debian. Thanks for your help. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Removing obsolete packages
On Mon, 17 Nov 1997, Dirk Bonne wrote: Lukas == Lukas Eppler [EMAIL PROTECTED] writes: Lukas Hi, When I remove a package, which contains files also Lukas contained in other packages, is it removed? Lukas Or, if I remove a file from perl-base, is /usr/bin/perl Lukas still there afterwards, because it's contained in the perl Lukas package? Do I risk perl and dpkg to be unable to run by Lukas removing perl-base? dpkg will refuse to remove a package if it removal would break another installed package (except if you force dpkg) Nooo... see what happened io# /etc/init.d/sendmail start /etc/init.d/sendmail: /usr/sbin/start-stop-daemon: No such file or directory io# head -3 /usr/sbin/start-stop-daemon #!/usr/bin/perl -- $version= '1.4.0.8'; # This line modified by Makefile io# ls /usr/bin/perl ls: /usr/bin/perl: No such file or directory That was exactly why I asked, to avoid this! I used dselect, no happy dpkg hacking. grunt. perl-base was marked Obsolete. -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Pentium error
There was a message in my newspaper this morning (Tages-Anzeiger,17.Nov.97, Switzerland, www.tages-anzeiger.ch) Which said the following (translated, my english isn't perfect) - New Pentium bug Intel has confirmed the Bug in their Pentium chips which was discovered recently, with which a PC can crash upon the machine code F00FC7C8. This should apply only to Pentium machines which are running under a multi user system (like Unix). There are already sold defender programs against this so-called F0-Bug, which search programs like virus scanners for the fatal code. -- This sounds like a don't use linux statement from intel. They don't even mention Windows NT, which is a multi user platform, too. And why the hell has this something to do with multi user systems? Is the mentioned code in use with debian linux? Original Text: (german) Neuer Pentium-Fehler Intel hat den kürzlich bekanntgewordenen Fehler in ihren Pentium-Chips bestätigt, mit dem PC durch den Maschinenbefehl F00FC7C8 zum Absturz gebracht werden können. Betroffen sollen allerdings nur Pentium-Rechner sein, die als Mehrbenutzersystem (z.B. Unix) eingesetzt werden. Bereits werden Gegenprogramme gegen diesen sogenannten F0-Bug angeboten, die wie Virensucher die Programme nach dem fatalen Befehl absuchen. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Removing obsolete packages
Hi, When I remove a package, which contains files also contained in other packages, is it removed? Or, if I remove a file from perl-base, is /usr/bin/perl still there afterwards, because it's contained in the perl package? Do I risk perl and dpkg to be unable to run by removing perl-base? The following files seem to be shared: perl, libnet, perl-base, perl-suid, netbase: /usr/lib/perl5 perl, libnet, perl-base, perl-suid: /usr/lib/perl5/i386-linux perl, libnet, perl-base, perl-suid: /usr/lib/perl5/i386-linux/5.00307 perl, libnet, perl-base: /usr/lib/perl5/i386-linux/5.00307/auto perl, perl-base: /usr/lib/perl5/i386-linux/5.00307/auto/Socket perl, perl-base: /usr/lib/perl5/File perl, perl-base: /usr/lib/perl5/auto perl, perl-base: /usr/lib/perl5/auto/DynaLoader Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: win95 fat32
On Sun, 26 Oct 1997, hmaddox wrote: Rather new to linux but a pretty through search found nothing on vfat32 from win 95 osr2 and debian linux. I get an invalid file system responce when i try to mount . ??? This subject was discussed recently; look at the archive at www.debian.org: Subject: fat32 in base-install: archive/latest/16474 Subject: Re: Mouse driver as loadable module?: archive/latest/312 I don't know if it's still there, it should, it was this month. Most informational was this: ... if that's the case, you can re-compile your kernel using the patch found at: http://bmrc.berkeley.edu/people/chaffee/fat32.html ...make sure to read and follow all the instructions on patching your kernel beforehand and which options to choose during config (RTFM! :) ). Seems that the kernels available in stable haven't applied the patch. Probably some people running unstable know more...? Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: dselect problem...
On 26 Oct 1997, Ben Pfaff wrote: David Puryear [EMAIL PROTECTED] writes: With deselect, I have the Access with ftp with right directory and all then I pick update which download new packages files from dists/unstable/main, non-free, and contrib like the way it should. But when I do Install I get this messages: Processing status file... Processing Package files... Couldn't find packages file for dists/unstable/main distribution (re-run Update) ... If I recall properly, I had this problem at one point and solved by upgrading to a newer dpkg-ftp. Try downloading the newest version from hamm with a regular ftp program and installing with `dpkg -i'. Then re-run Update and Install and see if that fixes the problem. Ups, this seems a very brute-force fix. What happens if you clear the list of available packages? or run dpkg --forget-old-unavail and rerun update? or you can try to [not] keep debian directory [y/n] at the end of the Install dialog. Dselect is quite self-reparatory, so fiddling around for a while with it already solved such problems for me. I think forcing someone to use unstable packages when having such kind of problems is not the best way. a lot of people are using stable without these problems. Are you sure you don't need extra packages when using dpkg-ftp from hamm? Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: make dep errors
On Wed, 22 Oct 1997, Thalia L. Hooker wrote: Hi, I am trying to build a new kernel after installing a new base system. After 'make config', I get this error message when I try to 'make dep': gcc -I/usr/src/kernel-source-2.0.30/include -O2 -fomit-frame-pointer -o scripts/mkdep scripts/mkdep.c scripts/mkdep.c:1: stdio.h: No such file or directory ... It's in the following packages: libc5-dev: /usr/include/stdio.h I think its the libc5-dev, and probably some others you need. I had a tar.gz of the stable/binary/devel on which I ran dpkg -i * ... Kind of a brute-force way... :-) Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: new debian user questions
Hi, Daniel Martin pointed out exactly what I wanted to say, but in better words. Thanks. On Wed, 22 Oct 1997, Daniel Martin wrote: Having read the original post, I was under the impression that this would be a desirable state, not necessarily an expression of the current state. On 23 Oct 1997, Manoj Srivastava wrote: Daniel So... Does diety attempt to integrate the configuration tools Daniel that do come with some of the Debian packages? And what about Daniel the question earlier of having the configuration menus that Daniel appear during initial setup available without booting from the Daniel install CD/floppy? Well, if there were a standard interface, Deity might try; but, as it stands, most packages arrange to invoke the configuration utility/script out of the postconfig script anyway; so the config script is run by dpkg with no assistace from Deity. This is clear. It should also be able to do a basic setup without 'dialog' installed. But you know, it's not always running without problems after installing. When the postconfig went wrong, there should be a standard place/method to rerun the config, which should be as clear as installing. I did an install of debian on a laptop of a friend this week. He was not the person to struggle with the system, he just wanted X, octave, lyx/latex and a running pcmcia-network card. I can tell you, I got a red head configuring the basics. Once because of the work, and also because I felt somewhat ashamed to have to configure so many things with vi - for a basic installation. I know where to look, no problem. But it takes time anyway. We all had problems installing the basic system, isn't it? Trying to configure a kernel on a base system to access the cd-rom. To have X running. To be able to deliver mail locally AND outbound. To have colors in xterm, to have ls displaying in color and with slashes after a directory. To let xdm look better. I did not even know there was an xbase-configure (I did try xbaseconfig, which would be the standard name, or not?). Remember, even when Deity is deployed, dpkg shall remain the underlying mechanism; and it will alsways be legal to bypass Deity and use raw dpkg to install packages. We can't, then, do things in a fashion that the configs only work when the installation is run by Deity. I would like to invoke dpkg --setup on a package to invoke the script, or let dselect do that. and maybe dpkg --display-setup to see what is done. Any enhancements to configuring packages, thus, has to be implemented by dpkg. Does this make sense? Yes it does. It is only a calling standard. A definition. It would be simple to follow it for package maintainers, and easy to implement in dselect/diety. Please don't take me too literal, I learned english too late to be able to express everything the right way. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: new debian user questions
On 21 Oct 1997 [EMAIL PROTECTED] wrote: Lukas Eppler writes: This is also better for debian, because every package comes with its own config tool. This is news to me. What config tool comes with cnews? with suck? with pppd? You're right, I should have said it the other way: every config tool comes in its own package :-) It would be great if the packages you mentioned had a *config. I seem to have only these: /usr/sbin/gpmconfig* /usr/sbin/paperconfig* /usr/sbin/kbdconfig* /usr/sbin/sendmailconfig* /usr/sbin/kernel-packageconfig* /usr/sbin/shadowconfig* /usr/sbin/liloconfig*/usr/sbin/tzconfig* It would be nice to have more of them. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: new debian user questions
On Mon, 20 Oct 1997, Will Lowe wrote: On Tue, 21 Oct 1997, Ted Harding wrote: ... I'd argue, in the *nix fashion, that a bunch of little, specific, WELL-DOCUMENTED admin tools is better -- one for net, smail, sendmail, etc... A great example is the tzconfig script -- it's small, does it's work well, and doesn't involve tremendous overhead or hours of work to understand. This is also better for debian, because every package comes with its own config tool. That way, the config tool can be specific enough for the package. But: In my opinion it should be acessible through dselect. Not just once when installing, but always. You'd step through the package you want to configure it. Maybe you can see the actual configuration there without having to run the script actually. But probably they did that already in the diety project in hamm. They're doing a lot of work on that. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: From Slackware to Debian
On Fri, 17 Oct 1997 [EMAIL PROTECTED] wrote: My install Problems. ... Well IT DONT WORK!! How I installed it - ... - I did NOT have the packages.gz file. This is the problem. Dselect needs to have a structure like this: path/to/debian/basedir/ \__stable \__binary-all \__binary-i386 \__Packages.gz \__non-free (with the same structure, optional) \__contrib (with the same structure, optional) You got to prepare a situation like above (You can hack around it with symlinks, I did that today, worked well). The Packages.gz can contain more entries than you have, so don't worry if you leave out X or other packages. But dselect gets all its information from there. (Think about an ftp install, would be too much to scan all directories and then getting the information for all packages separately). Then you step through menuentry 1, 2, 3, and then you got to cycle randomly between 4 (install) and 5 (configure). (*here's a difference between slack and deb: in slackware you configure the package before installing it, in debian you install it and then configure it. Okk, in slackware you config, make, install, config, config, config. *grin* ok, not funny. sorry.) * I copied the boot directory to the dos machine and ran rebooted from there. The Install seemed to run OK. I thought I selected all the required packages. (it seemed it be able to pick up all the available packages even when I did not have package.gz) I thought it would install the the packages with a * character ( in dselect). However when I rebooted it only had 18MB installed and no man no less no nothing. I think it's because there's no Package.gz. You can still select the packages, but they're somewhat like unavailable in this state. when you run 1. [U]pdate list of available packages after setting 0. [A]ccess Choose the access method to use. to mounted, you should be ready to go. I tried running dselect again and zippo nada nothing gets installed. I must be missing something simple What exactly looked the errors you got like? (run dselect | tee if you can't write so fast, there's no logfile *gulp* ) ** Another question is with the NE2000 install. ** An another problem is that once I Installed Linux I could not my DOS partition anymore. The install routine did not give me can't help with this. sorry. anyone? I was also planning to put several partitions on the hardisk there was some documentation with some suggested partition sizes for a 1.6 gig hardisk. I am relatively new to linux. I am asking back: What is bad about one big partition? I have it like that, I am flexible as hell, and have no clear problems with it. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Where is xload program?
On Thu, 16 Oct 1997, Daniel J. Mashao wrote: I am looking for xload package and I don't know where it is? It's in the xproc package. How can I find out where it is w/o having to waste bandwidth? There are Contents.gz files in the same locations as the binary-i386-dirs on every mirror. and secondly why don't they just leave xload alone? You can probably ask the maintainer. Maybe the old all-in-one-package got too big. Maintainer: Helmut Geyer [EMAIL PROTECTED] It seems that the source package is still procps for xproc and procps binaries. It got splitted not long ago. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: man page for nsswitch.conf
On 14 Oct 1997, Dale Martin wrote: Hello. As I have mentioned on this list several times, I have upgraded my machine to a fair amount of libc6 stuff. In general, it seems to be working pretty well, but a lot of the NIS stuff is messed up. I just wanted to remind you that libc6 is unstable and in development. So problems about it should go in [EMAIL PROTECTED] . Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Login's source absent
On Mon, 13 Oct 1997 [EMAIL PROTECTED] wrote: I wanted to see the source of a program login, and modify the greeting, and responses, however, I didn't find it under bo/source/base/[login] Is there any reason for it's absence? And where can I find the source for it? You don't need the source to modify the greeting. It's in /etc/issue for the local logins and /etc/issue.net for the others. You can use ansi sequences for it (colors and everything). Try fiddling around with setterm --color . When you're happy you can use (setterm ; echo...; setterm) /etc/issue to write a happy colorful greeting. Probably cp /etc/issue /etc/issue.old is a good Idea. After logging in, login displays /etc/motd. Can be colorful too, I think. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Problems
On Sat, 11 Oct 1997, Julien Pham wrote: First : my xterm has no color. Or, no, it has colors (when I do a ls --color in it, it works) but when I launch a text based appli (such as deselect for example) in it, it in in B W... Why ? I hassled a lot with this. I have no black-and-white-xterms, never seen. So I copied /etc/termcap/x/xterm-color over /etc/termcap/x/xterm after having made a backup of the black-and-white xterm description. No problems encountered until now. So, every normal xterm is known to be able to handle colors. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mail delivery notification
On Thu, 9 Oct 1997, Richard S. Gray wrote: How do I go about ensuring that an email message has been received? In other words, is there a way for me to get a notification mailed back to me that the message has been delevered and opened? When a mail server is set up properly, you can use the command 'finger [EMAIL PROTECTED]' to see when your recipient checked his mail for the last time. -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
my first package - xvgr
Hi! I just packaged xvgr for debian sucessfully - my first package. I'd like to tell all of you who never did that, but already have compiled software on their own that it is not very difficult. It's almost fun! (xvgr is a plotting tool for X. I am not sure if it is already in hamm, I just wanted to try out how it works.) Now, I need someone who can help me with some questions from time to time. I will have (and already have) questions which are not answered in info or man or /usr/doc. I need some 'daddy' to help me find the way :-) I could go to the debian-devel list, but I am afraid I would bother them with my stupid questions. And, I don't want to upgrade to unstable yet, so I may need someone who helps me with publishing. Let me advertise a bit more: If you think you may be able to debianise a program, try it. Let debian be growing. :-) Have fun, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: non-English Linux use
On Fri, 3 Oct 1997, Santiago Vila Doncel wrote: On Thu, 2 Oct 1997, G. Crimp wrote: #!/bin/sh /usr/bin/setfont iso01.f16 There are a lot of fonts to choose from. See /usr/share/consolefonts. This is somewhat away from the original question, but: When loading one of the iso-fonts, I lose the graphic characters used by dialog, make menuconfig/kernel, mc and others, an output from dialog --yesno hello 5 30 Ú¿ ³ hello³ ô ³ Yes No ³ ÀÙ as an example. when using the original charset, this comes out right, when using the iso font the lines are replaced by big A-umlauts and U-Circles. Probably 7-Bit-restrictions on mailservers prevent the image above to look right anyway, but you see what I mean. On my friends old slakware-linuxbox he manages to produce such dialogs with umlauts in the question...! It worked right out-of-the box. How is it done? What could be the problem? I don't suffer for that, I'm just interested. Thanks -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: mpeg and dat viewer
On Sat, 4 Oct 1997, Wiria Atmadja Kusuma wrote: can anyone be kind enough to point me what package to install in my linux box, so that I can view movie from it? and where can I get it. thanks You can try EIOM Pri Section Package Inst.verAvail.ver Description *** Opt non-free xanim 2.70.6.4-2 2.70.6.4-2 Plays Multimedia files which is on debian. It supports several formats. -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Problem installing Debian 1.3.1 on Thinkpad 760ED
On Fri, 3 Oct 1997, Hans M. Pedersen wrote: Hello ! I have been trying for about two days to install Debian 1.3.1 on a Thinkpad 760ED with XGA graphics - but without any luck :-( When I boot from the rescue disk (yes - I have used the floppy=thinkpad option), the root.bin and linux are loaded - and nothing more. After the newline in 'Loading linux..' the machine simply stops and are completely non-responsive. Try the Tecra Base Floppy Set which should be available on every debian mirror. The standard set contains a bzImage which Thinkpads cannot handle for some unknown reason. So you have to boot a zImage-Kernel. I heard the tecra set had a zImage kernel for such compatibility reasons. If that's not true (you can find out with a 'file *' on the bootdisk) you need to build one. Probably you'll have to copy the pcmcia- and kernel-sources to your laptop by hand afterwards to get pcmcia-support, maybe there is a pcmcia-support binary around for the targa kernel. Otherwise, look into the old messages here (old i.e. last week), there was a discussion about installing pcmcia-support on a 5-floppy-disk-base-installation. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: bo-updates
On Tue, 30 Sep 1997, Alex Yukhimets wrote: ... This would be a good solution, stable stays stable! And creating these sub-directories wouldn't be _that_ hard... or am I missing something? Yes, you are :) You don't need those subdirectories, but just one plain Packages.gz file in bo-updates/ So it would be even simpler!? So let's do it! Who can do it? -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: No Debian updates?
On Mon, 29 Sep 1997, joost witteveen wrote: On Mon, 29 Sep 1997, lukas eppler wrote: Do people who don't want to fiddle around with libc6-stuff wait half a year for an upgrade? ... The excuse we have this time is the libc6 change. This means _every_ package has to be rebuild, and possible modified somewhat, and this takes time. I'd say things are going well, though. Thank you for your kind reply, I lost my temper a bit, sorry. I may have underestimated the effort it takes to build up everything on libc6. Still, I imagine a debian release between stable and unstable. Stable is nice for the space shuttle or other critical purposes. But, let's say, the netscape installer package: There's no bug in the old version. But I think there should be a place outside the developers' corner for the new one ... Somewhere, where dselect can be used. ? Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
bo-updates
Hi! I may be wrong, but: When bo-updates had the same structure as stable, non-free and contrib, one could in fact use dselect to install the bo-updates, just by typing Enter space seperated list of distributions to get [stable non-free contrib]: stable non-free contrib bo-updates This would be a good solution, stable stays stable! And creating these sub-directories wouldn't be _that_ hard... or am I missing something? Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: No Debian updates?
On Mon, 29 Sep 1997, joost witteveen wrote: Im using Debian Linux for about a year now. Is it true that there were no updates in the stable non-free contrib files for 3 (in words: three) months now? (I am using ftp.debian.org as ftp-server). Well, stable is called stable for one reason! But I think there have been updates, only as little as possible. We don't go about adding packages/updates to stable just becase a new version was released, we only do it to fix serious bugs. Who says that? Do people who don't want to fiddle around with libc6-stuff wait half a year for an upgrade? I am afraid of installing unstable, I don't want to spend too much time to find bugs. But If I have to wait until Debian 2.0 is released I don't see the point in these quick and beautiful ftp upgrades. If I always have to use software which is four months behind, I can buy a cd, that's faster. The ftp/dselect method was, why I changed to debian. Why don't you try to have something stable and actual? Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Stable means not-changing?
On Thu, 25 Sep 1997 [EMAIL PROTECTED] wrote: Why not not installing libc6 coexisting with libc5, as described by Scott Ellis´ Mini-Howto which is weekly (?) posted on this list. It proved to be painless for me and has bash-2.01. It is a rather small step, making my system in no way unstable. This is surely a good idea. But you can't use dselect for this, you can't tell the testers of linux distributions to install some 'unstable' packages, whatever that means, before judging the power and actuality of debian before they publish. What would you think if we asked the maintainers to have a distribution between stable and unstable, somewhat like current ? It could be stable, but actual. Including libc6 _and_ 5, running the best packages of hamm, but keeping the basics with debian 1.3.1. When using the whole unstable distribution someone has to be a developer to accept having to fiddle around with the small bugs every now and then. But there could be a way to have a slowly improving system which is not always upgraded by hand. What I want to say is that when we don't use the capabilities of dselect together with dftp regularly, its more convenient to buy cd's every half a year. And when that happens, the debian way of installing software is not much better than the redhat or suse way, which would be a pity. Would it be too much work to maintain a current-distribution? -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Stable means not-changing?
Hi, Please tell me: Does stable mean not-changing or not-segfaulting? Stable was not changing for weeks. I remember times (around Debian 1.2) when stable changed when non-segfaulting upgrades came out. I liked that a lot. It gave me the feeling to have the most-recent not-segfaulting software which is available, something which can't be done by sharing software on cd. Are there no new stable packages, or do I have to wait until a major upgrade is ready? -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: disk activities on a notebook..
On Sat, 16 Aug 1997, David Puryear wrote: On 11-Aug-97 Lukas Eppler wrote: On Sat, 9 Aug 1997, Gonzalo A. Diethelm wrote: On Aug 8, 1997, at 07:05, David Puryear wrote: Does anyone know what is accessing the hard drive every 3 or four seconds? I finally got X to not access disk so much, but I don't know if this is right thing to do. Anyhow, removing cron and at packages did it for me. Now I am happy.g I found out what my problem was here: xntpd, and the always-said-to-be-sleeping apache :-) these two do do a regular access to the disk, even if there is no connection and no network ... is this a bug? The solution leaded into a new problem: I _do_ want to spin down my hd when not used, but now, my bios sets the laptop to sleep, which makes it inacessible from outside. Did somebody have the same problem, and solved it? p.s. If there is reason why I shouldn't remove these [cron, at] packages then please do explain, thanks. On most machines, cron is needed to rotate the log files in /var/log and does various tasks for news servers. Look in /var/log: almost every file has something to do with regular actions. I have cron activated, but it doesn't access the disk too much. Look in /etc/cron.daily and /etc/crontab what is done. You may want to comment out a few lines which are not necessary on your machine. Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: disk activities on a notebook..
On Sat, 9 Aug 1997, Gonzalo A. Diethelm wrote: On Aug 8, 1997, at 07:05, David Puryear wrote: Hi all, Does anyone know what is accessing the hard drive every 3 or four seconds? Let's see, how much physical memory do you have? If it is less than 16 MB, then X is going to swap a lot. That could explain it. I have 24 Megs and almost the same situation. Even shutting down X and killing cron doesn't help... Is there a good way to find out what is accessing the disk? lsof is too much and doesn't say when files are accessed, only which files... There are situations, where my hd spins down, but it managed only to stay calm for at maximum 2 seconds. ?? Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
trick for laptops
Hi, I was not happy with the low brightness of the standard characters on my laptop. On normal screens it may be nice to have only bold characters as bright as they could be, and the others dimmed; but on laptops, it would be better to have all chars really white: It happens a lot that there is too much light around you to read dimmed text. I found a command named setterm: mv /etc/issue /etc/issue.old (clear;echo; setterm -background blue -foreground white -bold on; echo Debian GNU \s 1.3 \n \l ;setterm -background black; setterm -bold on -store;echo;) /etc/issue If you run the above as root, you get a new /etc/issue, which looks nice, and sets the default to really white text. Because this file is printed out before every login process on the console, you don't have to bother which tty's are affected. There is also an issue.net file, which is printed when you log in over the net, so it changes the behaviour of local logins only. There is an /etc/issue file appended here as an example. I was just so glad about having found this trick that I wanted to share it. Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] [H[J [37m[44m[1mDebian GNU \s 1.3 \n \l [40m[1m[8]
supported serial cards?
Hi, Does somebody know a good serial card with at least eight ports, which is simple to set up with Debian? Thanks, -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
xserver-svga 3.3-3 on an lcd
There were a lot of problems I saw... hopefully solved now. My problem is different: I manage to get into X as before, but the server seems to add noise to the image when something changes on the screen. It begins with some stars (single pixels) before even the grey pattern appears, continuously pattering the screen while moving the cursor, like an old tv without an antenna. This happens with several Chipset entries. After some tries even the normal text font begins to show some strange pixels. (My t's showing up here are quite ugly...) What can I do to get my screen back? my old Device config in xf86config: (I used the best match) Section Device Identifier Trident TGUI9420DGi (generic) VendorName Unknown BoardName Unknown #VideoRam1024 Ramdac sc15025 Clockchip sc11412 EndSection this is my new try which shows the same snow: Section Device Identifier Trident TGUI9420DGi (generic) VendorName IBM BoardName Laptop card VideoRam1024 ChipSet tgui9320lcd EndSection this shows up when starting the server: XF86Config: /etc/X11/XF86Config (**) stands for supplied, (--) stands for probed/default values (**) XKB: disabled (**) XKB: keymap: xfree86(us) (overrides other XKB settings) (**) Mouse: type: PS/2, device: /dev/mouse, baudrate: 1200, 3 button emulation (timeout: 50ms) (**) SVGA: Graphics device ID: Trident TGUI9420DGi (generic) (**) SVGA: Monitor ID: My Monitor (--) SVGA: Mode 800x600 needs hsync freq of 48.08 kHz. Deleted. (**) FontPath set to /usr/X11R6/lib/X11/fonts/misc/,/usr/X11R6/lib/X11/fonts/75dpi/:unscaled,/usr/X11R6/lib/X11/fonts/Type1/,/usr/X11R6/lib/X11/fonts/Speedo/,/usr/X11R6/lib/X11/fonts/75dpi/ (--) SVGA: PCI: Trident TGUI 9320 rev 227, Memory @ 0x0800, 0x0820 (--) SVGA: Revision 3. (--) SVGA: Using Trident programmable clocks (**) SVGA: chipset: tgui9320lcd (**) SVGA: videoram: 1024k (**) SVGA: Using 8 bpp, Depth 8, Color weight: 666 (--) SVGA: Maximum allowed dot-clock: 80.000 MHz (**) SVGA: Mode 800x600: mode clock = 40.000 (**) SVGA: Mode 640x480: mode clock = 31.500 (--) SVGA: Virtual resolution set to 800x600 (--) SVGA: SpeedUp code selection modified because virtualX != 1024 (--) SVGA: Using Linear Frame Buffer at 0x0800, Size 1MB (--) SVGA: Using XAA (XFree86 Acceleration Architecture) (--) SVGA: XAA: No acceleration primitives defined. Gruss -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
sendmail and innd broken
I have upgraded to Debian 1.3 and compiled/installed a 2.0.30 kernel. Now sendmail and innd seem to have a problem: $ telnet localhost nntp Trying 127.0.0.1... Connected to io. Escape character is '^]'. then nothing happens for hours. the same with sendmail, but only if I'm disconnected (I connect with a pcmcia 10baseT-card which works perfectly). I start my sendmail with the -bd -om switch, to have my mail waiting in the mqueue until the /etc/pcmcia/network start script executes sendmail -q. this worked perfectly (now I can't test it, because I can't compose mails when I am disconnected.) my /etc/hosts file: 127.0.0.1 io localhost 195.48.69.100 heidelberg heidelberg.fear.ch (heidelberg is the mail gateway for inbound and outbound.) my /etc/resolv.conf file: search fear.ch nameserver 195.48.69.100 # eth0 begin domain fear.ch nameserver 195.48.69.100 # eth0 end My installation is stable. (For the Version numbers). Any help is appreciated. -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
/proc filesystem
There is a nice trick to avoid cut/copy/paste of large text portions: cat filename /proc/1234/fd/3 while 1234 is the process number of a telnet process. fd/3 is the file descriptor for the channel linked to the keyboard. Works perfectly on a suse distribution, seems not to work with debian! I get a bash: /proc/4125/fd/3: Permission denied probably helpful: on debian file gives $ file /proc/4125/fd/3 /proc/4125/fd/3: broken symbolic link to []:184934 while on the suse box it gives [EMAIL PROTECTED]:~$ file /proc/21144/fd/3 /proc/21144/fd/3: broken symbolic link to []:0 any clues? -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Able to boot zImages?
Hi, Less than 1% of machines which are able to run linux won't boot from a bzImage, but from a zImage. (This is what [EMAIL PROTECTED] told me, a german linux distributor.) I have this problem. Debian Rescue Disks base on a bzImage. I would like to install Debian, but the Rescue Disk crashes while loading. (This is the mentioned problem, I think.) If I install a zImage at the place of the old bzImage, the loader tells me that the image was invalid or damaged. Because the loading mechanism is rather different from the one used by Redhat or Slackware I do not know how to prepare such a boot disk. Suse spent a lot of time trying to solve that bzImage-Problem but didn't find a solution up to now. I do not need to solve that problem, but I would like to install debian. Do you know a way to install it without a boot disk? Or can you imagine how to prepare a simple zImage-Bootdisk to work with the other six Debian-Rescue-Disks? Thanks in advance -- Lukas Eppler (godot) http://www.fear.ch telnet://soil.fear.ch: talk:[EMAIL PROTECTED]