[03:23] <smoser> where does 'apt-install' in the server installer (debian-installer) come from?
[03:23] <smoser> i'm just wondering what changed that caused bug 912431
[03:30] <smoser> slangasek, cjwatson ? i'm sure one of you know.
[03:31] <smoser> we're hoping to do work on juju doing installs via cobbler during the rally, and this is currently broken
[03:34] <broder> smoser: i think it's in debian-installer-utils
[03:35] <broder> (source package0
[03:35] <broder> )
[03:49] <smoser> broder, you are correct of course. thank you.
[07:48] <pitti> slangasek: no, I haven't seen it myself either
[07:48] <pitti> Good morning
[09:44] <cjwatson> smoser: I will deal with your bug today
[09:45] <cjwatson> smoser: oh, you've already submitted an MP, cool
[10:11] <Laney> Riddell: Can you fix ~kubuntu-dev to have ~developer-membership-board as owner or admin please?
[10:25] <micahg> pitti: can you please copy firefox,mozvoikko,ubufox from {natty,oneiric}-proposed (bug 906389) to {natty-oneiric}-security
[10:33] <pitti> micahg: sure, copying to -updates as well
[10:33] <micahg> pitti: yep, that's fine :)
[10:36] <pitti> micahg: all done
[10:36] <micahg> pitti: thanks
[10:58] <chrisccoulson> yay \o/
[10:58] <chrisccoulson> now we're only 3.5 weeks until the firefox 10 release ;)
[10:59] <micahg> chrisccoulson: we're still ahead of Mozilla pushing out to everyone :)
[11:00] <chrisccoulson> heh
[11:00] <chrisccoulson> and, 3.5 weeks to fix this KDE patch!
[11:00] <chrisccoulson> g'ah
[11:01] <chrisccoulson> it's way too late in the beta cycle to be making significant code changes
[11:01] <micahg> I'm hoping we'll have an upstream powerpc FTBFS fix by then
[11:02] <chrisccoulson> if we care about it, then we should just fix it rather than waiting for somebody else to do it ;)
[11:02] <micahg> glandium was working on it, I'll check in with him next week (he had a partial fix, but something was wrong), I filed a tracking bug for it
[11:41] <sveinse> I'm here because I don't know where else to ask. I hope it not OT:
[11:41] <sveinse> I'm trying to apt-get into a staging directory (on a Ubuntu system). I've used the apt config option RootDir to prefix into the staging dir. Then I run apt-get update. It downloads the indexes to the staging dir, however it fails at the end as it tries to access /var/lib/dpkg/lock without the staging dir prefix. Bug?
[11:41] <sveinse> On a Natty system, sorry
[11:42] <seb128> cjwatson, hey, is the .maintscript use documented somewhere?
[11:42] <seb128> i.e the format of the files, the version of dpkg needed, the pre-depends stuff etc?
[11:43] <seb128> somewhere like in a manpage or a wiki, I keep looking for examples on my disk every time I set one, I didn't find a reference documentation
[11:43] <cjwatson> seb128: Pre-Depends documented in dpkg-maintscript-helper(1), file format documented in dh_installdeb(1)
[11:43] <sveinse> When strace-ing the apt-get, I see that it uses the proper stage dir prefix as setup in the config, except at the end where it tries to access the host's dpkg lock file.
[11:44] <cjwatson> and in fact the latter documents the substvar needed so that you don't need to know the precise pre-depends version
[11:45] <seb128> cjwatson, ok thanks
[12:14] <pitti> jibel: is bug 911813 covered by the daily auto-upgrade tests?
[12:15] <pitti> jibel: I think I know what's wrong, but as it takes ages to reproduce, could we just wait for the next auto-run to verify?
[12:15] <pitti> The fix I'm doing is necessary anyway, I'm just only 90% sufe it's sufficient
[12:15] <pitti> "sure"
[12:18] <pitti> jibel: ah, and it's a race condition, too; if gdm gets configured before lightm, it happens, otherwise not
[12:44] <jibel> pitti, no, the test didn't discovered it. It's on my todo to find why.
[12:44] <pitti> jibel: might just have been lucky, it's a race condition
[12:45] <jibel> pitti, very lucky then because I reproduced it more than once.
[13:46] <mdeslaur> cyphermox: happy new year: 912758
[16:08] <krinetic> just filed bug#912818
[16:33] <smoser> bug 912818
[16:42] <krinetic> changed it to affect gnome-settings-daemon
[16:53] <bdmurray> pitti: might bug 911834 be related to bug 905602?
[16:59] <pitti> bdmurray: at first sight it's a different problem; maybe over three corners
[16:59] <pitti> bdmurray: I'll follow up with a question
[17:00] <bdmurray> pitti: are there likely to be lots of dupes of 905602 though?
[17:00] <pitti> bdmurray: asked and sub'ed
[17:00] <pitti> bdmurray: I did a bug search for _decode_value, not that many actually
[17:00] <pitti> most of them duped themselves through apport
[17:00] <bdmurray> pitti: ah, great.
[17:03] <pitti> good night everyone, see you in Budapest!
[17:45] <SpamapS> you know what is *blatantly* missing from https://wiki.ubuntu.com/UbuntuMainInclusionRequirements  ? *documentation*
[17:48] <psusi> cjwatson: say, at what point do you think Ubuntu should drop grub-legacy?
[17:49] <cjwatson> psusi: I'm in no rush
[17:49] <psusi> why not?  grub2 has been in since hardy... ;)
[17:49] <cjwatson> there are still enough corner cases of one kind or another; I don't see a need to deliberately break them
[17:50] <ogra_> we should drop it as soon as we drop lilo ;)
[17:50] <psusi> hrm... is there a list of them so that grub2 can be fixed to handle them?
[17:50] <psusi> well, lilo still has upstream support doesn't it?  so that's just a matter of preference
[17:50] <cjwatson> not coherently.  xen support is a big one, which I know is in the works
[17:50] <cjwatson> but there are probably various odd bugs
[17:51] <psusi> xen support?  what's wrong there?  I thought we recently got a xen package that you just install and it sets up grub2 to load the xen hypervisor as the kernel and the kernel and initrd as modules and off it goes?
[17:52] <cjwatson> grub2 doesn't have any xen support
[17:52] <cjwatson> as in guest
[17:53] <cjwatson> I'm not talking about the hypervisor
[17:53] <psusi> I thought the guest didn't use grub... that the xen tools just loaded the kernel and initrd directly
[17:53] <cjwatson> pvgrub
[17:54] <psusi> ohh... hrm... interesting...
[17:54] <cjwatson> hence e.g. grub-legacy-ec2
[17:54] <cjwatson> anyway, I don't see any reason to rush.  It takes negligible amounts of my time to maintain and it's no longer the default so it shouldn't get in people's way particularly
[17:55] <cjwatson> "maintain" i.e. leave alone and occasionally fix build bugs or whatever
[17:57]  * psusi wants to whack the bug list ;)
[18:00] <cjwatson> they aren't hurting anyone
[18:00] <psusi> cjwatson: which partman package deals with copying partitions?
[18:01] <cjwatson> partman-partitioning has the UI for it; the actual copy is done by a COPY_PARTITION command to parted_server (in partman-base) which is mostly implemented in libparted
[18:01] <cjwatson> (whether that works in libparted 3 I don't know)
[18:37] <infinity> @pilot in
[18:37] <infinity> I knew I forgot something today.
[19:10] <slangasek> jodh: bug #912558> maybe it's noteworthy that in his screenshot before the crash, he's getting a lot of spam from udev due to a broken rule in /etc/udev/rules.d?
[19:14] <jodh> slangasek: can you point me at the screenshot you're looking at as the one on that bug doesn't mention udev (?)
[19:14] <jodh> slangasek: the good news is that daviey and I have -- we believe -- just identified the problem causing the assert failure :)
[19:15] <slangasek> jodh: ah, what's the problem?
[19:15] <slangasek> jodh: the one showing udev is this: http://bootie.daviey.com/~dave/erk/precise-init-crash.jpg
[19:15] <slangasek> apparently Daviey trimmed the context when posting to the bug!  Bad bug submitter! :)
[19:16] <jodh> slangasek: that seriously hurts my eyes :)
[19:17] <jodh> slangasek: if upstart attempts to run a single-line *non-existent* command with no shell meta-characters in it, the assertion fires.
[19:17]  * slangasek blinks
[19:17] <slangasek> how specific!
[19:17] <slangasek> so he has some non-standard job on his system that's triggering it?
[19:17] <jodh> slangasek: yeah!
[19:18] <slangasek> oh, because that's tied in with getting a different return value from exec(), right
[19:18] <jodh> slangasek: I have a fix for it but it'll need atleast 1 new build test and 1 new runtime test (so prolly tomorrow).
[19:18] <slangasek> ok
[19:18] <jodh> slangasek: yes, so the child process fails in that specific instance and the io watch therefore has an invalid fd to watch, hence the follow-on "bad file descriptor" error.
[19:20]  * slangasek nods
[19:21] <jodh> slangasek: so, you hit this bug if you've removed a package but the .conf files persist, as they refer to non-existent binaries.
[19:22] <jodh> slangasek: I'll update the bug with this and a script folks can run to convince me I've fully understood the problem.
[19:49] <SpamapS> jodh: nice work finding it (and hiding from your arch-nemesis, jhunt ;)
[19:53] <jodh> SpamapS: thanks. Yeah, he shadows me everywhere. Keep catching glimpses of him in the mirror... He *so* needs a haircut.
[20:18] <dupondje> Its still planned to removed vinage & rdesktop and switch to remmina & freerdp ?
[22:03] <SpamapS> does anyone else get a nice warm fuzzy when they turn on the test suite for a package build? Like.. you have suddenly done the security team and the users a nice big favor that you didn't have to? ;-)
[22:03] <kees> +1
[22:04] <SpamapS> now if only java test suites didn't take *days* to run. :-P
[22:04] <chrisccoulson> it's ok until the tests start hanging your builds ;)
[22:04] <SpamapS> yeah I think ZK's test suite may literally take days to run on arm. :-P
[22:08] <lifeless> SpamapS: you need a beowulf cluster of armhf then, obviously
[22:08] <SpamapS> lifeless: btw, we still can't triage bugs in the server team. :(
[22:08] <lifeless> SpamapS: the bug is still open and escalated, yes ?
[22:09] <SpamapS> lifeless: yes
[22:09] <lifeless> good good
[22:10] <SpamapS> I know there's nothing more we can do.. but.. SQL fail is so painful to see. :-P
[22:14] <jdstrand> SpamapS: every time :)
[22:14] <SpamapS> 60% of the time, it fails, every time
[22:15] <jdstrand> I was responding to the good feeling
[22:15] <jdstrand> :)
[22:16] <SpamapS> ah good
[22:16] <SpamapS>     [junit] Running org.apache.zookeeper.test.QuorumTest
[22:16] <SpamapS>     [junit] Tests run: 14, Failures: 0, Errors: 0, Time elapsed: 185.082 sec
[22:17] <SpamapS> paging performance? performance? no, sit down java