Re: dpenedencies in testing totally broken :(
lee wrote: > dependencies in testing are totally broken now :( Do not upgrade the > package lists until this has been fixed! > > If you upgrade your system, it will leave you with a broken system. I > was lucky to get it working to some extend, but dselect wants to > remove many packages I want to keep installed. Did you try if aptitude does a better job? It did so and had absolutely no problem to upgrade. Also make sure that you install the xorg package! Some package are still from xorg-6.9 but none of those should be any harm: libice-dev libxft-dev pm-dev and some small other ones, some not from xorg. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: eth* devices (again)
Rick Reynolds wrote:
> I've posted about this before and there have been several suggestions --
> none of which have seemed to work. A new wrinkle has recently showed up
> that I thought I would post.
>
> I have eth0 (lan), eth1 (wireless), and eth2 (firewire) defined on my
> system. I've had problems for a few months of eth1 and eth2 swapping
> places (and occasionally even eth0 becoming the wireless). I attempted
> the udev rule strings that have been suggested before, but they didn't
> seem to work -- while my rule was in place wireless became eth2 again,
> for instance. Here are the rules I've used before (anonymized a bit):
>
> KERNEL=="eth*", SYSFS{address}=="xx:xx:xx:xx:xx:xx", NAME="eth0"
> KERNEL=="eth*", SYSFS{address}=="xx:xx:xx:xx:xx:xx", NAME="eth1"
This CANNOT work!
If eth0 is taken by something, you cannot assign it again by such rules, so
all of them will fail.
Give your devices descriptive names:
eth_lan0
eth_wlan0
eth_fw0
Those won't conflict with the names that the kernel chooses, so you do not
conflict.
This applies to whatever method you choose to change the interface name.
HS
--
Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de
oder über pgp.net
PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: /etc/hostname /etc/dhclient.conf question
Grant Thomas wrote: > Quick question about the two files mentioned above. > > If I change my hostname, I also have to change a line in > /etc/dhclient.conf to update my hostname on the network. "have to" is definitely wrong. > Is there a Debian Way to synch both of these files? Is there a real reason to do so? Note that /etc/hostname must not represent the name on the current network. This is even impossible at times where more than one network is attached. HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: resolv.conf changing at boot
Don Jackson wrote: > I have a router (D-Link DI-604) and several computers on my LAN. The > router is 192.168.0.1. That is the address the resolv.conf file is > being changed to for nameserver on bootup. By restoring (editing) the > resolv.conf file to the proper nameservers of 214.134.xxx.yyy and > 214.134.xxx.zzz. then everything works fine (no significant delay for > DNS lookup)...until next time I reboot and they're back to 192.168.0.1 > in resolv.conf. (and the 214.134... entries missing). > > Can anyone point me to where/what is happening here? You use DHCP and your DHCP client changes /etc/resolv.conf because new name servers are announced. It is one of the main reasons to use DHCP: getting a working network setup > Maybe I have a > misunderstanding of wherein the problem lies. The DI-604 router shows > the 214.134... nameservers correctly when I look at it. These are > nameservers provided by my ISP. I have not changed anything recently > that I could imagine would affect this. > > (I do not have "resolvconf" installed -- as mentioned in some past > messages on this list.) But installing it would be the solution to your problem as you can define settings that are always present. Another solution would a setup of the router to your preferred settings. Also ask yourself what piece in the chain causes the DNS lookup delay that you experience. Then, either complain to your provider or to the vendor of your router. But people like work-arounds and hacks instead of going the direct route. Your handling of some obscure name server addresses is such one. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgp5B5xzs3yDw.pgp Description: PGP signature
Re: Debian, X and ATI Radeon X1300
Vitaliy Ischenko wrote: > The better choice for linux systems will be Nvidia (latest drivers > [closed-source] support 7400,7800 & 7900 > > P.S. That's my opinion :) nVidia released less information to create free drivers than ATI. Their binary-only drivers are probably good but what about other architectures, what if nVidia stops supporting the chipset you have on your card with drivers for newer versions of X? How about running a kernel without binary-only drivers? Why can't nVidia release documentation for older chipsets? And who needs the newer graphics oven anyway ;) HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: cdrecord + udev
Hugo Vanwoerkom wrote: > 1. introducing udev brought about a change in the transport layer > identification from ATAPI:0,1,0 to ATA:1,1,0. This is plain wrong as udev has NOTHING to do with your problem, really. You said that you also changed kernel version and THAT is the point where it changed. Also, do NOT use dev=ATAPI:x:y:z or dev=ATA:x:y:z with linux-2.6, always use dev=/dev/hd*. I know that cdrecord then prints some bogus messages but it is well-known that the author has a strange focus on the way that Solaris handles stuff. Not every text that cdrecord spits out has an actual meaning. However, the dev=/dev/hd* approach is what works best. > I found this by cdrecord > dev=help and cdrevord -scanbus dev=ATA. The previously posted -scanbus > w/o dev always produced the posted result. Sure, because then it scans the busses that use the SCSI layer. Those include real SCSI, USB, Firewire, SATA and other but NOT the IDE layer (the /dev/hd* devices). This is intentional behaviour. > 2. I realize that I haven't got a clue as to what cdrecord actually > does. Indeed. > But thank you Joerg Schilling for the code and I heard that there > are Debian issues. There are issues with plain cdrecord, too, that its author does not care about. Debian patches cdrecord, reduces some problem but not completely and as a result, the author says that Debian has issues? Judge yourself about him. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: bluetooth
Christoph Simon wrote: > On Sun, 2 Apr 2006 12:56:47 +0200 > Hendrik Sattler <[EMAIL PROTECTED]> wrote: >> obexftp(>= 0.19) or ussp-push > > This is debian stable. current version is obexftp 0.10.7 and no > ussp-push in apt-cache. Does this mean that obexftp in debian stable > is nonfunctional? No, it just hasn't OBEX-Push implemented, only OBEX-FTP. >> To send a file to a phone using OBEX-Push: >> obexftp -H -S -U none -b bl:ue:to:ot:ha:dd file.name > > Here, -HSU are invalid options to obexftp. I will assume (unless told > otherwise) the obexftp version in debian stable to be badly broken, as > the pairing was correctly intiated (as shown in daemon.log) but not > answered. Maybe your phone only implements OBEX-Push and not OBEX-FTP. In this case, the older version cannot work. But don't call it broken just because it's incompatible to your phone. > At the contrary it caused the celphone to reboot (!). Then your cellphone has buggy firmware. Just like a crashing application. > Thank's for the help. I'll search for a backport or try to compile the > newer version my self. Use obexftp-0.18 in this case (same features but for older libopenobex). HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Bluetooth modem
Frusch wrote: > How can I use a 56k bluetooth modem? > > I can find the modem with "hcitool" but a I can't create any connection. You should take a look at the supported bluetooth profiles as further steps depend on this. sdptool browse If it offers Serial Port Profile (SPP), make a connection to is using rfcomm and then use /dev/rfcomm* like a normal serial port. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: bluetooth
Christoph Simon wrote: > I was asked to set up a service which would allow anybody having a > bluetooth enabled device (celphone, pda, etc) to retrieve some file > using a debian box. Unfortunately, I'm really green with bluetooth, I > don't even own such a device myself. Thus I would like to ask if > somebody could give me the missing hint, maybe in form of a quick > roadmap, telling me that I need to set up this or that server, or what > ever. You need the packages: bluez-utils obexftp(>= 0.19) or ussp-push If you want a GUI: kdebluetooth To scan for bluetooth devices: hcitool scan To send a file to a phone using OBEX-Push: obexftp -H -S -U none -b bl:ue:to:ot:ha:dd file.name HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: udev can't create /dev/audio (nForce2 AC97 Audio Controler)
SZERVÁC Attila wrote: > Problem: udev can't create /dev/audio Why should it? see below > however creates it in .static No, nothing is created in .static just as its name says > please, help, what can i do ? How about installing the alsa packages and loading the proper driver? The most sensible thing to do, isn't it? HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: How to capture video with ATI graphics card?
Tim Beauregard wrote: > 6.9.0 and the xorg.conf uses the "ati" driver successfully at the moment) Then you have to take a look at the X server start log in /var/log. To record, you need avview and the km kernel module. And yes, compiling that just plain sucks. However, you may want to try without ffmpeg first to take a look if the video can actually be displayed. And yes: gatos (included in x.org >=6.9) is the only road to that goal, the binary-only driver from ATI doesn't do anything but TV-Out. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: problem with dhcp, might be firestarter related
Juha Pahkala wrote: > Mar 8 13:03:25 server dhcpd: Wrote 0 leases to leases file. > Mar 8 13:03:25 server dhcpd: Multiple interfaces match the same shared > network: lan eth1 > Mar 8 13:03:25 server dhcpd: Bind socket to interface: No such device > > So, somewhere in my config files there's still a dangling reference to > that old 'lan' interface that doesn't exist anymore! I've tried to grep > through all the config file in /etc and /var and /usr/share but haven't > found anything. This may come from the /etc/dhcp3/dhcpd.conf (or something like that). What does your "ifconfig" output say about the network configuration? HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpMqVY3gTis6.pgp Description: PGP signature
Re: Error rebooting after upgrade from 2.6.12-i36 to 2.6.15-i686
Arnau Rebassa Villalonga wrote: >I've installed a new Dell poweredge 850 server with SCSCI disks > formated as ext3 from a netinstall, it run smoothly and installed the > 2.6.12-i386 kernel image. After the aptitude dist-upgrade, I upgraded > the linux-image to the 2.6.15-686 and also to 2.6.15-686-smp (having > with both the same result) and when I reboot the machine it displays the > following errors: > >Alert! /dev/sda1 does not exist. Dropping to a shell! > >/bin/sh: can't access tty; job control turned off > > and a busybox shell appears. Anybody knows how to fix this? I also have this problem and I also have root on a SCSI disk. I guess this is something about this damn initramfs and an driver that does not get loaded. Also udev shows LOTS of errors prior to that. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: description of udev, sysfs, hotplug, hal, etc.
John Halton wrote: > On 2/22/06, Hendrik Sattler <[EMAIL PROTECTED]> wrote: >> However, none of this is in its final state. > > What is? To get the right combination of kernel and udev and hal and dbus somwhat implicates this. Stable interfaces are something different HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: description of udev, sysfs, hotplug, hal, etc.
David Berg wrote: > Does anyone know of a good overview of how all these programs work > together and what function they serve? I'm just looking for a very > coarse look at the programs so I can better understand the docs that > come with the specific programs. hotplug: called by the kernel, loads the proper driver is can run scipts udev: called by the kernel, dynamic filesystem in userspace that relies on tmpfs, lately replaces hotplug hal: makes use of udev to get the events and has lots of rules for lots of stuff, hal uses dbus to address loose clients dbus: makes sending and receiving of events possible (e.g. a bluetooth pin helper, hal and it's device management) There's not much more to it. How the seperate parts work is only important if you have to fiddle with it. However, none of this is in its final state. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: How to prevent kernel upgrade from apt-get?
[EMAIL PROTECTED] wrote: > I am running a couple of OpenMOSIX clusters, and OM depends on kernel > 2.4.26. I would like to use a script to run: > >apt-get upgrade > > on all of the cluster members, but "apt-get upgrade" wants to upgrade the > kernel: > ># apt-get -V upgrade >Reading Package Lists... Done >Building Dependency Tree... Done >The following packages will be upgraded: > kernel-image-2.4.27-2-386 (2.4.27-10 => 2.4.27-10sarge1) This does not upgrade any 2.4.26 kernel. You have 2.4.27-2-386 installed (currently running or not does not matter) and only this kernel image gets updated. This does not harm your current system if your boot loader is correctly configured to boot your chosen kernel by default. If you don't need 2.4.27-2-386, uninstall it. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Midnight Commander does not support UTF8 ?
Žáček Kryštof wrote: > I have recently switched my SID to UTF8 locales. It was as easy as running > dpkg-reconfigure localeconf (great !). > > Everything works fine as before. The only problem I have is the Midnight > Commander displaying nonsense - the columns are not correctly alligned, > characters randomly split around the screen. > > Can it be true that mc does not support UTF8 ? It can and thus mc does fail on UTF-8. Some other distris patch it to be somewhat more usable with UTF-8. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpMjgFeOyBh9.pgp Description: PGP signature
Re: Exim4 & SSL support
Marco wrote: > How to start another exim4 process with your options? > What I must change in my /etc/default/exim4 file? Edit the script in /etc/init.d HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Branded Servers that support Debian GNU/Linux
Rishi wrote: > Already did that and they said that Debian isn't certified, however, > it may work. Sure they say that. > That's why I thought I'd try the Debian User Community and check if > there are servers that support Debian Sarge for at least (a) disk > controllers and (b) network card. That's what you have to ask the vendor. Do not ask for support or if Debian works but ask for detailed hardware specs: means, telling you the model of the 'chips' (not vendor hardware names) used and not just some marketing bla bla. After that, see if 2.4.27 or 2.6.8 support this. If not, you may have to create your own installation media with a newer kernel or install another way (using knoppix and debootstrap or preparing the hard drive in another computer). > high resolution is something that would be nice, but I can do with > 800x600 if I have no choice. Huh? There is absolutely not graphics card out there that limits you to 800x600. Additionally, backports.org features X.org-6.9. And you can always switch to a different card. If you do not use rare RAID-stuff, it should be absolutely no problem to get that thing running. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
RE: RE: IDE PCI Advice Needed
Stan Banash wrote: > That is the path I am currently on. So far I have moved the primary drive > (20 GB) over to the IDE controller (Intel 82371 PIIX4 chipset) on the > motherboard. The 250 GB drive is set as the secondary on the Rocket 133 > IDE > controller. I am getting ready to install the 2.4.27-2 kernel from the > cdrom. I have already compiled the hpt302.o driver using my laptop > running > debian sarge (kernel 2.4.27-2). Once I have installed debian on the > Optiplex, I plan to move the driver over and run the 'insmod' command. > Hopefully that will get the card recognized. In the installer, press Alt+F2 to get a console. You can load the driver from a floppy disk from there. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Exim4 & SSL support
Marco wrote: > I have this in my /etc/exim4/exim4.conf.template file: > > log_selector = +tls_cipher +tls_peerdn > tls_advertise_hosts = * > tls_certificate = /etc/ssl/certs/exim4cert.pem > tls_privatekey = /etc/ssl/certs/exim4key.pem > > When exim is started, if I try a netstat -punta |grep exim4 command I > have this output : > tcp0 0 0.0.0.0:25 0.0.0.0:* > LISTEN 3781/exim4 > Exim don't use the port 465 (for smtps). > > Infact, if I try to connect with an TLS encryption is OK, but if I try > to connect with an SSL encryption (on port 25) > the connection fail. To support SSL directly, you must start another exim4 process with the options -tls-on-connect -oX See the exim4 manpage for details. However note, that older Outlook Express is the only software that needs this. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: system requirements for debian
Bob Hynes wrote: > I have debian running on a 500Mhz box with 256MB ram. It's pretty > slow...slower than windows would be. Is there anyone who know what I > could be doing to speed things up? turning off services, adding ram? I'm > also considering tying another version of Linux, but maybe this box is > just too slow at 500Mhz? "have debian running" is no sufficient description to make any further useful comment on this. Did you install X? KDE? Gnome? Any server software? What kernel version? What version of Debian anyway? Define the meaning of "slower than windows" with something that others can measure, too. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Debian vs. Debian vs. Debian vs. Debian
Felix Miata wrote: > Even after that fix, mc doesn't work right. Its > display on the ttys is subject to character shifting so that you're > rarely exactly sure where you're looking at valid display and where > you're seeing corruption. That's simply a matter of the locale: mc cannot cope with UTF-8 because all the layout calculation probably depend on the assumption that the string binary length matches the display length (I'd call those strlen() and strwidth()). That may be due to the fact that, although many application have this problem, there is not easy-to-use library that automatically handles this in a portable way (and no, the mcs* function are _not_ a solution). HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: redirecting where cucipop looks for mail
Jon Miller wrote: > I've had to move the mailbox (/var/spool/mail) from the installed location > to another location /db/mail. This was done due to space constraints. > Like to know if anyone knows how to reconfigure cucipop to point to > /db/mail instead of /var/spool/mail. Users are not getting their e-mail > in their clients. Don't know cucipop but I guess it's like many others of its kind: take a look at /etc/login.defs for the MAIL_* variables (or read pam_mail.so documentation if you use testing or unstable). In your case, set MAIL_DIR to /db and relogin. I'd suggest a maildir in the users home (if cucipop support that), e.g. /home/user/Maildir. In this case, set QMAIL_DIR and MAIL_FILE to Maildir/ and do not set MAIL_DIR. You also need to add the line to /etc/pam.d/common-session: session optionalpam_mail.so nopen dir=~/Maildir HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: hotplug problem
John and Holly Klug wrote: > Package: hotplug > Version: 0.0.20040329-22 > Kernel: 2.6.8-2-386 > Every time I connect my SanDisk 12 in 1 USB flash reader or disconnect > it, I get many stuck hotplug processes that I must kill. > > About 30 processes get stuck on each connect and disconnect. > > I use the flash memory on a KVM switch, and it is very annoying. Also, > the hotplug autodetect doesn't work for the flash memory, but I can > manually mount it. > > root 14510 14494 0 02:17 ?00:00:00 > /etc/hotplug.d/default/20-hal.hotplug scsi Does removing the "hal" package help? I found that hotplug+hal is a very unsafe combination in Debian stable. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpmCOJi0kz25.pgp Description: PGP signature
Re: k3b.
Mauro Sanna wrote: > Why there aren't k3b and konq-plugins in debian testing? Because of the c2a transition. You can use the version from unstable, currently. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgp2ahdahCPEo.pgp Description: PGP signature
Re: bug 336022 closed but not fixed(broken binutils)
NZG wrote:
> I'm having the same problem described in bug 336022.
> After upgrading binutils to the most recent testing version I cannot
> compile any kernel other than 2.6.14.
> This bug was closed with the recommendation to upgrade you binutils and/or
> use a more recent kernel.
>
> I have done both, but the problem remains.
>
>> standard input}:688: Error: suffix or operands invalid for `mov'
>> {standard input}:689: Error: suffix or operands invalid for `mov'
>> {standard input}:887: Error: suffix or operands invalid for `mov'
>> {standard input}:888: Error: suffix or operands invalid for `mov'
>> {standard input}:967: Error: suffix or operands invalid for `mov'
>> {standard input}:968: Error: suffix or operands invalid for `mov'
>> {standard input}:1051: Error: suffix or operands invalid for `mov'
>> {standard input}:1052: Error: suffix or operands invalid for `mov'
>> {standard input}:1061: Error: suffix or operands invalid for `mov'
>> {standard input}:1073: Error: suffix or operands invalid for `mov'
>
> I have tried using gcc-2.95, gcc-3.3, and gcc-4.0.
> They all do exactly the same thing.
Same problem here when trying to compile 2.4.25 for some older stuff. Using a
Sarge chroot fixed it. 'as' just complains more and that's the problem.
Either you stick with older binutils or you have to fix the process.c file in
the kernel tree by fixing the inline asm. Maybe there is a patch out there?
HS
--
Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de
oder über pgp.net
PingoS - Linux-User helfen Schulen: http://www.pingos.org
pgpDhoT5cUytn.pgp
Description: PGP signature
Re: wacom graphire4
kangja wrote: > thinking about getting a Wacom Graphire4. > I am using 'testing' with kernel 2.6.8 and XFree86 4.3. > Getting the pen/tablet to work would be a staightforward process? If the USB ids did not change, it may not actually differ from the functionality of a Graphire3. And linux-2.6.8 already supports a Graphire3. I am not so sure about XFree86-4.3, though. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpSl9kn2ujc4.pgp Description: PGP signature
Re: Getting the Euro symbol on to my keyboard
Andrew M.A. Cater wrote: > On Sat, Dec 17, 2005 at 08:22:19AM +, Alan Chandler wrote: >> I want to get the ability to generate the Euro symbol from my keyboard. >> The symbol itself is etched in to the "4" key (on the bottom right). Where? In X or in a VT? >> I found this >> >> http://www.debian.org/doc/manuals/debian-euro-support/ch-configure.en.html >> >> which is (a probably somewhat old) howto on adding the symbol. >> >> Apart from the fact that is assumes that you should use 'AltGr e' for the >> symbol (it points to a non existant EU document), I fall at the first >> hurdle, because it suggests running dpkg-reconfigure -plow locales and >> selecting your locale with the @euro after it. >> >> en_GB (being a UK resident) is what I want to set, but there is no @euro >> option available. That does not really matter, even the other @euro locales do make no sense anymore since 2002 (the Euro is the only sensible choice for those countries anyway). Choose an UTF-8 locale or something with latin9 (ISO-8859-15). In X, you then have to adjust your keyboard mapping to use the EuroSign symbol, e.g. with xmodmap. > Cheat - use the Irish keyboard layout. You want 8059-15 anyway. > You _can_ hand hack the keyboard mappings: I seem to recall > that the authors of the Spanish package to hack the Euro onto > Spanish keyboards had details. UTF-8 will also do it. > [Emacs under X displays ¤ as a Euro and alt-gr c as a ¢ but will > also still display $ and £] My emacx-nox in a KDE konsole displays the € just fine. If it displays an ¤ for you, then you must change the font. In an UTF-8 locale, this should not happen at all. With ISO-8859-1(5), ¤ and € have the same byte value. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpZOh6gOdqmj.pgp Description: PGP signature
Re: scsi device naming / driver built in kernel 2.6.x
Gebhardt Thomas wrote: > On Monday 12 December 2005 14:52, L.V.Gandhi wrote: >> There is a work out with udev See the article in >> http://www.debian-administration.org/articles/126. >> same principle can be applied for your USB hard drive. > (should probably read > ... can be applied for your SCSI hard drive) > > Sorry, I forgot to mention that I'm booting and mounting / > from /dev/sda. As far as I can see the udev solution needs > some user space tools that are available only after the boot, > i.e., after the disks have already been named. Or did I miss > something here? But there is an alternative: compile the host-adapter with the rootfs disk _into_ the kernel. This way, it will always be the first host adapter. Then, compile the other host adapter driver as module. udev or hotplug will automatically load it. If you know what the system is, only _very_ rare cases really need an initrd. Most system can go with without an initrd and thus with less pain. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpDIWLPxdA2a.pgp Description: PGP signature
Re: DMA not permitted since 2.6.14
David LIMA wrote: > Le dimanche 11 décembre 2005 à 23:52 +0100, Hendrik Sattler a écrit : >> David LIMA wrote: >> > >> > I'm running Debian SID with linux-image-2.6-686, since few days i'm >> > facing DMA problems after upgrading kernel to 2.6.14 (from stock). >> > Hdparm reports "operation not permitted" when I try do activate DMA. >> > >> > I have read http://lists.debian.org/debian-user/2005/11/msg00410.html >> > and recompile my kernel but still same issue. >> [...] >> > /dev/hda: >> > UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 udma6 >> > >> > /dev/hdc: >> > UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 udma6 >> > * signifies the current active mode >> > >> > Has someone any clue about this problem that slow down my system ? >> >> Actually, no clue what you are complaining about: both drives run with >> UDMA5, there not much more what you can enable with respect to DMA >> transfers. > > For sure, DMA is not running, my system is very slow each time I access > my disks (CPU eating). What IDE hardware do you have (lspci), what modules are loaded (lsmod). Maybe you use a generic IDE driver instead of one for your chipset. Since yaird or initramfs-tools are used with Debian's 2.6.14, which one are you using to create the initrd or initramfs content? HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpENW2EzuRgz.pgp Description: PGP signature
Re: DMA not permitted since 2.6.14
David LIMA wrote: > > I'm running Debian SID with linux-image-2.6-686, since few days i'm > facing DMA problems after upgrading kernel to 2.6.14 (from stock). > Hdparm reports "operation not permitted" when I try do activate DMA. > > I have read http://lists.debian.org/debian-user/2005/11/msg00410.html > and recompile my kernel but still same issue. [...] > /dev/hda: > UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 udma6 > > /dev/hdc: > UDMA modes: udma0 udma1 udma2 udma3 udma4 *udma5 udma6 > * signifies the current active mode > > Has someone any clue about this problem that slow down my system ? Actually, no clue what you are complaining about: both drives run with UDMA5, there not much more what you can enable with respect to DMA transfers. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: What is the use of the [EMAIL PROTECTED] locales?
Gonzalo HIGUERA DÍAZ wrote: > I was wondering whether there is any difference between the .utf8 and > the [EMAIL PROTECTED] locales. > > I thought @euro was used initially so that classic character sets > (such as Latin1) would not be trampled by modified ones made in order > to incorporate the euro character and other minor modifications (for > example Latin9, a euro-including Latin1 modification). However, > Unicode/ISO-10646 (and thus UTF-8), incorporates the euro sign, so I > am at a loss as why this tag is needed. What am I missing? See http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=274491 HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpCdte0pbbp1.pgp Description: PGP signature
Re: No package signature check for specific repository
Georgi Alexandrov wrote: > I'm using debian unstable and the main debian unstable repositories to > update my system. > I have added the ftp://ftp.nerim.net/debian-marillat/ repository to my > sources.list file. > Works great, but every time I deal with apt-get or aptitude I get > notified: > > "W: GPG error: ftp://ftp.nerim.net sid Release: The following signatures > couldn't be verified because the public key is not available: NO_PUBKEY > 07DC563D1F41B907" > > I want to turn off gpg check fot that particular respository, is that > possible? Why don't you just think about "because the public key is not available", get that key and make it available? Read "man apt-key". HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpBfethEKMmd.pgp Description: PGP signature
Re: Graphire3 under debian sarge
Henk Boom wrote: > I've installed the 2.6 kernel in sarge, then (after reinstalling my > nvidia drivers) did "m-a auto-install wacom" to install the wacom > module, then rebooted. Now, instead of acting properly as a USB mouse, > when I move the pen close to the tablet the cursor jitters wildly up > and down. "xsetwacom list dev" still yields no results. The tablet > still works properly as a regular USB mouse in 2.4. From drivers/usb/input/wacom.c: *v1.30.1 (pi) - Added Graphire3 support It may be possible that the version in 2.6.8 is older than that but backports.org has 2.6.14 available (you also need to get "yaird" from there in this case), 2.6.11 hat version 1.4.0. You must also _NOT_ use /dev/input/mice any longer because in this case, you will get all events twice. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpFYgm7JleRs.pgp Description: PGP signature
Re: CUPS and samba intergration problem
Am Donnerstag, 1. Dezember 2005 22:12 schrieb Tom Vier: > On Tue, Nov 29, 2005 at 07:27:34PM +0100, Hendrik Sattler wrote: > > Did you try without the last three command lines? Ususally, you do not > > need to set anything like that for cups, there is special support for > > cups in samba and it would be bad it did not have the right commands for > > all that. > > When i try without that, and comment out "passwd backend" (set to "guest" > before, which i don't think is valid) it gets this far: > > > [2005/12/01 16:09:32, 0] printing/print_cups.c:cups_queue_get(790) > Unable to get jobs for ipp://localhost/printers/printers - > client-error-not-found Again, this points to an inaccessable CUPS (that implements the Internet Printing Protocol) on the localhost network interface (127.0.0.1). Either this is not present (must be) or CUPS does not listen on localhost (also a must) or a firewall causes this (does not really make sense on localhost) or there exists not CUPS printer entry with the name "printers" (that most likely your problem because this sounds like a strange naming) > [2005/12/01 16:09:38, 1] smbd/service.c:make_connection_snum(642) > triad-laptop (192.168.255.82) connect to service printers initially as > user nobody (uid=65534, gid=65534) (pid 24695) > [2005/12/01 16:09:39, 0] printing/print_cups.c:cups_job_submit(643) > Unable to print file to printers - client-error-not-found > [2005/12/01 16:09:43, 1] smbd/service.c:close_cnum(830) > triad-laptop (192.168.255.82) closed connection to service printers Those are only there because of the failure. Again, check "netstat -lp" or "netstat -lnp" and make sure that CUPS listens on localhost. You can also check by opening http://localhost:631/printers yourself. Normally, problems like this a straight-forward to solve with some common sense (example here: thinking about the URL from the error message above). HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgp0mK6L91FHo.pgp Description: PGP signature
Re: Graphire3 under debian sarge
Henk Boom wrote: > *I'm having trouble installing my Graphire3 Wacom Tablet. I've installed > wacom-kernel-modules-2.4.27-2-386 after having it compiled by > wacom-kernel-source, and I have installed wacom-tools, but even after a > reboot my tablet seems only to be recognized as a regular USB mouse. I'm > using discover1 (not udev), and I don't know how to go about making this > work. I might also add that my regular mouse is plugged in as a USB mouse. > * HINT, HINT: current linux is at version 2.6.xx. Using a stone-aged version like 2.4.27 might _not_ work with current hardware. Clear text: install the 2.6.xx kernel, a package kernel-image-2.6-* should work (* might be 386, 686, k7 or another available sub-arch). If that does not work, look out for an even more recent kernel (current is linux-2.6.14.3). HS pgpxLkRder2Pr.pgp Description: PGP signature
Re: Debian Wireless Networks
Mark Crean wrote: > I recently went through a > similar experience with a laptop and an atheros-based card. First stop > was the Madwifi project's drivers but they didn't work very well. Depends on which driver you tried. Madwifi works fine here, only wpa_supplicant had to be recompiled to use WPA/WPA2. However, a pity that it is not a fully free driver (there is a non-free HAL coming with it). HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpoPdWfBUwop.pgp Description: PGP signature
Re: autoconf macro directory
Fritz Wettstein wrote: > I'm trying to build a driver from sources I've downloaded from > ftp://download.canon.jp/pub/driver/bj/linux/. The autogen.sh reports: > Please add the files > codeset.m4 gettext.m4 glibc21.m4 iconv.m4 isc-posix.m4 lcmessage.m4 > progtest.m4 > from the /usr/share/aclocal directory to your autoconf macro directory > or directly to your aclocal.m4 file. > > What's my autoconf macro directory? What does the sentence "directly to > your aclocal.m4 file" mean? If you really want to rebuild the build scripts (that should already be present, BTW) with that above, you need to check the following things: aclocal automake autoheader autoconf aclocal ist the tool for your question above. It usually does not need any options. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpnhKOHUZq2R.pgp Description: PGP signature
Re: CUPS and samba intergration problem
Am Donnerstag, 1. Dezember 2005 22:12 schrieb Tom Vier: > On Tue, Nov 29, 2005 at 07:27:34PM +0100, Hendrik Sattler wrote: > > Did you try without the last three command lines? Ususally, you do not > > need to set anything like that for cups, there is special support for > > cups in samba and it would be bad it did not have the right commands for > > all that. > > When i try without that, and comment out "passwd backend" (set to "guest" > before, which i don't think is valid) it gets this far: > > > [2005/12/01 16:09:32, 0] printing/print_cups.c:cups_queue_get(790) > Unable to get jobs for ipp://localhost/printers/printers - > client-error-not-found > [2005/12/01 16:09:38, 1] smbd/service.c:make_connection_snum(642) > triad-laptop (192.168.255.82) connect to service printers initially as > user nobody (uid=65534, gid=65534) (pid 24695) > [2005/12/01 16:09:39, 0] printing/print_cups.c:cups_job_submit(643) > Unable to print file to printers - client-error-not-found > [2005/12/01 16:09:43, 1] smbd/service.c:close_cnum(830) > triad-laptop (192.168.255.82) closed connection to service printers Looks like that CUPS does not listen on where you want to address it. What is "netstat -lp" (or "netstat -lnp") is telling you about the situation? HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpIbo1jbeqM7.pgp Description: PGP signature
Re: CUPS and samba intergration problem
Tom Vier wrote: > in the cups log, i see this: > get_jobs: resource name '/printers/lpstat -o printers' no good! > > This is my smb.cnf: > > [printers] > printing = cups > comment = All Printers > path = /var/spool/samba > printer admin = root > guest ok = Yes > printable = Yes > use client driver = Yes > browseable = yes > writable = no > create mode = 0700 > print command = lpr-cups -P %p -o raw %s # using client side printer > drivers > lpq command = lpstat -o %p > lprm command = cancel %p-%j Did you try without the last three command lines? Ususally, you do not need to set anything like that for cups, there is special support for cups in samba and it would be bad it did not have the right commands for all that. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: DHCP with Static Addresses
Kenneth Jacker wrote: > However, what I'd like to do somehow is make my RFC1819 addresses > remain the same at home The big question is: why do you actually care for the address? There are only rare cases where this is actually needed for a client. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpiqsukMnRj7.pgp Description: PGP signature
Re: 2.6.14 Problems (was Re: sarge: 2.6.14 kernel panics on boot)
David Baron wrote: > On Sunday 06 November 2005 19:32, > [EMAIL PROTECTED] wrote: >> Thats because devfs is creating the nodes in kernels < 2.6.14. >> With 2.6.14 DEVFS is no longer supported in the kernel, so you have >> to use another method to create device nodes like udev (the new way) >> or MAKEDEV( the old way, before devfs). > > Could this be why alsa audio is not working in this kernel? > > For those having trouble booting, my advice: Make your own initrd using > the old and trusty mkinitrd rather than the new yaird stuff. If you build your own kernel, there is rarely a reason to use an initrd. And if you don't use one, you don't have problems with it. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: SATA DVD not recognized
Basajaun wrote: > Somehow I fail to see how changing the source file of a module one has > to subsequently recompile can qualify as "works pretty well". Damn, we > are not in the dark ages were Linux users were suposed (required) to be > computer geeks. Compiling something (and compiling a kernel ist just that despite all stories that are made about it) is not geek at all. Especially since all you have to do is reading the manual that hopefully comes with it. > I am lately having some bad experiences with Debian, which, although > made me learn a lot, where a pain, and I keep seing people have similar > problems, mainly with SATA, SCSI emulation, 2.4 vs 2.6 kernels and the > system being able or not to see the HD or CD-ROM: > > 1) Debian can not be installed on a SATA HD from scratch with a 2.6 > kernel. Rather, one has to install with kernel 2.4, _then_ upgrade the > kernel. This is so because, under 2.6, SATA is seen as SCSI, and > somehow this makes the install program not see the IDE CD drive it is > being installed from. Worked here just fine: SATA disk and PATA dvd drive, installing sarge with linux-2.6 > 2) Booting on 2.4 gives SATA drives /dev/hdaX names. Booting on 2.6 > gives /dev/sdaX names. Result: /etc/fstab hell. I think that fiddling > with grub/menu.lst can fix it, but this shouldn't be necessary. You could use LABEL= directives in /etc/fstab to easily solve such problems. Or use symlinks (udev can help you here for linux-2.6). > 3) Unless some modules (ide-core, etc.) are inserted in the kernel or > included in the initrd.img, when booting on a SATA drive, the IDE > devices will go unnoticed, because the ata_piix module hijacks the ide0 > and ide1. > > All this can be "easily" fixed... but gr, it shouldn't be required! > I want Debian to boot and install flawlessly on both IDE and SATA, with > eider kernel 2.4 or 2.6. And this works. But what gave you the idea that both at the same time are supported, especially in a bit more complicated environments like yours? HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpYf6fiNXuL2.pgp Description: PGP signature
Re: Wacom Graphire 3 (usb) under Debian Sarge
Wolfgang Qual wrote: > recently, I got a nice Wacom Graphire 3 (usb) graphic table. As far as I > know, the driver for this device is already integrated in the Debian > Kernel.I adapted my XF86-Config-4 after having read quite a few pages on > the internet. However, it is still not possible to *work* with the graphic > table: the cursor is jumping (or a jumping white border is visible) around > when I hold the stylus pen still on the table! Did you read the wacom(4x) manpage? Look for Option Suppress there. > Also, klicking on the menue buttons will not work properly. Probably because of the steady movement. > Section "InputDevice" > Identifier "stylus" > Driver "wacom" > Option "CorePointer" > Option "Type" "stylus" > Option "USB" "on" > Option "Threshold" "10" > Option "Device" "/dev/input/wacom" > Option "Mode" "Absolute" > Option "PressCurve" "0,5,95,100" > Option "zMin" "0" > Option "zMax" "512" > Option "KeepShape" "on" > Option "debuglevel""10" > EndSection I did not use most of the options that you have above. Some of those are not even mentioned in the wacom(4x) manpage. Maybe those are a cause of your problem? You can only have one "CorePointer", but you can use the AlwaysCore option for the other ones. Not that if you do _not_ make them corepointer, you get some nice support in gimp (but also as corepointer). Do not use /dev/input/mice while using the wacom or all events from wacom input will appear twice. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org pgpW68MgNGhj7.pgp Description: PGP signature
Re: sound, and floppies
Mark Grieveson wrote: > Hello. I work in a shelter, and we obtained a donated computer which I > set up. The computer is a Pentium II, 350 MHz, with 192 mb ram. It has > the default install, with kernel-image-2.4.27-2-386 (and headers), > alsa-modules-2.4.27-2.386, and the default gnome desktop. Shisen-sho, > mahjongg, and supertux are a hit. However, there are some problems. > > I cannot get sound working properly. Upon booting, it indicates that > alsa is started, but "(not loaded)". Sound works in gnome applications > like gnometris, and on the internet, and other games, but when I load a > kde program (guarddog, or kolf) I get an error message indicating that > the sound server is pissed off, and is using the null output device (or > something) instead. Then, sound works in nothing. After running > alsaconf, I get sound working in the net, and working in games which are > not associated with either kde or gnome (in those games, the sound does > not work at this point, however). What's up with this? Shouldn't sound > work in all the programs? (my co-workers all suggested to me that > Windows 98 would be better, but I insisted on Debian. This does not > strengthen my case). Did you install arts? If yes, manually set to to alsa output in kcontrol. The sound driver autodetection in arts _never_ worked for me. And install Windows98? Today? With Debian at hand? Even Windows2k would be a better choice than Windows98! And Debian is surely an even better one :) > Floppies are difficult to mount. Only a few can be mounted in the Debian > computer. All of them were successfully read in the Windows computer at > work. Any hints? All responses appreciated. This might depend a lot on the floppy disk drive. You might want to test the one from your work with the computer that has debian on it. Additionally, you might want to use the mtools for easier access to floppy disks (see the floppy:/ target in konqueror). HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Moved to sarge: Great, but no sound!
Louis Woods wrote: > I don't know if this will solve your problem, but it worked for me: > After "alsactl store" didn't help (actually it maybe did, but I didn't > notice - see below) I started playing around with the sound plugins of > XMMS. Sidenote: all the time I was only testing sound with the > XMMS-Player. Anyway, I had a look at the plugins under > Options/Preferences/Audio IO Plugins/Output Plugin and saw that the > Output Plugin was set to "ALSA 1.2.10 output plugin [libALSA.so]". I > changed the Output Plugin to "eSound Output Plugin 1.2.10 > [libesdout.so]" and since then it all works they way I want. I don't > know why the alsa plugin isn't working, but for me this solved the > problem. Also Gnome-Sounds are working and rebooting isn't a problem > anymore. Sure, if you use an ENS-1371 card, that only support one channel, and you run ESD, then a normal ALSA output will wait. In this case: a) use alsa's dmix plugin b) use esd (as above) c) disable esd For the gnome-mixer effects: maybe it is still set to the old driver and needs a configuration reset. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Hotplug, udev, hal, etc.
jjluza wrote: > Anyway, Sid is not intend to work perfectly everywhere. it is intend to > test and stabilize new things. For people who want to have a stable > system, they should use Sarge. Me waits for the day that this major change gets to testing. Then it's not at unstable anymore. As I said before, I don't want to use udev because I simply don't need and I don't want a dynamic /dev. I would be perfectly fine with udev IF I could set it to a static /dev at installation time. No, reading README.Debian is not enough, I don't even want udev to start with its dynamic stuff. Why do we have debconf if not even a package like udev uses it? HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Hotplug, udev, hal, etc.
jjluza wrote: > Udev without hotplug works well here with kernel 2.6.13. Read again! 2.6.13 is greater than 2.6.12, so you completely missed the point! > All things keep on working very well : sound, nvidia, tv, cpufreq, usbkey > and others ... So, what the problem is ? Before hitting the maintainer, > you should give it a try first, shouldn't you ? "works here" never was a good argument. Did you try with linux-2.4.30 and linux-2.6.11 parallel-booting with linux-2.6.13? Additonally, why do you carry this to another list? At least you should ask if the others involved persons do read the other list! Additionally, your quoting style is called TOFU (text above, fullquote below). I stay with my opinion that the current udev package is an unacceptable drawback regarding compatibility because having both installed shouldn't be that hard to do: use hotplug for 2.4.xx and ask the admin what to do for 2.6.xx. HS > Le Dimanche 9 Octobre 2005 12:36, Hendrik Sattler a écrit : >> Am Sonntag, 9. Oktober 2005 11:57 schrieb Anders Ellenshøj Andersen: >> > Udev in unstable now conflicts with hotplug. From what I can gather >> > udev has eaten up all of hotplugs functionallity, but I don't really >> > have a clue about what is going on. Since some important KDE stuff >> > depends on all this to work, I'd sure like if someone who knows this >> > stuff could explain what the h... is going on. >> >> Sure, by upgrading, you loose all abilities to run linux-2.4.xx with >> hotplug functionality. You also agree to not run a linux kernel <2.6.12 >> because that won't work, too. Actually, you should run linux kernel >> 2.6.14 or higher. I know, doesn't exist, yet but that seems to be the >> attitude of the udev maintainer. >> Obviously, he lives in a perfect world, where all external kernel modules >> adapt to kernel changes very fast. >> >> Obviously, some DDs (e.g. the udev maintainer) gave up to make it >> possible to have more than one major kernel version installed (and still >> be able to run it properly). >> >> The udev package somewhat digs its own hole and I am glad that I did not >> change to udev. The udev package could also ask if you want to use >> hotplug or udev's coldplug, does it? (Can't tell because I don't have it >> installed).
RE: Basic routing problem
Peter Coppens wrote: >> I assume you missed to add a route on R for the net of A pointing >> to B. > Yes...that is probably what is wrong. Problem is I don't have enough > privileges on the router to do that. Seems I am stuck, sigh. You can do NAT for A on B or install a proxy on B. HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Moved to sarge: Great, but no sound!
Hendrik Boom wrote: > I've been told that hotplug is specific to USB devices, that can be > plugged in and out dymanically. This is wrong. Why do you rely on such false information instead of just taking a look?: $ ls -1 /etc/hotplug | grep agent dasd.agent firmware.agent ieee1394.agent input.agent net.agent pci.agent scsi.agent tape.agent usb.agent Looks a lot more than usb to me and in fact, it DOES more. > discover detects the devices that are less dynamically attached (such as > IDE drives and sound card and the like, and makes wure the right modules > are installed. But it only cares for modules, not for actions that might be assigned to them, like not just loading the ethernet card module but also bringing the device up. On every mashine that I freed from discover insanity, I had a few problems less. > They do different things. One looks actively (discover), the other is told by the kernel (hotplug). HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Moved to sarge: Great, but no sound!
Louis Woods wrote: > I just got sound working on my desktop. I first installed the following > packages: > > 1) alsa-base > 2) alsa-oss > 3) alsa-utils [...] > Now my only problem is that (same as above) I have to rerun alsaconf > everytime I reboot. Any solutions to how this can be fixed would be grand. Well, why should anyone here guess for you? You give NO details about loaded modules before/after alsaconf or anything else! > Above it was mentioned to install "discover". When attempting to do so > apt wants to remove "discover1". What is the difference between > "discover" and "discover1"? apt-cache show discover discover1 Ususally, getting rid of both of them fixes numerous problems: discover just sucks and you ususally do not need it. WHY would you want to do the same damn detection on every boot? Hotplug is ususally enough because you also use it at runtime and it does most of the things that discover does, too, only more integrated with the system. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: very slow usb 2.0
Am Sonntag, 25. September 2005 22:26 schrieb Roberto Winter: > First of all I do have hci-hcd, ehci-hcd, and usb-storage compiled in the > kernel. Not as modules though, but that doesn't make a difference. > The thing is that I also hav UB compiled and it is not a module, so I can't > simply unload it. Is there a way to tell the kernel not to load it? > How do I use hotplug for that? Can someone point me a good reference > maybe?... Can I use modutils blacklisting in that case? In this case, you have to recompile your kernel. One more point for using modules. One more point for only compiling in stuff that you know that you'll need it. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: very slow usb 2.0
Am Sonntag, 25. September 2005 15:28 schrieb John covici: > OK, possibly stupid question -- how do I prevent that driver from > being used when the hot plug subsystem would normally use it -- and if > I do prevent that will I mount /dev/sda,b,c ... instead? I have not > noticed a slowdown, but I didn't test very much either. use either hotplug or modutils blacklisting HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: ipw2200 compile problems + 2.6.12.6
Marco Calviani wrote: > i'm trying to compile the ipw2200 module (from Debian ipw2200-source > version 1.0.6-6) using a vanilla 2.6.12.6 kernel version. > I've compiled and installed ieee80211 (from Debian ieee80211 version > 1.0.3-3) correctly. [...] > /usr/bin/make -C driver KSRC=/usr/src/linux KVER=2.6.12.6-050925-vanilla > IEEE80211_INC=/usr/include/ieee80211 > make[3]: Entering directory `/usr/src/modules/ipw2200/driver' > mkdir -p /usr/src/modules/ipw2200/driver/tmp/.tmp_versions > cp /usr/include/ieee80211/*.mod > /usr/src/modules/ipw2200/driver/tmp/.tmp_versions > cp: cannot stat `/usr/include/ieee80211/*.mod': No such file or directory > make[3]: [modules] Error 1 (ignored) > /usr/bin/make -C /usr/src/linux M=/usr/src/modules/ipw2200/driver > MODVERDIR=/usr/src/modules/ipw2200/driver/tmp/.tmp_versions CC="gcc-4.0 -I > /usr/include/ieee80211" modules > make[4]: Entering directory `/usr/src/linux-2.6.12.6' > CC [M] /usr/src/modules/ipw2200/driver/ipw2200.o > /usr/src/modules/ipw2200/driver/ipw2200.c: In function 'ipw_adhoc_create': > /usr/src/modules/ipw2200/driver/ipw2200.c:5459: warning: implicit > declaration of function 'ieee80211_channel_to_index' Did you even check that /usr/include/ieee80211 exists? HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: very slow usb 2.0
Roberto Winter wrote: > ub(1.19): GetMaxLUN returned 0, using 1 LUNs > uba: uba1 Do NOT use the ub driver! Use the usb-storage driver instead. The ub driver is not meant for high-performance and it IS the cause of your problem. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: unable to initialize modem (Siemes MC35)
Marc Brünink wrote: > I'm trying to get a Siemens MC35 Modem up and running. > wvdialconf found the modem but everytime wvdial tried to initialize the > modem I got a "CME Error: 3" which stands for "Operation not allowed". > Every other AT command results in a 258: "phone is busy". > I connected to the serial port with minicom and got a 258 twice. After > I waited long enough eventually I got a cpas: 0 (without doing > anything!) which means the modem is ready. After this I was able to > connect to the internet with minicom and wvdial. > > I'm stuck. Does someone have a clue how to fix this issue? Or just an > idea where to look next? It is a GSM modem with a SIM card. And that SIM card needs a PIN. You have to enter that pin with _after_ the CME ERROR message: AT+CPIN=1234 Or simply let scmxx do it for you by doing something simple like setting the time. HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: How to setup a ftp/http server to enable network install in a lan?
Liu Zhen wrote: > I have several server to administrate, and I'm expected to administrate > more servers in the recent future. > I want to install debian3.1_r0a-ia64 to these servers. > I have downloaded cd ISOs from debian, and placed these ISOs on a http/ftp > server in my lan. > I can boot a server using debian's boot floppy disk, the server can get an > ip address from dhcp server in my LAN, and the installation process > proceeds normally until I'm asked to provide an IP address and a path to > continue the install. > I want to use the ISOs in my LAN to continue the install, but I can not > figure out how to setup a correct http/ftp path to enable the install. > Does anybody know how to setup a http/ftp server to enable a install in > LAN? Thank you. The ISO are not for http/ftp install but are for burning CDs/DVDs. If you do not want to download again and your webserver runs linux: you can loop mount the file for a start windows: there are programs that can make a drive letter from an image file HS -- Mein GPG-Key ist auf meiner Homepage verfügbar: http://www.hendrik-sattler.de oder über pgp.net PingoS - Linux-User helfen Schulen: http://www.pingos.org
Re: Bizarre ls Behavior
Martin McCormick wrote: > -r-xr-xr-x2 root root 301376 Jan 21 09:45 ls > Thanks for any hints as to why this directory doesn't list if > one is in it but does list if one is either root or doing ls from > outside the bin directory. Guessing: your $PATH contains "." before the "/bin" Bad thing! When in doubt: run /bin/ls instead of ls. HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: howto get popup for connection speed
hacker (of golf) wrote:
> I'm trying to create a popup to show the connection speed.
> I'm trying this:
>
> root :~# kdialog --display localhost:0 --msgbox "Connection Speed
> `tail -n1 /var/log/ppp-connect-errors | awk '{ print $6 }'` bps"
> kdialog: cannot connect to X server localhost:0
>
> but, as you can see, it won't connect to the xserver.
>
> How can I get this to work?
Prepend:
XAUTHORITY=${USER_WITH_X}/.Xauthority
export XAUTHORITY
Now you can call your command
HS
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Cd writing tool under Linux
Wim De Smet wrote: > normal user access to recording devices has been broken many times in > the past by changes in the kernel API. It often doesn't work reliably > and therefore it's safer to run K3B as root. If you're in a multi-user > environment that's not an option, but if you're the administrator and > you know what you're doing I wouldn't see why not. Please seperate cdrecord and K3B! There is absolutely, really, really NO reason to run K3B as root. To run cdrecord suid-root is another thing. And an administrator should know better than running K3B as root, anyway. HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: HP LJ 1000 printer unsupported under linux?
Deboo ^ wrote: > Is this printer unsupported under linux? Zu dem Thema gibt es einschlägige Seiten im Netz: http://www.linuxprinting.org/show_printer.cgi?recnum=HP-LaserJet_1000 Und da steht sogar, was du machen mußt, damit der Drucker läuft. Und damit man das nicht immer wieder schreiben muß... HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Secure Relaying -- a start
will trillich wrote:
> now if you get PAM to cooperate, let me know.
>
> plain:
> driver = plaintext
> public_name = BASIC
> # $3 =~ s/:/::/g
> # if pam($2:$3) {yes} else {no}
> server_condition = ${if pam{$2:${sg{$3}{:}{::}}}{yes}{no}}
> server_set_id = $2
>
> when i do the interactive tests, it works like a champ; when i
> try it from a remote client, nothing doing. still working on
> it...
For PAM, either run exim daemon as root or search at google for "pam_exim".
BTW: For plain auth it should be "public_name = PLAIN".
HS
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: exim and relaying -- for ONE user
Derrick 'dman' Hudson wrote: > On Thu, Jan 30, 2003 at 10:59:46AM +0100, Hendrik Sattler wrote: > | Derrick 'dman' Hudson wrote: > | > | > Note, however, that AUTH PLAIN isn't very secure. You should only > | > allow it if the client has first initiated a TLS connection. That > | > requires first setting up TLS. I don't know if exim 3 can restrict it > | > to a TLS session only, or how to do it. Either read the docs or > | > upgrade to exim 4 (I know how to check that in exim4). > | > | Exim3 can restrict it like exim4. > > That's good. What's the conf setting to achieve that? http://www.exim.org/exim-html-3.30/doc/html/spec_11.html#IDX636 So shortly: auth_over_tls_hosts = * ACL handling in exim4 might be better but above works: $ telnet abc 26 Trying 129.13.114.79... Connected to abc. Escape character is '^]'. 220 abc ESMTP Exim 3.35 #1 Thu, 30 Jan 2003 20:19:43 +0100 ehlo test 250-abc Hello xyz [xxx.xxx.xxx.xxx] 250-SIZE 250-PIPELINING 250-STARTTLS 250 HELP AUTH 503 STARTTLS required before AUTH > | You forgot the LOGIN method that is needed by some clients. > > I did leave it out. The configuration side is basically the same as > for PLAIN. Some docs I read said LOGIN was never actually > standardized, so I thought it was a good idea not to use it. IIRC old > netscape and old lookout only handle LOGIN, and one (or both) of those > won't recognize it unless the server incorrectly advertises it. Well, putting it into the config doesn't hurt, either. > | CRAM-MD5 should not be needed as TLS should really be secure enough, > | isn't it? ;) > > Depends on whether you want to use TLS or not. Well, TLS/SSL is way more common that CRAM-MD5. Additionally, CRAM-MD5 does not work with PAM. HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: exim and relaying -- for ONE user
Derrick 'dman' Hudson wrote: > Note, however, that AUTH PLAIN isn't very secure. You should only > allow it if the client has first initiated a TLS connection. That > requires first setting up TLS. I don't know if exim 3 can restrict it > to a TLS session only, or how to do it. Either read the docs or > upgrade to exim 4 (I know how to check that in exim4). Exim3 can restrict it like exim4. You forgot the LOGIN method that is needed by some clients. CRAM-MD5 should not be needed as TLS should really be secure enough, isn't it? ;) > An alternative to using exim's own lookup and crypt capabilities is to > defer to pam. There are several advantages of this, for one you can > use any backend (flat file, system account, LDAP, SQL, etc.) that pam > supports. If you use shadow passwords for system accounts and want > exim to use the same for SMTP AUTH you'll have to either run exim as > the 'shadow' group, or make the shadow file readable by the exim > group. To configure this method : Did you tried using pam_exim? It works great, letting exim continue to run as non-root and still using pam (using an external suid-root pam helper). HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: mozilla failing to SMTP-AUTH
martin f krafft wrote: > Mozilla Mail (1.2.1i-9) apparently can't do CRAM-MD5 or DIGEST-MD5 (or > can it?). So I'll let it do PLAIN through SSL and configure the SMTP > Server in Mozilla to contact port 465 (smtps) and to always use SSL. > > When I now try to send an email, the 465/tcp session doesn't get > further than the SYN request from client to server. The server never > responds with a SYN-ACK. > > However, when I do > > openssl s_client -connect server:465 > > I am presented with the SMTP welcome header. So it seems to work on > the server side. > > What could be the problem? Port 465 is only for old clients that do not use the STARTTLS command. Do not use it, Mozilla will do the STARTTLS command when you tell it to use SSL/TLS (simply don't try to be smarter than the program) before sending the AUTH command (a good configured server will not allow AUTH without previous STARTTLS). Also CRAM-MD5 does not make much sense if the password is already encrypted via SSL, except the case that you do not even trust that server (but then, why do you use it). HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: hostnames for dhcp clients with bind
Stefan Radomski wrote:
> Hi there,
>
> I managed to set up a dhcp server to hand out ip addresses to clients,
> now I want these clients to get a hostname depending on their MAC
> address to be resolved by bind.
>
> Has anyone successfully set up such a configuration and can give me a
> quick overview what needs to be done, or point me to some documentation
> on the internet?
What you want is dynDNS with ISC dhcpd3:
dhcp3-client - DHCP Client
dhcp3-common - Common files used by all the dhcp3* packages.
dhcp3-dev - API for accessing and modifying the DHCP server and client state
dhcp3-relay - DHCP Relay
dhcp3-server - DHCP server for automatic IP address assignment
It makes the DNS entries when a client registers with the DHCP server
(either the client provides its name and the server accepts it or you have
host directives with mac adresses that specify the name of the client:
host sdfg {
hardware ethernet xx:xx:xx:xx:xx:xx;
ddns-hostname "sdfg";
}
That even works with printers and bootp.
If you really want IP management with your bind server, use host statements
in the dhcpd.conf. You can use hostnames instead of ip adresses there, too.
But then you have to configure the name at the client _and_ the server which
defeats the point of it.
I even run such a setup (dynDNS) at home because it makes it absolutely easy
to use and maintain :)
HS
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Fwd: getting postfix + sasl to work
Derrick 'dman' Hudson wrote: > :-). Yes, I am the exim guru, but I'm also playing with postfix. I > wanted to better understand how it was designed and see how it's > configuration/configurability compared. I still have exim as the SMTP > server on my machine so that I can reject spam during the SMTP session > (using sa-exim), and currently postfix can't do that. However, I also > hvae the postfix package installed and postfix is /usr/sbin/sendmail > and performing local (and remove) delivery only through that > interface. (by "local" here I mean /var/mail/$USER, not using > maildrop or any other fancy MDA; exim is doing that) Well, personally, I favor exim so far. Mailman integration, delivery to /home/$USER/Maildir and now TLS and SMTP-Auth makes it pretty nice. Although I wonder about plans on exim4 in official Debian. Any news? Also, I am missing pam_exim in debian, as it allows to run exim as non-root and still use PAM for authentication. But exim also has a weird thing: I wanted to make my ssl key only readable to root.cert and having exim's user 'mail' in group cert. However that does not work. Strange, isn't it? The .key and the .crt file needs to be in group mail to let exim access it. Too bad :-/ Any ideas? HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: dhclient3 and event scripts
martin f krafft wrote: > before i go ahead and hack'n'tweak dhclient3 so that it calls one > script when it looses an IP for whatever reason and another when it > gets a new one, i'd like to ask the community if someone has already > done so. i am a little confused by the dhclient3 paradigm and would > much more prefer to build upon someone else's work who understood the > thing a bit better than i do. all i am trying to do is control IPsec, > pulling down a VPN tunnel when an IP is lost and renegotiating SAs > when a new IP is assigned. Well, you should already have all the help you need: /etc/dhcp3/ |-- dhclient-enter-hooks.d | `-- debug-enter |-- dhclient-exit-hooks.d | `-- debug-exit |-- dhclient-script |-- dhclient.conf read the top of debug-enter and debug-exit and all should be well, shouldn't it? HS PS: I am using sarge. If you use woody and do not have these files, try to recompile a more current package or get the necessary files by manually extracting them from a newer package. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: linux dsl router firewall
T.Maguin wrote: > Hallo, > > ich suche Leute, die mir KONSTRUKTIV bei meiner Seite helfen. Es geht im > wesentlichen um eine Anleitung zum Aufsetzen eines vernünftigen linux > Routers mit firewall für normal Sterbliche. > Wenn Ihr Verbesserungsvorschläge habt, dann kotzt mir nicht einfach vor > die Haustüre, sondern sagt mir. wie ich es besser machen kann. > Also schaut mal nach auf > > www.router-piewie.de.tf Siehe PM. Aber: 1. Hier ist englisch 2. diverse Newsgroups damit zuzumüllen kanns ja wohl nicht sein. Wenn, dann bitte als Crosspost mit passendem Reply-To: HS -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
