Re: about packages
On Sun, Jul 29, 2001 at 04:25:46PM +0800, Tao Liu wrote: When I type dselect, and select, it shows some Obsolete/local packages. Does that mean all these packages can be removed, if I have no local packages? It means that these packages are listed in the dpkg status database, but not in the available database. It is usually caused by packages being removed from the Packages file in the debian archive. If you build your own kernel-image.deb and install it, it will also be listed as obsolete/local, because your homebuilt deb is not known to the debian archive Packages file. Cheers, Joost
Re: about packages
On Sun, Jul 29, 2001 at 06:32:53PM +0200, Joost Kooij wrote: On Sun, Jul 29, 2001 at 04:25:46PM +0800, Tao Liu wrote: When I type dselect, and select, it shows some Obsolete/local packages. Does that mean all these packages can be removed, if I have no local packages? It means that these packages are listed in the dpkg status database, but not in the available database. It is usually caused by packages being removed from the Packages file in the debian archive. If you build your own kernel-image.deb and install it, it will also be listed as obsolete/local, because your homebuilt deb is not known to the debian archive Packages file. I'm sorry, that was not a complete answer to your original question. Any packages that you did not build yourself can probably safely be removed. Cheers, Joost
Re: rpc.statd puzzle
On Sun, Jul 29, 2001 at 02:14:32PM -0400, Eugene Tyurin wrote: I found this in my syslog: Jul 29 13:50:39 daBox rpc.statd[25571]: my_svc_run() - select: Bad file descriptor I don't have any NFS services running, nor do I mount any NFS filesystems. Any ideas and/or suggestions? I have nfs-common version 0.3.2-2 and 2.2.19 kernel with the reiserfs patch. If you don't have a need for any rpc services such as nfs, remove the portmapper package and all other packages that depend on it in dselect. Alternatively, remove the /etc/rc2.d/S*portmap* symlink and the same for the other rpc and nfs related services in runlevel 2. That way, you can selectively enable these services with telinit 3, but you won't be running them by default, which is runlevel 2. Generally, it's best to not install packages if you don't need them, and that goes especially for network services. Cheers, Joost
Re: problem sending mail
(please type a return after +/- 70 characters.) On Sat, Jul 28, 2001 at 12:44:44PM -, padmaja godbole wrote: I have a online site using jsp and I am sending mail from a jsp page .But There was error message saying-- error sending the mail. IO ERROR IN SENDING EMAIL:554 : Recipient address rejected: Relay access denied . and a returned mail from [EMAIL PROTECTED] with message-- This is the Postfix program at host bom4.vsnl.net.in. I'm sorry to have to inform you that the message returned below could not be delivered to one or more destinations. For further assistance, please send mail to postmaster If you do so, please include this problem report. You can delete your own text from the message returned below. The Postfix program [EMAIL PROTECTED]: host mail2.rediffmail.com[203.199.83.4] said: 553 sorry, your envelope sender is in my badmailfrom list (#5.7.1) pls help mw out to solve this problem You can't, because the receiving mta thinks you're a spammer. Cheers, Joost
Re: Fwd: Your site listed
On Sat, Jul 28, 2001 at 09:35:07AM -0500, d wrote: HEY, people what is this CARP? It's called spam. I have received some stuff like this before and just thought it was an ERROR. Hmmm, I wish it was. Welcome to the internet. Now I am begaining to wonder.. Is this sort of thingy acceptable practice or should we turn them in to some FROG or SALAMANDER some place FAR FAR AWAY? Please tell us how to do that if you find a way. You'd be mighty popular. Cheers, Joost
Re: Fwd: Your site listed
On Sat, Jul 28, 2001 at 04:39:32PM +0100, Anthony Campbell wrote: You think this is bad? I'm getting a message every 2 days or so with a 200K+ attachment in base 64 (which I can't read). It says something like: Hi. How are you? I'm sending this file for your comments. Thanks. See you later. And he does - about 6 times now. It's always from a different address so I can't filter it in any way. * rofl * It's another windows virus. It was in all the regular news media here in the christmas islands. The virus writers have come up with another brilliant and twisted for of automated social engineering. To convince new victims that this is not yet another attachment virus, it also sends along a random personal document from the current vector host. Another innovation is to scan not only the windows address book, but also the web pages that were visited by the host and remain still in the browser's cache. That probably explains why you got the mails with the so very peculiar personal touch. On a sidenote, I bet that there's lots of mailinglists that had personal files posted. Some of these mailing lists are archived automatically. It already appeared to me that anyone can readily download many popular windows virusses from mailing list archive websites. Or google's cache. Interesting times they are. Cheers, Joost
Re: How do I find my local IP assigned by my ISP when using pon, etc?
On Sat, Jul 28, 2001 at 09:07:13AM -0700, Randolph S. Kahle wrote: Thank you for the reply. I forgot to mention one complication, I am setting this machine up for someone who will not have root access (I will retain that). I am doing this so that they cannot mess up their own machine... The user will be able, from a user account, do a pon, poff, etc. to connect to the ISP. So, my challenge is to have the scripts run from user level security and install the firewall rules. How do I do this? Make sure that the user is in the dialout group, so she can run pon and poff to start and stop a dialin session determined by one of the files in /etc/ppp/peers/. When ppp has brought up a link, it starts the /etc/ppp/ip-up and passes it several parameters, among which is the assigned local ip address. When you install the ipmasq package, it installs a script in the /etc/ppp/ip-up.d/ directory, that is read in turn by /etc/ppp/ip-up. The ipmasq script will automatically setup your machine as a masquerading gateway. Cheers, Joost
Re: Root/User Priveleges
On Sat, Jul 28, 2001 at 03:09:26PM -0400, [EMAIL PROTECTED] wrote: How do I set priveleges for files/programs/commands for users? I'd like for my username to have close to root priveleges, so I can change system options, etc. I find that the NEVER BE ROOT! NEVER BE ROT!! philosophy is bit odd-- seeing as you have to SU to root to change system settings and there's just as much risk doing that. The idea is that you do su - every time you need to do something as root. When you are done, you exit the root shell, so you cannot make mistakes in it. If you've set up the system right, you'll hardly ever need to be tweaking aspects of the system. Get used to having a truly stable system, that behaves like you expect it to behave. Every time you change something, you lose a bit of that stability. If you don't care about regularly making mistakes that cause your system to need a clean reinstall, then do go ahead and use the root login always, whether it is required or not. Cheers, Joost
Re: dhcp ip addresses
On Sat, Jul 28, 2001 at 10:02:49AM +0800, Eric Boo wrote: Question: How are these ip address stored, if they are (under pump espcially), and how do I force pump to get another different ip if needed? Also, I can't seem to use pump on the command line after releasing it with pump -r. pump -R -i eth0 and simply pump both gave operation failed. Maybe I should go back to dhcpcd. Or is there an alternative? How's dh-client? Try to see if the pump process is still alive. IIRC, if you start pump the first time, it will fork and run in the background. You can communicate with the running pump by starting anothter pump process, that will use a local socket to talk to the background pump. Sometimes the background pump becomes unresponsive, I've found, and you need to kill it and restart pump. YMMV. It is also possible that you setup firewalling rules and that somehow these interact badly with pump. You can debug this easily by prepending a rule to the input and output chain that logs any packets on any interface and has no jump target. Then start pump and watch the syslog. Remember to remove the logging rules when you have captured enough pump packets, or else it will make your logfiles explode. Cheers, Joost
Re: Debian install issues!
(please press enter after +/-70 characters.) On Fri, Jul 27, 2001 at 08:33:49PM -0700, Brad Pillatsch wrote: Here's the problem I am having. This occurs on all the flavors of the latest stable release root disks. Basically when the root disk loads up into the install menus, specifically the first release notes window and the driver setup window, if I scroll down with the keyboard the system will show a black line in the middle of the window and then *hang* the system up to dry. I booted off floppies, I booted off a FAT32 disk with loadlin, and even ran memtest82 to see if I had memory probs. I bumped my clock back down to spec(celeron 300a) none of which seems to be working. I have installed debian in the past so I am assuming its an issue with the latest release. That is strange. Is it completely hung up in the kernel or is the installer deadlocked? Can you still switch consoles with leftalt-f1, -f2, etc.? Cheers, Joost
Re: libc6 2.2.3-7 post-inst error
On Sat, Jul 28, 2001 at 01:09:20PM +0800, Lindsay Allen wrote: On Mon, 23 Jul 2001, Lindsay Allen wrote: elm:# dpkg --configure libc6 Setting up libc6 (2.2.3-7) ... date: invalid date `Mon Jul 23 07:32:48 WST 2001' dpkg: error processing libc6 (--configure): subprocess post-installation script returned error exit status 1 Errors were encountered while processing: libc6 This happens on both my sid boxes. Is this just me? This problem continues with libc6_2.2.3-8 and I can't install any more packages until I get it fixed. The date `Mon Jul 23 07:32:48 WST 2001' was a perfectly good date in this part of the world. I tried running the post-inst script as root after commenting out the set -e line, but that was not effective. Add set -x to the script as well and run dpkg --configure --pending. Notice the commands that fail. Post them. Cheers, Joost
Re: Fwd: Your site listed
On Sat, Jul 28, 2001 at 04:54:52PM -0400, dman wrote: On Sat, Jul 28, 2001 at 05:22:28PM +0100, Colin Watson wrote: | Anthony Campbell [EMAIL PROTECTED] wrote: | [cc'ed according to mail-followup-to request] | On 28 Jul 2001, Joost Kooij wrote: | On Sat, Jul 28, 2001 at 09:35:07AM -0500, d wrote: | HEY, people what is this CARP? | | Hmm, seems like the SHOUTING meant a couple of letters were transposed | there. :) Maybe he just prefers fish wink? I have attached some for you, hope you like it. Cheers, Joost ___,--.___ __ __ | `~^^~' | | | | |`~^^~'| | | | $$$ $$$ $$$ | | | | $ $ $ $ $ $ $ $ | | | | $ $$$ $$$ $$$ | | | | $ $ $ $ $ $ $| | | | $$ $ $ $ $ $ )| | |`~^^~'| | | | original | | | | finest brand quality | | | | | | CANNED FISH | | | __ __ `~^^~'
Re: RFC 2015 (PGP/MIME) MTA compliance (was Re: Off Topic: Mailing Agents?)
On Sat, Jul 28, 2001 at 01:16:50PM -0700, Karsten M. Self wrote: ...so, arguably, at worst a mail client should display the body of the message but treat the signature as an attachment. HA HA HA People want to use their mail, not understand it. It should be *easy* to attach your personal documents to messages sent to addresses that happened to be listed on a webpage you just visited. Sigh, these unix people just don't get it, do they? pgp/mime will never work if users can't click on it to see what will happen next. Cheers, Joost
Re: Linux Sources
On Sat, Jul 28, 2001 at 04:25:55PM -0400, [EMAIL PROTECTED] wrote: I downloaded a copy of the Linux 2.4.7 sources as a tar-gz file, and I suppose I need to install them into /usr/src. However, the tar-gz's contents are not at all indicative of that, and doing a make config just gives me a bunch of questions to set up a kernel. Do I just make a kernel from this dir I'm working from, or do the files need to be in /usr/src? Install kernel-package and read all the files in /usr/doc/kernel-package. Cheers, Joost
Re: Chronically broken package - bugs? In tetex-bin AND apt-get?
On Sun, Jul 29, 2001 at 06:44:27AM +, hzi wrote: Just a remark: I'm not entirely sure if it wise to use any other than the default locale settings in the root account. What do you mean? All the bahavior here was set automatically by dselect. I haven't specified an installation directory different than the default, when you install tetex-bin. I meant the locale settings. Do env | egrep 'LANG|LC_' and if there are any variable assignments in the ouput, those are your locale settings. It is okay for regular users, and it should be okay for root maybe, but there is always a risk of unexpected formats. The bug is probably in tetex-bin, for having a broken prerm script. Post removal script Sorry, my tpyo. Cheers, Joost
Re: RFC 2015 (PGP/MIME) MTA compliance (was Re: Off Topic: Mailing Agents?)
On Sat, Jul 28, 2001 at 02:50:45PM -0700, Karsten M. Self wrote: Could you translate that into English please? Do you want fries with that, sir? phony smile Cheers, Joost
Re: RFC 2015 (PGP/MIME) MTA compliance (was Re: Off Topic: Mailing Agents?)
On Sat, Jul 28, 2001 at 05:09:47PM -0500, Nathan E Norman wrote: I think he's making a joke at the expense of all the people who have experienced the sircam worm. Woa! You made me grab for my Hendrix mp3's... Have you ever been experienced? I think the sircam victims had a pretty bad trip. Still, they're the lucky ones, they learnt not to click on things in the mail the gentle way. The next form of foot-and-mouth will be smarter and more damaging as well. The joke is at the the people who deny that this is only going to get worse with each new generation of viruses being smarter and each new generation of internet users being dumber when they first meet the net. Really I agree with Karsten's technical arguments, I just think that it is also important that people experience the relevance of standards. You can either cry about the facts or laugh at them. It won't make them go away either way, but at least they look better when you laugh. Cheers, Joost
Re: dselect forcing recommends?
On Thu, Jul 26, 2001 at 09:41:42PM -0400, Michael P. Soulier wrote: So, I'm in dselect, and it says that icewm-common recommends icepref. Unfortunately, it's not just a recommendation, because it won't resolve the conflict without it selected for install. You can tell dselect not to bother with the recommeded packages, by pressing 'D' and then 'Q' in the dependency resolution screen. 'D' sets the package selections back to your direct selections. 'Q' commits the current selections without further depends checking. 'X' exits, disregarding any changes in selections. If you are confused at any place in dselect, you can always keep hitting 'X' until you are back at the main menu, and no changes will have been made effectively. 'Enter' will do what 'Q' does, but first it checks the dependencies and if there are any unresolved ones found, you are prompted with a new dependency resolution screen. This is also explained more clearly in the help screen that always pops up before the dependency resolution screen and that you can get back to at any moment by pressing the '?' and 'i' keys. Is this a dselect bug? In the sense that you are asking, no. You just need to press 'Q', not 'Enter'. It is a feature. In another sense, it is clumsy at least that this has to be done every time when there is an unresolved recommends: in the active package selections. There are some patches for this, but I don't believe they are in a dselect near you just yet (maybe soon in unstable). Cheers, Joost
Re: dselect question
On Thu, Jul 26, 2001 at 09:55:38PM -0500, Lance Peterson wrote: I selected some a package with dselect and then it automatically selected a *bunch* of dependent packages. Then I decided not to install the original package, but all the other packages it thought were dependent still try and install every time I run dselect even though the original package has been deselected. Oops, you should have pressed 'R' or 'X' instead of 'Enter'. Now you've committed all the selections you made with dselect's interactive package selections management screen. Is there a file somewhere that I can purge that has all the pending for install packages so I can just wipe it out and start fresh again? How about marking the relevant packages for removal or purge in the dselect package selections list? It sounds intuitive enough to me that if you can select packages in dselect, then you can also unselect them. Admittedly, you'll have to look them up in the list, but that is not so hard, and you can play with the 'o' and 'O' keys for sort options if you like. And if you press '?', there is help at any time. When you are finished altering the selections, simply run install. Generally, walking the installed packages list in dselect once in a while and removing packages you don't see a need for, is a good thing. If you happen to try to remove something unexpectedly important, and a large list of packages is marked for subsequent removal in the dependency resolution screen, you can simply undo the removal request and all of the consequences, by pressing 'R' and 'Enter' (or 'Q' or 'X'). Try it. This way, you quickly get to know what packages are on your system for what reason. Cheers, Joost
Re: Chronically broken package - bugs? In tetex-bin AND apt-get?
On Fri, Jul 27, 2001 at 01:24:16PM -0300, Henry Lebowzki wrote: I'm sorry if I sound hysterical, but I have reasons for concern. In fact, I'm truly desperate: it has been over a three months that I try to tackle this problem whenever I have some time. I was never able to use TeX on my potato, because the package somehow got broken. And I mean *badly* broken. In dselect, it appear as *RI, which means that I should reinstall it. (Reading database ... 115823 files and directories currently installed.) Preparing to replace tetex-bin 1.0.6-7 (using .../tetex-bin_1.0.6-7_i386.deb) ... Unpacking replacement tetex-bin ... open /var/lib/xaw-wrappers/update-wrappers-history: Arquivo ou diret?rio n?o encontrado [file or directory not found] dpkg: warning - old post-removal script returned error exit status 2 dpkg - trying script from the new package instead ... open /var/lib/xaw-wrappers/update-wrappers-history: Arquivo ou diret?rio n?o encontrado [file or directory not found] dpkg: error processing /var/cache/apt/archives/tetex-bin_1.0.6-7_i386.deb (--unpack): subprocess new post-removal script returned error exit status 2 open /var/lib/xaw-wrappers/update-wrappers-history: Arquivo ou diret?rio n?o encontrado [file or directory not found] Just a remark: I'm not entirely sure if it wise to use any other than the default locale settings in the root account. I have no idea...I'm desperate, I've tried everything I know. Even dpkg with the force option... Try $EDITOR /var/lib/dpkg/info/tetex-bin.prerm and comment out the offending line or fix the statement so that it doesn't break. Please also consider filing a properly documented bug into the debian bug tracking system. With all the time and effort I put in solving this problem, I'm sttil at ground zero with this package. I am seriously considering the possibility that: 1) There is a bug in the tetex-bin package 2) There is a bug in apt-get (!) Why am I considering apt-get? Because if I try to install *anything*, apt-get simply goes crazy with stupid messages. For instance, when I apt-get install grep-dctrl: The bug is probably in tetex-bin, for having a broken prerm script. The problem on the packaging system level is not with apt-get, but in dpkg, in that it is not possible to deal with broken maintainer scripts in any other fashion than to edit the scripts. Cheers, Joost
Re: ACPI vs. APM
On Fri, Jul 27, 2001 at 03:27:43PM -0400, Steve Gran wrote: Hello all, I was hoping someone could help me out. I'm running kernel 2.4.7 on a an Ali MB w/ a P III chip. The BIOS boot messages tell me that this board uses ACPI as it's protocol, so I enabled that in the kernel, but it doesn't power down the monitor properly. I also have APM enabled, as that's where the various options are. I was under the impression that you could enable both in the kernel and the correct one would load, depending on the architecture. Was I wrong, or is it just a matter of telling LILO which to load? If I have to recompile, I will, but I would really be happier with some other solution. If somebody could also clear up the whole ACPI vs. APM thing (I know they are different methods of handling power by the motherboard - I meant the kernel usage of them) I would also appreciate it. I think that what you want is called vesa dpms. See the setterm(1) and xset(1) manpages for how to set this up on the console and in x11. Cheers, Joost
Re: Debian Install Problem
On Fri, Jul 27, 2001 at 04:28:17PM -0400, [EMAIL PROTECTED] wrote: I'm installing from a Debian CD-ROM I got from ISO Image 1 [ver. 2.2r3]. I partition my hard disk, install the OS and Modules, then the drivers, and usually that goes fine [drivers sometimes fails], and then I choose to install the base system, and 4 years later [not really] it tells me that there was a problem installing the base system, and I see snipets of errors on the sides of the screen: ut error, etc. If this doesn't happen, it just goes right back to the installation screen, with Install the Base System as my default option. If I try to go to Configure the Base System I get an error saying that I need to install it first. Every time I install the Base System it asks me if I want to overwrite the base system already written on the disk, so I know it was successful. I checked the surface of the disc numerous times, and it's just fine-- no scratches, fingerprints, etc. I even cleaned it, but it still does this. Sometimes the disk is bad, but you can't see it on the surface. What are the actual error messages? Press leftalt-f4 to see the installer error messages (it may also be f3, I forget so try all of them and see what you get). Leftalt-f1 will get you back to the installer menu. Cheers, Joost
Re: netscape 4.76 freezes
On Thu, Jul 26, 2001 at 03:43:45AM -0400, [EMAIL PROTECTED] wrote: every now and then (i REALLY dont know at what specific actions) netscape freezes up my machine. i cannot do anything but switch off and let fsck do its magic. has anyone of you experienced this also? i have smp activated in the kernel, to use both of my processors. maybe thats a problem? i read several mailing list archives and the problem does seem to be known. only i havent found a solution yet. Have you tried setting resource limits, so that netscape cannot make your machine run out of memory? man ulimit(1) and setrlimit(2). Cheers, Joost
Re: Off Topic: Mailing Agents?
On Thu, Jul 26, 2001 at 01:00:45PM -0600, John Galt wrote: On Thu, 26 Jul 2001, Peter S Galbraith wrote: The reason? My incoming mail goes through an MS-Exchange server, and it strips out the signature part and makes a mess of the mail header. There's no * ^X-Mailing-List: debian-user@lists.debian.org or * ^From [EMAIL PROTECTED] left to filter on. MS-Exchange sucks. What I've seen it do is this: you send a multipart/signed mime message through an exchange server, and it will munch up half the message headers, all the mime headers, the signature part and inline the signed text part, with a new mime declaration, introducing windows codepage 1251. I could never figure out what went wrong, because when I sent the exchange competence center people an smtp + pop3 dump illustrating the mangling, they were completely baffled about the possibility of sending email with telnet, so I gave up. Only MUA I tolerate with MS-Sexchange is telnet blah 110... retr and linux's terminal paging is perfect for me :) Umm, actually it doesn't or didn't. Fetchmail had to be patched to allow incorrect pop3 list responses from a certain vendor's mail server, because it reported not the actual message size, but the size of the compressed entry in the mail database. It is also explained in the fetchmail FAQ. When the company where this happened, tried to get some of the support they were paying (very big time) for, there was no answer. So two bsd gurus who were affected by this debugged it a bit and even found the likely cause of the problem sent a patch to the fetchmail bazaar. Shortly after the patched version of fetchmail hit the net, the submitters of the patch were contacted by an exchange developer. After some pounding with rfc's, he agreed that it was a bug and gave a magic registry hack to fix exchange, ie. one has to add a key called pop3 compatibility to be compliant with the rfc's. Since knowing about this, it has never been a doubt to me that open source has better support than closed source, even if that's what you're actually using. Cheers, Joost
Re: Modem problems after kernel upgrade (long, w/ Radeon VE sidenote)
On Mon, Jul 23, 2001 at 09:31:42PM -0500, Jeremy wrote: output of dmesg: Serial driver version 5.05a (2001-03-20) with MANY_PORTS SHARE_IRQ SERIAL_PCI ISAPNP enabled ttyS00 at 0x03f8 (irq = 4) is a 16550A ttyS01 at 0x02f8 (irq = 3) is a 16550A Redundant entry in serial pci_table. Please send the output of lspci -vv, this message (4793,4104,4793,215) and the manufacturer and name of serial board or modem board to [EMAIL PROTECTED] ^ you should perhaps do this ttyS04 at port 0xac00 (irq = 5) is a 16550A output of setserial /dev/ttyS2 uart 16550A port 0xac00 irq 5: Cannot set serial info: Address already in use Edit /etc/init.d/setserial to not autodetect ttyS2 and ttyS3 and setup ttyS2 with the above parameters specified. output of setserial /dev/ttyS04 uart 16550A port 0xac00 irq 5: /dev/ttyS04: No such file or directory Maybe it works after you MAKEDEV /dev/ttyS4, but it would be more interesting to get the pci_table problem solved. with anyone and if anyone has any suggestions. Please sanitize your emails when participating in a thread on a mailing list. There's no need to keep resending text that has already been received before by everyone on the list. Cheers, Joost
Re: Further Adventures of my ESS 1888
On Tue, Jul 24, 2001 at 12:51:31PM +1000, Kieren Diment wrote: I posted a message about my ESS 1888 on my Dell Laptop card last week, and am grateful for the most useful replies I got. Aagain, According to Windows95 it is an ESS 1688 Audiodrive (although according to Dell it is a ESS 1888) I/O Range 0240 - 024F 0388-038B 0330-0331 IRQs 5 and 9 DMA 00 and DMA 03 So I added the following to /etc/modutils/sound after reading the Kernel Sound documentation: alias char-major-14 sb post-install sb /sbin/modprobe -k ess1888 options sb io=0x240 irq=5 dma=0 mpu_io=0x330 esstype=1888 This only takes effect after running update-modules, which generates the /etc/modules.conf file that is the used by the kernel module autoloader (kmod). and the following to /etc/modules soundcore soundlow sound uart401 sb io=0x240 dma=0 irq=5 esstype=1888 This file is used by the debian init scripts, to load modules at boot time. It is a different mechanism. all goes fine until I get to insmod sb io=0x240 dma=0 irq=5 esstype=1888 when I get the following error: Using /lib/modules/2.2.19pre17/misc/sb.o /lib/modules/2.2.19pre17/misc/sb.o: init_module: Device or resource busy Hint: this error can be caused by incorrect module parameters, including invalid IO or IRQ parameters I'm not sure if it helps, but try modprobe instead of insmod. Can anyone suggest what I should do next. Did you verify that the card needs no plug-n-play setup? Cheers, Joost
Re: [OT] Making images from CDs
On Tue, Jul 24, 2001 at 12:39:06AM -0400, Thomas J. Hamman wrote: First, a simple question: What is the correct command-line way of copying a CD to an image file on the hard drive? Is a simple dd if=/dev/cdrom of=cd.img command sufficient? Or do I need to run Use cat if you like. cat /dev/cdrom cd.iso9660 Now, a tougher problem: I received an Input/output error while trying to copy a CD to an image file (using xcdroast, and I also got the error while trying dd). I tried another CD and it worked fine, so I'm guessing the problem is the particular CD. The CD does work fine, however. So, what I'm wondering is whether or not there is anyway around such an error, or whether it would be worthwhile to go ahead and burn the resulting image and hope it works. Are you sure it really is fine? Run: find /dev/cdrom -type f -exec cat {} \ /dev/null \; And watch the system console or log for errors. Cheers, Joost
Re: KDE Games
On Tue, Jul 24, 2001 at 10:38:13PM +1000, Steve Kowalik wrote: On Mon, Jul 23, 2001 at 09:31:43PM -0600, John Galt uttered: Let me guess: you're running X as root. Shame shame shame! I'll give you shame. X is setuid root. Well, shame on you too then, he meant the xsession, not the xserver. Cheers, Joost
Re: Problem with apt-get removing a package
On Mon, Jul 23, 2001 at 04:32:47PM +0100, Richard Gaywood wrote: On Mon, Jul 23, 2001 at 10:42:56AM -0400, Andrew Dixon wrote: I'd also recommend getting rid of Ximain and installing Gnome from the Debian .debs. It really does eliminate a lot of headaches. Know of any rough instructions on how to do this? Is it just a matter of taking Ximian out of sources.list, apt-get remove *gnome*, then apt-get install *gnome*? Not so long ago, I posted a way to do this to debian-user. Use google or the search engine on the debian mailing list archive web site. Cheers, Joost
Re: ALSA compile problems
On Wed, Jul 25, 2001 at 12:21:07AM +1000, Rob Weir wrote: Yep. Nothing about it, that I could see. Am I just stupid? Not for asking questions. Maybe it is a bug in alsa-driver-0.5, if the other modules built fine. Try to dig up more details and file a bug if you think there is one. Cheers, Joost
Re: [OT] Making images from CDs
On Tue, Jul 24, 2001 at 11:54:14AM -0400, Stefanus Du Toit wrote: On Tue, Jul 24, 2001 at 01:19:24PM +0200, Joost Kooij wrote: find /dev/cdrom -type f -exec cat {} \ /dev/null \; doing it on the device file won't get you anywhere :) Who said it has to be a device file? (tested, this time) :-) rm -rf /dev/cdrom mkdir /dev/cdrom mount -t iso9660 -o ro /dev/hdb /dev/cdrom find /dev/cdrom -type f -exec cat {} \ /dev/null \; It's not weird, it's unix. Cheers, Joost
Re: Debian Firewall
On Tue, Jul 24, 2001 at 10:11:19AM -0400, Case, Benjamin wrote: Security, Security, Security SSH Daemon NAT (Masq) Port Forwarding Graphical (web based ?) Network Analysis PPPoE support VPN support Convenient Method of Configuration (Web based, GUI based ?) [snip] What is the best apporach to creating this Firewall. Should I start with my own basic install of Debian and build from there ? Is there a floppy or CD based image worth trying that is based on Debian ? Install a debian base system. In the dselect package listing, remove all packages that are not needed on a firewall, like gcc, tetex and any bad stuff like telnetd or rwhod. Then select the packages you do want: ssh, ipmasq, pppoe, mrtg, perhaps a tiny httpd for the stats. Install the packages from the dselect menu. Repeat for any other packages you later find you need or don't need. I'm not very experienced with gui administration and I personally don't find it convenient at all. On a security sensitive system, you don't want to run anything more than strictly necessary, fancy configuration layers included. Just consider the various webinterfaces in embedded systems, like routers and network printers, and how these are accidentally hurt by iis sploit requests. Remember to netstat -at and to mercilessly remove any service that you did not put there yourself with the express intent to respond to arbitrary people on the internet. There exists a debian-firewall list, iirc. Try searching the archives of that list and posting there, it likely has a better yield. Cheers, Joost
Re: bash: man: command not found
On Tue, Jul 24, 2001 at 12:55:43PM -0700, Shriram Shrikumar wrote: wonder if someone can help me. was playing around with dselect trying to fix a package dependency issue, removed 2 many packages and now it says that 'man' can no longer be found. error message as in subject what needs to be installed for man ? manpages and man-db Cheers, Joost
Re: Non X-windows GUI web browser?
On Tue, Jul 24, 2001 at 07:58:24PM +0100, Frank Zimmermann wrote: Lance Peterson wrote: I'm trying to use Webmin to administer a Debian router/firewall, but I was hoping not to install X-windows in order to use a GUI browser. Is there a non-X GUI based web browser available for Debian that I can install to use Webmin without loading X-windows? I don't think you need to install X on your router. You fire up webin from your local machine and administer the router remotely. But then you'd have to make webmin listen to the network. If you only bind it to 127.0.0.1, you can still use the webmin interface in a browser run locally. For text mode browsers, see links, w3m and lynx. Cheers, Joost
Re: OT?:Proper owner of html files in Apache
On Tue, Jul 24, 2001 at 03:13:25PM -0400, Ken Januski wrote: What I'm trying to find out is if root.root is a good idea? I assume it is or it wouldn't be the default. It just seems odd to me to have to become root in order to write either a html or cgi page. You can setup the ownership of the webpages just like you like it. Read a book about unix permissions and ownership management and setup a nice scheme. It really depends mostly on your particular setup and needs. That is also one of the reasons debian sets no standards here, other than that the local admin sets the standard. Possible setups: root.root owned files, some user edits copies of the files in a local directory. When ready, the files are copied to the webroot by root. *.webwackers owned and group writable files, with all users who are supposed to be able to edit webcontent a member of that group. Put the sgid bit on the directories, if you like. This scheme can also be combined with the edit-a-copy scheme in the above. www-data.www-data owned files are evil, because then the webserver process can modify files. This is unwanted if the webserver process is somehow compromised and precisely the reason for the separate www-data userid, it is a dedicated nobody user. As all cgi scripts by default will also run as www-data, their output files are owned by www-data also, which is ugly for the above reasons, but hard to prevent. Because you can make virtual servers and scripts run under alternate userids of your own choice, your options are limited to your imagination. Cheers, Joost
Re: bash: man: command not found
On Tue, Jul 24, 2001 at 03:34:01PM -0700, Shriram Shrikumar wrote: was missing man-db, however, on trying to install, it shows the following error message trying to overwrite directory '/usr/share/locale/de' in package texinfo with nondirectory any ideas ? Can you give more error output? Cheers, Joost
Re: your mail
On Tue, Jul 24, 2001 at 05:01:15PM -, [EMAIL PROTECTED] wrote: Before to install the OS on my TP365XD, I would like to know if it's possible and if I need to take care of special issues. Thank you in advance for you response. Did you try searching the web for debian thinkpad 365 on google.com? Cheers, Joost
Re: forcing a pci nic to use a different irq?
On Tue, Jul 24, 2001 at 06:25:19PM -0500, Matthew Garman wrote: cat /proc/interrupts 11: 5986 XT-PIC sym53c8xx, eth0 append=ether=9,0,0,0,eth0 Then I *ran lilo*, rebooted, but the result was the same: the card still loads on IRQ 11. The card *used* to work fine. In fact it just started having this problem, and I can't figure out why. Perhaps you upgraded your kernel? Or you moved cards around? You can manually set interrupts in most bios setup menus, I believe. Also look at the setpci(8) manpage in the pciutils package. Cheers, Joost
Re: killing dead cdrecord processes
On Sun, Jul 22, 2001 at 02:46:55PM +1000, Drew Parsons wrote: I've been trying to burn a CD using gcombust. For some reason the burn got stuck halfway through (the output of cdrecord, which gcombust calls, said that all of a sudden one of the SCSI commands couldn't be understood. It's a HP USB CDWriter using the usb-storage module, which uses a series of scsi modules.) Well, so much for the stability of scsi over usb. I closed gcombust and tried running it again to blank the CD and start again. But it turns out that the old cdrecord processes are still running: $ ps aux | grep cdrecord root 1978 0.0 0.2 1412 232 pts/0D13:39 0:00 /usr/bin/cdrecord -reset dev=1,0,0 root 1983 0.0 5.8 5512 5512 pts/0DL 13:42 0:00 /usr/bin/cdrecord dev=1,0,0 -load root 2018 0.0 5.8 5512 5512 pts/0DL 13:53 0:00 /usr/bin/cdrecord -v -dummy -pad speed=4 dev=0,6,0 blank= all The state is D, which means uninterruptible: the processes do not respond to kill -9. They will, once they can be interrupted again. Is there any way short of rebooting that I can clear these processes out? Other than removing the cdr medium, resetting the peripheral, removing and reinserting the driver modules (if at all possible) or some sort of futzing with the usb interfaces, rebooting may be the only thing. Cheers, Joost
Re: lilo problem
On Sun, Jul 22, 2001 at 01:31:20AM -0700, Fallen Lord wrote: 1.)i just upgraded (force-overwrite) libc6-2.1.3 and libdb2 that comes with the debian 2.2r3 with the testing packages of libc6-2.2.3 and libdb2-2.7. Be very careful with any of the --force options to dpkg, they're not supposed to be needed, so if you use them, something is likely wrong somewhere and you might be making it even worse only. Use dselect for package management. Learn to understand the concepts behind the debian package management system and you'll find that dselect is a fine tool. Don't be scared by people who say that it is too hard to use, they just don't understand the packaging system. If you don't understand the packaging system, you're likely to nuke your system somehow some day anyway. after that i recompiled kernel 2.4.7pre3. when i ran lilo, it flashed out a message open /vmlinuz - no such file or directory when i used whereis open - open was still there. I couldn't use lilo because of Perhaps this was only an error message by lilo that the call to open /vmlinuz failed (i.e. open returned with a failure). The open that it complains about is probably not the open that you find when you use which. Check out the differences between the open(1), the open(2) and the fopen(3) manual pages. If your /etc/lilo.conf lists an image with the name /vmlinuz, and that file does not actually exist in your filesystem, lilo cannot setup the boot block to point to a place on your disk where that file is supposed to be (but isn't). If /vmlinuz is a symbolic link, eg. to /boot/vmlinuz-2.2.19, then the file, that the link is pointing to, must also exist. Read the lilo.conf manual page to find how to setup the right boot image for your system. this. Then i made the mistake of uninstalling lilo, so No need to remove lilo from your system, I would think? Unless you want to switch to grub as your bootloader, but I don't expect so and it would only create more confusion. when i rebooted the pc, LI instead of LILO: appeared (meaning LILO is damaged?). i'm triple-booting QNX4, You rebooted the machine whilst it did not have a valid lilo bootloader. H... One way of getting back in is to boot from a installation floppy with the boot parameters: rescue root=/dev/your debian root partition Then login as root and reinstall lilo, fix the /etc/lilo.conf and this time, run lilo -v -v (I always do that) so you can see more exactly what is going on. After all, the results of little mistakes are relatively large, so a little more attention is appropriate. You also need to understand the difference between /sbin/lilo, which is also called the boot block installer, and the little piece of code called the boot block, which is installed by /sbin/lilo onto special places on your harddisk. The boot block is what actually gets executed by the bios when your pc boots and it in turn loads the linux kernel. When /sbin/lilo is run, it updates the boot block's notion of where to find the kernels listed in /etc/lilo.conf. Win98 and Debian GNU/Linux 2.2r3 on my system. i went to the point of ridding lilo from the mbr just to use an OS (Win98 then loadlin-ing the kernel). How do I fix this? If you do a lot of rebooting and os switching, then loadlin.exe might not be such a bad idea after all, as long as you are diligent in making sure that a bootable kernel is always available as a file in the windows98 filesystem. Use the debian rescue floppy to boot into your system and read the lilo documentation (there's an awful lot of good details in /usr/doc/lilo if you care) and fix the boot loader. Alternatively, create a simple bootfloppy: cat /vmlinuz /dev/fd0 rdev /dev/fd0 $( rdev | awk '{ print $1 }' ) Voila, you have a bootdisk that always works, if the disk is any good, that is, and if you make it read-only. Make two if you want to be sure. And make sure that /vmlinuz is indeed the kernel that you want to boot. Use another filename if you like, there is nothing that says that your kernel must always be named vmlinuz or be in places like / or /boot. After all, if you cat it to a raw floppy, it doesn't have a name or a place either, since there is no filesystem. BTW, the rescue disk does have a FAT filesystem, so it can be read by windows as well, and the linux kernel is named linux. Be careful when you replace it, because it needs to be a kernel that can do special tricks, called initrd, so the installer can read the basic programs it needs from another file on the floppy. You do not need those for your own simple bootfloppy created by cat'ting your current kernel to /dev/fd0. 2.)Also, how do i install a kernel 2.4.x in potato? i hear that one needs to make kpkg the kernel source. can't i just copy the kernel image and install the sources (make bzImage; make modules; make modules_install) or do I really need to make kpkg the source just to make the system detect the modules? if so, how do i do it and what do i
Re: ipchains: cannot open file `/proc/net/ip_fwnames' (was: Re: No such file or directory - huh?!)
On Sun, Jul 22, 2001 at 10:59:15AM +0200, Gary Jones wrote: Joost Kooij wrote: On Sat, Jul 21, 2001 at 08:34:48PM +0200, Gary Jones wrote: ash-ock:/etc/init.d# ./firewall bash: ./firewall: No such file or directory [snip] What's going on? The script file is definitely there In the script, you are using a command with a tpyo in it or that is located in a place not in your current $PATH. Nope. See later for how I know why not... Perhaps the command is ipchains (/sbin/ipchains) and you are used to doing su to become root? In that case, next time do su -, so you get a propor root login, with all the sbins in $PATH. No, I ran that scripting session as root so that I wouldn't get anything silly like ownership issues. and Tim Moss wrote: The No such file could be referring to the shebang line. Does /bin/sh exist? Yes. If it wouldn't, all sorts of other things would also break, pretty violently. I still don't know what caused the problem. What I ended up doing was something like: cp ../init.d/firewall ../init.d/firewall.old cp ../init.d/network ../init.d/firewall jed ../init.d/firewall ../init.d/firewall.old and then copying the contents of 'firewall.old' into 'firewall'. After that I didn't get No such file or directory any more, though the original reason is still a mystery to me. Thanks for your help, though. Too bad that you do not have both files anymore, or you would have been able to at least make a diff of the working and the non-working version. Now I get: ash-ock:~# ipchains -F ipchains: cannot open file `/proc/net/ip_fwnames' [which is not surprising, since...] ash-ock:~# ls -la /proc/net/ip* -rw-r--r-- 1 root root0 Jul 22 10:29 /proc/net/ip_forward -rw-r--r-- 1 root root0 Jul 22 10:29 /proc/net/ip_input -rw-r--r-- 1 root root0 Jul 22 10:29 /proc/net/ip_output *sigh* I thought creating all the stuff required was the job of the install routine? Are you sure that you enabled firewalling support in your kernel configuration? I also get setsockopt : protocol not available when trying to set the policy. FWIW this is ipchains --version 1.3.4 (as per standard 'slink' distro, I believe) -- Gary Debian 2.1r4 (kernel v2.0.39); XFree86 3.3.6 ^^ That may also explain these other problems. Consider upgrading the machine to a newer debian release and a newer linux kernel. The facilities you are trying to use may not be supported very well or at all in the kernel and tools that you are using. Also, older debian releases do not get any official security updates. You need to run the stable release for those. Cheers, Joost
Re: Newbie Question
On Sun, Jul 22, 2001 at 11:30:13AM +0200, Chuan Guo wrote: a silly question: Are there some simple method, not dselect, to uninstall sth. in a group, i.e. Gnome, i'd like use only Blackbox in my small Notebook. Package dependency sets are a complicated beast. Dselect in fact makes it easier to manage packages, if you accept the dificulties that are at the conceptual level. but i have already installed gnome with tasksel, and use tasksel i cannot uninstall, or Problem is my test version Woody? You can, if you use dselect correctly. Read the dselect documentation. Use dselect. If you do not understand something particular about it, feel free to ask on debian-user. If you have an idea to make it better, feel free to send a patch to the bug tracking system. IMHO dselect is difficult is not very sensible to say, vi is more difficult. If vi is too difficult, then unix is not for you. Maybe computers, even. Cheers, Joost
Re: Keeping kernel compilation options
On Sun, Jul 22, 2001 at 05:14:22PM +0200, Mart van de Wege wrote: Sorry to butt in in this thread, but my question seems to belong here as well. I noticed yesterday that installing kernel-package on i386 does not install bin86, which is necessary to build the boot code. Now I know Joost is going to say that I should have used dselect :), but I think that bin86 You should have used dselect :) Cheers, Joost
Re: lprng for a home computer
On Mon, Jul 23, 2001 at 02:06:21AM +0900, Marshal Wong wrote: Philipp Lehman [EMAIL PROTECTED] writes: I can't help you with your lprng question, but a firewall actually makes sense even on a stand-alone workstation or laptop. You can filter in the input chain just like you'd do on a dedicated firewall host. I'm not an expert on firewalls, but if someone wanted to bring your computer to a grinding halt, i.e. DoS, they could just send a whole crap of packets, and firewall or no, the processor will have to spend all it's cycles dealing with these packets. If course, I guess it would happen if you didn't have a firewall too, wouldn't it? With any decent modern system, you'll likely be dos'ing the line, not the cpu, unless you have hundreds of ipchains with hundreds of rules each, which is unlikely to be the case on a personal machine firewall. A better solution anyway is to have a dedicated firewall machine. That way, you can install gnome and all the weird stuff that it needs, without having to fear that any of it is listening directly on an untrusted network. On a firewall, you can turn off all services except ip packet forwarding/masquerading. On your desktop, it would impede your productivity (read entertainment and spiffy gui). Having said that, it may nevertheless be a good thing to also employ some ipchains rules on your personal desktop. But it would mostly be useful for monitoring purposes, I think. So it would only be actually useful if you really regularly check those logs. Generally speaking though, if you know a bit of unix, don't bother with those personal firewall products, but give a 486 a second life instead. Cheers, Joost
Re: grave pppconfig bug
On Sun, Jul 22, 2001 at 12:25:23PM -0500, John Hasler wrote: Can anyone out there reproduce this? Package: pppconfig Version: 2.0.8 Severity: grave If I add a new connection pppconfig will allways select Number of all menu items. Quit doesn't work either. Only thing which works is quit. In the main Menu the same effect happens. It doesn't matter which item I select, it allways uses the first one. The menu selection works fine here, on sid, pppconfig version 2.0.8. It sounds like a dialog/curses/terminal problem. Have you verified that angle? Cheers, Joost
Re: Email line-length defaults to about 76; how to increase?
On Sun, Jul 22, 2001 at 01:48:57PM -0400, Jameson C . Burt wrote: My email lines get split after about 76 characters. How could I change this to something longer, or should email lines be split at 76 characters? Please do keep sending your regular email with line breaks at +/- 70 characters. It is not just a matter of style, it is about common courtesy. This limit causes problems whenever I email Linux syslog lines, which are seldom less than even 90 characters in length. I haven't been able to determine if this line-length limit is set by exim, procmail, or perhaps my mail user agent (balsa). AFAIK it is the user that is responsible for proper line formatting, but some mua's will do that automatically by default. It would not surprise me if balsa did, it being a featurefull mua. If you pipe or redirect your data to mail(1), neither it nor your mta should be changing the contents of your message. If you want to use balsa and still be assured of the integrity of the data you are sending, use mime to attach the data to the message or uuencode the data and include that in the message body. mail -s here's your data [EMAIL PROTECTED] datafile Cheers, Joost
Re: high resolution fonts?
On Sun, Jul 22, 2001 at 01:08:00PM -0500, Matthew Garman wrote: Is there a package that will give me some high resolution fonts, or at least allow my existing fonts to scale better? For example, I would like to design a simple logo for my website using the gimp. If I use my text tool to create *large* text, then the text comes out looking very pixelized (i.e. squared off or boxy, no smooth edges). Also, I noticed that in mozilla, any text surrounded by the h1 tag suffers from the same problem: the characters aren't smooth. Any ideas? For X11 in general, make sure that you have the xfonts-scalable installed or that you are running a truetype fontserver and have some truetype font sets installed. For gimp: look at the freefont and sharefont packages. The gimp package actually contains suggests: for these packages. Did you use dselect to install the gimp package? Raw apt-get ignores all recommends: and suggests:, but these are there for a reason. Use dselect. For the mozilla problem, configure it to use scalable fonts, and install scalable xfonts or install a truetype fontserver. Cheers, Joost
Re: ALSA compile problems
On Mon, Jul 23, 2001 at 12:21:39AM +1000, Rob Weir wrote: Does anybody have any idea about what to do? If it's a bug in make-kpkg, is it known or fixed anywhere? Did you look yet on http://bugs.debian.org/kernel-package ? Cheers, Joost
Re: Unable to upgrade libc6, cannot find why
On Sun, Jul 22, 2001 at 10:00:39AM -0400, J.F.Gratton wrote: --- Preparing to replace libc6 2.1.3-18 (using .../libc6_2.2.3-7_i386.deb) ... cp: invalid option -- L Try 'cp --help' for more information. dpkg: error processing /var/cache/apt/archives/libc6_2.2.3-7_i386.deb (--unpack): subprocess pre-installation script returned error exit status 1 --- It's already filed as an important bug in the libc6 package. Sorry if this seems long, I thought that putting the whole message might help people. That is fine, you provided the right data in fact. I went throught the /var/lib/dpkg/info/libc6.* to find any mentions of cp -L without finding anything. It broke during the unpacking, as the error messages say. In such case, dpkg will roll back to the previous version of the package. So the preinst script on your disk is the old one, without the cp -L. My system is now stuck half-installed as I cannot continue (libc6 being important as it is for other packages). Anyone knows where to go from there ? Thanks I'm afraid you're stuck until a fixed libc6 is uploaded. Maybe you can suspend dpkg just between unpacking the new preinst script and executing it, so you can fix it and then continue dpkg in the foregroung again, so it will run the fixed preinst script. It is a bad hack though and YMMV, ranging from fixing to completely hosing your system (no kidding). Your best bet is to sit on your hands until this is fixed in libc6. Cheers, Joost
Re: kernel 2.4: where's the FM?
On Sun, Jul 22, 2001 at 07:59:15PM -0400, dman wrote: I've just spent some time on kernel.org and elsewhere, but I can't find any FMs to R regarding the procedure to upgrade from kernel 2.2.19 to 2.4.x. Is an apt-get install enough (and point grub at the new image of course)? Where can I find an FM to R on this? linux/Documentation/Changes Cheers, Joost
Re: Removing a brokenly nstalled package?
On Sun, Jul 22, 2001 at 08:07:37PM -0400, Stan Brown wrote: By not paying enough attention to the depedncies that dselect picke, I would up with a broken install of aolserver. I don't need this package at all, but dpkg --remove aolserver fails. How can I remove all traces of this package? How does removing fail exactly? Can you unselect it in dselect and subsequently run remove from the menu? Cheers, Joost
Re: it keeps crashing
On Sat, Jul 21, 2001 at 09:43:39AM +0200, Guy Geens wrote: Martin == Martin F Krafft [EMAIL PROTECTED] writes: Martin instead, it keeps crashing on me... kernel panic in pid 0 Martin process swapper. however, memtest86 reports no errors for Martin the RAM chip, and badblocks, run with the destructive write Martin option, reports no bad blocks within the swap partition. My guess is that you simply have not enough memory. AFAIK 8MB is enough to boot linux, or it should be. As to why Martin is having crashes, I don't know. Try it with a smaller kernel image size, as Guy suggests, by leaving out options that you do not strictly need. Also be sure to use the latest available 2.2.x or 2.4.x kernel and if the crashes persist, read the ksymoops documentation and the linux-kernel mailing list faq, and post a decoded oops on that list. Cheers, Joost
Re: installing kernel-debs
On Fri, Jul 20, 2001 at 01:37:09PM +0200, Joerg Johannes wrote: I have, in the last few weeks, compiled a lot of kernels (with make-kpkg) Well, I have installed these kernel-image-blah.deb's, but without removing the old ones. dpkg -l shows ii kernel-image-2. # -- dpk -l does not show more of the package version ii kernel-image-2. ii kernel-image-2. ii kernel-image-2. ... This is broken dpkg behaviour. You can work around it by typing: COLUMS=200 dpkg -l foobar How can I safely remove the old ones (whose files were overwritten by the newer ones) ? No files should be overwritten, or else there is a bug in kernel-package. I could make a 2.4.5 kernel and remove all the 2.4.3.deb's, but will this work? won't dpkg complain about files that are not there but should be there? Unless you know how to play with flavours, you will not have multiple instances of the same kernel version installed on your system. What files do you mean that dpkg should complain about? Every kernel-image package has its own files, that must not be overwritten by any other packages. If you try to install a kernel-image of a kernel version that already has a kernel-image installed, then dpkg will treat the new install as an upgrade and cleanly replace all files. Cheers, Joost
Re: NFS alternative
On Fri, Jul 20, 2001 at 07:26:21PM -0700, Francois Gouget wrote: I don't know of a reference book that would be specifically about NFS but you can have a look at The Design and Implementation of the 4.4BSD Operating System by McKusik, Bostic, Karels and Quarterman. Chapter 9 (about 25 pages) is about NFS and should give you a feel for it. And the book as a whole is just awsome. Definitely a must have if you're interested in operating system design. Definitely a must if you want to understand unix in general. There is also a video set to go with the book, hosted by McKusick, in which he says about NFS that it really should mean no filesystem semantics. There are more cute anecdotes in it, like the one where McKusick expresses his regret about not having sided more strongly with Richard Stallman, when he proposed POSIX_ME_HARDER as the name for the variable that ended up with the name POSIXLY_CORRECT. Cheers, Joost
Re: updating to woody with custom kernel
On Fri, Jul 20, 2001 at 11:11:22PM -0700, Kurt Lieber wrote: On potato, I had to re-compile my kernel to get IP aliasing support. I used the newbiedocs over at sourceforge, and one of the instructions was to do the following: echo kernel-image-2.2.19 hold | dpkg --set-selections The newbiedocs have been misleading you. Use dselect for this. Which I assume tells dpkg not to ever upgrade kernel-image until I tell it otherwise. I'd like to upgrade my workstation from potato to woody, but I'm not sure how my re-compiled kernel fits in to all fo this. Not, so don't worry about it. So, what do I need to do to remove this flag? man dpkg tells me what the hold flag is for, but not how to remove it. I think I can bypass it with --force-things hold, but I have a feeling that's a sledgehammer approach. I'm hoping there's a kinder, gentler way. It is not a sledgehammer, it is the safety on the gun. Please use dselect as a user frontend to package management. Please also read the dpkg manpage, it mentions dselect prominently and also explains what the --force options do. Second question; when I upgrade to woody, is there any relatively easy way to have it compile in support for IP aliasing or will I have to recompile the kernel again after the upgrade is complete? You current kernel supports it, so what is the problem? You do not need to change your kernel when you upgrade you userland tools. Cheers, Joost
Re: Frame buffer at boot
On Sat, Jul 21, 2001 at 02:45:36AM -0700, Vineet Kumar wrote: The real question is: To whom is the message encrypted? Anybody on the Internet can access it, but probably only one person can read it. I suppose that's one way of keeping the spooks in the dark about who your covert contacts are... The spooks don't really put much effort into looking in your mail. The much maligned echelon eavesdropping spook system mostly notes who sends any mail to who. A few years ago, I visited the nsa webpage and followed some link to a publicly visible project. It was a huge computing system, to do calculations on sparse matrices, to be used for processing of astronomical data. Ever since, I've been wondering what eigenvector my ip address has. So you see, they don't need to know what you're writing, because they already do, as they know who you're writing it to. Let a thousand eigenvectors bloom, to paraphrase an infamous Chairman. Cheers, Joost
Re: [OT] Perl: exec and $variables
On Sat, Jul 21, 2001 at 01:04:40PM +0200, Sven Burgener wrote: I have a problem with some perl code. I know this is off-topic, but there are numerous knowledgeable people on deb-usr, so forgive me for posting this. Now to my problem. Given the following variable, my $BEGINREGEX = sprintf(\^!-- // begin of news\$no // !--\$\); Please tell us what you're trying to accomplish first. It is unclear what assumptions you are making. the following eval() call doesn't substitute $no for what it is and return the new string with the substituted $no in it: my $no = 1; my $bla = eval($BEGINREGEX); print $bla\n; $bla is empty for some reason. So how would I go about having $no substituted in $BEGINREGEX for whatever it happens to be set to at a particular point in the script? You probably do not want to use eval here, or at least not in this way. Cheers, Joost
Re: url-escaping a string in a shell script.
On Sat, Jul 21, 2001 at 01:37:26PM +0200, Martin F. Krafft wrote: hi, assuming that i have a string of the form 123%%%blabla*($( available in the shell script, how can i convert that into an escaped version for use with the HTTP protocol? i am not a perl wizard, or else i wouldn't ask. and maybe there is a better way. perl -MURI::Escape -ne 'chomp; print uri_escape($_), \n' A little scriptlet to do the same: #!/usr/bin/perl -w use URI::Escape; chomp, print uri_escape($_), \n while (); Cheers, Joost
Re: I did it a mess with perl
On Sat, Jul 21, 2001 at 01:22:19PM +, Victor wrote: perl -MCPAN -eshell *The bottom line* So, now I have that strange, confusing mixture of a perl 5.005 coming from a deb package and a perl 5.6.1 compiled from tarballs in CPAN under the same /usr/bin and /usr/lib. Obviously the perl command was overwritten during the compilation by version 5.6.1. *Your suggestion for a way out* Reinstall all the perl debs on your system. If you use the --reinstall option to apt-get, it is almost easy, even. Minimal: Because I need various DBDs which are not included in the deb files, is there away to fix perl as it is now in my PC? Reinstall all perl packages that dpkg thinks are still on your system. Tiresome: I luckily did a complete backup of my system tarring all the main dirs on another linux box via NFS. Therefore I could have a chance to recover my initial, working greatly, stable configuration. But With debian, you only need a backup of /etc and of /var/lib/dpkg/status. The rest can be installed completely. There are extra copies of the status file in /var/backups, too (useful when you have hosed only the status file and do not want to reconstruct it manually). what should I do to install a new DBI and many DBM and whatever else stuff from CPAN which are not included in the deb packages? Install dh-perl-make rtfm build your own libperl-dbd-foo.deb. dpkg -p dh-make-perl Package: dh-make-perl Priority: optional Section: devel Installed-Size: 92 Maintainer: Paolo Molaro [EMAIL PROTECTED] Architecture: all Version: 0.8 Depends: debhelper (= 3.0.0), libpod-parser-perl, perl, make, dpkg-dev, fakeroot Filename: pool/main/d/dh-make-perl/dh-make-perl_0.8_all.deb Size: 12988 MD5sum: 6cdf821df978d5da3f0eed7ad3096cac Description: Create debian packages from perl modules dh-make-perl will create the files required to build a debian source package out of a perl package. This works for most simple packages and is also useful for getting started with packaging perl modules. Given a perl package name, it can also automatically download it from CPAN. Cheers, Joost
Re: Again NE2000 network device
On Mon, Jul 24, 2000 at 01:28:01PM +0200, Bj?rn Fischer wrote: Please fix the clock in your windows settings. It is screwing up the mailbox sorting of many debian-user subscribers. Maybe as a Hint for other - still - Windows users: Don't rely on what Win tells you. I mean you can not rely on Windows anyway, but do not think hardwareinformation given by Win is correct. And don't trust the time information either. ;-) Cheers, Joost
Re: What's happened to the task- packages?
On Sat, Jul 21, 2001 at 07:10:23AM -0500, Brian McGroarty wrote: What's happened to the task- packages? Suddenly task-c-dev and the other programming-related task packages are listed as 'obsolete' on my system. Have these been replaced by something new? The task-* packages in their current form have been pulled from the archive. Properly maintained metapackages are to come in their place, some have already appeared. Cheers, Joost
Re: url-escaping a string in a shell script.
On Sat, Jul 21, 2001 at 03:11:06PM +0200, Martin F. Krafft wrote: also sprach Joost Kooij (on Sat, 21 Jul 2001 01:56:26PM +0200): perl -MURI::Escape -ne 'chomp; print uri_escape($_), \n' A little scriptlet to do the same: #!/usr/bin/perl -w use URI::Escape; chomp, print uri_escape($_), \n while (); almost, except that an input of $1$29492948$6uK7lvoFHD2wWI.P.yF111 is output as $1$29492948$6uK7lvoFHD2wWI.P.yF111 even though the '$' character needs to be escaped for HTTP... why? Why needs the '$' character be escaped for http? If you insist on escaping everything, use: uri_escape(foo, \0-\377) Of course, this is all explained in the URI::Escape manual page. Cheers, Joost
Re: url-escaping a string in a shell script.
On Sat, Jul 21, 2001 at 03:43:38PM +0200, Martin F. Krafft wrote: in fact, this doesn't seem to work at all: fishbowl:~ echo '$1$19496519$xnqy/01WTA6pfhLBqZT13.' | \ perl -MURI::Escape -ne 'chomp; print uri_escape($_), \n' $1$19496519$xnqy/01WTA6pfhLBqZT13. what am i doing wrong? You read the wrong rfc, the above characters are all allowed in http. Try it again, using spaces, '%', '#' and some control characters. Those will be escaped. Cheers, Joost
Re: video drives
On Tue, Aug 07, 2001 at 10:22:41PM +0900, Outa wrote: I need drives to install the graphical mode but so hard to find this driver .. I don't know where? the video card is trident cyber blade i1 .. comes together compaq presario 1600 xl 258... I'll hope for one solution You need to start here: http://linuxdoc.org/LDP/gs/gs.html Read it thoroughly, it will answer most of the questions that you have and will have. Also look at the other texts at http://linuxdoc.org/ no more thanks for the attention... send reply for this adress [EMAIL PROTECTED] Next time, you do some of the work too: set that yahoo address in the email From: address, so that all replies are automatically sent there. Cheers, Joost
Re: Debian books
On Sat, Jul 21, 2001 at 08:14:25AM -0400, alex wrote: Can someone tell me what they consider to be a good up to date Debian book for a beginner, one that doesn't assume that the reader has a background in Unix or DOS? Is there such a book? Most of the interesting literature is available online at http://linuxdoc.org/ Have a look at getting started and the network administrator's guide. One of the best introductions would be learn to use, understand and program the bash shell. Read the bash(1) manual page and the whole lot of shell scripts that many programs on your system are. They are of very high educational value, because you'll learn both about the shell and about the system. One very good book that is only available in the stores: Essential system administration, by aeleen frisch (o'reily). The book covers many different flavours of unix and examines linux only cursorily. Where are the books about Debian? I found dozens of up to date books for RedHat and just one out of date book for Debian (O'Reilly) in the 4 large bookstores that I visited. It's not that the Debian books were sold out, there just don't to be many published. Check out the documentation secion on http://www.debian.org/ Debian is very generic and of all the linux distributions that I know, it keeps in line with standard unix practices the most. Cheers, Joost
Re: [OT] Perl: exec and $variables
On Sat, Jul 21, 2001 at 02:31:58PM +0200, Sven Burgener wrote: On Sat, Jul 21, 2001 at 01:46:25PM +0200, Joost Kooij wrote: On Sat, Jul 21, 2001 at 01:04:40PM +0200, Sven Burgener wrote: my $BEGINREGEX = sprintf(\^!-- // begin of news\$no // !--\$\); Please tell us what you're trying to accomplish first. It is unclear what assumptions you are making. What I want is the variable $BEGINREGEX to contain a string like so: ^!-- // begin of news1 // !--$ or ^!-- // begin of news2 // !--$ The digit after the news should be whatever $no is set to at that point in the script. You are still not telling really what you want to accomplish, but I infer that you want to match lines like: !-- // begin of news1 // !-- To test if the entire $line matches it, you would write: $line =~ m(^!-- // begin of news1 // !--$); Notice that I used the m operator explicitly, so I can use an alternate regexp delimiter, or else I would have had to escape each of the slashes in your pattern. What is the need for the seperate variable $BEGINREGEX? It complicates things enormously when you want a variable $no to be evaluated whenever $BEGINREGEX is evaluated. The only sane way out is to completely reevaluate $BEGINREGEX after each change to $no. To do that successfully, you have to escape '$', '', and '\' and then escape some of the escapes, but others not, depending on wheter they should never be expanded, expanded in the eval or expanded when applying the regexp. I wouldn't touch that with a ten foot pole if I were you. If you succeed at it, you have great job security, and a maintenance nightmare. Easier is to not use a $BEGINREGEX at all: $line =~ m(^!-- // begin of news$no // !--$); should always work, for the current value of $no. my $no = 1; my $bla = eval($BEGINREGEX); print $bla\n; $bla is empty for some reason. You probably do not want to use eval here, or at least not in this way. What should I do then? It's simple, really. I am sure I am just making a stupid mistake. my $BEGINREGEX = sprintf(\^!-- // begin of news\$no // !--\$\); my $no = 99; my $bla = eval($BEGINREGEX); print regex string: $bla\n; What should be printed: regex string: ^!-- // begin of news99 // !--$ Why are you putting the sprintf in the regexp at all? The '^' and '$' anchors only work when at the begin, resp. at the end of the whole regexp. I think that the use of sprintf is unnecessary, and even complicates things enormously. But it isn't, so what am I doing wrong here? AFAICS, you're just not doing it in the most straightforward way. Try to use fixed regexps, and leave the $no in it, so it will be expanded every time perl uses the regexp. Cheers, Joost
Re: url-escaping a string in a shell script.
On Sat, Jul 21, 2001 at 05:07:23PM +0200, Martin F. Krafft wrote: also sprach Joost Kooij (on Sat, 21 Jul 2001 03:53:58PM +0200): You read the wrong rfc, the above characters are all allowed in http. Try it again, using spaces, '%', '#' and some control characters. Those will be escaped. The restricted set of characters consists of dig? its, letters, and a few graphic symbols chosen from those common to most of the character encodings and input facil? ities available to Internet users: A .. Z, a .. z, 0 .. 9, ;, /, ?, :, @, , =, +, $, ,, # reserved -, _, ., !, ~, *, ', (, ) [...] Some of the uric characters are reserved for use as delimiters or as part of certain URI components. These must be escaped if they are to be treated as ordinary data. Read RFC 2396 for further details. you can see that both '$' and '/' are restricted (reserved characters), and these are escaped by browsers and other HTTP clients during form submissions - which is essentially what i want to fake from the command line. They are not restricted, they are reserved, which the rfc explains as: you can use them, unless the particular uri of which they are a part gives them special meaning, in which case they must be escaped. Read section 2.2 of the rfc. http://www.rfc-editor.org/rfc/rfc2396.txt Of course, if you want to fake the behaviour of webbrowsers, avoid all standards like the plague. Certain browsers will escape all characters when sending a request to the server. Why do you think that .asp sites regularly have spaces in uris, without the designer being aware of that? Cheers, Joost
Re: Motherboards
On Sat, Jul 21, 2001 at 04:34:23PM +0100, Keith O'Connell wrote: I thought I would consult here as to the motherboard that will give the least compatibility problems with the various chip sets available and for on-board sound. I've yet to hear any bad stories about motherboards for athlon cpus with chipsets from amd. personally, I've not had any problems with motherboards with via chipsets, but there are some rumors out there about stability problems when running a kernel compiled with athlon-specific bulk move instructions. Cheers, Joost
Re: Wherer is dselect putting the kernel-source_2.4.6.deb
On Sat, Jul 21, 2001 at 11:14:31AM -0500, John Foster wrote: I have been having problems getting a new kernel compiled. I have already posted a query about whether dpkg or make-kpkg is broke in testing with no response. I downloaded a raw linux-2.4.6.tar.gz from kernel.org, put it in /usr/src/linux and tried to compile with make-kpkg buildpackage. I have never had a problem with this working before, but it faild 3 times. I then used dselect to install the new kernel-source.2.4.6.deb file from Debian, expecting it to install in /usr/src/kernel-source-2.4.6 but although dselect says it is installed on my system I can not find it or any directory matching *kernel-source* using mc find except in /var/tmp which are markers for dpkg. I did see the /kernel-source-2.4.6_docs installed from my installation from dselect. Anyone have any ideas here??? dpkg -L kernel-source-2.4.6 Cheers, Joost
Re: No such file or directory - huh?!
On Sat, Jul 21, 2001 at 08:34:48PM +0200, Gary Jones wrote: ash-ock:/etc/init.d# ./firewall bash: ./firewall: No such file or directory ash-ock:/etc/init.d# ./hostname.sh ash-ock:/etc/init.d# more ./firewall #! /bin/sh # Script to control packet filtering. [snip] What's going on? The script file is definitely there, I can 'more' it, 'jed' it, whatever I like except run it. I'm sure I'm missing something real simple here... In the script, you are using a command with a tpyo in it or that is located in a place not in your current $PATH. Perhaps the command is ipchains (/sbin/ipchains) and you are used to doing su to become root? In that case, next time do su -, so you get a propor root login, with all the sbins in $PATH. Cheers, Joost
Re: dist-upgrade from potato to woody
On Sat, Jul 21, 2001 at 09:17:27PM +0100, Graham Ward wrote: I just tried to upgrade my system from potato to woody. I believe these are the correct steps: (1) replace potato with woody everywhere in /etc/apt/sources.list (2) apt-get update (3) apt-get dist-upgrade. When I do step (3), I see (among other things) the message WARNING: The following essential packages will be removed This should NOT be done unless you know exactly what you are doing! sysvinit util-linux (due to sysvinit) 544 packages upgraded, 87 newly installed, 36 to remove and 6 not upgraded. Need to get 348MB of archives. After unpacking 173MB will be used. You are about to do something potentially harmful On the face of it, removing sysvinit looks like a bad idea, so I stopped at this point. Has something gone horribly wrong with my set-up, or is this in fact harmless? If you go ahead you system will most likely be hosed. Why aren't you using dselect? Both the dpkg and apt-get manual suggest you use dselect as a frontend to manage the package selections. For complex operations like distribution upgrades, you should really always use dselect. Here's what I would do in your current situation, I've added step (-1) to get your system back to its initial state: (-1) reset the available database to stable: place back potato everywhere in /etc/apt/sources.list and run: dpkg --clear-avail (0) prepare for the upgrade by running: dselect update select in the select screen, verify that you have no current unresolved dependencies and that your package selections are sane, eg all packages marked for installation are installed and at their latest versions. (1) replace potato with woody everywhere in /etc/apt/sources.list (2) update available list and verify the new dependencies by running: dselect update select In the selections screen, don't add new packages yourself, just press enter and let dselect ponder on the current selections. As there have been some replacements in packages and some changed dependencies between packages, dselect will prompt you with a list of packages involved in an unresolved dependency. Investigate the suggestions by dselect and accept these if reasonable. (3) download and install upgraded packages by running: dselect install Most dselect operations can also be done from the dselect main menu, which can be started by running dselect without command line arguments. Cheers, Joost
Re: Driver Update Disk
On Thu, Jul 19, 2001 at 05:36:06PM -0400, Manoj Jose wrote: Can we update a driver at the time of dabian installation?.. If yes how we can create driver update disk from source files.. Any idea?.. Try asking that question on debian-boot@lists.debian.org, where the people who know about this are more specifically reachable. Cheers, Joost
Re: What's happened to the task- packages?
On Sun, Jul 22, 2001 at 12:22:35AM +0200, Carel Fellinger wrote: On Sat, Jul 21, 2001 at 02:15:55PM +0200, Joost Kooij wrote: On Sat, Jul 21, 2001 at 07:10:23AM -0500, Brian McGroarty wrote: What's happened to the task- packages? Suddenly task-c-dev and the other programming-related task packages are listed as 'obsolete' on my system. Have these been replaced by something new? The task-* packages in their current form have been pulled from the archive. Properly maintained metapackages are to come in their place, some have already appeared. What list should one read to be aware of such changes? Try debian-devel, it is generally good to read that list if you are running unstable or testing. Supposedly, important announcements should be sent to debian-devel-announce, but many interesting topics are only ever mentioned on debian-devel. Cheers, Joost
Re: Keeping kernel compilation options
On Sat, Jul 21, 2001 at 03:31:01PM -0700, Ross Boylan wrote: What is the recommended way to keep your responses to the kernel configuration options when using the debian kernel package tools? Use kernel-package to build your kernels. It saves your .config in /boot/config-version so you always knwo the compile options to the installed kernel. I built a 2.4.2 kernel, and would now like to build 2.4.6. Surprise! 2.4.7 is out. I'm concerned that simply copying the .config file (name is from memory) is risky because options may get added or removed (even if it happens to be safe for the 2.4.2 - 2.4.6 move). Copy your old kernel's config the new linux kernel top level source directory, with the name .config, then type make oldconfig. It will skip all answers that were asked already when configuring the last kernel, but for all new options, the questions are asked. Do the package install scripts do anyhting clever (seems unlikely since they don't know where I actually built the kernal)? There are so many compelling reasons to use kernel-package that they are listed in a document in /usr/share/doc/kernel-package. Installing kernel-package will also pull in the other packages needed to build the kernel from source. Be sure to use dselect, or you will miss the suggested and recommended packages. While I'm in kernel land, is there a way to get the alsa drivers to be built automatically as part of the kernel build (again, using the alsa-source debian package)? I tried before, but didn't have much luck. Is it possible to, e.g., build the 2.4.6 kernel with alsa drivers while running 2.4.2? Or do the drivers end up targetted to the current kernel at the time they were built? Read the kernel-package on how to do this with the modules-image make-kpkg targets. Generally, kernel drivers should match the kernel they were compiled for, yes. Cheers, Joost
Re: system logs
On Fri, Jul 20, 2001 at 06:09:32PM +0100, john gennard wrote: Can someone please explain how/where savelog operates from? grep -r savelog /etc/cron* Cheers, Joost
Re: problem with menuconfig and ide controllers...
On Thu, Jul 19, 2001 at 01:59:30PM -0700, Carl Fenley wrote: However, both the cdm640.c and the cmd646.c source files came with the the current version of my kernel-source (2.2.12). ^^^ That is not a current version at all. I recommend you to get a more recent kernel source version. Looking at http://www.kernel.org/pub/linux/kernel/v2.2/ the current 2.2 kernel is 2.2.19, try to use that one. Cheers, Joost
Re: dselect trying to remove lots of stuff
On Fri, Jul 20, 2001 at 04:14:07PM -0400, Anthony Fox wrote: Seems I wasn't careful enough when using dselect. I have been trying to remove the gnome libs and binaries that I don't use on my system. Somehow, I must have selected the wrong package for purging. Dselect now wants to remove from my system, among other packages, XFree86 and KDE. I don't want these packages to be removed. Is there some way that I can make dselect forgot about previous selections? That is, is there some way for dselect to just start over? If you have changed the package selections, and confirmed these changes in dselect, they are set and you can't get the previous selection state back. What you can do, is to go back to dselect and in the package list, remark for installation all packages that have erroneously been marked for removal. Dselect displays both the current status and the desired status, so you can easily spot the packages that are still installed, but marked for removal. Cheers, Joost
Re: apache vhosts - quotas/details
On Wed, Jul 18, 2001 at 11:56:50PM +0200, Martin F. Krafft wrote: is it possible to configure apache to log the total amount of traffic (in Mb) that is generated by a virtual host? how would i set that up? Install the lire package. Cheers, Joost
Re: fetchmail, mutt, masqmail
On Thu, Jul 19, 2001 at 10:02:19AM +0200, Schoppitsch Dieter wrote: My data: * hostname: boneless * local mailaddress: [EMAIL PROTECTED] * command: fetchmail -v -S boneless * error message after ... reading message 1 of 2 (25915 octets): - SMTP connect to boneless failed - This message is in MIME format. Since your mail reader does not understand. Due to a similar problem in another mailing-list the answer should be the MTA is not listening while fetchmail tries to contact it - but masqmail is running (and hopefully listening) on my PC. Try telnetting to the smtp port on localhost and see what happens if you manually send a message into the mta. Perhaps it thinks that you are attempting to relay. Maybe you also need to convince fetchmail to deliver to a different address than the one set in the email header, which is likely your address at your isp. Try reading rfc821, rfc822 and the fetchmail manpage. It helps to understand. Cheers, Joost
Re: How do you edit the /etc/hosts with the smtp information - fetchmail problems
On Thu, Jul 19, 2001 at 04:28:54AM -0300, [EMAIL PROTECTED] wrote: I've been having prblems with fetchmail, getting a SMTP connect failed message. Since I read in the fetchmail FAQ that the problem could be that my SMTP port listener is down or inaccessible. Try: telnet 127.0.0.1 25 or telnet localhost smtp If you get a response, the mta is listening allright. If not, enable it, somehow. If it is up and running, it may be thinking that you are a relaying spammer, every time when you try to have fetchmail inject the mail from your isp into your private mail system. Perhaps putting user foo is user bar here in your fetchmailrc will help, if that is in fact the problem. You need to run fetchmail with maximum verbosity and then look at the logging where it went wrong. It's hard to say what to change if all you can tell is: it doesn't work. I need to know what doesn't work and why it doesn't work. The configuration of my /etc/hosts file has no line refering to SMTP. I reckon this is the problem. Can you guide me in writing /etc/hosts in order to enable SMTP? No, because it doesn't work like that. Perhaps you read something somewhere about /etc/hosts.allow or /etc/hosts.deby, but that is very unlikely to be a problem when you are connecting from localhost to localhost. Ignore it for now. Cheers, Joost
Re: How to give non-root user the right to start X
On Thu, Jul 19, 2001 at 04:17:51AM -0300, [EMAIL PROTECTED] wrote: In article [EMAIL PROTECTED], Ari Pollak [EMAIL PROTECTED] wrote: In /etc/X11/Xwrapper.config (if one exists), you want the first line to say: allowed_users=console What should one do if it doesn't exists? dpkg-reconfigure xserver-common Cheers, Joost
Re: accounting total traffic
On Thu, Jul 19, 2001 at 03:05:25AM +0200, Martin F. Krafft wrote: i need to get a semi-exact idea of the traffic through one of our servers, in and out. i understand that netstat -s gives me interface statistics, but there are two problems with it: (a) it lists packets only. as i understand, a packet is not always the same size. so that's no big use. (b) i absolutely need to cope with the disaster case in which the server goes down - netstat would loose all data... Use ipchains. It counts both packets and bytes. It makes tracking traffic by type simple, just create rules that match those connections without a jump target. Then setup a cron job that lists and resets the counters and mails the data to a collector address. You get all the robustness of the smtp mail system for free, if it is already there. Set the job interval appropriate for the amount of data loss is tolerable in case of an uncontrolled outage. Possibly, you can also get the stats data every minute and queue them up for an hourly mail sending. If you have this in place, then adding kernel logging from ipchains is a simple extension. On the receiver side, setup a dedicated user account that receives the stats mails and archives, indexes, processes and crossreferences the data. Write a reporting system that can publish reports on the web or by email. If you have it working, let the other hosts in your network send their data for processing. If you like this, then look into the lire package and see how much of this is already implemented. The lire people are happy with your patches, I'm sure. there is iptables/ipchains, but (b) still applies. i figure that there has to be a way to record these data without going higher up the provider hierarchy, right? any ideas? i don't like daily ipchains accounting mails and subsequent counter flushes... Why not? It would be the most straightforward implementation of what you're asking for. I can see that you don't like to create all the overhead yourself, so look into the lire package. Cheers, Joost
Re: free software likes http://remindme.arsdigita.com/
On Wed, Jul 18, 2001 at 09:48:03PM -0400, Alan Shutko wrote: Peter Kok [EMAIL PROTECTED] writes: I would like to know where I can get free software as function http://remindme.arsdigita.com/ Maybe http://www.arsdigita.com/acs-repository/? Perhaps http://www.openacs.org/ ? $ dpkg -p openacs Package: openacs Priority: optional Section: web Installed-Size: 2527 Maintainer: Brent A. Fulgham [EMAIL PROTECTED] Architecture: i386 Version: 3.2.2beta3-2 Depends: aolserver, postgresql, postgresql-pl, libpgtcl, perl5 Suggests: openacs-doc Filename: dists/woody/main/binary-i386/web/openacs_3.2.2beta3-2.deb Size: 2538650 MD5sum: dc7284162ed16807651642735b775181 Description: OpenACS This is the ArsDigita Community System ported to the Postgres RDBMS. It is a powerful platform upon which you can build full-featured, production-ready community sites complete with forums, interactivity, and other so-called dynamic content. It is implemented as a system of Tcl scripts written as add-ons for the AOLserver. . If it is just for a reminder service, it can't be too hard to cobble up a perl cgi script that schedules reminder jobs using at. Just be sure to use taint mode and check all the input very carefully, including the email addresses that are submitted (don't become a vehicle for harassment). Cheers, Joost
Re: Thin-X-Client-Laptop
On Thu, Jul 19, 2001 at 10:36:25AM +0200, Schoppitsch Dieter wrote: Hi Cajus, thanks for your hints - but it did not work - what did I do wrong? You are leaving all of the post you are replying to at the bottom of your message, without there being an apparent need for any of it. Please read a netiquette faq on proper email quoting. Thanks. * On the PC I checked /etc/X11/xdm/Xaccess there was a * * On the PC I started xdm (login window) * On the Laptop I started X with X -query 192.168.0.1 * On the Laptop I had the x-cursor only * After C-A-backspace on the laptop I saw the message XDM: too many retransmissions Try netstat -at | grep xdm on the pc running xdm. If xdm is not listening to the network, then it is unlikely that the laptop xserver ever gets a connection. Probably, you need to comment out the bottom line of the /etc/X11/xdm/xdm-config file on the pc acting as xdm server. PS: instead of *, better use *.your.domain in Xaccess. Cheers, Joost
Re: embarrassing X question
On Thu, Jul 19, 2001 at 08:27:49AM -0400, Richard Black wrote: For some reason, I can no longer remote login to another terminal and display stuff on mine! This started happening last week (with, possibly, the changes to gdm...) I have tried many different things. Typical is something like: [local machine] xhost + rlogin remote [remote machine] export DISPLAY=local:0.0 nedit But all I get is: NEdit: Can't open display The xfree86 packages have been changed to not accept tcp connections at all by default. Check out the -nolisten option in your xserver manual page. If you want to turn it back on, change /etc/X11/xdm/Xservers or /etc/X11/xinit/xserverrc, depending on how you start your xserver. Generally, don't use xhost, it is not safe. Instead use xauth. Cheers, Joost
Re: Thin-X-Client-Laptop
On Thu, Jul 19, 2001 at 03:29:37PM +0200, Joost Kooij wrote: Try netstat -at | grep xdm on the pc running xdm. If xdm is not Correction: as root do netstat -tap | grep xdm Cheers, Joost
Re: how-to configure a printer on my potato
On Fri, Jul 20, 2001 at 12:26:36AM +1000, Joel Mayes wrote: Paul Huygen [EMAIL PROTECTED] writes: 4) I have installed lpr and magicfilter and that runs fine. However, I understand that CUPS is a modern alternative for the two (am I right?) I as understand it yes, it also offers beter drivers/filters for new printer then lpr or lprNG and it's much easier to configure Neither lpr nor lprng is supposed to do any filtering, you have to configure them to let another program, like magicfilter or apsfilter, do that if necessary. The differences with cups are that cups implements the functionality of both the lpd and the filter in one, and that it has a gui for setting it up and managingthe printer queue. Apart from the gui, it does nothing new afaik. Cheers, Joost
Re: Help with ESS1888 setup.
On Thu, Jul 19, 2001 at 08:57:53PM +1000, Kieren Diment wrote: I am having great problems with configuring my Dell Laptop's sound card. Perhaps it needs plug-n-play to assign resources to it. According to Windows95 it is an ESS 1688 Audiodrive I/O Range 0240 - 024F 0388-038B 0330-0331 IRQs 5 and 9 DMA 00 and DMA 03 I have tried configuring sound into the kernel to no effect. I have tried the commercial OSS package with no sucess. It claims not to be able to compile soundcore properly. I am using kernel 2.2.19pre17. I have tried installing ALSA with no success. Can anyone help me based on the above information If it is a plug-n-play problem, install the isapnptools package, read /usr/share/doc/isapnptools/*. Run pnpdump, edit the output and save it as /etc/isapnp.conf, then run isapnp /etc/isapnp.conf. Then modprobe your sound driver with the correct options. Use modinfo -p modulename to see the list of supported options. When you have found the right configuration, create a file /etc/modutils/local, in which you put the lines: alias char-major-14 modulename options modulename moduleoptions Then run update-modules. Check that /etc/modules.conf now contains the above lines. Cheers, Joost
Re: Safe File Manager to run as root ?
On Wed, Jul 18, 2001 at 10:43:22AM -0500, Case, Benjamin wrote: Is there such thing as a GUI File Manager that any security and safety consious Debian users would use, as ROOT, to manage a file system (i.e. move, copy, change permissions) ?? Is it just a better practice to use CLI w/ suid to make those kind of changes. Learn to use the command line interface instead. You can do things on the command line that are almost unimaginable in a gui system, where on can only do what can be clicked on, roughly speaking. On the command line, you have more control, for a small price of having to think about every little thing that you type in. And besides, most operations on files involve processing the file contents, not moving the file about or changing its attributes. Another thing is that gui system management tools do not scale nicely over the network, generally speaking. You are best off to start with a book about the bash shell, and a general book about unix system administration. For reference, or if you do not want to pay for a book, there is the bash manual page, man bash, and at http://ibiblio.org/ , you can find various online books about linux. Cheers, Joost
Re: Help PLEASE!!! -- big ld problem
On Wed, Jul 18, 2001 at 08:23:49AM -0700, Harvey Werner wrote: Did you ever solve this problem? http://lists.debian.org/debian-user-9905/msg00160.html Or, did you have to reinstall your Linux system? If only the symlink is nuked, just boot with a rescue floppy, get a shell, mount the root partition and recreate the symlink. If the actual ld-x.y.z.so file is gone, you'll have to extract it from the proper deb and copy it back. In either case, the rescue disk is your friend. It is the same disk as the installation boot disk. Cheers, Joost
Re: Starting with postgreSQL and pgaccess
On Wed, Jul 18, 2001 at 03:09:20PM +, Victor wrote: Joost Kooij [debian-user] 17/07/01 21:55 +0200: Just create a valid user with your id, grant priviledges to create new databases and then after that you can do most or all things as regular user. First of all, thanks Joost, it now works for the postgres superuser only, not for other users I've defined. Now the problem of this absolute beginner is the definition of ID when using createuser. What exactly should I input at the prompt for it? My user ID as a linux user (the one I find in records in file /etc/passwd) or what else? I'm sorry, I should have mentioned the appropriate tools the first time. It is all explained in createuser(1). Also interesting are pg_passwd(1) and createdb(1). Run createuser as the postgres admin user. Cheers, Joost
Re: PS/2 mouse, what device?
On Wed, Jul 18, 2001 at 08:28:52PM +0200, Danie Roux wrote: This should be easy, but I can't get it: My keyboard is on /dev/psaux. Where will my ps2 mouse be? On the keyboard connector? ;-) Cheers, Joost
Re: locatedb question
On Wed, Jul 18, 2001 at 01:27:09PM -0400, Mike Dresser wrote: router:~# locate \* | wc -l 68558 router:~# updatedb router:~# locate \* | wc -l 91395 Every night, updatedb runs, and updates, removing something like 21000 files from the locatedb. Looking through the cron.daily, i see updatedb runs as nobody. Is there a particular danger in running this as other than nobody? On your router, likely there is no harm in having a full locatedb. On a true multi-user system, users want to be able to chmod go-rwx their directories and not have the names of files still available to random other users on the system. AFAIK that is the reason. Cheers, Joost
Re: .deb
On Tue, Jul 17, 2001 at 04:41:16PM +1000, Peter Donaldson wrote: This might sound like a dumb question. But how do i make a .deb file??? Start with man dpkg, then install dpkg-dev, read a few manpages, then go to www.debian.org and find the rest of the information scattered over the developers corner page. Don't forget to download hello.tar.gz, hello.diff.gz and hello.dsc. Then do dpkg-source -x hello.dsc and build a hello.deb by typing debian/rules binary in the unpacked hello source top level directory. Cheers, Joost
Re: zip drive?
On Tue, Jul 17, 2001 at 11:32:08AM +0100, Frank Zimmermann wrote: filnames. Well, most times I use vfat foramttd ZIPs and the few times I use something else I use mount -t hfs /dev/hdd /zip (why don't I need a partition number in this case?) You don't. I did mke2fs /dev/sda; mount -t ext2 /dev/sda /zip on a couple of zipdisks and it works fine. Cheers, Joost
Re: But ....
On Tue, Jul 17, 2001 at 12:29:38PM -0400, D-Man wrote: I think it means pick 'safe' settings for your video card and don't start up any fancy services so that, maybe, the user can fix what's broke. Remember that Windows doesn't know what a virtual console is and _always_ needs a graphics card and mouse. Perhaps it uses the standard bios interfaces only when in safe mode. IIRC the standard vga bios interface is limited to a 640x480 resolution. Cheers, Joost
Re: installing hardware
On Mon, Jul 16, 2001 at 10:53:01PM -0700, Dan Cox wrote: How do I install hardware? More specifically I have a Hayes Accura v90 modem. I understand that this is a winmodem??? maybe. I looked at linmodem.org and after some searching I found this site http://www.sfu.ca/~cth/ltmodem/ which I think has the driver I need. Anyway I downloaded the .deb and did the hole deb thing, but I don't know what to do now. Any help would be greatly appreciated. Your questions are at a such generic level that I think you should best go to http://linuxdoc.org and read at least the modem-HOWTO, the PPP-HOWTO and perhaps the NET-HOWTO. If you have specific questions that are not answered in the HOWTO's, feel free to ask here. Cheers, Joost
Re: PCMCIA Network driver.
On Tue, Jul 17, 2001 at 10:02:54AM -0500, Case, Benjamin wrote: I just got an ENCORE PCMCIA Netowrk Card. From their website I d/l the linux drivers. This consisted of these files: 8390.c, gen1, gen2, and PCNET_CS.c. as well as a readme with the following instructions: 16-bit 100/10M Fast Ethernet PCMCIA Adapter LINUX DRIVER INSTALL Note: this driver for linux 2.0.30 ^^ 1. copy driver to /FASTPCM # mcopy a:/* /FASTPCM 2. download pcmcia-cs-3.0.x.tar.gz from hyper.stanford.edu in the /pub/pcmcia directory readme PCMCIA-HOWTO file install it Don't do this. Use the debian pcmcia-cs package instead. 3. add the following lines into /etc/pcmcia/config card 16-bit 100/10M Fast Ethernet PCMCIA Adapter version PCMCIA, 100BASE bind pcnet_cs 4.# cd /FASTPCM # chmod +x gen1 gen2 5.# gen1 # gen2 All (and more) done automatically by the debian package. 6.# reboot Windows mindset. In short: why aren't you using the driver in the linux kernel? Most of these vendor drivers are cheap ripoffs from Donald Becker's drivers in the official linux kernel. Use the drivers in the standard kernel, those are up-to-date and reviewed. The manufacturer drivers for linux are nice publicity and add credibility to linux in the eyes of some (official vendor support), but are not to be actually used. 3com PCMCIA network card, so I think I have the SLOT configured properly. Is there anyway to simply compile that 8930.c file into 8930.o and 'insmod' it ? I have never used linux on a laptop, and I have never worked with PCMCIA before. I assume the list doesnt like attachments, so if anyone is intereseted in the files, I will be happy to provide them. As an alternative I will make this e-mail extrememly long and attach the text of the files below: Are you NUTS?!? This add nothing to the content of your message. It pushes a very big email through the phone line of hundreds of subscribers to this list. Go read a few netiquette faqs before ever doing that again. About your problem: install pcmcia-cs and pcmcia-modules and read the PCMCIA-HOWTO at http://linuxdoc.org Cheers, Joost
Re: install debian up redhat 6.2 ...
On Tue, Jul 17, 2001 at 09:42:24AM -0500, Saul Fabian wrote: I have a problem, ?How I can format my hard disk ? I want to change my redhat 6.2's linux to debian, but I can?t install debian. When I boot the debian installation disk, my system gives my this message: loading linux ... can anibody help my??? No need to format your harddisk. Just reuse the existing partitions and the debian installer will reformat these automatically as filesystems are created with mkfs.ext2. In the case of a /home filesystem, you don't have to zap it, you can simply mount the old filesystem during installation. Cheers, Joost
Re: Starting with postgreSQL and pgaccess
On Tue, Jul 17, 2001 at 04:30:57PM +, Victor wrote: I'm now having a go at using postgresql, the version included in debian 2.2r3. Now, while I've been able to create my first db and tables using psql under postgres user, No need for that. Just create a valid user with your id, grant priviledges to create new databases and then after that you can do most or all things as regular user. Postgresql is a real multi-user database system, you do not need to be the postgres administrative user all the time. It's like your linux system: you shouldn't have to be root all the time. I'm in trouble using pgaccess. After trying to open the same db giving its name and postgres as user it invariably answers: Connection to database failed ConnectDB().Connection refused: Is the postmaster running (with -i) at localhost and accepting connections on TCP/IP port 5432? You must enable postgresql to listen to tcp/ip sockets. By default it will only listen to local unix domain sockets. The version of Pgaccess that you have can only connect to postgresql through the tcp/ip socket interface. IIRC you have to change the file /etc/postgresql/pg_hba.conf Make it only listen to 127.0.0.1, that is all pgaccess on the local host requires. I've checked in /etc/services for port 5432 and it is there for postgres. That file is just an informative lookup table. It does not configure any of your services, it only lists well-know ports and commonly associated services. Cheers, Joost