Mailing list Identification (was Re: [expert] Mandrake and USB camera confusion)
On Thu, Dec 12, 2002 at 10:29:02AM -0800, Todd Lyons wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Wolfgang Bornath wrote on Thu, Dec 12, 2002 at 06:37:47PM +0100 : Hmm, my mail proggy acts like he should: In the header of the mailing list I see a From: and a Reply-to: line. When I hit 'g' (ReplyAll) the proggy asks: Reply to [EMAIL PROTECTED]? [Yes] No? So I just hit [Enter] and no doubles go out. When I hit 'n' the mail goes to the address in the From: line. Again no doubles. Add to your .muttrc: subscribe expert@ While we're at it, what's the recommended method for identifying this mailing list? I've got :0: * ^X-Loop: [EMAIL PROTECTED] mandrake-expert in my .procmailrc, but there's a noticable chunk of the messages which don't contain this header. (And I don't approve of filtering mailing lists according the To: or CC: as that won't differentiate between duplicate copies sent to the list and to myself.) -- T. Our three main weapons are Sarcasm, Ignorance, and Denial. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: Mailing list Identification (was Re: [expert] Mandrake and USB camera confusion)
On Fri, Dec 13, 2002 at 03:29:41PM -0800, Todd Lyons wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Tommy Wareing wrote on Fri, Dec 13, 2002 at 09:44:17AM + : Add to your .muttrc: subscribe expert@ I've got :0: * ^X-Loop: [EMAIL PROTECTED] mandrake-expert in my .procmailrc, but there's a noticable chunk of the messages which don't contain this header. Really? It always has this for me. Go look at http://www.cerritoslug.org/tutorials/mutt.html and look for the procmail section. There's a stanza for getting rid of duplicates. The only problem with it is that if someone cc's you and the list, then only the first one gets through (usually the direct email) so it shows up in your inbox instead of the list you're filtering to. Is it possible that the chunk of messages you're talking about is where you are cc'd on list mail? I'm not trying to delete duplicates ;-) I'm trying to ensure that anything I recieve via the list goes into one mailbox, and anything I recieve because it's addressed to me goes into a personal mailbox. (And therefore, if I'm cc'd as well, I'll get one copy in each of two boxes). The two most recent message without the X-Loop header were: From: Sylvestre Taburet [EMAIL PROTECTED] Subject: Re: [expert] Cups/Samba pdf generator. Date: Thu, 12 Dec 2002 17:37:45 +0100 Message-Id: atae0p$v16$[EMAIL PROTECTED] Sender: [EMAIL PROTECTED] X-Loop: expert@ which can't be filtered on the Sender: field. I'd guess X-Loop: got truncated due to originating within Mandrake. From: Gonzalo Avaria [EMAIL PROTECTED](by way of Gonzalo Avaria [EMAIL PROTECTED]) Date: Thu, 12 Dec 2002 12:36:51 -0300 Message-Id: [EMAIL PROTECTED] X-Loop: expert@ Subject: [expert] IP look for Sender: [EMAIL PROTECTED] which has been clipped for some reason I'm not going to guess at. Looks like I'd better remove linux-mandrake.com from my filter. -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] How can i copy 8G to an NT machine and keep.....
On Fri, Nov 29, 2002 at 03:02:23PM +0100, Wolfgang Bornath wrote: IMHO it may be a lack of available space in /tmp which makes the tar-process stop at 2G. It's a kernel/file system limit. Move the data in smaller chunks (ie. more than one file). I presume that the other option (using no intermediate filesystem storage) tar cf - . | rsh nt-machine tar xf - doesn't work when one of the machines is ruining windoze. -- T. Our three main weapons are Sarcasm, Ignorance, and Denial. Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] routing/DNS problems - wireless connection sharing
On Thu, Nov 28, 2002 at 07:49:05AM -0800, Praedor Tempus wrote: How do I get name resolution to work? I am not running a local DNS (and would prefer not to). I SHOULD be able to use the DNS settings of box 1 to get name resolution on box 2. What settings need to be looked at on which box to get this working? It HAS worked before but I have no idea how/why - while now it doesn't. Give box 2 the address of box 1 as the DNS server. Set up the iptables on box 1 to forward DNS traffic from box 2 to box 1's DNS server. I think the responses should automatically go back to the right place as the masquerading kicks in. Be careful doing this: if you accidentally redirect too much traffic, you might end up redirecting DNS responses back to the DNS server. Which would be bad... It's a long time since I've poked IP tables, and I don't have a box to hand to experiment on, so I can't tell you to try particular commands. And I may be barking up completely the wrong approach anyway... -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] man madness
On Sun, Nov 24, 2002 at 08:48:17PM -0500, Matthew O. Persico wrote: Mine (Mandrake 9.0) actually says /usr//etc/man.config. Which by convention tends to me /etc/man.config (stuff// being interpreted as /) Huh? I type ls -lart /usr//etc and I get zero files, because /usr/etc is empty on my machine. Where did you pick up that convention? Emacs, actually. It's a convention at the application level, not at the file system. So if the application doesn't support it, it doesn't support it. I suspect what's actually happened in this case is that the configuration has replaced the default location with /etc/man.config in the binary (so the binary doesn't support the above convention), but *appended* it in the man page (because the man build script gets it wrong). However, simply being aware this sort of thing can happen makes it much easier to read the thing correctly :-) However, the Mandrake 8.2 man comes with the manpath command: manpath - determine user's search path for man pages So you can simplify all this to: MANPATH=`manpath`:/new/dir/goes/here Yes, I could do that. However, the only reason I am playing with MANPATH at all is that I have developed some bash functions to swap versions of Perl. I modify PATH to get at the one I want. I was also modifying MANPATH. I may just decide to just use perldoc instead. Or maybe not. Depends on how much I want to muck around. I have real work to do to. Well, if you want to replace your MANPATH with the default value, and then append something to it: unset MANPATH MANPATH=`manpath`:/new/dir/goes/here But that'll upset anybody who's modified it from the value you last set it to. Much better to: case :$MANPATH: in ::) MANPATH=`manpath`:$NEWPERL/man ;; :$OLDPERL/man:) MANPATH=`echo $MANPATH | \ sed 's/$OLDPERL\/man/$NEWPERL\/man/'` ;; *) MANPATH=$MANPATH:$NEWPERL/man ;; esac -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] man madness
On Sat, Nov 23, 2002 at 08:23:18PM -0500, Matthew O. Persico wrote:
MANPATH=/opt/perl/lib/perl5/man:/usr/lib/perl5/man
Apparently, unlike Solaris and any other *NIX system I have used,
MANPATH is not additive - it REPLACES all other heuristics for
locating manpages.
Eh? You get exactly the same behaviour under Solaris. man defaults to
looking in somewhere (see your documentation for the exact
location). Setting MANPATH overrides this.
Hence, I gained manpages for Perl and lost every
other manpage in the system.
So, I unset MANPATH and did man man and it said to check out
/usr/etc/man.config and man man.conf(5). Well, neither the file nor
the manpage exist on my system.
Mine (Mandrake 9.0) actually says /usr//etc/man.config. Which by
convention tends to me /etc/man.config (stuff// being interpreted as /)
How does one ADD man paths to the lookup heuristics?
MANPATH = $MANPATH:/new/dir/goes/here:/another/new/dir
This does not work. A default Mandrake install does not initialize
MANPATH at all. By setting MANPATH, I override all the settings in
/etc/man.config. Yes, it's weird. No, I've not seen this in any *NIX
system I've ever worked on. But there it is.
You're using a different version of Solaris from me then (and I've got
2.5, 7 and 8). I've always known it to work like this.
Now, I could, someplace in /etc/profile.d/foobar.sh do this:
MANPATH=$(manpath)
From the parameter expansion section of the bash manpage:
MANPATH=${MANPATH:-default}:/new/dir/goes/here:/another/new/dir
will do what you want (using default if MANPATH isn't set).
So the only problem is how to determine the default. Our Solaris setup
consists of
MANPATH=${MANPATH:-/usr/man}:/usr/local/man
However, the Mandrake 8.2 man comes with the manpath command:
manpath - determine user's search path for man pages
So you can simplify all this to:
MANPATH=`manpath`:/new/dir/goes/here
--
Tommy
Want to buy your Pack or Services from MandrakeSoft?
Go to http://www.mandrakestore.com
Re: [expert] automount lmontel
Given that Laurent Montel is one of the Mandrake development team, and his name is attached to many of the RPMs (particularly kdegraphics), I think it's actually a left-over configuration option. - Original Message - From: Bryan Whitehead [EMAIL PROTECTED] To: [EMAIL PROTECTED] Sent: Wednesday, November 20, 2002 12:01 AM Subject: Re: [expert] automount lmontel Brian Schroeder wrote: I am getting numerous messages (in /var/log/messages) from automount, trying to mount /home/lmontel. How can I find where this is coming from and stop it? automount[1163]: attempting to mount entry /home/lmontel automount[1975]: mount: special device /local_home/lmontel does not exist (Note: the automount map has the entry * localhost:/local_home/ to catch anything not explicitly mentioned earlier) Brian. _ Add photos to your e-mail with MSN 8. Get 2 months FREE*. http://join.msn.com/?page=features/featuredemail Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com If you running KDM with the little penguin faces it will look in every user home directory for what face it should use. That will cause automount to mount every user home directory on your machine. If user lmontel does not exist on your machine, then delete the account. You could also just make a directory in /local_home called lmontel :) -- Bryan Whitehead SysAdmin - JPL - Interferometry Systems and Technology Phone: 818 354 2903 [EMAIL PROTECTED] Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] Dial-in ppp and Bastille
I've got a dial-in ppp server running on my home server, which also runs Bastille firewall. Unfortunately, my default dial-out connection (which its unlikely to use, as I've got broadband) is ppp0, and so is the incoming connection. Obviously, with only one modem, only on of these can be active as once, so this conflict doesn't cause any problem in practice, until it comes to setting up the Bastille configuration. I'd like to configure it so that the dial-in connection is treated as an internal connection, but other ppp connections (ie. outgoing connections) are treated as untrusted external interfaces. So how do I tell the pppd daemon to give a specific name to one of the connections (preferably the incoming one), which I can then pass on to Bastille? -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Wireless woes: New initscript package seems broken, and high loads kills connection.
On Fri, Nov 15, 2002 at 10:07:28PM +0100, Teemu Torma wrote: The grep statement in check_link_down does not work. I just commented it out... - grep -q -E $1: +[09]+ +0. /proc/net/wireless return 0 My /proc/net/wireless contains: Inter-| sta-| Quality| Discarded packets | Missed face | tus | link level noise | nwid crypt frag retry misc | beacon eth1: 74. 228. 154. 0 44 0 0139 0 The eth1: line fails to match the 74. against the 0. of the regex. Which hopefully helps somebody out. Todd? -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Wireless woes: New initscript package seems broken, and high loads kills connection.
On Fri, Nov 15, 2002 at 09:52:37PM +, Tommy Wareing wrote: On Fri, Nov 15, 2002 at 10:07:28PM +0100, Teemu Torma wrote: The grep statement in check_link_down does not work. I just commented it out... - grep -q -E $1: +[09]+ +0. /proc/net/wireless return 0 My /proc/net/wireless contains: Inter-| sta-| Quality| Discarded packets | Missed face | tus | link level noise | nwid crypt frag retry misc | beacon eth1: 74. 228. 154. 0 44 0 0139 0 The eth1: line fails to match the 74. against the 0. of the regex. Which hopefully helps somebody out. Todd? However, while check_link_down is running, the line is: eth1: 0. 154. 154. 0 0 0 0 0 0 which does match. And kills the link :-( -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] Irda laptop Palm
On Mon, Nov 04, 2002 at 09:13:05PM +0100, Joan Tur wrote: I was wondering if it is possible to sync a Palm m515 through Irda. I don't know if irda is working in my laptop (I haven't used it)... how can I check Irda's support is installed and working ?? Thanks in advance ! ;) - -- Joan Tur. Eivissa-Spain AOL quini2k, ICQ 11407395 www.ClubIbosim.org Linux: usuari registrat 190.783 I've just set up my USB Infrared adaptor. And it works! To see if you've got the linux side working: $ lsmod | grep ir irda 87152 1 [irda-usb] You may need other modules, depending on the eaxct nature of your IrDA unit (I need irda-usb, for example), but you at least need this one. $ dmesg | grep IrDA IrDA: Registered device irda0 Wahey! My machine knows it's got an IrDA interface. Make a careful note of the device name. $ ps -ef | grep irattach root 2237 1 0 23:57 ?00:00:00 irattach irda0 -s If this process doesn't exist, then just start it. The parameter is, of course, copied from the device at the previous stage. If you're running the irda service, then this device is set in /etc/sysconfig/irda. $ cat /proc/net/irda/discovery IrLMP: Discovery log: nickname: SIEMENS S35, hint: 0x9024, saddr: 0x1231c29e, daddr: 0x08365522 And there's my mobile phone. Even better. Now to find some suitable applications... ;-) -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] (YA) USB Mass storage question
On Wed, Oct 09, 2002 at 06:38:54PM +0100, bascule wrote: try looking in /dev/scsi/host0/bus0/ or similar, i'm not sure what other scsi devices you might have you might have to change the host0 to host1 and/or the bus0 to bus1 etc. somehwere under the /dev/scsi/ tree should be the device file for your camera that you can mount, Oh, it's been showing up in /dev/scsi/host1 quite cheerfully, but that doesn't seem to be helping. you may also have a symlimk of the form /dev/sdb1 that you can use, if this is your only 'scsi' device then perhaps sda1? for scsi devices there are several device files associated with them, not all of them are block devices, Ahah, it's /dev/sda1, but... sorry to be so vague, just look for device files that aren't there when the camera is not present It's worse than that: I need to look for device files that only exist after I try to use them! ls /dev | grep sda produces no output mount /dev/sda1 /mnt/camera works ls /dev | grep sda now lists /dev/sda1 (and then it all works). If the camera is plugged in at the time the usb system is started, then it does log /dev/sda1 in /var/log/messages, but doesn't if the camera is plugged in later. And it's not possible to stop and restart the usb service while the machine's running, since some bright spark has actually commented the call to usb_stop out of the init script. Grr (and a quick edit). -- T. Our three main weapons are Sarcasm, Ignorance, and Denial Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] ML 9.0 - postfix calls *all* connections unknown
On Wed, Oct 09, 2002 at 12:16:23AM -0500, Ray Warren wrote: On 8 Oct 2002 at 14:00, David Guntner wrote: I found out from someone else that apparently you now need to have a copy of your hosts and resolv.conf files in /var/spool/postfix/etc for some reason. The default installation now appears to be chrooted and you need to check the configuration setup for a chroot environment rather than a standard setup.I don't think postfix can see anything in /etc so everything Postfix needs has to duplicated in /var/spool/postfix/etc. More secure but more setup needed. Anybody know how to make this work with linuxconf profiles? I move my laptop between my home and office, and need different DNS configurations at both. Linuxconf handles this adequately. Or it used to, until this change to postfix, which now drops dead until I update the /var/spool files by hand, and restart the service. -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] ML 9.0 - postfix calls *all* connections unknown
On Thu, Oct 10, 2002 at 03:38:01PM -0700, Todd Lyons wrote: Tommy Wareing wrote on Thu, Oct 10, 2002 at 11:26:56PM +0100 : everything Postfix needs has to duplicated in /var/spool/postfix/etc. More secure but more setup needed. Anybody know how to make this work with linuxconf profiles? I move my laptop between my home and office, and need different DNS configurations at both. Linuxconf handles this adequately. Or it used to, until this change to postfix, which now drops dead until I update the /var/spool files by hand, and restart the service. Untested, so you're my guinea pi^W^W tester: cp /etc/hosts /var/spool/postfix/etc rm /etc/hosts cd /etc ln -s ../var/spool/postfix/etc/hosts hosts Don't know if linuxconf will complain about that being a link, but if it doesn't, it should work. The question is, does linuxconf copy files around or does it modify the file in place? If it copies them, it probably won't work the way you want. But if it modifies the files in place, it probably will work the way you want. Actually it's resolv.conf I need to target, not hosts. But that's a side issue. Yes this does work. Cute. I'd considered the link the other way around (from postfix to 'normal'), but that would break from within the chroot anyay, so I'd just cursed. There's the question of security: this now allows a breached postfix to rewrite resolv.conf for the rest of my machine, although only that single file. So it's safer than not chroot-ing, but not as good as fixing all the applications which write to resolv.conf (linuxconf and dhcp being the two I can think of, but there may be sufficient others to mean this isn't viable). Of course, if we ever have two different chroot'd applications, this won't be viable, so in the long term, it's still a potential problem. But at least I can continue to outfunction all our office Solaris boxes ;-) -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] (YA) USB Mass storage question
I've got a Sony DCR-TRV25E which I can connect via a usb cable. It's not in the list of working USB devices, but it is a mass storage device. It's detected happily enough: Oct 9 11:54:50 nickel kernel: hub.c: USB new device connect on bus1/2, assigned device number 4 Oct 9 11:54:50 nickel kernel: usb.c: USB device 4 (vend/prod 0x54c/0x2e) is not claimed by any active driver. Oct 9 11:54:54 nickel automount[1490]: attempting to mount entry /net/prodigy Oct 9 11:54:54 nickel /etc/hotplug/usb.agent: Setup usb-storage for USB product 54c/2e/200 Oct 9 11:54:54 nickel automount[3737]: showmount: can't get address for prodigy Oct 9 11:54:54 nickel automount[3737]: lookup(program): lookup for prodigy failed Oct 9 11:54:54 nickel kernel: Initializing USB Mass Storage driver... Oct 9 11:54:54 nickel kernel: usb.c: registered new driver usb-storage Oct 9 11:54:54 nickel kernel: scsi1 : SCSI emulation for USB Mass Storage devices Oct 9 11:54:54 nickel kernel: Vendor: Sony Model: Sony CamcorderRev: 2.00 Oct 9 11:54:54 nickel kernel: Type: Direct-Access ANSI SCSI revision: 02 Oct 9 11:54:54 nickel kernel: USB Mass Storage support registered. and the usb modules are loaded: usb-storage51952 0 scsi_mod 90372 3 [usb-storage sg ide-scsi] usb-uhci 21676 0 (unused) usbcore58304 1 [usb-storage usb-uhci] The camera does show up as /dev/sg1, and cdrecord sees it: Cdrecord 1.11a32 (i586-mandrake-linux-gnu) Copyright (C) 1995-2002 Jörg Schilling scsidev: '1,0,0' scsibus: 1 target: 0 lun: 0 Linux sg driver version: 3.1.24 Using libscg version 'schily-0.6' Device type: Removable Disk Version: 2 Response Format: 1 Vendor_info: 'Sony' Identifikation : 'Sony Camcorder ' Revision : '2.00' Device seems to be: Generic CCS Disk. but when I try to: [root@nickel tmp]# mount /dev/sg1 /mnt/camera mount: /dev/sg1 is not a block device Is this a result of not being a working device, or do I need to do something else? -- Tommy Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
[expert] pcmcia breaks after powering down
I've been using 8.2 quite happily on my laptop (Acer aspire 1203XC) for a while. I've just upgraded it to 9.0, and started having trouble with the pcmcia service. Specifically, it hangs the machine whilst trying to modprobe yenta_socket. No messages, no diagnostics, absolutely nothing that I can report (there may be stuff happening, but since the machine's hung, I can't find out). This used to work fine. The strange, but fortunate, part is that if I reboot from the install CD, this finds and initialises the pcmcia cards fine, and I can then reboot normally. After that, it seems to be fine until I power down again. This is, um... a bit of a faff. Any ideas? -- T. Our three main weapons are Sarcasm, Ignorance, and Denial Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
Re: [expert] NFS write problems - linuxconf sucks?
On Wed, Oct 02, 2002 at 03:25:59PM -0700, Todd Lyons wrote:
SainTiss wrote on Thu, Oct 03, 2002 at 12:19:42AM +0200 :
Hi,
I'm trying to set up NFS here... I've noticed linuxconf is kind of
f*cked up... It tries to execute /etc/init.d/network
{status|reload|start|stop} or something, which is of course a syntax
error...
The {status|reload|start|stop} is probably an error message from the
script that is showing the valid commands.
The Linuxconf howto/FAQ says:
The network script (/etc/rc.d/init.d) is a little special. For the
network script, it uses the tag probe: true. This tells Linuxconf
to let the script finds out by itself if it needs a restart or not
(restart of whatever). When probe is true, Linuxconf calls the
script with the probe argument
/etc/rc.d/init.d/network probe
this normally prints nothing when everything is up to date. If
something is not right, the script output simply reload. Linuxconf
picks this string and if you choose to activate the changes, it
calls the script with this string.
Somebody has removed the probe handling section of the script (it was
there in 8.2) and left in the # probe: true line.
--
T.
Our three main weapons are Sarcasm, Ignorance, and Denial
Want to buy your Pack or Services from MandrakeSoft?
Go to http://www.mandrakestore.com
