[07:48] <dkessel> Guten Morgen!
[07:50] <elfy> morning dkessel :)
[07:52] <elfy> DanChapman: if you get a chance could you install xubuntu in a vm and see if you're affected by bug 1284635
[07:52] <elfy> you being an English people that I know :)
[07:54] <elfy> or lubuntu or ubuntustudio ...
[07:54] <elfy> all affected by it here
[08:11] <elfy> seems to be ibus - purge that and no problem with keyboard layout
[08:25] <jibel> good morning
[08:25] <pitti> hey jibel
[08:25] <pitti> jibel: FTR, yesterday's a-u-t fixes made the "services" test actually work in VM
[08:25] <jibel> bonjour pitti
[08:26] <pitti> jibel: this morning I fixed the races and some missing blacklisting, so they are back to green
[08:26] <jibel> pitti, thanks!
[08:26] <pitti> http://d-jenkins.ubuntu-ci:8080/view/Upgrade/
[08:26] <pitti> jibel: so we are back to the mysteriously failing q->s tests
[08:26] <jibel> pitti, I looked at q->s but it doesn't make sense
[08:26] <pitti> (whoopsie not running)
[08:26] <pitti> jibel: that's still due to ubiquity-dm?
[08:27] <jibel> pitti, yes, when the initial container is provisioned ubuntu-dekstop (I guess) pulls ubiquity, casper and their friends
[08:28] <jibel> pitti, from apt history the command is: apt-get install --allow-unauthenticated -y ubuntu-desktop language-pack-en linux-generic
[08:28] <jibel> pitti, but if I create a base container of quantal on wazn and run this command, then it's fine (ie. no ubiquity)
[08:29] <jibel> and I don't find the difference when I do it manually or with a-u-t
[08:29] <pitti> indeed, why would u-desktop pull in ubiquity..
[08:29] <jibel> I must certainly be missing something obvious
[08:29] <pitti> jibel: it doesn't happen in the other cases, is that maybe hardcoded in some scenario package lists?
[08:31] <jibel> pitti, it is not, look at /var/lib/lxc/upgrader-ubuntu-quantal-saucy-desktop-amd64/rootfs/var/log/apt/history.log
[08:31] <jibel> Start-Date: 2014-03-12  21:52:57
[08:31] <jibel> Commandline: apt-get install --allow-unauthenticated -y ubuntu-desktop language-pack-en linux-generic
[08:31] <jibel> and on the install list ubiqutiy is present
[08:34] <pitti> jibel: but it's listed as upgrade, or am I reading that wrong?
[08:34] <pitti> ah no, that's the apt pinning/automatic stuff
[08:35] <pitti> this is just wrong
[08:35] <jibel> ubiquity:amd64 (2.12.16, automatic)
[08:35] <jibel> automatic ie as a dep of something
[08:36] <pitti> it doesn't happen with that apt-get install in a quantal schroot
[08:36] <jibel> pitti, nor in a quantal container created on wazn
[08:36] <pitti> jibel: how do we build the base VM?
[08:37] <pitti> i. e. from any image which already has ubiquity?
[08:38] <jibel> pitti, it is create with lxc.create() (python bindings)
[08:38] <pitti> hm, just the "ubuntu" LXC template
[08:38] <jibel> with the ubuntu template
[08:39] <jibel> +d
[08:39]  * pitti runs sudo lxc-create -n quantal -t ubuntu -- -r quantal
[08:39] <pitti> and while that's running, grabbing some breakfast
[08:39] <jibel> pitti, I'll enable debug mode of the resolver on wazn if I cannot reproduce loally
[08:39] <jibel> +c
[08:40]  * jibel drops a bag of letters that people can use to complete my sentences
[08:49] <pitti> jibel: hm, apt-get install shows no ubiquity there, as expected; but I ran this with trusty's LXC, alderamin is saucy
[08:49] <jibel> pitti, right but I did the same on wazn and no ubiquity
[08:49] <pitti> but I don't see anything in wazn's /usr/share/lxc/templates/lxc-ubuntu that would install ubiquity
[08:51] <jibel> pitti, I'll enable pkgproblemresolver temporarily on wazn to at least understand how this package gets there
[08:51] <jibel> I also cleared the cache of lxc
[08:52] <pitti> I'm still trying to interpret Install: ubiquity-frontend-gtk:amd64 (2.12.16, 2.15.26)
[08:52] <pitti> those are clearly versions of quantal-updats to saucy
[08:52] <pitti> so perhaps "available versions"
[08:53] <pitti> jibel: ah no, I was completely wrong: this is only *installed* during upgrade, it isn't already present in the quantal base VM, right?
[08:54] <pitti> jibel: i. e. that log contradicts what you said earlier (that the initial apt-get install u-desktop pulls it in )
[08:54] <pitti> so I guess I'm  even more confused and probably need to reset and start over
[08:54] <jibel> pitti, it is installed during bootstrap, isn't it?
[08:55] <pitti> jibel: yes, apparently (in bootstrap.log)
[08:56] <pitti> jibel: but why does history.log have it as "install" with versions available from saucy?
[08:56] <pitti> that indicates that history.log is from dist-upgrade, when saucy sources are already enabled
[08:56] <jibel> ⟫ grep -n "Get:.*ubiquity" bootstrap.log
[08:56] <jibel> Get:1195 http://archive.ubuntu.com/ubuntu/ quantal-updates/main ubiquity-frontend-gtk amd64 2.12.16 [66.4 kB]
[08:57] <jibel> from http://d-jenkins.ubuntu-ci:8080/view/Upgrade/job/upgrade-ubuntu-quantal-saucy-desktop-amd64/62/artifact/results/bootstrap.log
[08:57] <jibel> it is the version from quantal
[08:58] <pitti> right
[08:58] <pitti> jibel: do we know, is that still from self.lxc_container.create()?
[09:00] <pitti> ah no, obviously not
[09:00] <pitti> it copied an existing rootfs from teh cache
[09:00] <pitti> this is BasePkg = ubuntu-desktop
[09:01] <jibel> pitti, no, it is after create, in UpgradeTestBackend line 150
[09:01] <pitti> i. e. from line 161
[09:03] <jibel> line 161 is only executed if AdditionalPkgs is set which is not for this profile
[09:03] <jibel> so it only installs BasePkgs
[09:04] <pitti> jibel: oh, I was looking at AutoUpgradeTester/UpgradeTestBackendLXC.py
[09:04]  * jibel pulls latest rev, and remove his previous comment
[09:04] <jibel> pitti, 161, you're right :)
[09:04] <pitti> jibel: ah yes, we were looking at the same thing
[09:04] <pitti> jibel: I guess this morning's race fix changed the lines, sorry
[09:05] <pitti> the "additonal:" typo ought to make it fairly easy to search in the log
[09:05] <pitti> but no occurrece
[09:05] <pitti> occurrence
[09:05] <pitti> (occurence? ocurrence? bah)
[09:06] <pitti> "no match"
[09:06] <jibel> occurrence
[09:13] <jibel> pitti, same problem locally on a trusty host
[09:13] <jibel> and no cache
[09:13] <jibel> with: sudo ./bin/auto-upgrade-tester -b UpgradeTestBackendLXC share/profiles/trusty/ubuntu-quantal-saucy-desktop-amd64
[09:47] <jibel> pitti, pkgproblem resolver isn't helpful at all, but at least I can reproduce even if I still cannot make sense of it
[09:47] <jibel> ~# apt-get install --simulate ubuntu-desktop|grep "Inst ubiquity"
[09:47] <jibel> Inst ubiquity-frontend-gtk (2.12.16 Ubuntu:12.10/quantal-updates [amd64]) []
[09:47] <pitti> jibel: can you check dpkg --get-selections after boostrapping the container?
[09:48] <pitti> jibel: i. e. is there perhaps anything marked for "want install" which shouldn't be?
[09:48] <pitti> jibel: does apt-get -f install after boostrapping want to install anything?
[09:48] <jibel> http://paste.ubuntu.com/7083778/ nothing marked for install
[09:49] <pitti> hm, LGTM
[09:49] <jibel> and apt-get -f install says: 0 mis à jour, 0 nouvellement installés, 0 à enlever et 0 non mis à jour.
[09:49] <jibel> ie. nothing to do
[09:50] <pitti> jibel: apt-cache policy ubuntu-desktop?
[09:50] <pitti> jibel: and do the apt sources look sane, or is there something which there shouldn't be?
[09:51] <jibel> # LANG=C apt-cache policy ubuntu-desktop
[09:51] <jibel> ubuntu-desktop:
[09:51] <jibel>   Installed: (none)
[09:51] <jibel>   Candidate: 1.287
[09:51] <jibel> and apt sources look sane
[09:52] <pitti> ok, so something in quantal actualy pulls in ubiquity
[09:52] <jibel> quantal/-updates/-security + main restricted
[09:52] <pitti> jibel: does it also happen with sudo apt-get install ubuntu-desktop^
[09:52] <pitti> jibel: i. e. installing the "ubuntu-desktop" task, not the package
[09:52] <jibel> indeed, but the resolver says nothing
[09:52] <pitti> (that's what the installer says)
[09:52] <pitti> err, does
[09:53] <pitti> ubiquity-frontend-gtk doesn't have an ubuntu-desktop task: in quantal
[09:53] <jibel> the task doesn't install the installer
[10:02] <pitti> jibel: (sorry, multiplexing)
[10:02] <pitti> $ apt-get install -s ubuntu-desktop|grep ubiqui
[10:02] <pitti> so that also doesn't install ubiquity in a schroot
[10:02] <jibel> pitti, no problem, I found the reason
[10:02] <pitti> jibel: but perhaps installing a task is better.. oh?
[10:02] <jibel> pitti, right, I fix the profile too
[10:02] <pitti> .. anyway, as that's more closely resembling what happens with installation
[10:03] <pitti> jibel: what was the problem?
[10:04] <jibel> pitti, gnome-control-center recommends indicator-power recommends indicator-applet|indicator-renderer and ubiquity-frontend-gtk provides indicator-renderer
[10:04] <pitti> aah
[10:04] <pitti> jibel: so that's why the task works, as the seeded stuff is handled first before alternatives
[10:04] <pitti> jibel: but why on earth wouldn't that happen in a schroot install?
[10:05] <jibel> pitti, maybe the initial set of pacakges installed in a chroot is a bit different and apt calculates a different installation path
[10:05] <jibel> (just a guess)
[10:06] <pitti> maybe, picking alternatives sounds a bit like throwing dice
[10:06] <pitti> jibel: ah no, I know
[10:06] <pitti> jibel: indicator-applet is in universe
[10:06] <pitti> jibel: and I bet your VMs only have main
[10:06] <pitti> jibel: so in my schroot it would just install indicator-applet
[10:06] <jibel> pitti, correct, let me try with universe
[10:07] <pitti> jibel: but anyway, I think the correct answer here is "install the task"
[10:07] <jibel> pitti, indeed, I'll fix that
[10:07] <pitti> jibel: as that's essentially what seeds and germinate are already figuring out
[10:07] <pitti> jibel: splendid; *phew*, what a weird bug to see from "whoopsie not running"..
[10:07] <jibel> phew, I couldn't sleep without know why on earth it installed the installer :)
[10:08] <pitti> it's always astonishing which kinds of bugs you find with new tests which were meant for something entirely different :)
[10:13] <jibel> pitti, just to close the topic and confirm your point with universe it works fine
[10:13] <jibel> pitti, thanks for your help
[10:13] <pitti> jibel: de rien
[10:13] <pitti> great
[11:27] <davmor2> Morning all
[13:55] <elopio> davmor2: do you have your phone with qt5.2?
[13:55] <davmor2> elopio: I will have after lunch that I'm going for now
[13:57] <elopio> davmor2: ok, ping me when you are back please.
[14:44] <davmor2> elopio: back
[14:45] <elopio> davmor2: can you please open the clock app and go to the timer tab
[14:45] <elopio> I see some NaNs there, would like you to confirm.
[14:45] <davmor2> elopio: indeed
[14:46] <davmor2> I think they are meant to be presets let me double check it
[14:46] <elopio> davmor2: and does it gets stuck and crashes for you?
[14:48] <davmor2> elopio: should be softboiled egg, al dente, Aeropress coffee not nan(a) nan(a) nan(a) batman
[14:48] <elopio> :)
[14:48] <elopio> thanks davmor2.
[14:49] <davmor2> elopio: clock doesn't lock up though
[15:29] <elopio> rvr, alesage: lets skip our meeting today, because the webbrowser uds sessions is important for us.
[15:30] <rvr> elopio: Ok
[15:30] <alesage> elopio, agreed
[17:55] <pitti> jibel: hm, so I have a working rabbitmq/swift autopkgtest worker script now, in just ~ 150 lines (including configuration and CLI parsing, etc.)
[17:55] <pitti> jibel: well, conceptually working, of course far from production ready
[17:56] <pitti> but I must say, I'm quite pleased (aside from not knowing how to get unique dir names)
[17:56] <elopio> rvr: I replied to your review in https://code.launchpad.net/~elopio/ubuntu-ui-toolkit/initctl_env_var/+merge/208612
[17:56] <rvr> elopio: Reading
[18:02] <jibel> pitti, that's excellent. Is your PoC available somewhere?
[18:58] <Guest33169> Hi
[18:58] <Guest33169> How do we make debian pkg with simple methods
[18:59] <Guest33169> debuild too
[19:03] <balloons> Guest33169, http://askubuntu.com/questions/1345/what-is-the-simplest-debian-packaging-guide
[19:19] <elopio> I'm going to take a long lunch.
[19:19] <elopio> bb in ~3 hours.