[00:50] <asac> pease promote libeina-svn-06 to main ;) ... the source already is in main, but this one is now needed by evas ;)
[00:50] <asac> please
[00:52] <asac> james_w: ^^ do you have powers?
[00:53] <asac> thx
[01:18] <cjwatson> asac: done
[01:23] <ebroder> Is there a way for a program to catch an Upstart event (event, not job)? I want to have a program do something when an event comes from the upstart-udev-bridge
[02:26] <un214> What I'd like to know is who was the dope who thought that setting the ionice of a process to idle would keep it from interfering with the disk activity of other processes.
[02:26] <micahg> !coc | un214
[02:28] <ion> un214: If it doesn’t, the kernel should be fixed.
[02:28] <un214> it can't work on rotating disks
[02:28] <ion> Sure it can.
[02:29] <un214> seek time
[02:29] <ebroder> un214: That's why you have a buffer cache and an I/O scheduler
[02:30] <ion> If higher-priority processes do frequent disk IO, the grace time is never reached and the idle-priority process never gets to do its IO. Now, if that’s not what happens, the kernel should be fixed. Perhaps the grace time is too low.
[02:30] <un214> and replacing mountall with fsck -A -P && mount -a took minutes off boot time fsck
[02:30] <un214> mountall correctly set the second fsck ionice to idle and it didn't help
[02:32] <ion> I take it /sys/block/sda/queue/scheduler says cfq is active?
[02:32] <un214> yes
[02:34] <un214> oh duh
[02:34] <un214> the actual problem is mountall set both fscks to idle priority
[02:35] <un214> there's a third fsck on a second disk that actually runs first
[02:36] <ion> Ok, that should not happen. Please add --debug >/dev/mountall.log 2>&1 to the end of the exec line in /etc/init/mountall.conf, start with forcefsck and provide /dev/mountall.log, thanks.
[02:36] <un214> can't do it anymore, I ripped the whole boot setup out
[02:36] <ion> Then it’ll be difficult to fix the bug.
[02:37] <un214> it's immediately obvious where the bug is coming from
[02:37] <un214> mountall ignores the pass parameter in /etc/fstab
[02:38] <ion> Did i get this correctly: you have two physical disks, and *all* of the fscks for one of them had the idle iopriority while one or more fscks were running for the other disk?
[02:38] <eggonlea> Hi all, I have a question about how to install conflict configuration file. E.g. package A would install a default /etc/init/a.conf; but package B would enhance package A and therefore introduce a modified /etc/init/a.conf. What's the recommended way to do this?
[02:38] <un214> yes you did
[02:40] <eggonlea> Is there any way to hack package B only and leave package A untouched (because it's transparent to package A totally)?
[02:41] <ion> The debug log from mountall would make it possible to pinpoint where the problem is. Whenever you are able to provide one, please contact me again, so i can debug the issue.
[02:43] <RAOF> eggonlea: Policy forbids packages from modifying conf files owned by another packge.
[02:43] <ebroder> eggonlea: You could look at config-package-dev (http://debathena.mit.edu/config-package-dev/)
[02:44] <RAOF> eggonlea: There are two ways this is handled: (1) By having package A have a /etc/foo.d directory, where package B can drop a config snippet in, or (b) by having package A provide a programatic mechanism for modifying its config, and having package B use that.
[02:44] <ion> eggonlea: If /etc/init/b.conf has “start on starting a”, whenever a is going to be started, b starts first, blocking the startup of a until running.
[02:44] <ion> Dunno if that’s what you’re after.
[02:45] <un214> ion: it's safe to run mountall directly from a maintenance shell right?
[02:45] <eggonlea> RAOF: thanks! Would take a look at config-package-dev
[02:46] <eggonlea> ion: This should work on my case!
[02:46] <ion> un214: It’ll probably never notice the disks until udev is running, and udev should be only started after mountall has mounted the virtual filesystems.
[02:46] <ion> You could mount them manually, start udev and then start mountall. That *might* work. :-P
[03:59] <un214> ion: I uploaded log to https://bugs.launchpad.net/ubuntu/+source/mountall/+bug/599624 -- looks like ionice triggered but was ineffective
[04:23] <ion> un214: Ok, thanks. I’ll see what i can figure out.
[04:32] <getpwnam> what do people think about adding a speed limit/pause to the Synaptic download dialog? I was in the middle of a post-installation ~1GB download when my wife tried to use Skype, but Synaptic was hogging the network so the call quality was very poor. It would've been nice to have a download speed limiter so I wouldn't have to cancel..
[04:33] <ion> apt-cache show trickle perhaps
[07:14] <tjaalton> sigh, is it ext4 failing or why do some package installations on current lucid fail by complaining "corrupt tar", and the succeed the next time
[07:14] <tjaalton> happens at least with large in-house debs
[07:14] <pitti> Good morning
[07:16] <tjaalton> morning pitti
[07:16] <pitti> hey tjaalton
[07:20] <ajmitch> morning pitti
[07:20] <tjaalton> hmm, I see a lot of ext4 changes in the latest lucid kernel.. maybe that's why I could do a full install before but not anymore
[07:20] <tjaalton> on one go that is
[07:49] <bilalakhtar> dholbach: good morning
[07:49] <dholbach> good morning
[07:49] <dholbach> hey bilalakhtar
[07:49]  * bilalakhtar wished before dholbach could
[08:17] <dholbach> pitti: oops, I should've read emails before doing sponsoring :)
[08:18] <pitti> hey dholbach
[08:18] <pitti> dholbach: did you break everything? :-)
[08:18] <dholbach> no, just upload a couple of long-waiting main sponsoring requests
[08:18] <dholbach> if there's blame, I'll decide to ignore it though :)
[08:19] <dholbach> (and for today I'm done sponsoring)
[08:33] <ajmitch> dholbach: soft freeze means you only get hit softly when you upload stuff, right?
[08:37] <dholbach> ajmitch: not sure if the release team really dares to do that :-P
[08:44] <directhex> time to upload mono 2.6.7!
[11:43] <dholbach> pitti: do you have any idea why a source package maybe have been accepted but binary packages rejected?
[11:43] <dholbach> https://launchpad.net/ubuntu/maverick/+source/mtdev/1.0.6-0ubuntu1
[11:44] <pitti> dholbach: I mailed Chase
[11:44] <dholbach> I'm not sure I understand this :)
[11:44] <dholbach> ah
[11:44] <pitti> dholbach: the library -dev shipped a .la, and I don't want this to infect the archive
[11:44] <pitti> so I asked him to remove the .la and reupload
[11:44] <dholbach> ara: ^
[11:45] <dholbach> ara: if you prepare a branch for the change I'll test and sponsor it
[11:48] <ara> dholbach, ok, will do later, thanks!
[11:48] <dholbach> rock!
[11:59] <asac> cjwatson: i would like to remove packages in a downstream seed. is that possible with some magic atm? like adding a "!xorg" to a linaro-xxx seed still using desktop-common to get rid of that part?
[12:42] <dupondje> mmm, xterm removed libutemper again :(
[12:42] <dupondje> https://bugs.launchpad.net/ubuntu/+source/libutempter/+bug/589103
[12:48] <alkisg> Hi, I'm trying to upload fixed tuxpaint/tuxtype packages to lucid-proposed, is there a wiki page that describes that process?
[12:49] <alkisg> (trying to get an SRU for bug #572994 and bug #512285)
[12:51] <geser> the SRU wiki page describes the process but beyond this uploading to -proposed is nothing special and works like any other upload
[12:51] <alkisg> OK, so I'll just modify my ~/.dput.cf and upload. Thanks!
[12:52] <geser> why modify?
[12:52] <tumbleweed> alkisg: err, you shouldn't need any modifications
[12:52] <alkisg> To add a [lucid-proposed] stanza there, just to make the dput command line easier :)
[12:53] <geser> you need lucid-proposed in the changelog and upload to "ubuntu" as usual
[12:53] <alkisg> "upload to "ubuntu" as usual" ==> sorry I've never done that so it's not usual for me :)
[12:53] <alkisg> I've uploaded to my ppa, but nowhere else
[12:54] <tumbleweed> alkisg: do you have the rights to upload this then?
[12:54] <alkisg> tumbleweed: I believe so, I'm in edubuntu council which should indirectly put me in ubuntu-developers
[12:55] <Laney> Is the edubuntu council supposed to confer developer rights?
[12:56] <alkisg> I'm not sure. I'm also an edubuntu-developer, if that makes a difference.
[12:56] <tumbleweed> alkisg: you need to be MOTU to upload arbitrary universe packages, if the package is in the edubuntu packageset then you can probably upload
[12:56] <Riddell> ev, rgreening: how would you feel about a usb-creator upload?  I just fixed the kde frontend which had an issue that stopped it starting with the new PyQt
[12:56] <tumbleweed> (easiest way to tell is to try) :)
[12:57] <alkisg> Thanks tumbleweed, /me tries to find some wiki page about "uploading to ubuntu as usual" :)
[12:57] <tumbleweed> alkisg: dput ubuntu foo.changes
[12:57] <alkisg> Thanks
[12:58] <Laney> alkisg: Have you had sponsored uploads before?
[12:58] <alkisg> Laney, no
[12:58] <alkisg> (this particular package requires no source changes, just a version bump)
[12:58] <rgreening> Riddell: is there a bug/patch? ev does releases, but I can commit to bzr and bug/poke him to release
[12:58] <Laney> hmmmmm
[12:59] <Riddell> rgreening: I've already committed to bzr
[12:59] <rgreening> k
[13:19] <cjwatson> asac: sorry, it is not.  The blacklist syntax ("!") is a much bigger hammer that has effects you don't want - see germinate(1) for discussion
[13:25] <asac> cjwatson: so what can i do to achieve this? is only way to fork desktop-common and modify and then regularly sync?
[13:27] <cjwatson> asac: the upstream seed needs to be refactored
[13:27] <cjwatson> (or else you need to fork)
[13:28] <ivoks> i would appriciate guidance in one problem; is predepend justified if i have a package that requires mysql up and running before that package is configured? is there another way out (triggers or something)?
[13:28] <cjwatson> depends is sufficient to have something happen before the package is configured
[13:29] <ivoks> but depending on mysql doesn't guarantee that package would be able to populate database
[13:29] <ivoks> mysql could be stopped during package's configure stage
[13:29] <cjwatson> pre-depends won't help in the least
[13:30] <ivoks> ok, is there other way out?
[13:30] <cjwatson> don't know, look for other packages that depend on mysql for this kind of thing maybe
[13:31] <ivoks> ok, thanks
[13:41] <ivoks> well, everything else is broken too :)
[13:42] <ivoks> 'broken'
[13:44] <happyaron> kees: ping
[13:47] <LukeD_> hi guys, I have a question about building gcc on x86_64, can anyone point me to the right ubuntu irc to ask about it?
[14:01] <zul> can someone from the foundations team review the upstart scripts in bug #612975 and bug #612958
[14:07] <geser> ivoks: wouldn't that require that the app and the mysql db are on the same host?
[14:07] <ivoks> geser: it would
[14:08] <ivoks> geser: which isn't that bad since dbconfig-common with low priority assumes localhost
[14:08] <ivoks> er... high priority
[14:09] <ivoks> having mysql triggers which would be populated by dbconfig-common sounds like a solution
[14:10] <geser> ivoks: wouldn't moving e.g. mysql-client to recommends in dbconfig-common fix it too? make mysql available by default but leave the opportunity to have a remote db and no need to have mysql installed just for the dependency
[14:12] <ivoks> but if you want db on localhost, dbconfig will fail if mysql-server isn't installed
[14:12] <ivoks> dbconfig really depends on mysql-client cause it needs it to populate database :)
[14:15] <geser> I'm not sure if recommends are handled the same in installation order but recommending mysql-server in dbconfig-common doesn't work?
[14:16] <ivoks> haven't tried it
[14:20] <geser> IIRC depends should be enough for your case as IIRC the dependencies of a package get configured before the package itself
[14:20] <geser> and what should happen if mysql is already installed but the admin stopped it?
[14:23] <ivoks> same thing
[14:23] <ivoks> depends is not enough
[14:23] <ivoks> you get package installed and configured before mysql is configured
[14:24] <ivoks> and if it's a service, it won't start and just die on postinstall
[14:24] <ivoks> anyway, i'll think more about it later... take care
[14:39] <smoser> cjwatson, are you around ?
[15:03] <dpm> hi pitti, I believe you are the maintainer of langpack-locales. May I ask you to have a look at bug 124987? There is a patch available, but I'm not sure it can be used as is
[15:03] <dpm> (I know they should eventually send the patch upstream, though)
[15:29] <Damascene> hi, when is the meeting
[15:36] <ScottK> Who is our "d-shlibs maintainer"? http://launchpadlibrarian.net/52996671/buildlog_ubuntu-maverick-armel.uw-imap_8:2007e~dfsg-3.1ubuntu1_FAILEDTOBUILD.txt.gz
[15:45] <cjwatson> smoser: what's up?
[15:45] <Chipzz> zul:
[15:45] <Chipzz> +[ -f /etc/default/dhcp3-relay ] && ./etc/default/dhcp3-rlay
[15:45] <smoser> would you be opposed to carrying a 'mbchk' in grub2? I need a reliable way to determine "is this a multiboot file", and I couldn't find any way to do so with grub2, but grub1's mbchk does exactly that.  I could write something, but don't really see the point in copying.
[15:45] <Chipzz> typo
[15:45] <zul> Chipzz: thanks
[15:45] <Chipzz> you're missing a space after the .
[15:46] <Chipzz> and a second one
[15:46] <Chipzz> rlay instead of relay
[15:46] <Chipzz> also
[15:47] <Chipzz> I see no upstart script for the version built with ldap support
[15:49] <cjwatson> smoser: I'm opposed to carrying it specifically in Ubuntu, but perhaps you could get in touch with upstream about it?
[15:49] <cjwatson> smoser: I don't imagine anyone would object to porting it over (maybe as grub-mbchk)
[15:49] <smoser> upstream as in grub? or debian? i'm fine with bringing it up to either.
[15:49] <cjwatson> smoser: upstream upstream
[15:50] <smoser> yeah. ok.
[16:05] <rsavoye> did something happen with OpenJDK packaging in the last 24 hours ? Icedtea can't find the source package anymore
[16:06] <pitti> dpm: can do
[16:06] <dpm> pitti, awesome, thanks!
[16:24] <ara> pitti, I am removing the .la from the libmtdev-dev package, but I would like to take the opportunity to learn more about the -dev packages, .la files and how they are considered harmful
[16:24] <ara> pitti, where can I find information about it?
[16:35] <Keybuk> mvo: why would update-manager hang when installing updates?
[16:36] <seb128> Keybuk, hang or being very slow and use cpu?
[16:36] <Keybuk> seems to freeze entirely after clicking "Install"
[16:37] <Keybuk> (not my computer, family member's)
[16:37] <seb128> ok, dunno then
[16:37] <Keybuk> no, me neither
[16:43] <mvo> Keybuk: might be a policykit issue? does ps afx show anything interessting?
[16:44] <Keybuk> mvo: no idea, I'm not entirely sure I can get mother to open a terminal
[16:45] <mvo> Keybuk: *ick* is that lucid?
[16:45] <Keybuk> yeah
[16:48] <alkisg> Keybuk: vnc can be handy in such cases :) x11vnc -connect <your-ip> to bypass NATs, firewalls etc.
[16:49] <Keybuk> is that something she'd run that'd give me her screen?
[17:01] <pitti> ara: they usually specify too many dependencies, so that auto-generated Depends: fields of packages carry way too many dependencies; this is a problem for library transitions
[17:01] <pitti> ara: and those .la files are generally not needed on Linux; it's just a PITA to get rid of them
[17:01] <ara> pitti, thanks for the explanation :-)
[17:02] <Keybuk> pitti: *disagree*
[17:02] <rsavoye> No package 'mozilla-plugin' found
[17:02] <rsavoye> did something change yesterday ? I'm getting tons of new failures with packages trying to build icedtea and OpenJDK
[17:10] <chrisccoulson> rsavoye, it's probably a silly question, but do you have xulrunner-dev installed? that provides mozilla-plugin.pc
[17:11] <rsavoye> arg.... look slike build-dep missed some things...
[17:11] <rsavoye> I'm also having problems (on maverick) finding the OpenJDK source
[17:12] <rsavoye> I guess this is what I get by doing a full reinstall on my beagleboard...
[17:13] <ion> keybuk: Bug #458299 perhaps.
[17:13] <chrisccoulson> mvo - my girlfriend sees the "update-manager hanging when pressing the install button" on lucid too
[17:13] <chrisccoulson> but it only happens on her account (it works fine on mine)
[17:14] <chrisccoulson> and it only happens when u-m is autospawned from u-n (if i open u-m from the terminal, then it works fine)
[17:14] <ion> keybuk: That one can make the system slow enough to be utterly unusable whenever something triggers the reload of apparmor profiles when upgrading.
[17:15] <mvo> chrisccoulson: oh, i know - no policykit auth agent running?
[17:15] <mvo> ^--- seb128
[17:15] <chrisccoulson> mvo - that shouldn't matter for lucid though should it? (it's still using gksu)
[17:15] <mvo> seb128: for what users is the polkit auth agent started?
[17:15] <ion> keybuk: Oh, it seems a fix has been uploaded. I’m not sure whether i’ve encountered the bug since that or not.
[17:15] <mvo> chrisccoulson: hrm, right
[17:16] <chrisccoulson> mvo - the agent is started for all users
[17:16] <mvo> chrisccoulson: ok, that was not the one then. she is in the admin group, right?
[17:16] <seb128> mvo, right, the issue is only for the maverick version I think
[17:16] <mvo> seb128: yeah, sorry. I had a similar issue in mav
[17:16] <seb128> mvo, and it doesn't hang just goes back to the ui as if you hadn't clicked
[17:16] <chrisccoulson> mvo - yeah, she's in the admin group (but she wasn't originally)
[17:16] <mvo> chrisccoulson: hm, hm, if its reproducable, can we debug it together at some point (does not have to be now)?
[17:17] <chrisccoulson> mvo - yeah, sure. i'll log in to her account at some point and see if i can get it to happen again
[17:18] <chrisccoulson> i keep asking her why she's not updating the computer, and she keeps telling me it's because "It doesn't do anything when i press the install button" ;)
[17:18] <mvo> chrisccoulson: please, maybe we can even track it down and fix for 10.04.1
[17:18] <mvo> chrisccoulson: just a ps afx dump will be helpful for a start
[17:19] <ion> keybuk: In any case, see if she has an “apparmor_parser: page allocation failure” line in syslog.
[17:20] <ion> keybuk: Oh, i misread your initial line, thinking her whole system hangs.
[17:21] <dholbach> pitti: the mtdev .la fix was uploaded :)
[17:21] <Keybuk> no, just update-manager
[17:24] <mvo> Keybuk: just update-manager in ps afx ? or what does the "just" referes to?
[17:25] <Keybuk> just update-manager hangs
[17:25] <Keybuk> haven't got a ps :-/
[17:25] <Keybuk> she's gone out now :p
[17:26] <mvo> Keybuk: hrm, thats bad, so everything else works fine. did it happen the first time for her?
[17:26] <chrisccoulson> mvo - ok, i can't get it to reproduce when i actually want it too ;)
[17:27] <mvo> chrisccoulson: the best kind of bugs :(
[18:01] <rsavoye> openjdk-6-source has a package dependency problem
[18:02] <rsavoye> penjdk-6-source : Depends: openjdk-6-jre (>= 6b20~pre2-0ubuntu2) but 6b18-1.8-2ubuntu2 is to be installed
[18:03] <micahg> rsavoye: what chipset do you have?
[18:05] <rsavoye> this is on an XM board, running maverick
[18:05] <micahg> rsavoye: is that armel?
[18:06] <rsavoye> yeah
[18:06] <micahg> rsavoye: armel's currently broke for openjdk
[18:06] <rsavoye> I know, I'm trying to fix it...
[18:07] <rsavoye> I had been running a patched kernel, then last night installed ubuntu-maverick-alpha2-minimal-armel.tar.7z
[18:07] <rsavoye> major issues with build-dep and icedtea6-plugin
[18:07] <apw> Riddell, i hear you are archive admin of the day ... we have some packages which are in need of removing and seem to be hanging about: details at the end of bug #595949
[18:07] <rsavoye> which was working yesterday. COurse maybe I screwed something up too
[18:07] <micahg> rsavoye: so, you probably want to pull down the .deb source
[18:07] <micahg> rsavoye: apt-get source openjdk-6
[18:07] <rsavoye> I did , it installed, but I get other erro5rs
[18:08] <rsavoye> E: Unable to find a source package for openjdk-6
[18:08] <Riddell> apw: ok will do
[18:08] <rsavoye> which was there yesterday...
[18:08] <micahg> rsavoye: you need to enable the deb-src repo in sources.list
[18:10] <rsavoye> I did, but maybe got the URL wrong
[18:10] <rsavoye> I just grabbed the deb-src from my ia32 maverick machine, which used to work...
[18:11] <micahg> rsavoye: you can dget this: http://archive.ubuntu.com/ubuntu/pool/main/o/openjdk-6/openjdk-6_6b20~pre2-0ubuntu2.dsc
[18:16] <rsavoye> what's weirder is I build icedtea with --with-openjdk-src-dir=, so who cares about the source package ?
[18:16] <rsavoye> I'll dig into it, I must be doing something wrong on my end
[18:17] <rsavoye> the problem being apt-get build-dep is broke, so I have to install them by hand
[18:17] <rsavoye> unfortunately, all the package error messages are for Fedora 13...
[18:18] <tantiv> For some reason only partimage-i386 (not amd64) is in the Ubuntu respository.... two years ago there was an amd64 bug which is probably why it was never put into amd64.  However the most recent version (including the source in the Ubuntu repository) fully supports amd64 and I tested it after running dpkg-buildpackage myself.  Who should I go to, to get it released into the amd64 repository?
[18:39] <rsavoye> my problem turned out to be a bad deb-src URL, now build-dep works again
[20:28] <seb128> StevenK, can you set https://code.edge.launchpad.net/~bluetooth/ubuntu/lucid/obexd/main/+merge/24151 to merged? it seems only teams members can do that
[20:37] <pitti> ogra: if you are still online, new efl launcher could be uploaded in 15 mins
[20:56] <ogra_cmpc> pitti, i hope asac sees that :P
[21:13] <pitti> ogra_cmpc: no, he's already offline
[21:14] <ogra_cmpc> pitti, i know, but he has everything prepared
[21:14] <ogra_cmpc> i dont even have the branch here
[21:48] <seb128> ev: hi, could you get bug #538744 review? not sure if the bug or change still make sense but would be nice to clear it from the sponsoring queue
[21:51] <seb128> ev: same for bug #546445 if you can
[21:51] <ev> seb128: responded
[21:51] <ev> okay
[21:51] <seb128> ev: thanks!
[23:16] <SpamapS> mathiaz: v2 of rubygems draft sent.
[23:25] <mathiaz> SpamapS: gotcha
[23:26] <SpamapS> mathiaz: still feels a bit long, but I think removing a lot of the FHS quotation helped narrow the focus