[00:17] <soreau> Would someone be able to tell me how to use wild card characters with apt, ie. 'aptitude <action> *package*' ?
[00:17] <ebroder> 'aptitude action ~npackage'
[00:18] <mathiaz> soreau: use quotes ",' around your package otherwise you'll get shell globing
[00:18] <soreau> Thank you ebroder
[00:18] <soreau> mathiaz: What?
[00:19] <ebroder> Oh yeah - mathiaz is right. aptitude will handle * without the ~n
[00:19] <ebroder> But your shell is eating the * if you don't do '*package*' <- with the quotes
[00:21] <directhex> om nom nom
[00:22] <soreau> ebroder: ~n works but I don't understand how to do the same with quotes
[00:22] <Laney> mmm asterisk
[00:22] <ebroder> soreau: aptitude install '*package*'
[00:22] <ebroder> Leave the quotes there
[00:23] <ebroder> Go read http://tldp.org/LDP/Bash-Beginners-Guide/html/sect_03_03.html
[00:23] <soreau> ebroder: I am actually using version number and doing that just says it can't find package "*<version>*"
[06:19] <dholbach> good morning
[09:00] <chrisccoulson> does anyone know if it is possible to stop the apport retracer from marking new bug reports as duplicates of a particular bug report? i'm investigating a gnome-settings-daemon crash, which is getting lots and lots of duplicates, but it has a similar backtrace to an older (now fixed) crash, and apport is continually marking the new reports as duplicates of the old (fixed) report
[09:00] <chrisccoulson> there is currently someone manually changing the duplicate link on every new report
[09:01] <\sh> guys, after / mount, after the fscks, when fsck breaks, it tells me that it writes a log file under /var/log/<bla>/ but at this stage, /var/log is not even correctly available, because the mounts were not going through...where does it write the logfiles then?
[09:01] <\sh> (disregarding the fact of people using flat directory structures with a single partition)
[09:03] <dholbach> chrisccoulson: pitti would know, but he's on holidays
[09:04] <chrisccoulson> dholbach - thanks
[09:04] <chrisccoulson> its unfortunate, because it scrubs the stacetrace from every new report
[09:05] <dholbach> maybe drop him an email
[09:05] <chrisccoulson> dholbach - i don't want to disturb him on his vacation ;)
[09:05] <chrisccoulson> it's nearly the end of the week now anyway, so he'll be back in a couple of working days
[09:06] <dholbach> I dunno what he was planning, if he has a similar vacation as I had, then you'd be out of luck :)
[09:06] <ogra> he might have taken two weeks ...
[09:06] <chrisccoulson> ogra - yeah, he's taken 2 weeks
[09:06] <chrisccoulson> this is the second week already though
[09:06] <ogra> and if he reads mails during holidays its his own fault ;)
[09:07] <ogra> its not like you ring his mobile or so ;)
[09:08] <chrisccoulson> heh, yeah, he should switch his mobile off when on vacation ;)
[09:08] <chrisccoulson> i had my employer phone me up once when on vacation
[09:08] <chrisccoulson> i was not very happy ;)
[09:10] <highvoltage> I guess it depends on your kind of job, in some situations it's acceptable for your boss to phone you during holidays.
[09:11] <chrisccoulson> not in my job it's not ;)
[09:23] <glatzor> morning mvo
[09:30] <mvo> hey glatzor!
[09:39] <glatzor> mvo, I wrote a transaction cache using a metaclass
[09:39] <mvo> glatzor: sweet, what do you think about the dettach() code?
[09:43] <glatzor> mvo, merged in (I rewrote your patch to fix a spelling error and use another naming convention signal_matchers)
[09:43] <glatzor> thanks a lot
[09:43] <glatzor> mvo,  I also added a time out to the daemon transaction
[09:44] <glatzor> you now have got 30 seconds between requesting a transaction and performing an action with it
[09:44] <glatzor> furthermore the transaction will be removed after it is finished from the bus
[09:48] <mvo> glatzor: nice, so all leakings should be gone now - thanks a lot :)
[09:48] <glatzor> mvo, by the way how can I track a bug ony for a single ubuntu version? I fixed a bug in karmic which is still present in jaunty.
[09:48] <mvo> glatzor: you can add a karmic bugtask and a jaunty one
[09:49] <mvo> glatzor: and close them individually
[09:49] <mvo> glatzor: what is the bugnumber?
[09:52] <glatzor> https://bugs.edge.launchpad.net/ubuntu/+source/packagekit/+bug/369081
[09:55] <mvo> glatzor: I added them (via nominate for release
[10:03] <lifeless> mvo: hi
[10:03] <lifeless> mvo: I'm sitting on my jaunty desktop; building a few things and installing the debs
[10:04] <lifeless> mvo: /every time I install one/ update-manager pops up in my face about some [dangerous X-edger stuff]
[10:04] <lifeless> mvo: not to mention that my CPU churns while it re-reads the DB :(
[10:04] <lifeless> mvo: any advice ?
[10:09] <mvo> lifeless: could you please try installing the update-notifier from jaunty-proposed?
[10:09] <mvo> lifeless: that sounds like bug 369198)
[10:09] <jms> i wonder why the display config doodad in jaunty refuses to push 1600x1200 to my external monitor. is there an ubuntu x channel?
[10:10] <lifeless> mvo: thanks!
[10:11] <evand> tkamppeter: Should foomatic-db-hpijs be removed from the seeds now that foomatic-db replaces it?
[10:12] <tkamppeter> evand, yes, packages which should be removed are foomatic-db-hpijs and hal-cups-utils.
[10:12] <evand> tkamppeter: okay, I'll update the seeds then
[10:12] <RAOF> jms: Yes; #ubuntu-x
[10:12] <tkamppeter> evand: Thanks.
[10:12] <evand> sure thing
[10:13] <jms> ta RAOF
[10:13] <jms> hm, no right-click join channel in smuxi
[10:13] <evand> tkamppeter: I take it the hpijs and hplip packages can stay, correct?
[10:15] <tkamppeter> Yes, hplip is the principal driver package for HP printers, perfectly maintained by HP, and hpijs is a binary sub-package of hplip.
[10:15] <tkamppeter> So both have to stay.
[10:16] <evand> great, thanks!
[10:17] <lifeless> mvo: :( it still turns up, or do I need to kill a process?
[10:18] <mvo> lifeless: yeah, killall update-notifier and start it again
[10:18] <tkamppeter> evand: foomatic-db recommends hpijs and hplip recommends hplip-cups. This should assure that the HPLIP printer drivers get installed.
[10:19] <evand> ah, noted
[11:30] <Unggnu> hi all
[11:30] <Unggnu> I like the Netbook edition very much but it lacks one thing. Is it somehow possible to use a background image? I would like to use this version for my mothers laptop but she always has an image of the whole family is background which she would probably miss.
[11:31] <Unggnu> So that the netbook menu is transparent or something like that.
[11:36] <Unggnu> The netbook menu application also doesn't restart if it crashes in Karmic.
[11:39] <glatzor> mvo, hello. the new install|broken|delete_count are broken in the latest python-apt
[11:41] <mvo> glatzor: fixed in my tree, thanks
[11:42] <glatzor> mvo, will you also merge revno 304?
[11:47] <glatzor> mvo, my last uses of the internal _depcache object are for the problem resolver/action group factory, IsInstBroken and IsAutoInstalled
[11:48] <glatzor> mvo, for actiongroup jak already implemented the context manager
[11:55] <mvo> glatzor: yes, I can merge that as well
[11:57] <glatzor> mvo, see you i Have to change trains!
[12:00] <Riddell> tkamppeter: CD builds are broken with  foomatic-db: Recommends: cjet but it is not installable  any fix on the way?
[12:01] <slangasek> ccheney`: something appears to have regressed in openoffice.org-l10n since alpha4; openoffice.org-*-en-{gb,za} are being pulled onto the CD again
[12:03]  * mvo waves to glatzor
[12:06] <mvo> hey slangasek!
[12:06]  * slangasek waves
[12:47] <ogra> StevenK, there is a linux-image-imx51 and friends in binary NEW, could you punch that so it gets published ?
[12:50] <ogra> or al-maisan_ ^^ (either of you is archive admin of the day it seems :) )
[13:10] <al-maisan> ogra: I am still a learner :) and was advised to stay away from the NEW queue initially :)
[13:10] <ogra> oh, even binary ?
[13:10] <ogra> i'll wait for someone else then
[13:36] <al-maisan> ogra: I ran lintian on linux-image-imx51 and friends; please see http://pastebin.com/df22969e
[13:36] <al-maisan> ogra: is that what you'd expect to see?
[13:37] <slangasek> al-maisan: well, the 'unknown-architecture' warning is self-evidently spurious :)
[13:37] <ogra> heh, i have never run lintian on a kernel package from the kernel team, so i dont know
[13:37] <al-maisan> aha :)
[13:37] <ogra> linux-image-2.6.31-100-imx51: no-copyright-file is a bit funny
[13:38] <slangasek> "funny" - in that if it's true, it makes the package undistributable
[13:38] <slangasek> and it's true :(
[13:38] <ogra> well, its using the std kernel team scripts .... weird
[13:39] <ogra> i would exepect the outcome to be the same as other kernel packages
[13:39] <al-maisan> so, what now? should the binary be rejected?
[13:41] <slangasek> al-maisan: yes, unfortunately
[13:41]  * amitk al-maisan 
[13:41] <amitk> oops
[13:41] <al-maisan> :)
[13:41] <ogra> al-maisan, wonderful catch though :)
[13:42] <al-maisan> ogra: I am just a beginner as I said ..
[13:43]  * al-maisan did what the https://wiki.ubuntu.com/ArchiveAdministration page said ;)
[13:43] <ogra> and you did well :)
[13:46]  * ogra tickles the publisher ... come on do your ports mirroring
[13:52] <liw> mvo, could you have quick peek at #416362 and the attachment at http://launchpadlibrarian.net/30571823/log ? " __init__() got an unexpected keyword argument 'deptype'" in /usr/lib/python2.6/dist-packages/apt/package.py looks like a python-apt problem to me
[13:53] <mvo> liw: yes, my error - I upload a new python-apt now
[13:53] <liw> mvo, shall I reassign this to python-apt?
[13:54] <mvo> liw: yes
[13:55] <liw> mvo, done, thanks
[13:55]  * liw wnats to be as effective as mvo (and joeyh) at fixing bugs, when he grows up
[13:57] <mvo> liw: as long as you don't become as efficient at adding them ;)
[13:57] <mvo> liw: sorry for the trouble
[14:03] <liw> mvo, no worries, it was very little trouble
[14:18]  * liw sighs at comment changes in conffiles
[14:18] <beuno> do I still have time to update a package for karmic?
[14:18] <ogra> the system should really learn to ignore comments
[14:19] <ion> liw: In reality, you sigh at how badly dpkg handles (read: does not handle) the merging of conffile changes. :-P
[14:19] <liw> ogra, that's practical as soon as we standardize on one configuration file format
[14:19] <ion> ogra: No, it should just do a standard three-way merge.
[14:19] <liw> beuno, feature freeze is only next week
[14:19] <ogra> ion, blindly ? ouch
[14:19] <ion> ogra: Of course not blindly.
[14:20] <beuno> liw, cool, thanks. I will make a release of loggerhead today then, so we're not stuck with 1.10 for karmic
[14:20] <liw> ion, please don't tell me what I sigh at, thanks; until dpkg can be improved in this, it is silly of the maintainer to make comment-only changes in conffiles
[14:20] <ogra> well, thats what it does, no ? it asks you to review the changes
[14:20] <ogra> though in comments thats really pointless
[14:21] <ogra> having something like COMMENTCHAR that ucf can read and you can set on a per package base should make it ognore such commented lines completely imho
[14:21] <ogra> *ignore
[14:30] <tkamppeter> Riddell, what is the exact problem with cjet? Is the problem that it is in Universe or are dependencies not fulfilled?
[14:31] <Riddell> tkamppeter: it's in universe
[14:33] <tkamppeter> Riddell, so it would already help if I move it to Suggests?
[14:33] <Riddell> tkamppeter: yes that should sort it
[14:37] <blackxored> hello, what's the best way to use the debian-policy, debootstrap, build-essential and the like from debian unstable in ubuntu?
[14:48] <mterry> Hey, slangasek, I wanted to talk to you about what exactly feature freeze means for some work I'm doing on ubiquity.  Ping me back when you get a chance
[14:49] <slangasek> mterry: this is a contentless pong :-P
[14:50] <mterry> slangasek, :)  OK, so I'm working on (A) having ubiquity support plugin pages (drop a python script in a dir and it gets picked up) and (B) porting it internally to use plugins.  I suspect this triggers FF concerns because of the public API involved?
[14:50] <mterry> slangasek, even though it's a new API
[14:50] <mterry> slangasek, FF would only care about point (A), right?  Not (B)
[14:52] <slangasek> mterry: mm, I would consider both of those to be covered by FF
[14:52] <slangasek> I would define "features" in this case as "things that aren't bugfixes", so code reorg qualifies :)
[14:53] <mterry> slangasek, ah, k
[15:01] <tkamppeter> Riddell: Fixed foomatic-db uploaded.
[15:03] <Riddell> tkamppeter: great thanks
[15:04] <Riddell> blackxored: you should be able to debootstrap a debian chroot
[15:04] <Riddell> mterry: you saw that gwt got accepted?
[15:05] <blackxored> Riddell, no, that wasn't what I asked I need to work with the latest debian-policy, the latest debootstrap, I mean, only a subset of packages, my lvm chroot is for building
[15:05] <mterry> Riddell, no?  gwt?
[15:05] <ogra> blackxored, you asked for the best way, Riddell told you "inside a chroot"
[15:06] <blackxored> ogra, truly but I know there was a way of pinning or assigning priorities to specific packages from different repos I just don't remember
[15:08] <Riddell> mterry: I'm getting my terry's confused, I want ttx
[15:08] <mterry> Riddell, :)  np
[15:08] <ttx> Riddell: Thierry :)
[15:08] <ttx> Riddell: Yes, I saw that.
[15:10] <ttx> Riddell: still stuck in binaryNEW though.
[15:13] <Riddell> ttx: I'm sure StevenK or al-maisan will get to it shortly, it's their archive admin day
[15:15] <al-maisan> ttx: sorry .. what is the binary/NEW upload you're waiting for?
[15:16] <ttx> al-maisan: gwt and eucalyptus-commons-ext. No hurry though...
[15:16] <al-maisan> hmm .. right.
[15:27] <dholbach> tkamppeter: can you replace ubuntu-devel@lists.ubuntu.com with ubuntu-devel-discuss@lists.ubuntu.com in one of your next foomatic-db uploads?
[15:27] <dholbach> (in debian/control)
[15:28] <dholbach> I get a mail from archive.ubuntu.com in the moderation queue every time :)
[15:36] <EagleScreen> the default init boot parameter in a Debian system is "init=/sbin/init", which is the default in Ubuntu?
[15:37] <ion> init=/sbin/init
[15:37] <EagleScreen> in Ubuntu /sbin/init is not an executable
[15:37] <ogra> huh ?
[15:37] <ion> Then your system is majorly broken.
[15:37] <ccheney`> slangasek: ok will check it out
[15:39] <EagleScreen> /sbin/init is a shared library in Ubuntu, but an executable in Debian
[15:40]  * ogra wonders what EagleScreen looks at ... definately not ubuntu
[15:40] <soren> EagleScreen: "md5sum /sbin/init" please.
[15:41] <EagleScreen> 591e0d59e1355aeffbf57ffbb4581007  init
[15:41] <soren> Is this in Karmic?
[15:42] <EagleScreen> is in juanty, but it is isntalled using debootstrap + tasksel
[15:42] <ccheney`> amd64 jaunty init
[15:42] <EagleScreen> yes, amd64
[15:42] <soren> EagleScreen: What makes you say it's a library?
[15:43] <slangasek> that's the correct md5sum for the amd64/jaunty upstart /sbin/init
[15:44] <slangasek> which is not a library
[15:44] <EagleScreen> Dolphin says: Shared library, and put it an icon with 1's and 0's, but in a Debian init, dolphin Says: executable, and put it an icon of a metalic gear
[15:44] <ogra> sounds like a dolphin bug
[15:45] <EagleScreen> then should i can use root=/sbin/init for Ubuntu?
[15:45] <ogra> or whatever tells dolphin about filetypes
[15:45] <james_w> the magic is "shared object", not "executable"
[15:45] <james_w> but it doesn't make it invalid
[15:45] <ogra> that would try to use /sbin/init as your root device
[15:46] <virtuald> $ file /sbin/init
[15:46] <virtuald> /sbin/init: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV), dynamically linked (uses shared libs), for GNU/Linux 2.6.15, stripped
[15:46] <ogra> what exactly do you try to achieve ?
[15:46] <james_w> EagleScreen: you should use init=/sbin/init, but I think you don't need to, as it's the default
[15:46] <ogra> right
[15:46] <ogra> you dont need to
[15:46] <ccheney`> the only difference i see from file output of that /sbin/init (not virtuald's) is that it is build for 2.6.8 support vs 2.6.15 of other things
[15:46] <EagleScreen> i try to boot with grub a debootstrapped root tree
[15:47] <ogra> well, you definately dont need init= at all
[15:47] <EagleScreen> kernel developpers suggested me that y cna try playing with init parameter
[15:47] <ccheney`> EagleScreen: do you have a kernel installed? debootstrap doesn't install one by default
[15:48] <ccheney`> EagleScreen: it also doesn't setup fstab iirc
[15:48] <EagleScreen> yes i have a kernel installed, and also ubuntu-minimal, ubuntu-standard and a couple of configuration files ready
[15:48] <tkamppeter> dholbach: Updated foomatic-db.
[15:49] <slangasek> EagleScreen: this is really far out of scope for this channel; you should ask these questions on #ubuntu
[15:49] <dholbach> tkamppeter: gracias!
[15:49] <slangasek> (and preferably, start your question with an error message showing what's wrong instead of speculation about init= options...)
[15:49] <EagleScreen> yes, i know, but thay drop me here lol
[15:50] <EagleScreen> okay, thanks, you have given em an idea
[15:50] <jerrcs> nobody really helps in #ubuntu..
[15:51] <jerrcs> If it's a more technical issue it's very hard to get a reply.
[15:51] <jerrcs> I've been asking my question about I/O errors in dmesg for hours now.. I tried last night too.
[15:54] <ccheney`> slangasek: looking into why openoffice.org-l10n-en-gb is installed and the only things that rdepend on it at all are openoffice.org-help-en-gb and language-support-translations-en and for openoffice.org-help-en-gb it is openoffice.org-l10n-en-gb and language-support-translations-en, so now idea at all how it is getting pulled onto the cd
[15:55] <slangasek> ccheney`: ok, perhaps it's related to OOo being out-of-date; let me check
[15:55] <ccheney`> slangasek: ok
[15:55] <ccheney`> slangasek: its all in the archive now, OOo itself had some new pieces that just got accepted today i think
[15:55] <slangasek> ccheney`: I know, I accepted them. :)
[15:55] <ccheney`> slangasek: ah, thanks! :-)
[15:56] <ccheney`> slangasek: er language-support-translations-en specifically depends on that stuff, not even just recommends
[15:56] <ccheney`> slangasek: is that package being pulled into the cd?
[15:56] <slangasek> yes
[15:57] <slangasek> and OOo may be the reason for that
[15:57] <ccheney`> er?
[15:57] <ccheney`> the package is from july 15 2009
[15:57] <ccheney`> so nothing new about it that i can see
[15:58] <slangasek> I mean, OOo being out-of-date may have been the reason for that
[15:58] <ccheney`> oh ok
[15:58] <ccheney`> on my system language-support-translations-en isn't installed, so maybe it normally isn't
[15:59]  * ogra wishes he even had OOo
[15:59] <slangasek> yes, it normally isn't
[15:59] <ccheney`> ok
[15:59] <ogra> silly slow publisher on ports :(
[16:01] <ccheney`> anyone know if ia64 buildd is going to get more space soon? OOo ftbfs on it due to not enough space to unpack
[16:01] <ogra> NCommander might know ... he often cares for the exots :)
[16:01] <ccheney`> oh yea
[16:01] <ccheney`> NCommander: ping
[16:15] <slangasek> ccheney`: yep, getting OOo back in has brought the alternate CDs back down in size, cheers
[16:15] <slangasek> now to find out why exim4 is on the CD :P
[16:19] <slangasek> hmm, 'at'
[16:21]  * ccheney` is on the phone with the bank, apparently he got hit by that 130Mil account credit card theft
[16:21] <liw> does anything require at to be in the default install? few people seem to use it, and they could add it separately
[16:21] <liw> ccheney`, ouch, good luck
[16:22] <slangasek> liw: Unix pride? :)
[16:22] <slangasek> the real bug here is an at merge that turned Suggests: postfix | m-t-a into Recommends: exim4 | m-t-a
[16:23] <liw> slangasek, I'm not sure pride is sufficient justification ;-)
[16:23] <slangasek> in fact, this was a bad merge, not merely an oversight
[16:25] <lamont> ccheney`: happen to know which buildd that was?
[16:25] <lamont> I'll go look at both though
[16:26]  * ogra starts getting angry with the publisher ... OOo built 3h ago and still doesnt show up on ports ...
[16:26] <slangasek> ogra: which arch?
[16:27] <slangasek> ogra: if armel, that's because it's in NEW
[16:27] <ogra> gah
[16:27] <slangasek> and now it's out of new
[16:27] <ogra> i should have looked there first :P
[16:27] <ogra> instead of hitting reload on the archive dir for 3h
[16:28] <liw> what we have here is a failure to communic... complain loudly enough, early enough, that someone with the answer notices
[16:29] <ogra> liw, well, i know actually that it takes about 2h
[16:29] <slangasek> it's not supposed to take 2h.
[16:29] <ogra> it does on ports
[16:29] <liw> ogra, actually, in an ideal world, you could have told some system somewhere that you're interested in this and it would've told you what's up
[16:29] <slangasek> then you should be talking to IS :)
[16:30] <ogra> well, when i complained during A4 preparation you said it would be a known issue that its slow
[16:30] <ogra> so i kept quiet
[16:30] <slangasek> the publisher currently takes only ~30m to run (thanks, cprov!); that means the average time from binary accept to mirroring is 1h
[16:31] <slangasek> and the mirroring should take only a fraction of that
[16:31] <slangasek> ah, yes, it was a known issue during A4 that was being actively addressed:)
[16:32] <ogra> well, i'll check if its actually true with one of my packages then before i complain, i guess i cant count OO.o now after it was stuck on NEW
[16:32] <ogra> but good to know its supposed to be fixed
[16:33] <slangasek> you should be able to tell with OOo
[16:33] <mathiaz> when should a system account used by a daemon be removed? during the remove phase or the purge phase?
[16:34] <slangasek> it was accepted 7 minutes ago, so it'll hit the next publisher run in 29 minutes, and should be visible on ports about 45 minutes after that
[16:34] <slangasek> mathiaz: never :)
[16:35] <mathiaz> slangasek: so system account are created but never deleted?
[16:35] <liw> mathiaz, it is possible the account owns files (possibly on other systems via NFS but offline during package removal), which means many people (slangasek and I) think it should never be deleted automatically
[16:35] <slangasek> yep
[16:36] <liw> if someone were to want to create tools for managing system accounts (create upon install, remove at purge if local admin says it's OK), that would be much appreciated, thanks
[16:36] <liw> alternatively, someone might want to write a computer-janitor plugin ;-)
[16:38] <slangasek> liw: I think the Windows approach (never deallocate an RID once it's been allocated) is much better. :)
[16:38] <slangasek> simpler
[16:40] <mathiaz> Is it ok to remove user-added configuration files during a purge?
[16:40] <slangasek> gray area; IMHO yes
[16:40] <mathiaz> ie rm -rf /etc/config_dir/ where /etc/config_dir/ is part of the dirs package
[16:41] <liw> I wouldn't do that, but if it saves a lot of effort, I wouldn't object strongly
[16:41] <ion> The user should have backups if she doesn’t want purge to, well, purge. :-P
[16:42] <mathiaz> And data stored in /var/lib/daemon_data/ should be deleted ( rm -rf /var/lib/daemon_data) on remove?
[16:43] <liw> mathiaz, I don't think so
[16:43] <mathiaz> on purge?
[16:43] <liw> though I forget what the Policy says
[16:43] <liw> not even on purge
[16:44] <mathiaz> or never..
[16:44] <slangasek> I don't think Policy says about /var/lib/daemon_Data
[16:44] <slangasek> some packages ask
[16:44] <liw> it would be awkward to delete, say, a news spool just because you purge one NNTP server
[16:44] <slangasek> which is a poor proxy really, because there's always a chance that the question is asked far away from the point when you're doing the actual purge
[16:44] <mathiaz> in that use case if you'd reinstall the package you'd have the old data come back
[16:47] <slangasek> stgraber: does ltsp-server-standalone really need to Recommend: dhcp3-server?  This costs us 300k on the Ubuntu alternate CD
[16:48] <liw> mathiaz, hm, if re-installing package will restore data in /var/lib, why is the data there? what data is it?
[16:48] <mathiaz> liw: let's say a database
[16:48] <mathiaz> liw: on pkg install you need to initialize the database
[16:49] <mathiaz> liw: now if you re-install a package that had been purged you'd still have the old database lying around
[16:49] <liw> mathiaz, I think the relevant question is: does the /var/lib/foo directory contain user data that cannot be easily re-created; if so, don't remove it ever
[16:49] <liw> or that's what I would do (I hope)
[16:50] <liw> the preinst/postinst script can handle an old database (perhaps by renaming it)
[16:55] <mathiaz> And how about log files (/var/log/foo)? Should the directory be rm -rf on remove/purge?
[17:03] <slangasek> mathiaz: that one's covered in policy.. :)
[17:05] <mathiaz> slangasek: section 10.8 :)
[17:06] <slangasek> mathiaz: yep
[17:20]  * jtimberman thinks mathiaz is looking at stompserver's removal behavior!
[17:20] <jtimberman> ;)
[17:20] <mathiaz> jtimberman: yeah - more comments to come soon :)
[17:20] <jtimberman> mathiaz: already opening up my editor ;)
[17:33] <jtimberman> mathiaz: ah, simply looks like we don't need the postrm script at all then.
[17:34] <jtimberman> oh, rather, the purge action should remove /var/log/stompserver, and the other actions not do anything.
[17:34] <mathiaz> jtimberman: yop - and may be rm -f /etc/stompserver/
[17:34] <mathiaz> jtimberman: yop - and may be rm -rf /etc/stompserver/
[17:34] <mathiaz> jtimberman: on purge
[17:38]  * ogra hugs slangasek ... OO.o is there exactly how you predicted
[17:39] <jtimberman> updated
[17:53] <lyhana8> hi, I just install kubuntu 9.04 few days ago and my touchpad doesn't work anymore (worked on hardy), cat /dev/input/mouse* give nothing
[17:56] <jtimberman> mathiaz: that's looking better now. purging the package leaves /var/lib/stompserver alone if there's anything in it, otherwise removes it if empty. logrotate script is created properly (and removed) as well, of course.
[17:57] <mathiaz> jtimberman: yop
[17:57] <mathiaz> jtimberman: howeevr /var/lib/stompserver won't be deleted on a production system as it will contain data
[17:57] <jtimberman> mathiaz: right, if that option is set in the config file :)
[17:58] <jtimberman> with that patch included ;)
[17:58] <cdE|Woozy> is there some magic required to post to the ubuntu mailinglists? I've now posted three mails to two different mailinglists and apparently nothing ever appeared on either mailinglist =X
[17:59] <ScottK> slangasek: The last uploads for the boost 1.35 transition are done.
[17:59] <slangasek> ScottK: woot
[17:59] <ScottK> cdE|Woozy: Normally you have to be subscribed.
[17:59] <ScottK> slangasek: As an added bonus, so far they all build.
[18:01] <mathiaz> jtimberman: in the logrotate file: /etc/nit.d/stompserver force-reload
[18:02] <mathiaz> jtimberman: there is an i missing in there
[18:02] <ion> Rather: service stompserver force-reload, since it will make sure the init script’s environment won’t get polluted etc.
[18:03] <jtimberman> mathiaz: fixed and uploaded.
[18:09] <cdE|Woozy> whooho, it worked. thanks ScottK
[18:09] <jtimberman> ion: /etc/init.d/foo is mentioned in the policy so thats what i put..
[18:18] <lyhana8> anyone could explain me why my touchpad doesn't work anyomore after moving from hardy to jaunty ?
[18:30] <dholbach> Ubuntu Global Jam meeting in 30m in #ubuntu-meeting
[18:41] <mathiaz> slangasek: I've modified the slapd user creation to use /nonexistent as the home directory.
[18:41] <mathiaz> slangasek: does it make sense to have daemon user to have non existing home directory?
[18:41] <mathiaz> slangasek: or it would be better to keep the home directory of slapd to /var/lib/ldap?
[18:50] <cr3> what/who might be a good reference about udev? not writing rules, but where/how values are retrieved by udevadm
[20:03] <ogra> tkamppeter, does foomatic-db-hpijs need to be unseeded ? (ubuntu-desktop currently depends on it but foomatic-db conflicts with it)
[20:05] <Chase_> I've got a package I've developed and I'm interested in publishing it under the GPL
[20:06] <Chase_> I saw that the copyrights may be an issue
[20:06] <Chase_> I use Qt's QSslSocket, which uses openssl indirectly
[20:07] <ogra> oh, evand did unseed it already
[20:07] <Chase_> I had seen on the ubuntu wiki somewhere that GPL apps linked directly or indirectly to openssl would be flagged
[20:07] <Chase_> is that really an issue since I'm using the qt libs?
[20:09] <Chase_> Specifically, this is where I saw that issue: https://wiki.ubuntu.com/PackagingGuide/Basic#Copyright
[20:09] <Chase_> second to last bullet point of "Common errors:"
[20:17] <Usama> hello, there is outdated information on the Ubuntu web site
[20:17] <Usama> is this the place to report it?
[20:18] <mathiaz> Usama: https://bugs.launchpad.net/ubuntu-website/
[20:19] <mathiaz> Usama: ^^ this is place to report a bug related to the ubuntu web site
[20:21] <Usama> ok I'll post the problem here because I'might have no time to report it there http://www.ubuntu.com/getubuntu/downloadmirrors#dvd the links only shows hardy but if you went to higher level in the domain you will get the latest ftp://swtsrv.informatik.uni-mannheim.de/pub/linux/distributions/ubuntu-dvd-release
[20:50] <tkamppeter> ogra:  foomatic-db-hpijs must get unseeded, also hal-cups-utils
[20:51] <ogra> tkamppeter, yes, my fault, i missed that evand already unseeded it
[21:08] <Usama> https://bugs.launchpad.net/ubuntu-website/+bug/416622
[22:34] <lifeless> doko: hi
[23:20] <chordogg> hello all, there is a behavior in ubuntu that i dont care for, i was wondering if you knew how to change it. when i switch to say tty1, and log in. if i logout of that shell, ubuntu switches me back to tty7 when i really just want to stay in my current tty...i can't hunt down what controls this behavior, anyone know?
[23:28] <slangasek> mathiaz: I don't think we should use nonexistent directories as home dirs for users; something might create them later and give unexpected results
[23:37] <jono> any idea which package the Preferred Applications tool is in?
[23:39] <mathiaz> slangasek: right - I'm going to change that. /var/lib/ldap wil be created anyway as part of the slapd package installation
[23:39] <mathiaz> slangasek: the reason to use /nonexistent was that at first I though /var/lib/ldap/ wouldn't need to be created anymore
[23:40] <slangasek> ah
[23:41] <chrisccoulson> jono - gnome-control-center
[23:41] <jono> thanks chrisccoulson
[23:55] <stgraber> slangasek: yes, we need to have dhcp3-server. It used to be a hard depend but as some people are using dnsmasq instead, I moved it to a recommend.
[23:55] <stgraber> slangasek: if we don't have it, your thin clients won't be able to boot after install.
[23:56] <stgraber> (sorry for the late answer :))