Re: Root partition stuck in read-only mode.
[ ... ] try Timo's rescue CD (debian based): http://rescuecd.sourceforge.net/download.html If I'm not mistaken his bootimage uses initrd, and does some autodetection of drives. If the autodetection fails, it's possible to get a shell early in the boot process and load the needed scsi modules. I've never done that so I don't really know how it works, but I expect modprobe driver-for-your-scsi-card would suffice. [ ... ] Look in /boot/config-2.4.20-686-smp, it lists all the config params used to compile that kernel. Look for scsi-generic and scsi-disk and the apropriate module for your scsi card. Thank you very much. This all looks useful, and I'll be giving it a try a little later. I hope that my next message here will be a summary of what I did to get this all working correctly. -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: columbia -- what really happened
the hell... Brooks R. Robinson wrote: | How the US can justify spending so much money on Space while 33 | million US citizens live below the poverty line amazes me. [ ... ] [ ... ] Having poor people join the capitalist middle class does not in any way lead to a socialist worker's paradise. So as a good little leftist, you ought to be against anything that would improve the lot of the poor in a capitalist country, because the more miserable and downtrodden they are, the more likely they'll join the revolution. [ ... ] Yes, all this is indeed OT, despite how interesting it is. I can tell that we're getting very close to the point where Hitler is going to be mentioned, so that we can then put Godwin's law into effect. Oh! ... I already mentioned him. OOPS ... -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Root partition stuck in read-only mode.
[EMAIL PROTECTED] wrote: On Sat, Feb 01, 2003 at 05:30:36PM -0500, Lloyd Zusman wrote: [ ... ] Is there something special I have to do to tell my kernel that there are SCSI disks at the time I make the boot floppy? That all depends:) It's a long time since I made boot floppies (prefer bootable CD's) but if I'm not mistaken mkboot uses your current kernel. So if all the scsi stuff is compiled in all's swell. But if (part of) the SCSI stuff is compiled as modules you need to use initrd to get to those modules in time. Thank you. Well, I guess I'll have to go back to the drawing board to figure out how to make a usable boot floppy ... or perhaps a boot CD ... I really don't care which one I have, as long as I have _some_ source for a boot record besides my hard disk. I guess I'm going to have to gain the knowledge necessary for using initrd during the process of building a boot disk/CD, since mkboot and the yard suite are currently at the limits of my knowledge. Without my Linux box working for more than a few minutes at a time, I have at best intermittent internet access, so could someone point me to a step-by-step cookbook that describes how to make a bootable floppy or CD, when initrd is needed to load added SCSI modules? Also, something that tells me how to find out which SCSI modules I need, and which of these are and aren't compiled into my kernel would also be useful. I have the standard, unmodified 2.4.20-686-smp kernel, by the way. I'm not asking anyone to do my work for me ... I'd just appreciate a pointer to a reliable source of information that I could get to quickly and then download ... preferably something that presents all the needed data in one place so I don't have to follow links all over the net during the limited time that I can get access. Once I get this source of info, I'll take it from there. Thanks again to all of you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Transfering my system to a new disk (was: Root partition stuck in read-only mode.)
Quoting Nathan E Norman [EMAIL PROTECTED]: [ ... ] Gack! How old is the drive? Any chance the issue is termination or a misconfigured jumper? You're using an Adaptec 2940? I'll check the jumper later, when I open up my box. Yes, I believe it's a 2940 [ ... ] Lloyd Zusman [EMAIL PROTECTED] writes: Thanks again for the earlier help, and thanks in advance for any help you might be able to offer for this new system-copy task. This may prove difficult. Why not just remove /dev/sda once you've copied stuff over? Then, if you run LILO boot wiuth a rescue floppy and/or CD, type linux root=/dev/sda1 at the prompt, and rerun lilo once you're up. If you use GRUB, it would be a good idea to make a grub boot floppy; then you can boot to any kernel on any drive in the system! Google for this or ask me; I'll look around for my link. Well, I'm using LILO. I want to create a proper rescue disk, and I used the mkboot utility to do so. It told me that it created a proper disk, but when I tried to boot off of it, I got a kernel panic with a message that stated that there was not a proper root file system. I did the boot by typing linux and also linux root=/dev/sda2. Both resulted in the same error. I don't know if this matters, but there are no /boot/boot*.b files on my system. The mkboot utility seems to be looking for a /boot/boot.b file, although it doesn't give me an error. But is that a problem? [ ... ] -- Nathan Norman - Incanus Networking mailto:[EMAIL PROTECTED] prepBut nI vrbLike adjHungarian! qWhat's artThe adjBig nProblem? -- alec flett @netscape -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Transfering my system to a new disk (was: Root partition stuck in read-only mode.)
Quoting Lloyd Zusman [EMAIL PROTECTED]: [ ... ] Well, I'm using LILO. I want to create a proper rescue disk, and I used the mkboot utility to do so. It told me that it created a proper disk, but when I tried to boot off of it, I got a kernel panic with a message that stated that there was not a proper root file system. I did the boot by typing linux and also linux root=/dev/sda2. Both resulted in the same error. One more piece of info: I do have a /boot/boot.0800 file, so I also tried to boot from the diskette with linux root=0800. That didn't work, either. But without a boot diskette, my system comes up just fine ... ??? [ ... ] -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Root partition stuck in read-only mode.
Quoting Carel Fellinger [EMAIL PROTECTED]: On Fri, Jan 31, 2003 at 11:17:12PM -0500, Lloyd Zusman wrote: [ ... ] I have some more info about my problem that might be useful. Kernel panic: VFS: Unable to mount root fs on 08:02 ... Would this indicate a hardware problem? Or is there still hope that this can be fixed with software? ...more likely your kernel doesn't know about SCSI disks, so software. I've been booting up and running off of SCSI disks day in and day out for weeks on gthe same kernel, when I boot off the disk itself and not a boot floppy. This read-only problem only started a couple days ago, and I haven't changed my kernel in 3 weeks or so. So how could my kernel forget about SCSI disks when I create a boot floppy? I did it via mkboot and also via the yard suite. Same results each time. Is there something special I have to do to tell my kernel that there are SCSI disks at the time I make the boot floppy? -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Root partition stuck in read-only mode.
I'm having a problem with my testing+unstable system, and I'm not sure whether it's a hardware or software problem. It has rendered my system virtually useless. But before I start fooling around with hardware, I'd like to see if my problem might be familiar to any of you, and if perhaps there is some sort of known fix for it, software-wise. The problem is that yesterday, my root file partition started booting up read-only. Because /var and /tmp are on that partition, my system is virtually useless. I haven't made any kernel changes for more than two weeks, and I have booted up numerous times with my 2.4.20-686-smp kernel without a single problem before yesterday. I have a Dell 530 box with dual Xeon 1700MHz processors. I have two 17G SCSI disks: SEAGATE model ST318451LW, Rev: 0003, Type: Direct-Access. I have an Adaptec controler (I'm not sure what model), and I went through the SCSI verification (accessible at boot-up time via Ctrl-A), and both disks verify fine ... no bad blocks, no errors of any kind. I get numerous read-only file system errors on boot-up, and then the system freezes for a long time. But finally, it unfreezes and the boot continues limpingly, until I finally get a login prompt on the console. As you probably surmise, I cannot bring up X, but at least I can run simple things from the console. A while ago, I had changed my /etc/fstab by commenting out and replacing my root partition mount, as follows: #/dev/sda2 / ext3 errors=remount-ro 0 1 /dev/sda2/ ext3 0 1 Since the errors=remount-ro line is commented out, I don't see how this could be causing my problem ... or could it? I'm writing you from my webmail account via my (ARGGH!) Windows-based laptop, but at least my Linux machine is nearby and I can run a few diagnostics on it. Does any of this sound familiar? Can any of you suggest some diagnostics that I can run that might help me zero in on the problem? As I mentioned, I'm not sure if this is a hardware or software problem ... but I want to eliminate the software before I start taking my machine apart. Thanks in advance for any help you folks can give me on this. -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Root partition stuck in read-only mode.
Thank you VERY much for this help. I have a question though ... see below: Lloyd Zusman said: #/dev/sda2 / ext3 errors=remount-ro 0 1 /dev/sda2/ ext3 0 1 ^^ in the options field try putting 'defaults' you can also jump to runlevel 1 'init 1', and when you login type: mount / -o remount,rw I did this, but I got the following error ... mount: /dev/sda2 already mounted or / busy mount: according to mtab, /dev/sda2 is already mounted on / Do you know a way around this? Thanks again ... - Lloyd then mount -a then you should be set ..to edit fstab and stuff. Does any of this sound familiar? Can any of you suggest some diagnostics that I can run that might help me zero in on the problem? I think perhaps the lack of anything in the options field made the system think that your mounting it with the option of '0', which may mean read-only for all I know :) Also linux will remount a filesystem read only if there is excessive errors, either filesystem or disk/controller based. but this process doesn't affect how the disk is mounted the next time the system boots. You will see a message like Remounting filesystem read-only after seeing a bunch of errors on the screen if this occurs. It's only occured to me on a failing disk. nate -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Root partition stuck in read-only mode.
Shawn Lamson [EMAIL PROTECTED] wrote... On Fri, 31 Jan 2003 21:08:56 -0500 (EST) Lloyd Zusman [EMAIL PROTECTED] wrote: [ ... ] in the options field try putting 'defaults' you can also jump to runlevel 1 'init 1', and when you login type: mount / -o remount,rw I did this, but I got the following error ... mount: /dev/sda2 already mounted or / busy mount: according to mtab, /dev/sda2 is already mounted on / Do you know a way around this? I'm not sure but I think it should be mount -o remount,rw / Well, that didn't work, either. Here's the error: EXT3-fs: Unrecognized mount option 0 mount: / not mounted already, or bad option And it is indeed mounted. -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Root partition stuck in read-only mode.
I have some more info about my problem that might be useful. I booted up off of disk 1 of my Woody installation CDROM set. My root partition is /dev/sda2, so I entered the following at the boot: prompt ... rescue root=/dev/sda2 However, this errored out quickly. I got a couple screens' worth of messages, of which I believe the final few lines are significant: Partition check: apm: BIOS version 1.2 Flags 0x03 (Driver version 1.13) apm: disabled on user request request_module[block-major-8]: Root fs not mounted VFS: Cannot open root device 08:02 Kernel panic: VFS: Unable to mount root fs on 08:02 The system froze at this point. Would this indicate a hardware problem? Or is there still hope that this can be fixed with software? Thanks again to all of you, for all your help. -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Root partition stuck in read-only mode.
Thanks for your help. This solved the immediate problem ... Quoting Nathan E Norman [EMAIL PROTECTED]: [ ... ] 1) You cannot remount / because other filesystems are mounted on it (I surmise this because you don't mention /usr or /home). 2) / is readonly either because you provided no options at all, or there are errors. Here's what I would do: 1) unb0rk your /etc/fstab. the errors=remount-ro needs to be there. It's now unbOrk-able, because I can't alter it. It's on a read-only file system. But even so, I decided to continue ... 2) reboot. pause lilo or grub and boot with linux emergency (replace linux with the label of your default image) 3) enter the root passwd when prompted. 4) run fsck /dev/sda2 I performed steps 2, 3, and 4, and even though I was on a read-only file system, I figured I'd see what happens with the fsck. When I ran it, it came back right away and said that /dev/sda2 is clean. But hope springs eternal, so I'm continuing ... 5) run mount -n -w -o remount /dev/sda2 / Now, it worked! I now have a proper writable filesystem, and so I unbOrked /dev/fstab. 6) type exit Did it ... and now, my system booted up just fine. So ... now that things are sort of back to normal, my question is this: what caused the filesystem to become read-only to begin with? Could it be hardware errors? The fact that the fsck found no errors seems to point to this as a possible cause, correct? At any rate, thanks again! -- Nathan Norman - Incanus Networking mailto:[EMAIL PROTECTED] Warning: dates in calendar are closer than they appear. -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Transfering my system to a new disk (was: Root partition stuck in read-only mode.)
Quoting Nathan E Norman [EMAIL PROTECTED]: On Fri, Jan 31, 2003 at 11:59:28PM -0500, Lloyd Zusman wrote: [ ... ] So ... now that things are sort of back to normal, my question is this: what caused the filesystem to become read-only to begin with? Could it be hardware errors? The fact that the fsck found no errors seems to point to this as a possible cause, correct? No, hardware errors would cause fs corruption I think. Probably the fact that the options field was blank caused the problem. In the future, make sure that field says defaults if you have nothing else to put in there :-) Well, it does say defaults now, and I rebooted with it set that way. But a few minutes ago I started getting disk error after disk error. So I think that I do indeed have some hardware problems, after all [sigh]. So now, I have a completely new question: I belive that the error is on my first drive, the /dev/sda drive on SCSI channel 0. My second drive (/dev/sdb, SCSI channel 1) seems fine. Assuming I can boot up and run reliably for a while, I want to copy everything from /dev/sda to the top level of /dev/sdb. Then, I'd like to reset my system to boot off of the second drive (which has plenty of empty space). I think I should do something like this (but I'm not sure if this is correct): 1. My /dev/sdb only has one partition, and I have mounted it as /opt. So what I'll do is create an /opt subdirectory on that drive and move everything from /dev/sdb1 there. 2. I should use tar to copy all the filesystems under root, except for /opt, over to the /dev/sdb1 partition. This should copy links and permissions correctly. [Is that true?] 3. I should change /etc/fstab on the second partition to only have one huge partition (i.e., get rid of the /usr, /home, and other mount points). 4. I need to create a proper boot sector on the /dev/sdb1 partition. How do I do that? 5. I configure my system to boot off of /dev/sdb1 instead of off the faulty drive. Does this sound reasonable? If not, can anyone suggest what I'm missing? And I'd be grateful if someone could explain how to do steps 4 and 5. Thanks again for the earlier help, and thanks in advance for any help you might be able to offer for this new system-copy task. At any rate, thanks again! Glad it worked for you. Random fs weirdness sucks. -- Nathan Norman - Incanus Networking mailto:[EMAIL PROTECTED] Profanity is the one language all programmers know best. -- Lloyd Zusman [EMAIL PROTECTED] God bless you. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Some myths regarding apt pinning
Vincent Lefevre [EMAIL PROTECTED] writes: On Sat, Jan 25, 2003 at 13:25:50 +0100, Thomas Hood wrote: If I am not mistaken, it is possible to avoid this worst case scenario by appropriately setting up apt's preferences. Suppose I set the priorities of distributions as follows stable 900 testing 800 unstable 700 and, starting with a woody system, upgrade a single package foo to version vvv from unstable apt-get -t unstable install foo which pulls in unstable libc6. Later when I do apt-get dist-upgrade apt will upgrade most packages from stable but will upgrade foo from unstable, or from testing if version vvv has made it into testing; and likewise libc6. There would still be security problems for installed packages that are in testing: the upgrade to unstable (to get the fixed package) won't be automatical. So in that case, what would be the solution? I guess we would have to know what testing packages have fixes in unstable, and then use the -t unstable option to apt-get ... correct? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-get dist-upgrade doesn't downgrade to stable
Osamu Aoki [EMAIL PROTECTED] writes: On Thu, Jan 23, 2003 at 07:18:18PM -0500, Lloyd Zusman wrote: [ ... ] I learned this in this list but reading manual page of apt-get (8) ... --default-release This option controls the default input to the policy engine, it creates a default pin at priority 990 using the specified release string. The preferences file may further override this setting. In short, this option lets you have simple control over which distribution packages will be retrieved from. Some common examples might be -t '2.1*' or -t unstable. Configuration Item: APT::Default-Release This is where it is documented. Tricky :-) Yes, I see it now, and it's less tricky than I thought. I overlooked it the first few times I read the manual, though, because at that time, I didn't yet understand about /etc/apt/apt.conf. Thank you. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Verification of apt-get config for mixed system.
Dale Hair [EMAIL PROTECTED] writes: On Thu, 2003-01-23 at 10:57, Lloyd Zusman wrote: [ ... ] That is what I do, but as I said earlier a dist-upgrade could remove some packages you want to keep so the -s option will tell you what it will do before you regret doing it. A dist-upgrade is not always a wise thing to do when you dabble in unstable. I think pinning a package is more for keeping a package from being upgraded. If you are running testing you could keep a package at stable, or you could pin it to a specific version in testing. Even though I haven't explicitly stated this in any messages, I always do a run with '-s' as a precaution before a real run, and I agree with you about its value. But even so, it's also still valuable to me to not have to remember to put the '-t' option on the 'apt-get' command line, and now I understand how to set things up so I don't have to do that, except for when I'm first installing a package from unstable. Thanks to you and others for all your help. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Reconfigure of localepurge has no effect
After installing a series of packages, I received the following message during the configuration process (see below). I followed the instructions in this message and ran this: dpkg-reconfigure localepurge I performed this reconfiguration and changed the status of the new locale file mentioned in the message I received. However, every package installation I perform still causes that message to be printed. It's as if the reconfigure of localepurge didn't have any effect. And I've run the localepurge reconfiguration repeatedly, each time this message comes up, but it never seems to have any effect. Any ideas? Thanks in advance. Here's the message that I get: localepurge: checking for new locale files ... Some new locales have appeared on your system: mn They will not be touched at all until you reconfigure localepurge with the following command: dpkg-reconfigure localepurge -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Some myths regarding apt pinning
Erik Steffl [EMAIL PROTECTED] writes: [ ... ] but the point is that pinning is not very good because you either bring a number of important packages from unstable (libc6, perl etc) or you simply cannot use it. reading of the manual page and checking the apt-listchanges does not solve the problem. i.e. you recommend pinning, person reads the manpage, tries pinning and finds out that it was pretty much pointless excercise because it would upgrade large part of the system to unstable. or yet another wording: Adrian Bunk wasn't complaining about system actually upgrading packages but about system trying to upgrade packages. erik I want to be sure that I understand the significance of this. Are you saying that pinning a certain package, say randompackage, to unstable in /etc/apt/preferences is worse than doing this the first time that randompackage is installed? ... apt-get -t unstable install randompackage Or do these two methods have equally undesirable effects? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Reconfigure of localepurge has no effect
[ ... ] Would you please run script /tmp/localepurge.typescript (the command /usr/bin/script is part of the bsdutils package) and within this very script session run first localepurge -d, then dpkg-reconfigure localepurge, and then again localepurge -d. After that, exit the session and please send me the resulting /tmp/localepurge.typescript (preferably g'zipped ;) along with the separate output of the command dpkg -l * | grep ^ii to [EMAIL PROTECTED]. This would be most helpful helpful to find out what goes wrong! Thanks, P. *8^) No problem. I'm glad to help you with this. Attached are typescript.gz with the script output, and grep.gz with the grepped output of dpkg (the exact command sequence you specified). I hope this helps. Also, if you'd like me to run any more tests, feel free to let me know, as I'm glad to help you debug this. I've had years of Unix experience, and so I know all about 'script' and all sorts of other tools. Sincerely, - Lloyd typescript.gz Description: typescript grep.gz Description: grep -- Lloyd Zusman [EMAIL PROTECTED]
Re: Reconfigure of localepurge has no effect
OOPS. I'm sorry I sent my previous message in this thread to the mailing list. I intended to send it to Paul Selig, and forgot to change the address before typing ^C^C in gnus. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
How to comment out lines in /etc/apt/preferences?
Is there a way to comment out lines in /etc/apt/preferences? I tried various comment characters, but none worked. I couldn't find anything about this in the documentation. Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-get dist-upgrade doesn't downgrade to stable
Osamu Aoki [EMAIL PROTECTED] writes: On Wed, Jan 22, 2003 at 11:08:09PM -0500, Lloyd Zusman wrote: ... [ ... ] apt-get update apt-get -t testing dist-upgrade ^^ Why -t is set here -t set priority of that testing to 990 overriding 1001 This is what went wrong. # apt-get update apt-get dist-upgrade This is what you needed :-) Yes, it indeed is what I needed. It worked great. Thank you! Is it also true that setting /etc/apt/apt.conf to contain `APT::Default-Release testing;' also sets testing to 990? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Verification of apt-get config for mixed system.
I'd like to maintain a testing system, with only one program (and its dependencies) coming from unstable. I think I understand what to do, but I'm hoping that someone here could confirm that my guessed-at setup is correct ... before I start making lots of apt-get install runs with this configuration and then possibly mess up my system. What I want to do is always just enter apt-get install package, without having to use the -t distribution parameter. This way, I don't have to keep remembering when to use and not use -t So ... here's what I have: In /etc/apt/apt.conf ... APT::Default-Release testing; In /etc/apt/preferences ... Package: mplayer* Pin: release a=unstable Pin-Priority: 700 Package: * Pin: release a=testing Pin-Priority: 650 Package: * Pin: release a=unstable Pin-Priority: 600 In other words, all packages come by default from testing except mplayer, which comes by default from unstable. My desired results are: apt-get install mplayer-686 = installs latest from unstable apt-get install other-package = installs latest from testing, unless other-package only exists in unstable, in which case it installs from unstable Do I have it right? Thanks. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to comment out lines in /etc/apt/preferences?
Ron Johnson [EMAIL PROTECTED] writes: On Thu, 2003-01-23 at 05:39, Lloyd Zusman wrote: Is there a way to comment out lines in /etc/apt/preferences? I tried various comment characters, but none worked. I couldn't find anything about this in the documentation. The hash (#) always works for me... But the hash doesn't work for me on /etc/apt/preferences ... # cat /etc/apt/preferences Package: mplayer* Pin: release a=unstable Pin-Priority: 700 #Package: ruby* #Pin: release a=unstable #Pin-Priority: 700 Package: * Pin: release a=testing Pin-Priority: 650 Package: * Pin: release a=unstable Pin-Priority: 600 # apt-get update ... normal `apt-get update' output, followed by: E: Invalid record in the preferences file, no Package header If I remove the hashes before lines in the ruby* section, I no longer get that error. This seems to imply that the hash mark is not considered to be a proper comment character in /etc/apt/preferences. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to comment out lines in /etc/apt/preferences?
Dale Hair [EMAIL PROTECTED] writes: On Thu, 2003-01-23 at 07:06, Lloyd Zusman wrote: [ ... ] # apt-get update ... normal `apt-get update' output, followed by: E: Invalid record in the preferences file, no Package header If I remove the hashes before lines in the ruby* section, I no longer get that error. This seems to imply that the hash mark is not considered to be a proper comment character in /etc/apt/preferences. Leave a space between # and the line you're commenting out. That still doesn't work for me ... # cat /etc/apt/preferences Package: mplayer* Pin: release a=unstable Pin-Priority: 700 # Package: ruby* # Pin: release a=unstable # Pin-Priority: 700 Package: * Pin: release a=testing Pin-Priority: 650 Package: * Pin: release a=unstable Pin-Priority: 600 # apt-get update ... normal `apt-get update' output, followed by: E: Invalid record in the preferences file, no Package header Once again, when I remove the leading # from each of the ruby* lines, everything works fine. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Verification of apt-get config for mixed system.
Dale Hair [EMAIL PROTECTED] writes: On Thu, 2003-01-23 at 06:32, Lloyd Zusman wrote: [ ... ] Do I have it right? [ ... ] This should work but I don't pin individual packages on mine just the distributions. Once a package is installed from unstable apt-get upgrade or dist-upgrade will upgrade those from unstable and testing will follow testing. My preferences file contains Package: * Pin: release a=testing Pin-Priority: 600 Package: * Pin: release a=unstable Pin-Priority: 550 I always use the -s option for apt-get dist-upgrade to show me what will happen before I actually perform the upgrade. I use gnome from unstable and for the past week dist-upgrade wants to remove gnome and a few others so I used apt-get upgrade. OK. So I guess a better way to do what I want is to remove the pinned packages like you said, and then simply use -t unstable when installing a package that I'd like to get from that distribution. And from that time forward (unless I do a downgrade or fool around with my apt configuration), when I do an upgrade or an install, I'll always get those packages from unstable and the others from testing. Is that correct? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to comment out lines in /etc/apt/preferences?
andrej hocevar [EMAIL PROTECTED] writes: On Thu, Jan 23, 2003 at 09:59:10AM -0500, Lloyd Zusman wrote: That still doesn't work for me ... # cat /etc/apt/preferences Package: mplayer* Pin: release a=unstable Pin-Priority: 700 # Package: ruby* # Pin: release a=unstable # Pin-Priority: 700 Package: * Pin: release a=testing Pin-Priority: 650 Package: * Pin: release a=unstable Pin-Priority: 600 You should remove at least one of those empty lines -- either before or after the comments -- or both of them. Yep. That worked. Finally! Thank you. So ... now I'm off to see if someone has already submitted a bug report about this less-than-ideal commenting behavior; and if not, I'll submit one myself. andrej [ ... ] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to comment out lines in /etc/apt/preferences?
will trillich [EMAIL PROTECTED] writes: On Thu, Jan 23, 2003 at 05:07:04PM -0200, andrej hocevar wrote: [ ... ] You should remove at least one of those empty lines -- either before or after the comments -- or both of them. hmm. if that's the case, it looks like you might be able to ADD a has to one of the blank lines, to accomplish the same trick. no? Yes, that works, as well. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-get dist-upgrade doesn't downgrade to stable
Osamu Aoki [EMAIL PROTECTED] writes: On Thu, Jan 23, 2003 at 06:57:53AM -0500, Lloyd Zusman wrote: [ ... ] Is it also true that setting /etc/apt/apt.conf to contain `APT::Default-Release testing;' also sets testing to 990? Yes, that is my understanding. man apt_preferences 100 to 1000 Standard priorities. 990 is the priority set by the --target-release apt-get(8) option. Yes ... I read that, too. But it doesn't say anything specifically about the `APT::Default-Release testing;' in apt.conf, and I have learned the hard way not to make assumptions about the way that various settings in various places interract with one another in the 'apt' system. But using the empirical method, I have determined (I think!) that `APT::Default-Release' does indeed seem to correspond to the same numerical priority as the --target-release option. Thank you for your helpful posts. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Do all mirrors in mirrors_full contain all distributions?
I'm trying to understand the use of the `mirrors_full' file in the `netselect-apt' utility, and I realize that there's a gap in my knowledge. I couldn't locate the information to fill this gap in any documentation that I've been able to find, so I'm asking here. Is each and every site listed in the `packages over HTTP:' lines in `mirrors_full' guaranteed to contain all distributions (`stable', `testing', `unstable', etc.), or do certain sites only contain a subset of these distributions? If the latter is true, what is the most net-efficient method (i.e., requiring the fewest queries over the internet) for determining which sites in `mirrors_full' contain any given distribution? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-get dist-upgrade doesn't downgrade to stable
Simon Law [EMAIL PROTECTED] writes: On Wed, Jan 22, 2003 at 06:15:53PM +0200, Mohammed Sameer wrote: -BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, I have a mixed system (stable+unstable) and wanted to downgrade to stable. here's my /etc/apt/preferences Package: * Pin: release a=stable Pin-Priority: 1001 and /etc/apt/apt.conf APT::Default-Release stable; Make sure you have stable in your sources.list. Take out APT::Default-Release stable;. Then apt-get update apt-get -t stable dist-upgrade. You should see it downgrade. Simon I had a similar problem, and so I appreciate this information. But my understanding is still a bit shaky. Could someone point me to an appropriate doc which explains why `APT::Default-Release stable;' has to be removed from /etc/apt/apt.conf' in order for the downgrade to take place? ... or else perhaps could someone post a short explanation here? Since we are talking about downgrading to `stable', I don't understand why the presence of `APT::Default-Release stable;' would prevent that. Thanks. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt-get dist-upgrade doesn't downgrade to stable
Dale Hair [EMAIL PROTECTED] writes: On Wed, 2003-01-22 at 12:22, Lloyd Zusman wrote: Simon Law [EMAIL PROTECTED] writes: On Wed, Jan 22, 2003 at 06:15:53PM +0200, Mohammed Sameer wrote: [ ... ] here's my /etc/apt/preferences Package: * Pin: release a=stable Pin-Priority: 1001 and /etc/apt/apt.conf APT::Default-Release stable; Make sure you have stable in your sources.list. Take out APT::Default-Release stable;. Then apt-get update apt-get -t stable dist-upgrade. You should see it downgrade. Simon [ ... ] I just tried exactly what Mohammed said he did and my testing/unstable machine wanted to remove 176 packages and downgrade 586 to stable. I also tried it with apt.conf set to testing with the same results. The preferences file seems to have priority over the apt.conf file. Something is strange on my system. I did the same thing, except I wanted to downgrade from a mixed testing+unstable system to a pure testing system. I followed the same steps as above, except replacing the string 'stable' with 'testing'. Nevertheless, nothing got downgraded (see below for details). I know that I previously installed perl-5.8.0 from the 'unstable' distribution, and that it is my default perl interpreter (I can tell by running perl -V). But the following did not do anything, and after these steps, I was still using perl-5.8.0 instead of perl-5.6.1 from 'testing' ... contents of /etc/apt/preferences Package: * Pin: release a=testing Pin-Priority: 1001 /etc/apt/apt.conf was completely deleted Then ... apt-get update apt-get -t testing dist-upgrade My output (after proper output from the apt-get update portion): Reading Package Lists... Done Building Dependency Tree... Done Calculating Upgrade... Done 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. So no downgrade (nor anything else) took place. Does anyone know what else might be affecting this, besides the contents of /etc/apt/apt.conf and /etc/apt/preferences? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Ignoring ps's pid in ps
Mike Dresser [EMAIL PROTECTED] writes: On Mon, 20 Jan 2003, Lloyd Zusman wrote: Does anyone know of a way to use the standard 'ps' to get a process listing without showing the pid of that instance of 'ps' itself, WITHOUT any piping being performed or auxiliary processes being done, and without using a different utility, such as 'pgrep'? how close does ps --deselect come? Well, perhaps ... but I can't figure out how to use it to get what I want. Suppose that I have several 'ps' instances running, and I do ... ps -C ps u Let's use the term my process to refer to the process that gets invoked by this command line. Is there a way to use the --deselect option to show all of the 'ps' instances except my process? In other words, how would I distinguish my process from all the other 'ps' processes? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Doing a mass downgrade
I'm using the priorities in /etc/apt/preferences to control the software base from which I'm updating my system. I had 'testing' set to the highest priority for a long time. But recently, I made the mistake of setting 'unstable' to be the highest priority, which meant that my daily upgrade runs from aptitude started bringing in a lot of 'sid' packages. Well, needless to say, my system is now unstable, and now I'd like to back out all of these 'unstable' packages. The problem is that I'm not sure which packages I have upgraded to 'unstable', and which ones are still either 'stable' or 'testing' packages. What is the recommended way to do a mass downgrade so that all my 'unstable' packages get downgraded to 'testing' versions? I have already reset the priorities in /etc/apt/preferences so that 'testing' has the highest priority, 'stable' has the next highest, and 'unstable' has the lowest (my current /etc/apt/preferences listing follows). So now what should I do to perform the mass downgrade? I apologize if this question was answered here recently. One side effect of my newly created system instability is that I lost lots of my archived email messages, including many from this mailing list. Thanks in advance for any pointers and suggestions you can offer. Current /etc/apt/preferences Package: * Pin: release a=stable Pin-Priority: 500 Package: * Pin: release a=testing Pin-Priority: 600 Package: * Pin: release a=unstable Pin-Priority: 5 -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Doing a mass downgrade
Dale Hair [EMAIL PROTECTED] writes: [ ... ] Giving testing a priority of 1001 will downgrade to testing. Thank you. I just did that. And now, which `apt-get' or `aptitude' command(s) should I invoke to perform this downgrade? When I go into `aptitude', I don't see anything obvious which shows me that I can perform a downgrade (everything looks the same as it did before I changed the `testing' priority to 1001). Thanks again, in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Doing a mass downgrade
Dale Hair [EMAIL PROTECTED] writes: On Sat, 2003-01-18 at 12:03, Lloyd Zusman wrote: Dale Hair [EMAIL PROTECTED] writes: [ ... ] Giving testing a priority of 1001 will downgrade to testing. Thank you. I just did that. And now, which `apt-get' or `aptitude' command(s) should I invoke to perform this downgrade? [ ... ] apt-get dist-upgrade I did this, but nothing happened. The command session is shown below. I know that a number of unstable packages had been previously installed. Any ideas? Thanks. # apt-get dist-upgrade Reading Package Lists... Done Building Dependency Tree... Done Calculating Upgrade... Done 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Doing a mass downgrade
Dale Hair [EMAIL PROTECTED] writes: On Sat, 2003-01-18 at 13:04, Lloyd Zusman wrote: [ ... ] I did this, but nothing happened. The command session is shown below. I know that a number of unstable packages had been previously installed. Any ideas? Thanks. # apt-get dist-upgrade Reading Package Lists... Done Building Dependency Tree... Done Calculating Upgrade... Done 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. [ ... ] If your /etc/apt/preferences has Package: * Pin: release a=testing Pin-Priority: 1001 and unstable with a priority lower than 1000, then you must not have any packages from unstable. Well then, I guess I must have just gotten lucky since I started downloading with 'unstable' having the highest priority, such that my only upgrades happened to come from the 'stable' or 'testing' categories. Thanks for your help. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Curious...Are most of you in tech-related careers/schooling?
Scott --sidewalking-- [EMAIL PROTECTED] writes: [ ... ] I wonder if all (or most) of you are in similar careers and that is why you are so proficient with compiling and testing and tweaking all of this stuff. Or is it just a hobby that has gone on for so long that you have advanced your knowledge of Linux/Debian to these levels that all of you are at? Just curious... Well, as for me, I wrote my first computer program in 1969, and we all used puchcards and paper tape and printouts, as well as console entry switches on the computer itself. We programmed in Assembly Language, FORTRAN, PL/I, and some people used a new, state-of-the-art experimental language called Basic, which actually ran on an interactive terminal (a 10 character-per-second teletype that also accepted paper tape). Some people used COBOL, ALGOL, and other languages, but I didn't see those until later. Anyway, back then, everyone compiled source code without the benefit of anything resembling a graphical user interface. Then, in 1975, I got to play with a fairly new operating system that was making the rounds at various universities. It was called Unix. Unix contained constructs that made all of us hackers (that was a positive term back then ... it meant geeks) drool, because it had a really cool language called C which was a higher-level language that was also written to be able to do byte and bit manipulation just like Assembly Language. And it had really neat things like auto-increment/auto-decrement (both pre- and post- !) and something etherial that sent us to software nirvana called pointers. C wasn't the first language to have some of these features, but the way it was put together into one language was a work of art! Unix kept getting more and more popular, and soon it had taken over most University computer departments because it was so cool. Thousands or maybe millions of software nerds were studying and improving Unix. Linux is a version of Unix that first came out in the mid-1990's. So by now, it's part of the nearly 30-year evolution of the original Unix. Microsoft Windows has been around for around 18 years, but not in its current form until around 1995. That means that Unix has a 20-year jump on Windows in terms of sophistication, quality, reliability, etc. In other words, MS Windows is a Johnny-come-lately, while Linux, and its direct ancestor Unix, are a standard that has been around for a long time. Many of us love Linux because it's a robust, reliable, powerful, lean and mean OS, as opposed to the hacky bloatware that comes out of Microsoft. Many of us love Linux because it's free, and because thousands of people all over the world are contstantly volunteering their time to write and improve it, and software for it. And free means there is no monopoly. Bill Gates and Microsoft are trying to set up a monopoly, and if they succeed, software innovation will stand still, since there will no competition to force the innovation. Look at Microsoft Word: there is no longer any significant competition for it, and so it has stagnated into a piece of mega-bloatware junk, containing hundreds of known bugs which have not been fixed in 10 years because there is no market incentive to do so. That's the result of a monopoly. On the other hand, the Linux world is filled with people working on and improving the software. There are updated debian packages almost every day that I download. For free! So to me, the differences between Windows and Linux fall into three categories: 1. Quality -- Unix and its descendants such as Linux are robust, sophisiticated, and very stable. MS Windows is just now barely approaching a minimum level of stability. MS Windows is a bloated piece of junk that is trying to be everything to everyone with a PC, and hence, it doesn't do hardly anything well. Linux is more efficient and leaner. 2. Religion -- Many people in the Linux world (me included) are morally opposed to the monopolistic practices that Microsoft engages in, and we believe that working on free software like Linux is a way to fight against the devil. 3. Fun -- For a programmer, working under Unix tends to be a lot of FUN! That's because there is so much power in the OS and in all the software tools that other fun-loving programmers have created. Working in the MS world is much more of a pain in the ass for lots of us. For me, Linux (and the various Unix offshoots) is like playing with my toys; MS Windows is like going to work every day in the salt mines. Well, enough of my ranting. But just wait until you read the replies of someone who is _really_ biased! :) :o) Scott (sidewalking) -- Lloyd Zusman [EMAIL
Re: WTF is popping up that frigging window?
Robert L. Harris [EMAIL PROTECTED] writes: Recently my machine started popping up a window that I don't remember telling it to do. When I select a URL in my konqueror or mozilla windows this pops up asking how I want to open the URL. When I double click a URL in an Eterm it does the same thing. How do I turn this stupid thing off? It's driving me nuts. Robert If you're using KDE, this is probably being controlled by an application called `klipper'. It should be represented on your KDE Panel by a little icon that looks like tha clipboard with a letter K on it. You can either remove this completely, or else you can right-click on it and select Preferences to configure its behavior. HTH -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
aptitude's analog to apt-get ... -t?
I'm trying to figure out what, if anything, is aptitude's analog to using the -t option with apt-get. In other words ... Given an /etc/apt/preferences file which prioritizes stable, testing, and unstable distribtions, I can get apt-get to override this default priority by using its -t flag. For example, if my /etc/apt/preferences file specifies that testing is the default distribution (by giving it the highest priority), I can override this and install a package from, say, the unstable distribution as follows: apt-get install -t unstable the-package-name Sometimes I'd like to start up aptitude in its interactive mode and, in a manner similar to using the -t flag with apt-get, have it perform all of its package operations on one of the distributions which isn't my default. Does anyone know how I can make aptitude do this, short of manually changing the priorities in /etc/apt/preferences before starting aptitude? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Preferred procedure for getting extra php4 functionality?
I noticed that the standard debian php4 package doesn't seem to be built with the --enable-mime-magic option, and therefore, it doesn't provide the capability to use the mime_content_type() function. What's the preferred debian procedure for getting extra php functionality like this? Download the php4 source packages and rebuild them myself? Send a message to the package maintainer(s)? Or ... ??? Thanks. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Preferred procedure for getting extra php4 functionality?
Colin Watson [EMAIL PROTECTED] writes: On Thu, Jan 02, 2003 at 11:43:42AM -0500, Lloyd Zusman wrote: I noticed that the standard debian php4 package doesn't seem to be built with the --enable-mime-magic option, and therefore, it doesn't provide the capability to use the mime_content_type() function. What's the preferred debian procedure for getting extra php functionality like this? File a wishlist bug report against php4. See http://www.debian.org/Bugs/Reporting. Hecho. (Spanish for done). Thanks. Cheers, -- Colin Watson [[EMAIL PROTECTED]] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Getting a Higher Version w/out upgrading
Stig Are M. Botterli [EMAIL PROTECTED] writes: In article [EMAIL PROTECTED], Hal Vaughan wrote: When thinking about this, I wonder: Is there a way to easily specify and add versions later than in stable for only specific packages? In other words, is there some way I can do an apt-get and specify Perl 5.8 (assuming it's already packaged) without having to move to the testing or unstable branch? Add the following line to /etc/apt/apt.conf: APT::Default-Release stable; With this, apt-get will default to installing packages from stable. To install a more up-to-date .deb if available, simply throw in '-t testing' or '-t unstable' as an argument to apt-get. Works brilliantly. [ ... ] I presume that this can also be done just as easily by putting a file with that line into the /etc/apt/apt.conf.d directory, correct? Also ... does anyone know if the aptitude program also abides by this APT::Default-Release property in /etc/apt/apt.conf? Thanks. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Sometimes getting sound fine under woody, but not with xine
I'm a bit mystified about a problem that I've been pursuing on and off for the past couple weeks or so. I've finally thrown up my hands and decided to ask about this here. I'm running woody, and in many cases, sound works just fine. For example, I hear the music that KDE desktop plays when it comes up (even though when I get some time, I want to figure out how to disable this!). Also, I have installed the mplayer debian package, and that program plays mpeg's and avi's just fine with sound, using its oss driver (ao=oss in the mplayer config file). However, when I try to use xine and its cousins (such as sinek) to view these same avi's and mpeg's, I never get any sound, and I get error messages like this (long lines wrapped to fit in this message): audio_oss_out: Opening audio device... audio_oss_out: using device /dev/dsp1 audio_oss_out: opening audio device /dev/dsp1 failed: No such device load_plugins: audio output plugin /usr/lib/xine/plugins/xineplug_ao_out_oss.so: init_audio_out_plugin failed. main: the specified audio driver 'oss' failed I indeed have a /dev/dsp1: % ls -Flasd /dev/dsp* 0 crw-rw1 root audio 14, 3 Dec 28 01:37 /dev/dsp 0 crw-rw1 root audio 14, 19 Dec 28 01:37 /dev/dsp1 0 crw-rw1 root audio 14, 35 Dec 28 01:37 /dev/dsp2 0 crw-rw1 root audio 14, 51 Dec 28 01:37 /dev/dsp3 I don't get any sound even when I run xine and sinek as root, by the way. The following line appears in /etc/group: audio:x:29:root[, other user names] % uname -a (long line wrapped ...) Linux home 2.4.20-686-smp #1 SMP Sat Dec 21 16:41:15 EST 2002 i686 unknown unknown GNU/Linux And below are the results of lsmod. Notice the entries for i810_audio, ac97_codec, audio, sound, and soundcore in that list. There has got to be something simple that I'm overlooking here. Thanks in advance for any help you might be able to provide. % lsmod Module Size Used byNot tainted iptable_filter 1768 0 (autoclean) (unused) ip_tables 11160 1 [iptable_filter] i810_audio 23912 0 ac97_codec 9864 0 [i810_audio] audio 38172 0 (unused) sound 55884 0 (unused) soundcore 4036 4 [i810_audio audio sound] msr 1448 0 (unused) microcode 3808 0 (unused) cpuid 1032 0 (unused) netlink_dev 1860 0 (unused) ipx18788 0 umsdos 24800 0 (unused) msdos 4908 0 [umsdos] fat30488 0 [umsdos msdos] smbfs 34320 0 (unused) romfs 3812 0 (unused) binfmt_misc 5952 1 binfmt_aout 4068 0 usb-storage55900 0 (unused) visor 8716 0 (unused) kl5kusb105 9048 0 (unused) ipaq5056 0 (unused) usbserial 17724 0 [visor kl5kusb105 ipaq] usbnet 10146 0 (unused) printer 7008 0 (unused) microtek5520 0 (unused) mdc800 7292 0 (unused) ibmcam 39456 0 (unused) usbvideo 23000 0 [ibmcam] videodev5696 0 [usbvideo] dc2xx 3200 0 (unused) isa-pnp28484 0 (unused) tulip 39712 1 raw1394 6968 0 (unused) ohci1394 15976 0 (unused) ieee1394 31468 0 [raw1394 ohci1394] agpgart34528 0 (unused) parport_pc 21224 1 (autoclean) lp 6016 0 (unused) parport24320 1 [parport_pc lp] busmouse3292 0 (unused) nbd15396 0 (unused) keybdev 1664 0 (unused) usbkbd 2908 0 (unused) input 3520 0 [keybdev usbkbd] usb-uhci 0 0 (unused) usbcore58112 1 [audio usb-storage visor kl5kusb105 ipaq usbserial usbnet printer microtek mdc800 ibmcam usbvideo dc2xx usbkbd usb-uhci] rtc 6588 1 (autoclean) sd_mod 10764 12 (autoclean) ext3 60832 5 (autoclean) jbd40040 5 (autoclean) [ext3] aic7xxx 111640 6 (autoclean) scsi_mod 91480 2 (autoclean) [usb-storage microtek sd_mod aic7xxx] unix 15784 108 (autoclean) -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sometimes getting sound fine under woody, but not with xine
nate [EMAIL PROTECTED] writes: Lloyd Zusman said: audio_oss_out: Opening audio device... audio_oss_out: using device /dev/dsp1 audio_oss_out: opening audio device /dev/dsp1 failed: No such device load_plugins: audio output plugin /usr/lib/xine/plugins/xineplug_ao_out_oss.so: init_audio_out_plugin failed. main: the specified audio driver 'oss' failed while I cannot provide a solution I can say that I too have experienced this. On my IBM thinkpad T20 with 2.2.19, getting xine to work with audio was hell. I never figured out how I got it working since everytime I rebooted the system I had to start from scratch again, maybe I only got it working by blind luck. I upgraded to 2.4.20 a couple weeks ago, ran xine for the first time last night and audio worked perfectly for about 30 seconds then the X server froze(though the audio was still comming through clearly). Haven't tried it other then the one time. I notice your using 2.4.19 now.. perhaps switching to ALSA would help? I tried switching to ALSA and had no luck either under 2.2.19. Though XMMS, mpg123, and games like simcity 3000 worked flawlessly as far as sound goes. never a glitch. Thanks for your reply. At least now I don't feel so alone with this problem. Actually, I'm using 2.4.20-686-smp (see the uname -a entry in my original message). But the same problem has been there for me in 2.4.{18,19}-686-smp, as well. perhaps a bug in the driver..or in xine ? I don't recall trying other video players on my laptop, only xine. nate Well, since I can hear KDE's startup music, and since mplayer's oss driver properly produces sounds for me, I do think that it's somehow a driver problem. But the mystery is whether it's an xine driver problem, a kernel driver problem (perhaps mplayer and KDE use their own sound drivers which bypass the kernel), or what ... ??? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[SOLVED] Re: Sometimes getting sound fine under woody, but not withxine
Stephen Gran [EMAIL PROTECTED] writes: [ ... ] I don't remember xine's options all that clearly, and I can't check right now (I'm ssh'ed in in console only mode), but /dev/dsp1 looks suspicious to me. Do you have two sound cards, or two audio outs? I would try to force xine to send it's audio out to /dev/dsp, where it will normally be dealt with by the oss wrapper in KDE's arts server. You just provided me with the key to the solution. For some reason, the default ~/.xine/config file that got generated by running the default installation of xine-ui contained the following lines: # /dev/dsp# device to use for oss output, -1 = auto_detect # numeric, default: -1 audio.oss_device_num:1 The 1 after the colon means that /dev/dsp1 should be used. I changed the line as follows, and now the program tries to use /dev/dsp, and the sound works fine: audio.oss_device_num:-1 I know I probably just could have commented out that line altogether to have the same effect, but I like to spell things out instead of relying on defaults. At any rate, all is well now, and thank you very much. As for the KDE thing, I think you can disable sound in the control panel somewhere. I'll look for it. It hasn't been all that important to me, but now that the other problem is fixed, it's time to take care of this one. HTH, and sorry I'm not clearer - hopefully it gives you a starting point, though. As you can see, it did. Thanks again! -- -- | Stephen Gran | There is no time like the present for | | [EMAIL PROTECTED] | postponing what you ought to be doing. | | http://www.lobefin.net/~steve | | -- -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Are openquicktime and quicktime4linux mutually exclusive?
I'm trying to get quicktime support, and I'm noticing all sorts of apt-get install conflicts between the various openquicktime packages and the various quicktime4linux packages. Is it safe to say that I should install only one, and not the other of these? If so, do you folks have any thoughts, opinions, or recommendations about each of these? If not, then could someone point me to some documentation as to which packages of each that I need to install, in which order, in order to avoid installation conflicts? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Safe Sex Shell + Condom = Total Protection+=
Lin Hua [EMAIL PROTECTED] writes: Safe Sex Shell + Condom = Total Protection [ ... ] I know that there are already debian packages for the shells csh, tcsh, bash, zsh, etc. But this is a new shell, and I think that we should package this one up as well, possibly as sssh. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: spamassassin out of memory
Small but significant correction. See below ... Anthony Campbell [EMAIL PROTECTED] writes: On 16 Dec 2002, Nicolas SABOURET wrote: Salut, J'utilise spamassassin pour filtrer mon courier ? moi, ? l'aide d'un .forward et d'une r?gle procmail. Quand j'ai trop de mail d'un coup (genre apr?s un we), il p?te les plombs et j'obtiens des : Out of memory: Killed process 12786 (spamassassin) Ma question (angoiss?e) est : quand il fait ?a, est-ce que je perds le mail qu'il ?tait en train de filtrer ? PS: J'y connais rien ? procmail et spamassassin. Merci d'avance, Nico. -- Nicolas SABOURET LIMSI-CNRS, BP133, 91403 Orsay, France http://www.limsi.fr/Individu/nico Translation: I use spamassassin to filter my mail using .forward and a procmail rule. When I have too much mail it suddenly, usually after a we [?], it crashes and I get out of memory; killed process... My desperate question is: when this happens do I use all the mail which ^^^ I think you meant to type lose instead of use, above. it was filtering at the time? Did things get turned around in your brain because of the phrase use it or lose it? :) -- [EMAIL PROTECTED]|| http://www.acampbell.org.uk using Linux GNU/Debian || for book reviews, electronic Windows-free zone || books and skeptical articles -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Default fonts for applications?
Rob Weir [EMAIL PROTECTED] writes: On Sat, Dec 14, 2002 at 11:25:33PM -0500, Lloyd Zusman wrote: [ ... ] [ ... ] For example, mozilla (version 1.0.0) comes up with that default font on its menu bar. I've looked in app-defaults and in the various X startup files, but I can't figure out how to change that particular font. I don't really know where Mozilla gets it default font from, sorry. To be clear, I want to emphasize that this same font is is used for quite a few other X-based apps that I run, such as gqview. None of gqview (IIRC), is a gtk-based program, so it's font will be set in ~/.gtkrc. This file is pretty simple, and it should be obvious how to modify it. Use 'xfontsel' to find a font, if you don't know it's full X font resource string. Perhaps I didn't make my original question clear. Many X apps, including mozilla (which is NOT a gtk app) and gqview (which IS a gtk app), plus lots of other X apps (gtk and non-gtk alike) use this exact, same, default font that I'm referring to. This font is NOT application-specific or toolkit-specific, and as I mentioned, it's not controllable via the KDE control panel. Therefore it seems to me that there must be some sort of default X font that gets used when an application doesn't specify its own font. If so, where is this default font set? Does it result from the FontPath setting in my XF86Config-4 file, perhaps? Is there maybe some generic Xdefaults setting that controls this? Is it hard-coded into the X server? Or ... ??? -rob -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Finally figured it out (was: Default fonts for applications?)
Lloyd Zusman [EMAIL PROTECTED] writes: [ ... ] Files: $HOME/.gtkrc-kde /usr/share/gnome/gtkrc /usr/share/themes/Default/gtk-2.0-key/gtkrc /usr/share/themes/Default/gtk-2.0/gtkrc /usr/share/themes/Default/gtk/gtkrc /etc/gtk/gtkrc.en Statements: style default-text { fontset = -misc-fixed-bold-r-*-*-*-100-*-*-*-*-iso8859-* } class GtkWidget style default-text class * style default What am I missing here? Well, I finally figured it out: 1. $HOME/.gtkrc-kde gets overwritten every time I log in, so my changes disappeared before they could take effect. 2. I noticed the following environment variable setting: GTK_RC_FILES=/etc/gtk/gtkrc:/home/ljz/.gtkrc:/home/ljz/.gtkrc-kde So, I created a $HOME/.gtkrc that contained the style and class statements. That file doesn't get overwritten. 3. The last class line should have been this: class * style default-text (I had accidentally typed default). Based on these insights, I was able to get the fonts I desire in mozilla, gqview, and lots of other apps. Thanks again to you folks who pointed out that mozilla uses gtk after all, which got me back on the right track. And now ... is any of this documented anywhere? If so, where? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Finally figured it out
Rob Weir [EMAIL PROTECTED] writes: [ ... ] 2. I noticed the following environment variable setting: GTK_RC_FILES=/etc/gtk/gtkrc:/home/ljz/.gtkrc:/home/ljz/.gtkrc-kde So, I created a $HOME/.gtkrc that contained the style and class statements. That file doesn't get overwritten. Again, this is KDE playing games. I think I mentioned the ~/.gtkrc file in my original response. Yes, indeed you did. But I got confused because there are no .gtkrc's in any HOME directories on my system, only .gtkrc-kde files. Because of that, I incorrectly assumed that I have to stick with .gtkrc-kde in my KDE environment. Silly me. [ ... ] And now ... is any of this documented anywhere? If so, where? Not sure, aside from the GTK source, I guess...I'm sure it's come up (at least) on the GTK lists over at gtk.org, so maybe try there? I'll give it a shot. This is really a pain to figure out without suitable docs. Thanks again. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Default fonts for applications?
I'm running woody with kde. I've used the kde control panel to specify the fonts I'd like to use, and these settings indeed work for various kde objects, for window manager decorations, etc. However, many X applications come up with a default font that I don't like, and I want to change it. However, I can't figure out where this default font resides. For example, mozilla (version 1.0.0) comes up with that default font on its menu bar. I've looked in app-defaults and in the various X startup files, but I can't figure out how to change that particular font. To be clear, I want to emphasize that this same font is is used for quite a few other X-based apps that I run, such as gqview. None of these apps have entries in the app-defaults directory, which leads me to believe that this is some sort of default font. Anyway, any pointers you folks could give me as to where this default font is set would be greatly appreciated. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Newbie trying to install mod_ssl
Jokke Heikkilä [EMAIL PROTECTED] writes: I've installed Woody and wanted to add mod_ssl to my apache conf. I downloaded the tarball from modssl.org and unpacked. When installing the mod, the installer asks for the path for my Apache source tree. Where is it? I sort of thought that it would be /usr/lib/apache/src , but there's no folder by that name. jokke h. Unless you have a special need to compile mod_ssl in a non-standard way, I think you can avoid building it altogether. Instead, you can just install the apache-ssl debian package, which already has SSL support built in. It provides a separate apache-ssl executable which allows this ssl-enabled Apache server to run in parallel with the standard one. I'm running the servers from both the apache and apache-ssl packages on my system, and everything works fine. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[REPOST] Only minmal modules loaded with usermin?
[NOTE: I posted something very similar a few days ago, but I haven't seen any replies. Therefore, in case my original query didn't go out properly, I'm trying again with this almost identical message. Please forgive me if this has already shown up for you.] I've installed both webmin and usermin under woody, but there seems to be some sort of configuration problem with usermin: not all of its installed and configured modules are showing up. I installed all the webmin.* and usermin.* packages, including `webmin-usermin'. I have gone into Usermin Configuration via webmin, into the Available Modules section, and I have selected all 22 modules to be made visible to users. I have then gone into Module Restrictions and made all of these modules available to the users in question. Under Allowed Users and Groups, I have selected Allow all users. All other usermin configuration settings are reasonable. Then, I go to the usermin page at https://my.domain.name:2, and I log in. All I see are the following 4 modules: Disk Quotas Plan File Scheduled Commands Scheduled Cron Jobs This happens no matter what user I log in as. Where are the other 18 modules? Even after shutting down and restarting webmin and usermin, I can't see more than these 4 usermin modules when I log in. And yes, I made sure to Save everything under the Usermin Configuration section of webmin. I know that this has worked, because all my choices have been saved and reappear every time I go to that configuration section. Any ideas? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sources of changes to /etc/motd and of other login messages?
martin f krafft [EMAIL PROTECTED] writes: also sprach Colin Watson [EMAIL PROTECTED] [2002.12.11.1244 +0100]: Also, what causes the following prompt to print out at login, after /etc/motd has been displayed? ... Last login: Wed Dec 11 06:25:26 2002 from 204.212.175.30 Not sure ... PAM: % grep lastlog /etc/pam.d/login sessionoptional pam_lastlog.so wow, i knew something that colin didn't... I mentioned in my original message about pam_lastlog.so. I pointed out that I've set it as follows inside of /etc/pam.d/login, but I'm still getting the Last login: ... line when I log in: sessionoptional pam_lastlog.so silent Apparently, this line is getting ignored, and something else is causing that Last login: ... message to get printed. But what ... ??? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sources of changes to /etc/motd and of other login messages?
martin f krafft [EMAIL PROTECTED] writes: also sprach Lloyd Zusman [EMAIL PROTECTED] [2002.12.12.1719 +0100]: I mentioned in my original message about pam_lastlog.so. Sorry, didn't see that... No problem. Apparently, this line is getting ignored, and something else is causing that Last login: ... message to get printed. on the console or through ssh? Both. have you checked /etc/login.defs? Yes. Everything is normal. The following appears at the bottom of that file, in the OBSOLETED BY PAM section (among about a dozen or so commented-out tags): #LASTLOG_ENAB Incidentally, the following line appears both in /etc/pam.d/login and /etc/pam.d/ssh: sessionoptional pam_lastlog.so silent No other file in /etc/pam.d has any mention of pam_lastlog.so in it. What 'login' are you using? I'm not sure. It's the standard /bin/login that comes with woody. How can I tell? -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sources of changes to /etc/motd and of other login messages?
Colin Watson [EMAIL PROTECTED] writes: On Thu, Dec 12, 2002 at 03:54:46PM +0100, martin f krafft wrote: also sprach Colin Watson [EMAIL PROTECTED] [2002.12.11.1244 +0100]: PAM: % grep lastlog /etc/pam.d/login sessionoptional pam_lastlog.so He said he'd added the 'silent' option, though, in the original post. This isn't via ssh, is it? There's a bug filed about this, #169938. Actually, it *is* via ssh. About an hour ago I wrote a reply to Martin Krafft here, saying that that the problem occurred both in login and ssh ... but I was wrong about login. I came back here to correct that post, and I just noticed this message of yours. I just checked bug 169938, and on the basis of that report I set PrintLastLog no in /etc/ssh/sshd_conf. That fixed the problem. I suppose that recompiling ssh with --without-lastlog might also have done the trick (as long as I also make sure that I use --with-pam). I got thrown off the track when I mistakenly thought that I saw this same problem when using 'login' as well as 'ssh'. Martin Krafft's question a little earlier today got me thinking that I had better double-check this, and I'm glad I did. Thanks to the both of you. Problem solved. I'm happy. :) -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Sources of changes to /etc/motd and of other login messages?
What process or utility changes the contents of /etc/motd? Occasionally, I see an extra system identification line that has been placed into that file. Also, what causes the following prompt to print out at login, after /etc/motd has been displayed? ... Last login: Wed Dec 11 06:25:26 2002 from 204.212.175.30 I have put the following into /etc/pam.d/login: sessionoptional pam_lastlog.so silent However, that Last login: ... message still shows up. This happens despite the fact that LASTLOG_ENAB is commented out in /etc/login.defs, in the OBSOLETED BY PAM section. Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Sources of changes to /etc/motd and of other login messages?
Colin Watson [EMAIL PROTECTED] writes: On Wed, Dec 11, 2002 at 06:34:48AM -0500, Lloyd Zusman wrote: [ ... ] [cjwatson@arborlon /etc/init.d]$ grep motd * bootmisc.sh:# Update /etc/motd. bootmisc.sh:uname -a /etc/motd.tmp bootmisc.sh:sed 1d /etc/motd /etc/motd.tmp bootmisc.sh:mv /etc/motd.tmp /etc/motd [cjwatson@arborlon /etc/init.d]$ dlocate bootmisc.sh sysvinit: /etc/init.d/bootmisc.sh I didn't think this was connected to system initialization at boot time, but perhaps the system has been rebooted more frequently than I thought. Thanks. [ ... ] Cheers, -- Colin Watson [[EMAIL PROTECTED]] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Only minimal modules loaded with usermin?
I've installed both webmin and usermin under woody, but there seems to be some sort of configuration problem with usermin. I installed all the webmin.* and usermin.* packages, including `webmin-usermin'. I have gone into Usermin Configuration via webmin, into the Available Modules section, and I have selected all 22 modules to be made visible to users. I have then gone into Module Restrictions and made all of these modules available to the users in question. Under Allowed Users and Groups, I have selected Allow all users. All other usermin configuration settings are reasonable. Then, I go to the usermin page at https://my.domain.name:2, and I log in. All I see are the following 4 modules: Disk Quotas Plan File Scheduled Commands Scheduled Cron Jobs This happens no matter what user I log in as. Where are the other 18 modules? Even after shutting down and restarting webmin and usermin, I can't see more than these 4 usermin modules when I log in. And yes, I made sure to Save everything under the Usermin Configuration section of webmin. I know that this has worked, because all my choices have been saved and reappear every time I go to that configuration section. Any ideas? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Mappings between CPAN packages and lib.*-perl packages
Colin Watson [EMAIL PROTECTED] writes: On Sat, Dec 07, 2002 at 01:46:11AM -0500, Lloyd Zusman wrote: Is it correct to assume that any given debian package whose name is of the form lib.*-perl corresponds to one and only one CPAN module, Not quite, but more or less. and that the algorithm for determining the corresponding debian package name from a given CPAN module name is this? ... - Add a leading lib and a trailing -perl to the CPAN module name. - Replace all occurrences of double colon ('::') with a hyphen ('-'). - Convert to lower case. That's pretty much it, yes, and it works 99% of the time. However, for one reason or another it might not always be absolutely reliable: occasionally packages have been aggregated or split, or are using an older naming scheme, or whatever. If I wanted something absolutely reliable, I would apply s,::,/,g instead and use either 'auto-apt search -f' or 'apt-file search'. (Both require an update step as root every so often.) Yep, got it. Thank you. And here's the first cut of a perl script I wrote to implement this. #!/usr/bin/perl # -*- perl -*- $0 =~ s:^.*/::; my $program = $0; unless (scalar(@ARGV) 0) { die usage: $program perl-module ...\n; # notreached } my $aptFile = '/usr/bin/apt-file'; my %output = (); foreach my $module (@ARGV) { $module =~ s|::|/|g; unless (open(AF, $aptFile search $module 2/dev/null |)) { die $program: unable to invoke: $aptFile\n; # notreached } while (defined(my $line = AF)) { chomp($line); unless ($line =~ m/:.*perl.*\.pm/) { next; } $output{$line} = 1; } close(AF); } foreach my $item (sort keys %output) { print $item\n; } exit(0); __END__ -- Colin Watson [[EMAIL PROTECTED]] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Mappings between CPAN packages and lib.*-perl packages
Colin Watson [EMAIL PROTECTED] writes: On Sat, Dec 07, 2002 at 01:46:11AM -0500, Lloyd Zusman wrote: Is it correct to assume that any given debian package whose name is of the form lib.*-perl corresponds to one and only one CPAN module, Not quite, but more or less. and that the algorithm for determining the corresponding debian package name from a given CPAN module name is this? ... - Add a leading lib and a trailing -perl to the CPAN module name. - Replace all occurrences of double colon ('::') with a hyphen ('-'). - Convert to lower case. That's pretty much it, yes, and it works 99% of the time. However, for one reason or another it might not always be absolutely reliable: occasionally packages have been aggregated or split, or are using an older naming scheme, or whatever. If I wanted something absolutely reliable, I would apply s,::,/,g instead and use either 'auto-apt search -f' or 'apt-file search'. (Both require an update step as root every so often.) OK, I understand. Thank you. And here's the first cut of a perl script I wrote to implement this. #!/usr/bin/perl # -*- perl -*- $0 =~ s:^.*/::; my $program = $0; unless (scalar(@ARGV) 0) { die usage: $program perl-module ...\n; # notreached } my $aptFile = '/usr/bin/apt-file'; my %output = (); foreach my $module (@ARGV) { $module =~ s|::|/|g; unless (open(AF, $aptFile search $module 2/dev/null |)) { die $program: unable to invoke: $aptFile\n; # notreached } while (defined(my $line = AF)) { chomp($line); unless ($line =~ m/:.*perl.*\.pm/) { next; } $output{$line} = 1; } close(AF); } foreach my $item (sort keys %output) { print $item\n; } exit(0); __END__ -- Colin Watson [[EMAIL PROTECTED]] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Starting a non-root daemon a minute or so after boot-up?
I'm looking for a way to start a non-root daemon a minute or so after boot-up. I tried the following (see the end of this message), but the program didn't start, and I didn't get a log file with any error indications. I'm sure that I can figure out a variation of the code below that would work, but instead of making use of some sort of hack, I'd rather use a utility that's designed for this purpose, if such a thing exists. The 'fcron' program looked promising, since it's able to measure time relative to its own startup. However, I can't figure out how to configure fcron to run a program once and only once. Here's what I previous tried, which didn't work ... - Create a file called 'localstart' in /etc/init.d. - Make sure this file is executable, owned by root, etc. - Make a symlink to this file from S00localstart in /etc/rc{2,3,4,5}.d - Here are the contents of /etc/init.d/localstart, which for some yet-to-be-determined reason didn't work: #!/bin/sh case ${1} in start) { /bin/sleep 120 /bin/su theuser -c \ /usr/bin/fetchmail -s -f /home/theuser/etc/fetchmailrc } /var/log/localstart 21 ;; esac Like I mentioned, I'm quite able to debug the code above and eventually to get it to work. But I'm hoping I can find a utility that might allow me to do this in a more elegant way. Any ideas? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Doing an apt-get --purge remove after an earlier remove?
Colin Watson [EMAIL PROTECTED] writes: On Thu, Dec 05, 2002 at 07:50:44PM -0500, Lloyd Zusman wrote: [ ... ] OK. So then I won't worry about --purge not working here. I guess I'll just dig into the installation scripts and try to figure out what they created, so I can get rid of it all manually. Erm, if it's been removed and there are no conffiles and no post-removal script, then there should be nothing left to remove. Otherwise it's a bug. Could you elaborate on what package you're talking about? Well, there are several, but the package that comes to mind right now is dpsyco. I installed it, but subsequently decided that I didn't want it, so I then removed it. Even after the removal and the --purge discussed in this thread, I still have the following files and directories on my system (see below). The one which causes a noticeable problem is the first one (90dpsyco), since it runs and prints and error message every time I do an apt-get install of other packages. It's a benign error (due to || true in DPkg::Post-Invoke), but I believe that it shouldn't be happening at all. Results of locate dpsyco ... /etc/apt/apt.conf.d/90dpsyco /etc/dpsyco /var/cache/apt/archives/dpsyco-base_1.0.10_all.deb /var/cache/apt/archives/dpsyco_1.0.10_all.deb /var/lib/dpkg/info/dpsyco-base.list /var/lib/dpkg/info/dpsyco-lib.list /var/lib/dpkg/info/dpsyco-lib.postrm Results of ls -l /etc/dpsyco ... total 20 -rw-r--r--1 root root 2573 Dec 3 00:46 adduser.conf -rw-r--r--1 root root 1390 Apr 8 2002 defaults.conf -rw-r--r--1 root root 312 Apr 8 2002 groups.conf -r--r-1 root root 736 Dec 4 19:46 sudoers -rw-r--r--1 root root 330 Apr 8 2002 users.conf Results of dlocate dpsyco ... dpsyco-lib: /etc/dpsyco dpsyco-lib: /etc/dpsyco/defaults.conf dpsyco-base: /etc/apt/apt.conf.d/90dpsyco dpsyco-base: /etc/dpsyco/groups.conf dpsyco-base: /etc/dpsyco/users.conf Results of cat /etc/apt/apt.conf.d/90dpsyco ... // All dpsyco packages should be updated after installation. DPkg::Post-Invoke {/usr/sbin/update-dpsyco || true;}; -- Colin Watson [[EMAIL PROTECTED]] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Doing an apt-get --purge remove after an earlier remove?
Colin Watson [EMAIL PROTECTED] writes: On Fri, Dec 06, 2002 at 07:34:40AM -0500, Lloyd Zusman wrote: Colin Watson [EMAIL PROTECTED] writes: That does indeed suggest that it's those packages you need to purge. Aha! Yes, that did the trick. I incorrectly assumed that dpsyco-lib and dpsyco-base would be dependent on dpsyco, and therefore should have been removed during the dpsyco purge. The dependencies were apparently the other way around. Right. Also, the technical meaning of foo depends on bar for Debian packages is that foo will not be configured until bar is configured. Once a package is removed (even if not purged), its dependencies are no longer looked at. Cheers, Got it. So does that mean that dependencies are not looked at during the removal process itself? ... or is it only after the removal process that they can't be used (because they've then disappeared)? -- Colin Watson [[EMAIL PROTECTED]] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Mappings between CPAN packages and lib.*-perl packages
Is it correct to assume that any given debian package whose name is of the form lib.*-perl corresponds to one and only one CPAN module, and that the algorithm for determining the corresponding debian package name from a given CPAN module name is this? ... - Add a leading lib and a trailing -perl to the CPAN module name. - Replace all occurrences of double colon ('::') with a hyphen ('-'). - Convert to lower case. If so, then I presume that I can search for any given CPAN package names as follows (using Perl code here): foreach my $cpanModuleName (@ARGV) { (my $debianPackageName = lib${cpanModuleName}-perl) =~ s/::/-/g; system(apt-cache search --names-only '^${debianPackageName}\$'); } This little program seems to work, but are my assumptions correct? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: shell game
Travis Crump [EMAIL PROTECTED] writes: Drew Cohan wrote: Hi, Using a bash shell script (/bin/sh), I need to know how to check to see if certain files exist in a directory, as in, Are there any jpegs in this directory?. I've tried various things (like using -s, -f with test and a do/for loop) but nothing seems to work right. The closest I can come is if test `ls /opt/images/*.jpg | wc -l` -gt 0 then ... Unfortunately, this gives me the error message ls: /opt/images/*.jpg: No such file or directory when there are no jpegs in /opt/images. So what am I missing here? From the there has got to be a better way department: ls /opt/images/*.jpg /dev/null 21 echo there are jpegs Here's one of several ways to do it in bash that won't involve forking a subprocess, in case you want to conserve system resources: shopt -s nullglob result=; for result in /opt/images/*.jpg; do break; done After this, the '$result' variable will be empty unless there are one or more files which match the pattern. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Doing an apt-get --purge remove after an earlier remove?
I have searched the various docs and mailing list archives, and I couldn't find anything about my question. But please forgive me if I have missed something obvious in the docs, because I've only started using debian a few days ago. My question: sometimes when I have wanted to purge a previously installed package, I have accidentally left out the --purge option to the apt-get command. In other words, I intended to type this: apt-get --purge remove the-package ... but I inadvertantly typed this: apt-get remove the-package In these cases, is there any way that I can now do the equivalent of the --purge without re-installing the package and re-issuing the remove command with the --purge flag? I know that I can probably search through the package database to figure out which items to manually remove from my system, but I'm hoping that there is some utility that I can use to perform this task for me. Any ideas? Thanks in advance. -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Doing an apt-get --purge remove after an earlier remove?
Colin Watson [EMAIL PROTECTED] writes: On Thu, Dec 05, 2002 at 09:34:01AM -0500, Lloyd Zusman wrote: In other words, I intended to type this: apt-get --purge remove the-package ... but I inadvertantly typed this: apt-get remove the-package In these cases, is there any way that I can now do the equivalent of the --purge without re-installing the package and re-issuing the remove command with the --purge flag? Use 'dpkg --purge the-package'. Cheers, Thank you, but sadly, that doesn't work for me. When I run it, I get this response. dpkg - warning: ignoring request to remove the-package which isn't installed. ... and none of the remaining pieces of the-package are removed. Is there perhaps some sort of global configuration option that I need to set, in order to make this work? -- Colin Watson [[EMAIL PROTECTED]] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Doing an apt-get --purge remove after an earlier remove?
Colin Watson [EMAIL PROTECTED] writes: On Thu, Dec 05, 2002 at 10:14:01AM -0500, Lloyd Zusman wrote: [ ... ] dpkg - warning: ignoring request to remove the-package which isn't installed. ... and none of the remaining pieces of the-package are removed. Then there's nothing more to do. The purge step only makes a difference if the package has a post-removal script; it seems that this one doesn't. Is there perhaps some sort of global configuration option that I need to set, in order to make this work? Nope. Cheers, OK. So then I won't worry about --purge not working here. I guess I'll just dig into the installation scripts and try to figure out what they created, so I can get rid of it all manually. I don't suppose there are any tools that aid in doing this, aside from the standard Unix text-processing utilties. As I learn more about the structure of these packages, maybe I'll write such a tool. Thanks for your help. -- Colin Watson [[EMAIL PROTECTED]] -- Lloyd Zusman [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]