[00:22] <RAOF> Is there a work-around for the sbuild breakage?
[00:23] <ScottK> Use pbuilder?
[00:23] <ScottK> AFAIK pbuilder works, you just can't create a brand new chroot now.
[00:24] <ScottK> (You could be able to make one in natty and upgrade it)
[00:25] <RAOF> I think I'll defer that work until sbuild has been broken for longer :)
[00:26] <RAOF> Oh, hey.  Look!  A natty-proposed queue.
[00:32] <TheMuso> RAOF: Yeah hit the same thing myself.
[00:32] <RAOF> Yeah, I saw.  Which is why I'd hoped someone mighmt have magically fixed it for me :)
[00:49] <pook1e> can anyone here help me with a pbuilder dependency?
[00:55] <lifeless> !ask | pook1e
[01:32] <pook1e> i'm trying to build indicator-network with pbuilder, but it's telling me pbuilder-statisfydepends-dummy depends on libnotify-dev (>= 0.7.0), but i have 0.5.0-2ubuntu1 installed, and so pbuilder-satisfydepends fails
[01:32] <pook1e> does anyone know how i can fix this?
[01:33] <pook1e> ls
[01:34] <mwhudson> pook1e: which series are you trying to build for?
[01:34] <micahg> that's libnotify from natty
[01:34] <micahg> or maverick for that matter
[01:34] <pook1e> how can i get the newest version?
[01:35] <pook1e> i used --extrapackages libnotify-dev when i was creating
[01:35] <TheMuso> t/c
[01:36] <RAOF> pook1e: Unless you're building for oneiric the version of libnotify-dev is insufficient to satisfy that dependency.
[01:38] <pook1e> I'm trying to build for natty.
[01:38] <RAOF> Then that won't work unless you backport libnotify.
[01:38] <RAOF> (At least)
[01:42] <pook1e> How can I do that?
[01:43] <pook1e> I'm creating my pbuilder environment with pbuilder-dist natty create --extrapackages libnotify-dev
[01:47] <RAOF> What is your actual goal here?
[01:48] <pook1e> I'm fixing a bug in indicator-network
[01:49] <pook1e> And now I'm trying to build the modified version to test it out.
[01:49] <RAOF> Then you probably need to test it on oneiric; I think the indicator stack is sufficiently different to make backporting stuff to natty and testing there not very worthwhile.
[01:53] <RAOF> It should be easy enough to set up oneiric in a VM if you don't want to install it bare-metal.
[01:54] <pook1e> So pbuilder-dist oneiric create --extrapackages libnotify-dev, then pbuilder-dist oneiric build?
[01:54] <RAOF> You don't need the --extrapackages; pbuilder will automatically install the build-depends for you.
[01:55] <RAOF> Apart from that, yes.
[01:58] <pook1e> Alright, I'll try that out.
[01:58] <pook1e> Thanks for your help RAOF, much appreciated.
[02:53] <mwhudson> maybe this is a crazy question
[02:53] <mwhudson> but if i require a package to be able to run debian/rules clean
[02:54] <mwhudson> is there some way i can build it in pbuilder?
[02:57] <RAOF> mwhudson: You can pass --extrapackages to pbuilder.  Isn't that a packaging bug, though?
[02:57] <RAOF> Oh, or do you want to build a *source* package from the current tree, and can't install that package?
[02:57] <mwhudson> RAOF: yes, debuild -S is failing
[02:58] <RAOF> Hawkward.
[02:58] <mwhudson> also with an extra p
[02:58] <mwhudson> (well, i just installed the package, in this case)
[02:58] <mwhudson> but i was wondering more generally
[02:59] <RAOF> There's no technical limitation on that as far as I'm aware, but I've never really wanted it so I haven't investigated whether it's implemented at all; I'd guess the answer is “no”
[02:59] <StevenK> mwhudson: You can run pbuilder login, install packages there and run debuild -S inside the chroot
[02:59] <mwhudson> StevenK: ah yeah, i guess that works
[02:59] <mwhudson> (kinda what building recipes does on lp)
[03:00] <StevenK> Hacky, but some packages require it
[03:03] <mwhudson> after a mere 4 and a bit years at canonical, i'm finally learning stuff about packaging :)
[03:04] <StevenK> mwhudson: Plenty of people in your timezone who are happy to help :-)
[03:04] <ajmitch> plenty of people even in your loco channel who can help out :)
[03:05] <mwhudson> well, if you guys could just stop inventing new ways to package python projects.....
[03:06] <ajmitch> this is the last change, honest...
[03:06] <StevenK> Haha
[03:07] <mwhudson> (and/or backport dh_python2 to lucid)
[03:07] <mwhudson> i suppose waiting for the next LTS is the lazy way to make that problem go away
[03:07] <ajmitch> I think that had been done somewhere, just not in lucid-backports
[03:07] <ScottK> mwhudson: I think barry is supposed to do that.
[03:07] <ScottK> ajmitch: If there's a good backport done, I'd be glad to have it in lucid-backports, I just don't care to do the work to package it myself.
[03:08] <ajmitch> the main problem is more the number of things it could potentially break
[03:16] <lifeless> + pkgme
[03:22] <ScottK> Shouldn/t
[03:22] <ScottK> It should only affect stuff that's build with it, so that's limited to -backports
[03:23] <ScottK> lifeless: +stdeb
[03:23] <lifeless> indeed
[03:24] <lifeless> and there is a python make-deb thing as well upstream
[03:24] <ScottK> That's different then stdeb?
[03:25] <ScottK> Automatic packaging of python stuff seemed like a really odd problem for pkgme to take on since it was already solved about as well as it could be with automation, IMO.
[04:27] <pitti_> Good morning
[04:35] <TheMuso> pitti: Do you have powers to re-enable cd image builds? Seems they are still disabled after the alpha 2 release.
[04:36] <pitti> TheMuso: they are enabled, but have failed to build ever since due to libc6
[04:36] <pitti> i. e. debootstrap failure
[04:44] <TheMuso> oh ok
[04:46] <infinity> I need to do some image mangling this week, so if debootstrap remains sad, I might do something about it.
[04:46] <infinity> (Happy to heave others beat me to it)
[04:47] <infinity> s/heave/have/
[05:38] <rsalveti> pitti: infinity: bug 807974
[05:38] <pitti> rsalveti: thanks
[05:52] <rsalveti> pitti: if you just enable var/run and var/lock at base-files it works fine again
[05:52] <rsalveti> pitti: https://code.launchpad.net/~rsalveti/ubuntu/oneiric/base-files/run-transition/+merge/67444
[05:52] <rsalveti> but then don't know exactly why this was removed
[05:52] <pitti> it's /run/ now
[05:52] <rsalveti> and how ubuntu is supposed to deal with /var/run
[05:53] <pitti> so I guess something in eglibc needs to be updated
[05:53] <rsalveti> pitti: I know, but why remove the directory?
[05:53] <pitti> /var/run/ should presumably become a symlink now?
[05:53] <rsalveti> pitti: not only eglibc, but I guess we have similar packages at the archive
[05:53] <rsalveti> pitti: should be a bind mount
[05:53] <rsalveti> at least this is how fedora is doing
[05:53] <pitti> I still have /var/run/ here, but this is an upgrade
[05:53] <rsalveti> to have a bind mount, it makes sense to have the directory there
[05:53] <pitti> either way
[05:54] <rsalveti> so that's why for me it's wrong to just remove it from base-files
[05:54] <rsalveti> as it's going to be bind mounted later
[05:54] <rsalveti> but yes, the libc6 package is also wrong
[06:14] <Chipzz> rsalveti: maybe that's just the slightly sadistic side of me speaking, but on the bright side, having the dir missing will make a lot of software that yet needs to be fixed show up, and bugs can be filed
[06:16] <pitti> but for the final release we'll have a /var/run/ either way
[06:17] <pitti> we can't break the world in just one cycle
[06:17] <slangasek> moreover, /run is not a standard (yet) - things that don't have a strict need for writing during early boot shouldn't be forced to switch yet
[06:18] <slangasek> the missing /var/run is a bug, plain and simple
[06:19] <rsalveti> slangasek: shouldn't /var/run be a bind mount of /run?
[06:19] <rsalveti> like fedora is doing
[06:19] <rsalveti> instead of a syslink?
[06:19] <Chipzz> hrrrm, s/having the dir missing/having the dir missing for a couple of days/ ...
[06:19] <slangasek> rsalveti: that's an implementation detail
[06:20] <rsalveti> but easier to do the transition, I believe
[06:20] <Chipzz> slangasek: but since it's a bind mount you might as well make everything write to /run
[06:20] <slangasek> easiest is to follow Debian's lead on this; the transition has already been done in unstable
[06:20] <slangasek> Chipzz: no
[06:20] <Chipzz> I see little point in having 2 dirs with the same purpose
[06:21] <slangasek> the FHS governs where things are supposed to write, and until there's a new FHS with clear and consistent guidance for the use of /run vs. /var/run, anything that uses /var/run today should be staying there
[06:22] <rsalveti> yeah, agree
[06:39] <buxy> Are there known kernel problems that could result in #807619 ? It's really weird
[06:48] <dholbach> good morning
[07:51] <Satoris> What package do I need to install to get the "bzr mu" command?
[07:52] <nigelb> Satoris: er, what does it do?
[07:52] <Satoris> It's used to update packaging when a new source release happens.
[07:52] <nigelb> probably bzr-buildeb I think
[07:54] <Satoris> That did it. Thanks.
[07:54] <nigelb> :)
[08:01] <sladen> slangasek: thanks for the icons.  Was that your doing? (Should I cancel the RT?)
[08:15] <slangasek> sladen: yes, I updated them; cdimage.u.c / releases.u.c contents are managed by the Ubuntu team, not by IS
[08:20] <Amoz> dholbach, morning!
[08:20] <dholbach> hey Amoz
[08:22] <panda|ac100> dholbach: ow
[08:22] <dholbach> hi panda|ac100
[08:39] <Amoz> is it appropriate to attach a patch to a bug in the reportbug utility?
[08:40] <Amoz> I'm trying to send a bugreport and fix upstreams.. and I'm kind of lost
[09:17] <Quintasan> cjwatson: ping
[09:25] <cjwatson> Quintasan: yes?  (best to include content with your pings to save on round-trips)
[09:27] <Quintasan> cjwatson: Oh, sorry for that. I was wondering if kde-baseapps is in kubuntu packageset since I still can't upload it (I am a Kubuntu Developer), is there any way I can check this without bothering you?
[09:28] <cjwatson> Quintasan: ScottK was due to give me a list of stuff to fix
[09:28] <cjwatson> Quintasan: lp:ubuntu-archive-tools, the edit_acl.py program there can answer queries like this
[09:29] <Quintasan> cjwatson: Thanks, if it's okay with you I think I can complete this list instead of having ScottK do it
[09:34] <cjwatson> Quintasan: sure - see http://irclogs.ubuntu.com/2011/07/09/%23ubuntu-devel.html#t21:24
[09:36] <cjwatson> Quintasan: I've kicked an auto-update now, which should help too (though it didn't touch kde-baseapps)
[09:40] <seb128> cjwatson, hey, just a question following the other day when I asked if you did autosyncs, during other cycles we used to have an email on u-d or u-d-a saying that the freeze is in effect, do you know if those are supposed to be announced?
[09:40] <nigelb> aa
[09:40] <nigelb> ~/
[09:40] <nigelb> ~/
[09:40] <nigelb> err, sorry for the spam
[09:41] <seb128> cjwatson, it was somewhat useful for me since I don't know the cycle calendar in details and don't always think to go and check there
[09:41] <seb128> I guess the reminder was useful for others as well
[09:42] <Quintasan> cjwatson: http://paste.kde.org/94267/ -> Here is what I got is missing, I will give you some more later since I assume you can't add non-existing packages yet to packageset
[09:50] <cjwatson> seb128: I guess we should have done, seems a bit late now though
[09:50] <seb128> cjwatson, ok
[09:50] <cjwatson> Quintasan: your assumption is correct, yes
[09:50] <seb128> (well at least I know we are in DIF now which I didn't realize before you said it ;-)
[09:50] <cjwatson> Quintasan: what previous packages do those replace?
[09:51] <Quintasan> cjwatson: kde-workspace would replace kdebase-workspace
[09:52] <Quintasan> kde-runtime replaces kdebase-runtime
[09:52] <Quintasan> kate was orginally a part of kdesdk
[09:53] <Quintasan> kwordquiz was a part of kdeedu (now separate source package)
[09:53] <cjwatson> kwordquiz was already in the package set (I think I only just added it though)
[09:53] <cjwatson> done the rest, thanks
[09:53] <Quintasan> cjwatson: Thanks
[10:08] <mvo> apachelogger: I'm just looking at your dbusworker software-properties stuff and I would like to work on this a bit, I noticed you use "slip.dbus" - what packages does provide that?
[10:08] <Amoz> dholbach, I've done some more markup stuff on the UPG
[10:09] <dholbach> Amoz, oh nice!
[10:09] <apachelogger> mvo: I think we did not have it packaged when I worked on it
[10:09] <dholbach> Amoz, sorry, I didn't get around to having a look at it, I just became too busy - today I won't have time for this either :-/
[10:09] <apachelogger> it's some redhat polkit helper IIRC
[10:09] <Amoz> dholbach, no worries
[10:09] <mvo> apachelogger: aha, ok. would you mind if I leave it out for now then in my branch?
[10:09] <apachelogger> nope
[10:09] <dholbach> Amoz, do you think you can push to   lp:~fougner/ubuntu-packaging-guide/retheming  and propose a merge when you're happy with it?
[10:10] <dholbach> Amoz, that'll give everybody a chance to have a look at it and comment
[10:10] <mvo> thanks!
[10:10] <dholbach> Amoz, I'll make sure I'll add all the boring bits (links to where everything is from, explain what was the actual changes and what was borrowed from somewhere else)
[10:10] <Amoz> dholbach, I'm not happy with it but...
[10:11] <dholbach> Amoz, it'll be a work-in-progress
[10:11] <dholbach> and give others a chance to help out and make it even prettier
[10:11] <dholbach> it's a GREAT start already
[10:23] <Amoz> dholbach, uhm, the branch is stacking on my "old" branch, does it matter?
[10:23] <dholbach> Amoz, no, that should be fine
[10:24] <Amoz> dholbach, well in that case it's up and kicking
[10:24] <dholbach> fantastic
[10:24] <dholbach> did you propose it for a merge?
[10:24] <Amoz> oh
[10:25] <Amoz> I will
[10:26] <Amoz> dholbach, proposed
[10:27] <dholbach> awesome
[10:27] <dholbach> it should be in my inbox soon and I'll do my best to tend to it soon
[10:28] <Daviey> talking of email, is thunderbird in a bad way for everyone else?
[10:28] <dholbach> Daviey, it's fine for me
[10:29] <dholbach> Daviey, oh, I assume you're talking oneiric - I still use it in a vm, ignore me :)
[10:29] <Amoz> cool, will the new theme be showing at your ppl.ubuntu.com site then?
[10:29] <davmor2> dholbach: so no shrinking filters window that is un-resizeable or 186% cpu usage then?
[10:30] <dholbach> Amoz, that's the plan once it's merged - also once it looks good and works nicely for all of us, we'll move it to a more official home
[10:31] <dholbach> davmor2, at least not on natty - no idea about oneiric
[10:31] <Amoz> nice
[10:32] <davmor2> dholbach: no in natty I get none of that behaviour when I've tested it, only on my oneiric test box
[10:32] <dholbach> davmor2, ignore me then :)
[10:32] <davmor2> which is admittedly my old netbook
[11:08] <Amoz> mmm..I wonder if the new asus u36jc plays well with ubuntu..
[11:41] <lifeless> mvo: sbeattie: btw, conflict-checker seems rather unhappy :)
[11:45] <mvo> lifeless: I'm checking it out currently
[11:46] <lifeless> \o/
[11:48] <mvo> lifeless: should be fine again
[11:48] <lifeless> what was it?
[11:50] <mvo> lifeless: dapper is end-of-life
[11:50] <lifeless> ahha
[12:25] <mantiena-baltix> Hi all, I have question about using aptd in postinst or configure script of debian package (now I creating deb package, which should add new repositories to /etc/apt/sources.list.d folder and seeking the best way how to do this)
[12:39] <mantiena-baltix> could I use aptd --add-repository "deb http://additional.repository.com ubuntu natty" in postinst or configure script of debian package or maybe someone know better way how to add add new repositories to /etc/apt/sources.list.d/ folder ?
[12:40] <mantiena-baltix> I mean aptdcon  --add-repository
[12:51] <directhex> mantiena-baltix, you can just put a file in that folder. i.e. your package includes /etc/apt/sources.list.d/moo.list as a regular file
[13:41] <mantiena-baltix> directhex: there is one problem during upgrades if my package simply includes /etc/apt/sources.list.d/moo.list as a regular file - user often  gets a message about changed config file, because /etc/apt/sources.list.d/moo.list automatically becomes a conffile when a deb package is built
[13:45] <mantiena-baltix> directhex: also I wanna make this package installable on different versions of ubuntu, so I wanna generate repository line "on the fly" during package installation, thats why I think to use aptdcon --add-repository "deb http://additional.repository.com ubuntu $codename" in postinst or configure script of debian package
[13:46] <mantiena-baltix> Maybe this is a crazy idea, maybe I should do in another way, I'm not sure
[13:47] <mantiena-baltix> Current realization of my package uses static *.list files, see http://bazaar.launchpad.net/~baltix-members/baltix-default-settings/ubuntu/files/head:/etc/apt/sources.list.d/
[14:32] <geser> mantiena-baltix: do you use the same package for all Ubuntu releases? and how do you plan to handle upgrades of the package (or is aptd intelligent enough to not create duplicate entries?)
[14:34] <mantiena-baltix> geser: yes, aptd is smart enough :)
[14:34] <mantiena-baltix> geser: thats one of reasons why I wanna use aptd
[14:51] <mantiena-baltix> So, are here some developers against using aptdcon --add-repository "deb http://additional.repository.com ubuntu $codename" in postinst or configure script of debian package?
[14:52] <mantiena-baltix> s/some/any
[14:55] <Laney> what file does that touch? what happens if I edit said file?
[14:56] <cjwatson> does that require dbus to be available from the postinst?
[14:56] <cjwatson> if so that could easily cause problems in chroots
[14:57] <jpds> How is it different from add-apt-repository ?
[15:02] <tkamppeter> pitti, hi
[15:02] <pitti> hey tkamppeter
[15:03] <tkamppeter> It is about bug 799442
[15:04] <pitti> cups*ys* oh my, that's old
[15:04] <tkamppeter> pitti, the problem is here that a manufacturer-supplied printer driver is still based on the package name "cupsys" which was abolished 2 years ago.
[15:04] <pitti> we renamed it more than 3 years ago
[15:04] <tkamppeter> pitti, now the user is not able to make this driver work.
[15:05] <pitti> does a three year old driver even still work?
[15:05] <pitti> tkamppeter: I guess it's time for Canon to do a rebuild?
[15:05] <tkamppeter> pitti, I do not know why manufacturers offer so old drivers and do not update them.
[15:06] <pitti> tkamppeter: a workaround for this guy might be to use equivs to generate a "fake" cupsys package
[15:06] <pitti> so that apt is satisfied
[15:06] <tkamppeter> pitti, seems that I have to recommend to this user to buy a new printer, HP or Epson.
[15:06] <pitti> heh
[15:06] <pitti> no free driver for this at all?
[15:07] <tkamppeter> pitti, can you add instructions to that bug report for the user to use equivs so that he can make his printer work?
[15:10] <mantiena-baltix> jpds:  "aptdcon --add-repository" is smarter than "add-apt-repository", because aptd doesn't add duplicate lines, it's very important for my, because sometimes users already have some of repositories, which my package wanna add
[15:12] <mantiena-baltix> cjwatson: AFAIK aptd requires dbus, I'm not sure, you should know better than I :)
[15:12] <pitti> tkamppeter: done
[15:16] <tkamppeter> pitti, thanks.
[15:17] <cjwatson> mantiena-baltix: I don't know offhand, you're the one proposing this so you should check
[15:17] <cjwatson> mantiena-baltix: I don't think postinsts should depend on being able to talk to a running dbus-daemon in the appropriate filesystem context
[15:17] <Riddell> dholbach: what's the plan for completing the ubuntu packaging guide?
[15:17] <cjwatson> mantiena-baltix: perhaps you should try to use the relevant backend parts of aptdaemon directly, rather than via aptdcon, if aptdcon is expecting to talk to a dbus backend
[15:18] <cjwatson> it's all in python so it shouldn't be too hard to call just the relevant bit
[15:23] <mantiena-baltix> cjwatson: it's ok to use python scripts in postinst?
[15:24] <cjwatson> sure
[15:26] <dholbach> Riddell, there is a number of bugs filed where we track how we massage the old content we have (like from the wiki) into usable articles
[15:26] <dholbach> Riddell, we'll hopefully soon move it to a more official place soon, getting it themed in ubuntu colours is something that's currently being worked on (among other things)
[15:27] <dholbach> not sure if that answers your question
[15:29] <tkamppeter> pitti, can you have a look at bug 807261? For some time now s-c-p asks for the user's password, even if he is in lpadmin.
[15:41] <tkamppeter> pitti, I had an interruption, did you answer my last message?
[15:50] <dholbach> https://wiki.ubuntu.com/UbuntuDeveloperWeek Day 1 to kick off in 10 minutes in #ubuntu-classroom
[15:51] <Riddell> dholbach: I think the main issue is if bug 792381 is to be fixed or won't fixed because that needs a fair amount of restructuring
[15:51] <dholbach> Riddell, I have UDW starting in a bit, so not much time to discuss it right now - I'll get back to you about it - can you subscribe to it?
[15:52] <Riddell> k
[15:55] <doko> bdmurray: thanks for cleaning up the -fsys-tarfile issues. how many are these?
[15:55] <bdmurray> doko: about 100
[15:57] <bdmurray> pitti: I want to include the casper package version from apport bug reports with LiveMediaVersion in them.  Is the generic.py hook the best place for that?
[16:00] <pitti> bdmurray: /usr/share/apport/general-hooks/ubuntu.py would be, as that has the ubuntu specific bits
[16:01] <pitti> tkamppeter: need to run, will look at the bug and reply there
[16:35] <mpt> sladen, hi, did you see this? https://lists.ubuntu.com/archives/ubuntu-devel-discuss/2011-June/012686.html
[16:38] <sladen> mpt: ta
[16:47] <debfx> pitti: the kubuntu seed branch has been changed to ~kubuntu-dev/ubuntu-seeds/kubuntu.oneiric
[16:52] <RoAkSoAx> pitti: howdy!! I just upgraded to the latest udev (as was the last package i need to upgrade in oneiric) and have lost keyboard/mouse in an x201
[16:53] <RoAkSoAx> any ideas?
[16:57] <RoAkSoAx> pitti: never mind, seems to be something with lightdm
[17:54] <thebishop> is there any plan to get synaptics clickpad support in 11.10?  It doesn't work as of Alpha2
[17:54] <thebishop> apparently it's been in xorg upstream for a whlie
[18:29] <james_w> pitti, https://code.launchpad.net/~ubuntu-branches/ubuntu/oneiric/gnome-keyring/oneiric-201107080812/+merge/67292 <- seems your upload didn't actually change the watch file as the changelog stated?
[19:56] <bambee> who did upgrade xorg recently? http://paste.ubuntu.com/642125/
[20:14] <Laney> any AA around to quickly sync tangerine? bug 808723
[20:14] <Laney> need it to start the lib transition
[20:17] <infinity> Laney: I can poke at it before I lunch.
[20:17] <Laney> infinity: thanks
[20:17] <Laney> the library transition should hopefully go much faster
[20:27] <cnd> I've got a daily build recipe for a package, but it says "No suitable builders": https://code.launchpad.net/~utouch-team/+recipe/libgrip-daily
[20:27] <cnd> what's going on?
[20:29] <Laney> they are down for a minute apparently
[20:33] <cnd> Laney, ok
[20:33] <cnd> thanks
[20:56] <ScottK> cnd: Generally #launchpad is a better place for such questions.
[20:58] <cnd> ScottK, ahh, I see
[21:09] <seif> any idea why the keyboard and mouse on my thinkpad stopped working on oneirick with the last update
[21:09] <hyperair> bad kernel update?
[21:11] <ion> seif: Dunno, but if you can disconnect and reconnect input devices after starting X, that’s a workaround. :-P
[21:12] <bryceh> seif, I do
[21:12] <seif> i found a bug
[21:12] <bryceh> seif, it's the new /run directory
[21:12] <seif> bryceh, yeah
[21:12] <seif> there is a workaround
[21:13] <bryceh> try sudo rm -rf /run
[21:13] <bryceh> it's bug #807306
[21:13] <bryceh> which basically is a dupe of bug #807974
[21:15] <ion> Hmm, that might also be the reason for pulseaudio not finding audio devices from udev.
[21:16] <bryceh> or at least sudo rm -rf /run/udev
[21:16] <bryceh> that might be a better workaround
[21:17]  * slangasek wonders why he's not seeing any of these problems locally
[21:18] <bryceh> slangasek, Sarvatt couldn't reproduce on 6 systems either, but cnd did
[21:19] <cnd> slangasek, I haven't "fixed" my system yet
[21:19] <cnd> I can help you prod a broken system
[21:20] <bryceh> ok, enough for me on that bug... pilot time
[21:20] <cnd> my machine also is an upgrade from maverick development through natty development to oneiric development
[21:20] <bryceh> @pilot in
[21:43] <bdmurray> RAOF: could you comment on bug 797894 please?
[21:43] <Laney> infinity: thanks a lot
[21:43] <Laney> here comes the lib uploads
[21:47] <kirkland> slangasek: howdy, around?
[21:48] <kirkland> slangasek: nevermind, I'll take it to email
[21:50] <bryceh> slangasek, seen https://code.launchpad.net/~rsalveti/ubuntu/oneiric/base-files/run-transition/+merge/67444 ?
[21:51] <bryceh> slangasek, want me to merge that?
[21:53] <slangasek> bryceh: if you test that it dtrt, sure :)
[21:55] <infinity> slangasek: It just adds the directories back, which is a reasonable stop-gap for now.
[21:55] <infinity> slangasek: I'll admit that I'm failing to understand why we can't just symlink and call it done.
[21:55] <infinity> (crazy talk of bindmounting and such...)
[21:55] <slangasek> yeah, I can't think of any way that readding the directories will break things
[21:55] <rsalveti> guess we can just create a symlink, but don't know if the transition is that simple
[21:56] <slangasek> still, should be tested
[21:56] <slangasek> rsalveti: the distinction between symlink and bind mount is generally moot, except in the case where you're poking a chroot from outside and get bitten by wide links
[21:56] <slangasek> but a symlink is a lot easier to code up
[21:56] <rsalveti> sure
[21:57] <infinity> rsalveti: The way dpkg treats symlinks to directories makes it a meaningless distinction (assuming anything still ships something in /var/run, which they shouldn't have been doing for years now)
[21:57] <slangasek> regardless, I'm not inclined to own an implementation of any /run code here - I want us to do what Debian did :)
[21:57] <rsalveti> :-)
[21:57] <infinity> Yeah, I'm happy with that, so long as "what Debian did" actually works smoothly for us before release. :P
[21:57] <rsalveti> slangasek: and what is the remaining merge at sysvinit you said it's complicated?
[21:57] <slangasek> workin' on it
[21:57] <slangasek> initscripts merge is a pain
[21:58] <rsalveti> guess we'll also have to remove the remaining directory and create the symlink while updating the package
[21:58] <broder> rsalveti: the debian update covered that too
[21:58] <slangasek> rsalveti: some 36-odd conflicts from the initial branch merge; half of these due to reorg of the source package (so, contents conflicts that require merging by hand); I'm about 3/4 way through the merge now
[21:59] <rsalveti> slangasek: oh, ok
[21:59] <rsalveti> so guess we can get both directories back for now, so we can make debootstrap to work again
[21:59] <rsalveti> and once you merge sysvinit we can then finally remove it
[21:59] <rsalveti> properly
[21:59] <slangasek> yep
[21:59] <infinity> Working debootstrap would make me a cappy hamper.
[21:59] <infinity> I was going to do that today if no one beat me to it. :P
[22:37] <barry> kirkland ping
[22:37] <kirkland> barry: pong
[22:37] <serge_afk> RoAkSoAx: have you looked at the udev changelog?
[22:38] <barry> kirkland, hi.  do you have a little time to help me with an ecryptfs problem?
[22:38] <kirkland> barry: sure, what's up?
[22:38] <barry> kirkland, this is on a natty box that has been working great for a long time.  just today i rebooted (after probably 2-3 weeks of uptime) and my encrypted home dir will no longer mount
[22:39] <barry> ecryptfs-mount-private says "encrypted private directory is not setup properly"
[22:39] <kirkland> barry: hmm, okay
[22:39] <barry> kirkland, um, i have no idea what's wrong or how to fix it
[22:39] <kirkland> barry: are you on the box as root now?
[22:39] <barry> yep, in a vt
[22:39] <kirkland> barry: or at least on the box as a user
[22:39] <barry> well, it's a different box than i'm typing on :)
[22:40] <barry> kirkland, yep, root on a vt
[22:40] <kirkland> barry: cool, grep sysfs /etc/mtab
[22:40] <barry> kirkland, non /sys sysfs rw,noexec,nosuid,nodev 0 0
[22:41] <barry> kirkland, s/non/none/
[22:41] <kirkland> barry: okay, so it's not that issue
[22:41] <RoAkSoAx> serge_afk: yes
[22:41] <kirkland> barry: hmm, well, i just dist-upgraded my natty box today for the first time in 3 weeks
[22:42] <kirkland> barry: and i haven't rebooted yet
[22:42] <barry> kirkland, this is still natty ;)
[22:42] <kirkland> barry: give me a minute, and let me see if i reproduce the issue here
[22:42] <kirkland> barry: yeah, i'm still natty on this box too
[22:42] <kirkland> barry: rebooting ...
[22:42] <barry> kirkland, thanks
[22:42] <kirkland> barry: i'm not sure if I want to reproduce it, or not :-)
[22:43] <barry> ;)
[22:45] <kirkland> barry: okay, I'm up, no problem
[22:45] <kirkland> barry: let's drill down into what might be wrong with yours
[22:45] <barry> kirkland, cool, thanks
[22:45] <serge_afk> RoAkSoAx: wait, the system is failing in oniric, or in uptodate natty?
[22:46] <kirkland> barry: as your non-root user, clear your keyring with 'keyctl clear @u'
[22:46] <RoAkSoAx> serge_afk: oneiric
[22:46] <serge_afk> ok
[22:46] <barry> kirkland, done
[22:46] <serge_afk> the bzr tree is, alas, not uptodate :(
[22:46] <infinity> cjwatson: Was there a reason (other than CPU usage) to have p.c.c/~u-a only run germinate every 4 hours?
[22:46] <kirkland> barry: and sh -x /usr/bin/ecryptfs-mount-private
[22:46] <kirkland> barry: should prompt you for your password
[22:47] <cjwatson> infinity: I think mostly CPU usage.  It takes a fair while to run and it seemed unfriendly to run it every hour.
[22:47] <barry> kirkland, it did not, it just failed with the previous error
[22:47] <cjwatson> infinity: germinate is packaged, so people in a rush can always run it locally
[22:47] <kirkland> barry: does that -x output give you any indication where that failure was?
[22:48] <barry> kirkland, it looks like it's failing at the [ -f /home/barry/ecryptfs/wrapped-passphrase ... ] test
[22:48] <infinity> cjwatson: True.  I'm looking at in-DC use (livefs builds), where I'd prefer pulling from a pre-generated source, if I can get away with it.
[22:49] <RoAkSoAx> serge_afk: i had the same issue as this: bug #807306 but followed advivce of moving /run/udev to /run/udev.old and mouse/keyboard working again, so I'm presumming it might be related as i think it all happen at the same time more or less
[22:49] <barry> kirkland, it looks like .ecryptfs -> /home/ecryptfs/barry/ecryptfs but that file doesn't exist
[22:49] <barry> er, /home/.ecryptfs/barry/.ecryptfs
[22:49] <barry> kirkland, the only thing in /home/.ecryptfs/barry is .Private
[22:50] <kirkland> barry: hmm, there should definitely be a ".ecryptfs" and a ".Private" in that directory
[22:50] <barry> kirkland, ls -la turns up only .Private :(
[22:50] <kirkland> barry: http://paste.ubuntu.com/642227/
[22:51] <kirkland> barry: whoa
[22:51] <kirkland> barry: sudo find /home -type d -name .ecryptfs
[22:51] <barry> kirkland, should i panic? :)
[22:51] <kirkland> barry: not if you have your mount passphrase written down (as you were reminded several times)
[22:52] <barry> kirkland, and if i know where i put it :(
[22:53] <kirkland> barry: do you have it in /home/barry/.ecryptfs ?
[22:53] <barry> kirkland, that find turns up nothing
[22:53] <kirkland> barry: or is that just a symlink?
[22:53] <kirkland> barry: where the heck has it been getting your ecryptfs configuration from?
[22:53] <barry> kirkland, it's just a symlink
[22:53] <kirkland> barry: do you have a /var/lib/ecryptfs?
[22:54] <kirkland> barry: that's where we put it in Jaunty era
[22:54] <barry> kirkland, excellent question, and i have no clue where the real file went
[22:54] <kirkland> barry: but it hasn't been there in eons
[22:54] <barry> kirkland, no /var/lib/ecryptfs
[22:55] <kirkland> barry: yeah, i did not expect a /var/lib/ecryptfs
[22:55] <kirkland> barry: but you damn well should have a /home/.ecryptfs/barry/.ecryptfs ...
[22:55] <serge_afk> RoAkSoAx: so mv /run/udev /run/udev.old works for you?
[22:56] <barry> kirkland, i'm guessing that a wrapped-passphrase file from another machine wll not help me, right?
[22:56] <RoAkSoAx> serge_afk: I did that and keyboard/mouse worked for me, but still networking for KVM is not
[22:56] <kirkland> barry: no unless the 128 bits of randomness inside are identical ...
[22:57] <kirkland> barry: is there any way possible that dir could have just disappeared?
[22:57] <RoAkSoAx> serge_afk: it is as if I cant find the DHCP server
[22:57] <kirkland> barry: there's no ecryptfs updates published in https://launchpad.net/ubuntu/+source/ecryptfs-utils
[22:57] <barry> kirkland, i can't think of how that could have happened.  certainly nothing that i did explicitly (that i know of or remember)
[22:58] <kirkland> barry: fwiw (and it's too late now), I escrow my keys at keyescrow.net
[22:58] <barry> kirkland, i suppose of course, that a hardware failure is possible, but nothing that i've noticed or seen in the log files
[22:59] <kirkland> barry: yeah, you grepped /var/log?
[22:59] <kirkland> barry: do permissions and ownerships look right in /home/.ecryptfs ?
[23:00] <barry> kirkland, /var/log grep turns up nothing
[23:00] <kirkland> barry: let me check the ecryptfs bugs in launchpad and see if there's an outbreak of something
[23:00] <barry> kirkland, should /home/.ecryptfs/barry be owned by me or root?
[23:01] <kirkland> barry: you
[23:01] <barry> kirkland, it's owned by root
[23:01] <kirkland> this is when /me wishes the bug "heat" indicator was worth a damn in LP
[23:01] <barry> kirkland, heh, yeah
[23:02] <Sarvatt> serge_afk: mv /run/udev /run/udev.old only works until udev gets updated again and recreates /run/udev because /run exists thanks to base-files and it uses that over /dev/.udev in that case
[23:02] <kirkland> barry: http://paste.ubuntu.com/642231/
[23:02] <kirkland> barry: did you encrypt all of $HOME, or just $HOME/Private ?
[23:02] <barry> kirkland, all of home iirc, and yeah, my /home/.ecryptfs doesn't look like that at all
[23:03] <barry> 755 perms on . .. and barry, root:root on all
[23:03] <barry> kirkland, the only thing i can think of is that i've done some schroot updates today before the reboot, and updated some of my virtual machines, but none of that should have leaked out into my host environment
[23:04]  * barry is starting to get a bad feeling
[23:04] <kirkland> barry: hmm ... i think there's an open bug against schroots and ecryptfs, let me check
[23:05] <kirkland> barry: https://bugs.launchpad.net/ubuntu/+source/ecryptfs-utils/+bug/769595
[23:06] <serge_afk> RoAkSoAx: it might be interesting to see if the cmds at http://paste.ubuntu.com/642236/ work for you under natty and oneiric
[23:06] <cjwatson> infinity: the germinate-output stuff on people.c.c is explicitly only informational - for example, it's only run for one architecture (i386).  I'd really rather that anything else do its own germinate run rather than relying on it.
[23:06] <barry> kirkland, is it possible that a bug caused an oneiric chroot update to screw the host $HOME?
[23:07] <rsalveti> bryceh: what's the udev bug?
[23:07] <rsalveti> have the bug number?
[23:07] <kirkland> barry: i'm pondering how that could possibly hurt you
[23:09] <barry> kirkland, i'm just pulling this out of my butt, what if an oneiric bug in ecryptfs somehow caused the perms and contents of ~/.ecryptfs/barry to get hosed, but because of some mount problem it didn't affect the chroot but instead the host fs
[23:09] <kirkland> barry: do you have an bind mounts right now?
[23:10] <Sarvatt> rsalveti: https://launchpad.net/bugs/807306 its a different bug than the one you were trying to fix though, this one is caused by udev using /run over /dev/.udev if /run exists which it does because base-files created it and is breaking things. debian fixed it in udev 167-2 (http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=620995)
[23:11] <kirkland> barry: i think some schroots do a bind or an rbind of /home, don't they?
[23:11] <kirkland> barry: sorry I don't know more about schroots, as I rarely use them
[23:11] <barry> kirkland, i think they do, but there's nothing in effect right now afaict
[23:12] <barry> kirkland, so, am i f*cked?  if i can't find the passphrase, is there no way to recover from a livecd?
[23:13] <kirkland> barry: :-(  there wouldn't be much security to ecryptfs if you could recover with a livecd and no passphrase
[23:13] <barry> kirkland, i guess my question was, knowing my password isn't enough, right?
[23:13] <rsalveti> Sarvatt: oh, ok, it's fixed at udev
[23:14] <kirkland> barry: it's possible that that wrapped-passphrase file is still somewhere in your filesystem
[23:14] <kirkland> barry: if the bits haven't been overwritten yet
[23:14] <barry> kirkland, ah good point.  i can't do much more harm searching for it
[23:14] <kirkland> barry: i don't know how big your disk is, but a dd image of it to an nfs share would be the next thing I'd do
[23:15] <kirkland> barry: so that you could do some forensics on it without peturbing it further
[23:16] <kirkland> barry: you can regenerate everything else in .ecryptfs;  you really just gotta find wrapped-passphrase
[23:16] <barry> kirkland, it's only 277G used out of a1.5T drive.  i can mount a usb drive and dd it there.  on the fortunate side, this was purely a dev box so nothing unrecoverable is on it.  just a pita to rebuild
[23:16] <kirkland> barry: or find a copy of the written down, decrypted contents of it
[23:16] <kirkland> barry: understandable;  i really need to figure out how/why that dir got removed
[23:17] <kirkland> barry: in case it's systemic
[23:17] <barry> kirkland, okay cool.  thanks for the help.  if i can't find it then i guess it's oneiric time for me
[23:17] <kirkland> barry: please record if you reinstall :-)
[23:17] <barry> kirkland, yep, and i'm happy to help.
[23:17] <kirkland> barry: and use keyescrow.net ;-)
[23:17] <kirkland> barry: or email it to yourself
[23:17] <kirkland> barry: or something :-)
[23:17] <barry> kirkland, most definitely
[23:18] <kirkland> barry: good luck;  let me know if i can help further
[23:18] <barry> kirkland, but i am suspecting some bad interaction between sbuild/schroot/oneiric
[23:18] <kirkland> barry: do some deep finds on your filesystem for that file
[23:18] <barry> kirkland, will do, thanks for your help
[23:18] <kirkland> barry: hmm, you might also talk to jdstrand
[23:18] <kirkland> barry: i *think* he might have seen something like this once before
[23:18] <kirkland> not sure why i'm thinking that
[23:18] <kirkland> barry: but check with him
[23:18]  * kirkland heads to dinner
[23:18] <barry> kirkland, will do
[23:34] <bryceh> slangasek, btw the merge proposal I pointed to earlier didn't actually fix it.  (probalby still a good idea, just not sufficient)
[23:37] <bryceh> slangasek, really I just noticed that branch while patch piloting; I'm not going to be chasing the /run bug down any further.  I do think it could use a short term fix, as a lot of people are losing their keyboards and mice.
[23:47] <rsalveti> bryceh: I believe for that we should have the same fix as debian
[23:48] <rsalveti> don't know how hard would be to merge the debian changes into our udev, didn't look