[06:02] <pitti> stgraber: ah, I was missing to set $UPSTART_SESSION, but it still doesn't get me any further wrt. ubuntu-app-launch; but I'll keep playing, thanks!
[06:21] <pitti> tedg, stgraber, dobey: so I set $DISPLAY (and tested with xeyes), started a session bus, exported $DBUS_SESSION_BUS_ADDRESS, started /sbin/init --user, exported $UPSTART_SESSION; initctl list works fine now, and I get some jobs launched
[06:22] <pitti> tedg, stgraber, dobey: but ubuntu-app-launch `ubuntu-app-triplet com.ubuntu.calculator` still fails with: Unable to find job 'application-legacy': GDBus.Error:org.freedesktop.DBus.Error.ServiceUnknown: The name com.ubuntu.Upstart was not provided by any .service files
[06:22] <pitti> /usr/share/upstart/sessions/application-legacy.conf is clearly present
[06:22] <pitti> application-legacy stop/waiting
[06:22] <pitti> $ initctl status application-legacy
[06:23] <pitti> initctl: Unbekannter Parameter: APP_ID
[06:23] <pitti> but I suppose $APP_ID is coming from ubuntu-app-launch
[06:24] <pitti> oh, and in the above "prep" list I forgot: mkdir /tmp/rd; XDG_RUNTIME_DIR=/tmp/rd, so that's done as well
[06:27] <pitti> I clearly see com.ubuntu.Upstart in d-feet too
[06:33] <dholbach> good morning
[07:02] <pitti> tedg: ah, I'm further now, untangling the session d-bus handling and env vars; it now fails with
[07:02] <pitti> $ ubuntu-app-launch `ubuntu-app-triplet com.ubuntu.calculator`
[07:02] <pitti>  
[07:02] <pitti> ** (process:6108): WARNING **: Unable to find keyfile for application 'com.ubuntu.calculator_calculator_1.3.277'
[07:02] <pitti> which key file is that looking for?
[07:05] <pitti> tedg: err, it's looking in /home/ubuntu/.cache/ubuntu-app-launch/desktop/com.ubuntu.calculator_calculator_1.3.277.desktop, /home/ubuntu/.local/share/applications/com.ubuntu.calculator_calculator_1.3.277.desktop, /usr/local/share/applications/com.ubuntu.calculator_calculator_1.3.277.desktop, etc.
[07:05] <pitti> tedg: shouldn't it be looking in /opt/click/... ?
[07:10] <pitti> tedg: ok, on the phone I see the symlinks in .cache/ubuntu-app-launch/desktop/ to /usr/share/click/preinstalled/; how do these symlinks get built?
[07:11] <AskUbuntu> reasons why you shouldnt install ubuntu-touch-1.0 | http://askubuntu.com/q/487021
[07:40] <pitti> tedg, dobey: yay! got it to work in a container, with http://paste.ubuntu.com/7689161/
[07:40] <pitti> tedg: so my remaining question is how to replace the .desktop symlinking with something "official"
[07:41] <pitti> tedg: or rather, why isn't u-a-l looking into the click's pkgdir for the .desktop?
[07:48] <mandel> Elleo, morning! I have just pushed an update to the qml for udm that uses a QVariantMap instead of a QMap<QString, QString> can you please test it??
[07:49] <mandel> Elleo, I have a small idea of your segfault, if you get one let me know (I have not fixed it on purpose )
[08:22] <Elleo> mandel: okay, great; building now
[08:23] <mandel> Elleo, superb, let me know how it goes, if you get the segfault I think is a matter of ensuring that the property does not use the download ptr before it is created, not hard at all
[08:23] <mandel> Elleo, I need to make a couple of changes in that qml api that gatox did not consider, but I want to unblock you first
[08:28] <pitti> tedg: hm, nevermind; must have been a weird artifact of my intermediate steps; it creates the link automatically now
[08:33] <mandel> Elleo, seb128 guys, you might know where I have to file the following bug => https://www.youtube.com/watch?v=xMY436yzjzk
[08:34] <mandel> Elleo, seb128 I have no idea of the project to target the bug too
[08:34] <mandel> to*
[08:34] <seb128> mandel, not sure, ubuntu-keyboard or mir/unity8 I guess
[08:35] <popey> know bug that the launcher is under the osk
[08:35] <popey> IMO the keyboard should dismiss whenever you tap outside an input field.
[08:35] <popey> you shouldn't need to manually swipe it away IMO
[08:35] <mandel> popey, yes, or at least let me mve it out of the way, do you know the bug number?
[08:35] <mandel> popey, to at least add a +1
[08:35] <popey> looking
[08:36] <pitti> dobey, tedg: so http://paste.ubuntu.com/7689335/ fully works now, great! thanks for your help
[08:36] <popey> mandel: bug 1090358
[08:36] <JamesTait> Good morning all; happy Monday and happy Let It Go Day! :-D
[08:36] <mandel> popey, thx
[08:36] <popey> Saviq: that bug ^ appears to be untouched for months.
[08:36] <Elleo> mandel: swiping the keyboard down got broken by the qt 5.3 upgrade, that's logged here: https://bugs.launchpad.net/ubuntu-keyboard/+bug/1332624
[08:37] <Elleo> mandel: it's on my list of things to look at today
[08:37] <mandel> Elleo, oh, great :)
[08:37] <AskUbuntu> Tasks-app - where is Trello | http://askubuntu.com/q/487039
[08:41] <Saviq> popey, oh it's being touched
[08:42] <Saviq> popey, with the current approach for composition there's no way to make it happen (launcher is part of shell, osk needs to be above shell)
[08:42] <Saviq> popey, but it will Just Work™ with the upcoming Qt compositor
[08:42] <popey> ☻
[08:43] <Elleo> popey: do you know which launchpad project contains the G+ app? I'm pretty sure this bug https://launchpad.net/bugs/1332828 is just that it isn't asking for the context_exchange permission (and oSoMoN wasn't hitting the problem because he was testing in the browser, which already has permission)
[08:45] <popey> Elleo: https://launchpad.net/click-webapps
[08:45] <Elleo> popey: thanks
[08:45] <popey> np
[08:57] <Wellark> brendand: would you have time to review https://code.launchpad.net/~unity-api-team/ubuntu-system-settings/morenetworks/+merge/220422 ?
[08:59] <Wellark> brendand: would need an Approved for the code. Will submit the MP checklist when packages are ready
[09:07] <brendand> Wellark, it's not so cool that we can land two features without a single test. let me have a look at it a bit and i'll make some suggestions. i understand some of the code might be a little tricky to test, but there are ways
[10:01] <Cimi> I have an issue with my nexus 4 and one pc
[10:02] <Cimi> I try connecting it and it keeps disconnecting with error in dmesg
[10:02] <Cimi> works fine on another pc
[10:02] <Cimi> other devices work fine on the same pc that has issues with the nexus 4
[10:02] <Cimi> (it always worked, apart that today is monday and not friday)
[10:03] <Cimi> errors like device not accepting address 4, error -71
[10:06] <Cimi> something like http://paste.ubuntu.com/7689670/
[10:09] <Cimi> maybe I used it too much
[10:24] <Elleo> mandel: haven't been able to test that branch with the browser yet (having to recompile the oxide download branch for qt5.3), but I did write a quick test app and was still able to trigger segfaults: http://paste.ubuntu.com/7689730/
[10:24] <Elleo> mandel: definitely seems to be due to accessing the m_download pointer early, as I can work around it by setting autoStart false, then only setting the headers after the downloadId changes
[10:26] <mandel> Elleo, as I suspected, the idea od the auto start makes a number of assumptions, don't worry I have an idea on how to deal with it, just wanted to make sure that it was indeed the issue
[10:26] <mandel> Elleo, is kinda ugly to do it in such a way, but hey.. is nicer for qml people :)
[10:26] <Laurynas> hi ;]
[10:27] <Laurynas> i wished to ask, if it is possible to install Ubuntu touch only on Nexus devices as listed in support list or to any android device?
[10:27] <mandel> Elleo, the rest of the properties do have the same issue
[10:27] <Elleo> mandel: ah, hadn't needed to set any of the other properties before
[10:28] <mandel> Elleo, well, I'll add this as a diff but but will fix it in the same branch
[10:28] <Elleo> okay
[10:28] <mandel> Elleo, I'll bully gatox in spanish for doing it like that hehehe
[10:28] <Elleo> heh
[10:59] <mandel> Elleo, can you try your test again with the updated I did in the branch?
[10:59] <tvoss> popey, ping
[10:59] <Elleo> mandel: sure
[11:00] <popey> tvoss: pong
[11:00] <ogra_> (advantage: popey)
[11:00]  * popey puts wimbledon on in the background
[11:00] <popey> properly summer now
[11:00] <ogra_> :)
[11:00] <lotuspsychje> gallery app fixed on update on my n7 now tnx
[11:01] <mandel> Elleo, it should be storing the properties in memory and will set them as soon as we have the internal object added, it should be doing the request with the headers you set
[11:01] <tvoss> popey, just playing around with the calculator. Trying: 1 + 1, then hitting "=" multiple times -> calculation takes like 3 seconds
[11:01] <popey> oof
[11:01] <tvoss> popey, seems like it is not storing intermediate results but instead walks the stack each and every single time
[11:01] <ogra_> math isnt easy !
[11:02] <popey> tvoss: thanks, will file a bug
[11:02] <tvoss> popey, thx
[11:02] <Elleo> mandel: okay, cool
[11:02] <Elleo> mandel: just got to wait for it to rebuild
[11:03] <mandel> Elleo, I need to add tests to that qml.. there are none :-/
[11:03] <mandel> mierda
[11:23] <pitti> tedg: so I still have trouble with nonexisting .cache/ symlinks; what is supposed to create them?
[11:46] <pitti> tedg: ok, I filed bug 1333215 with some more info, a reproducer, and a curious observation
[11:53] <pitti> tedg: ok, I think I fully figured it out now, bug updated
[11:56] <jibel> Saviq, I added a test case to bug 1333142
[11:56] <jibel> browsing app previews is enough to make it crash
[11:58] <Saviq> mhr3, looking at the trace this would be yours ↑
[12:02] <mhr3> Saviq, :/ not much that can be done with such a broken trace
[12:02] <Saviq> mhr3, yeah, but there's steps!
[12:07] <jibel> mhr3, I cannot produce a better trace, the stack is always corrupted (bug 1333142)
[12:08] <jibel> er, bug 1325503
[12:08] <mhr3> Saviq, can't rep, the only thing that seems to break is image loading after a while
[12:08] <Saviq> mhr3, :|
[12:09] <mhr3> breaks == queue of stuff to download gets too large and i'm impatient
[12:10] <mhr3> jibel, install the dbg symbols then?
[12:56] <mterry> kgunn, so are you leery of landing the wizard with the current delay-before-unity8 (DBU)?
[12:57] <kgunn> mterry: not leery, we should land, i just was just curious if there might be a simple way to address...
[12:58] <kgunn> mterry: i did a little negative testing and it seemed ok...did you as well ?
[12:58] <mterry> kgunn, there are things we could do to address the black screen after wizard.  (That's the wizard shutting down, setting up language, etc.)  The part with the spinner is just optimizing unity8 startup
[12:59] <kgunn> e.g. wrong password,
[13:00] <mterry> kgunn, yeah I've poked around with it a bit
[13:00] <K1773R> What is the current status on the sound issue with N5?
[13:00] <kgunn> mterry: its definitely functional, maybe we should land and tune after the fact
[13:02] <mterry> kgunn, that's my preference too.  There is known work (black screen after and privacy screen) to do on it.  But better to do small fix-up landings than keep the whole thing out of image
[13:03] <kgunn> +1
[13:03] <mterry> Last piece is to get the testing scripts updated to disable the wizard before running AP tests.  I'll talk to fginther today about it
[13:10] <dobey> pitti: huh. ubuntu-app-launch actually worked for you there?
[13:10] <pitti> dobey: yes, I figured it out now, modulo the above bug 1333215 (workaround is in the bug)
[13:11] <pitti> dobey: that also has the complete instructions how to launch a click app in a container
[13:11] <dobey> pitti: i kept getting "Permission denied" errors with ubuntu-app-launch :(
[13:11] <pitti> dobey: hmm, I encountered a lot of stumbling blocks, but not that one..
[13:12] <pitti> dobey: did you try to bind mount /run perhaps, and ran into ubuntu (container) vs. dobey (host) UID mismatches?
[13:13] <dobey> pitti: no, only /tmp/.X11-unix and /home are bind mounted in the container
[13:13] <pitti> dobey: ah, so supposedly the ubuntu user in the container fails to write into your $HOME? I didn't bind-mount $HOME
[13:13] <dobey> pitti: part of the problem was that nested apparmor profiles don't work, but i commented out the bits that try to load the profile
[13:13] <pitti> dobey: I tried with bind-mounting /tmp/.X11-unix (then I see calculator), and with xvfb (for running autopilot tests)
[13:14] <dobey> pitti: no, i am the same UID in the container
[13:14] <pitti> dobey: heh yes, I have a "rm /sbin/apparmor_parser" to avoid click install failure
[13:14] <pitti> ugly, but I didn't find anything else, except for running click install twice (it works the second time)
[13:15] <dobey> pitti: well i've got the click installed and everything, and i can run the app manually, but get permission error from ubuntu-app-launch, in the application-click job
[13:15] <pitti> dobey: hm, strace it and see what it's trying to do?
[13:15] <dobey> yeah, i tried but i couldn't find the failure :(
[13:17] <pitti> dobey: ah, so it's from the upstart job itself, not from u-a-l
[13:17] <pitti> dobey: anything in .cache/upstart/whatever.log ?
[13:17] <pitti> that has the output from the job
[13:21] <dobey> pitti: maybe it couldn't open the display or something
[13:22] <pitti> dobey: that worked for me (with --bdir /tmp/)
[13:22] <dobey> pitti: i'll have to get my lxc back up and check again. i was doing all this on friday, and that was before i upgraded my CPU and monitor :)
[13:50] <K1773R> What is the current status on the sound issue with N5?
[14:06] <jdstrand> why are you removing apparmor_parser? that (or reimplementing apparmor_parser to return 0) is only valid if you are testing installation of packages. as soon as you want to run a click, this is no longer a valid test environment
[14:06] <jdstrand> (because the app should be running in an environment that replicates what the user sees, which means under confinement)
[14:07] <ogra_> K1773R, rsalveti is working on changing the audio backend architecture ... if you are lucky it fixes it for N5
[14:07] <jdstrand> not to mention, I would think that UAL would fail because it can't change_profile into the (currently nonexistent/not loaded) profile
[14:07] <rsalveti> should fix, but I need to get some time to test on a n5 first
[14:11] <jdstrand> pitti, dobey: ^
[14:12] <pitti> jdstrand: yeah, this is mostly just for me for testing click tests in LXC
[14:12] <pitti> jdstrand: as LXC containers don't seem to implement the apparmor bits
[14:13] <dobey> jdstrand: because apparmor is broken in lxc
[14:13] <pitti> jdstrand: but at least while I develop autopkgtest click support (and maybe even for merge proposals), using LXC is soo much faster than using the emulator or real iron
[14:13] <dobey> pitti: lxc has apparmor, it just doens't work because it's nested under the host
[14:14] <pitti> jdstrand: and as the click_apparmor hook package isn't uninstallable, I just ripped out that as that's what the hook checks for
[14:14] <dobey> and there's no support for nested apparmor yet
[14:14] <dobey> jjohansen1 was working on it, iirc
[14:14] <pitti> dobey: right, whatever; "doesn't work" from my high-level perspective :)
[14:14] <jdstrand> he is
[14:14] <jdstrand> he has been for a while, it is not an easy problem to solve, but we are solving it
[14:14] <pitti> jdstrand: oh, great to hear
[14:15] <jdstrand> after abstract socket mediation lands, he'll be back to it
[14:15] <pitti> jdstrand: so this works in a schroot where it sees "no apparmor support", but in LXC it has just enough to claim that it can support it, but doesn't, and that makes click install fail
[14:15] <jdstrand> it may be delivered this cycle, more likely 15.04 for full lxc stacking which will make it work the way you want
[14:16] <jdstrand> I'm surprised removing apparmor_parser worked
[14:16] <jdstrand> I would've thought you would've needed to redirect the binary and have the new one return 0
[14:17] <jdstrand> incidentally, it may be possible to setup a special container that is not itself confined by apparmor so that you could runs things in the container with apparmor if you got creative with aliases in the container. I'll let jjohansen1 comment on that if it is something you are interested in
[14:18] <jdstrand> (ie, it might be possible to do that today)
[14:18] <pitti> jdstrand: apparmor/click.py apparmor_available() does if not os.path.exists(parser): raise AppArmorException
[14:18] <pitti> jdstrand: where parser == /sbin/apparmor_parser
[14:18] <pitti> jdstrand: I saw that, found that I can't easily take away /sys/kernel/security/apparmor, but can remove apparmor_parser
[14:19] <jdstrand> ok, so it exits with error but then chugs along
[14:19] <pitti> jdstrand: I already tried with aa.profile = unconfined, that also doesn't work
[14:19] <pitti> jdstrand: maybe apparmor_available() should additionally check for LXC ([ -e /run/container_type ]) until apparmor supports it?
[14:20] <pitti> I'd certainly prefer that over the rm hack
[14:22] <dobey> well i didn't rm apparmor_parser
[14:23] <dobey> i rm'd the hooks for click, so it just doesn't try to run those hooks on click install
[14:23] <dobey> and i commented out a line in the application-click hook that calls apparmor
[14:25] <jdstrand> seems not having click-apparmor installed is likely the way to go, no?
[14:25] <jdstrand> to have it installed and fail open doesn't seem right
[14:26] <pitti> jdstrand: can't, ubuntu-app-launch has a hard dep on it
[14:26] <jdstrand> anyway, we have a work item to think about and adjust this as necessary
[14:26] <pitti> (if we can fix that to be a Recommends:, that also sounds acceptable)
[14:27] <jdstrand> that is what I was getting at before. if click-apparmor can't load the profile, UAL can't change in to it
[14:27] <dobey> there are a lot of things i'v eseen in the ubuntu-touch or ubuntu-sdk package hierarchy that are hard depds, but should be recommends. :-/
[14:28] <dobey> i think UAL has a hard dep though, because of the upstart job
[14:28] <jdstrand> right. it ships two though, one for non-click and one for click
[14:29] <jdstrand> I think those come together in the same package, so depends is likely correct
[14:29] <dobey> right
[14:31] <dobey> trying to get upstart sorted in my lxc again
[14:38] <mdeslaur> ogra_: FYI, we're doing a silo that includes a new override file in ubuntu-touch-session for apparmor that now has an upstart job
[14:42] <Wellark> brendand: could we have a meeting tomorrow about the testing situation with NM related code?
[14:42] <Wellark> something like max. 30 minutes
[14:42] <brendand> Wellark, yes. i have a meeting with jgdx tomorrow anyway
[14:43] <brendand> Wellark, you can join that one
[14:43] <Wellark> brendand: great. what time?
[14:43] <Wellark> could you invite me and thostr_ and Satoris?
[14:44] <brendand> Wellark, 11UTC
[14:46] <Wellark> brendand: oh, thostr_ has a meeting overlapping..
[14:47] <ogra_> mdeslaur, to speed things up ?
[14:47] <mdeslaur> ogra_: to make sure that apparmor starts before confined apps try and start
[14:47] <mdeslaur> it was racy without an upstart job
[14:47] <ogra_> oh, that thing ... i remember
[14:47] <ogra_> cool, thanks !
[14:47] <mdeslaur> yeah, should have no impact
[14:48] <ogra_> yeah, it should have "improvement" :)
[14:49] <Wellark> brendand, jgdx: I proposed a meeting time for tomorrow
[14:50] <Wellark> please reply to the invite if the time is not suitable
[14:50] <jgdx> Wellark, ack
[14:51]  * dobey wonders if tassader found anything with the screen blanking issue
[15:01] <Wellark> pmcgowan: hi, could we get someone to look at this code wise https://code.launchpad.net/~unity-api-team/ubuntu-system-settings/morenetworks/+merge/220422
[15:01] <Wellark> we will figure out the testing side tomorrow
[15:02] <pmcgowan> Wellark, sure
[15:03] <Wellark> pmcgowan: thanks
[15:24] <pmcgowan> seb128, could you review Wellark's MR above or suggest who might?
[15:25] <seb128> pmcgowan, seems like brendand was reviewing it
[15:25] <seb128> he commented on it at least
[15:25] <seb128> I suggested cyphermox previously, but he's off this week
[15:25] <pmcgowan> seb128, he only asked for tests
[15:25] <pmcgowan> seb128, need a dev
[15:25] <seb128> tedg maybe can help there?
[15:26] <seb128> that's lot of code to review :/
[15:27] <seb128> especially for somebody who doesn't know n-m well
[15:29] <pmcgowan> seb128, maybe awe_ could have a look?
[15:29] <tedg> Hmm, or Satoris
[15:30] <tedg> I believe he was working with Wellark on that.
[15:30] <pmcgowan> I think Satoris wrote much of it
[15:30] <awe_> pmcgowan, which MR?
[15:30] <awe_> flight-mode?
[15:30] <pmcgowan> no network config
[15:30] <pmcgowan> awe_, https://code.launchpad.net/~unity-api-team/ubuntu-system-settings/morenetworks/+merge/220422
[15:31] <awe_> Wellark, what happened to the flight-mode silo?
[15:31] <awe_> pmcgowan, I could test it, but I'm probably not a good person to review the code
[15:32] <tedg> Ah, looking, probably charles or I can review that.
[15:33] <ogra_> tedg, i think i poked you about it already (but i'm not sure i really did)..... i have a WI to poke you about cleaning up application logs so that we only keep the current log around for a running app (since it eats a lot of diskspace to keep *all* of them)
[15:33] <Wellark> awe_: seems it got flushed
[15:34] <Wellark> thostr_: could we get it back? --^
[15:34] <tedg> ogra_, Hmm, seems "only current running" is too little no?
[15:34] <awe_> Wellark, also I did you see my latest comments about the signal strength?
[15:35] <thostr_> Wellark: is the urfkill thing fixed?
[15:35] <tedg> ogra_, Can we get logrotate to be more aggressive with application logs?
[15:35] <awe_> Wellark, to be clear there are issues with urfkill, however the bug reported Wellark was not confirmed
[15:35] <ogra_> tedg, well, feel free to do whatever you like ;)
[15:35] <tedg> ogra_, It'd be nice to have logs for a bit.
[15:35] <ogra_> the current situation is bad ...
[15:36] <awe_> ...and there's an issue that I reported with the FlightMode UI itself while testing the silo
[15:36] <tedg> ogra_, I have to wipe my device so often for silo testing I don't notice :-)
[15:36] <ogra_> tedg, (i'm caring for system logs, app logs are for you :) )
[15:36] <thostr_> Wellark: ^ have you fixed that one?
[15:36] <tedg> ogra_, Did you end up using cron for those?
[15:37] <ogra_> tedg, i'm planing to seed anacron for now ... long term we should have our own daemon that hooks into the android /dev/alarm infrastructure
[15:37] <ogra_> (even though i know lool is unhappy with me seeding a cron daemon)
[15:38] <tedg> ogra_, I understand, not ideal, but practical.
[15:38] <ogra_> right and reliable
[15:38] <tedg> ogra_, We're not worried about just application logs though, all user service jobs, right?
[15:39] <ogra_> yeah
[15:39] <Wellark> awe_, thostr_: we need to see the actual KillSwitch object states to conclude if it's an UI bug or not
[15:39] <ogra_> essentially all logs ... no matter where they occur
[15:39] <awe_> Wellark, I need a reproducible test scenario
[15:39] <awe_> which you haven't come up with yet
[15:39] <tedg> ogra_, Hmm, okay, seems like the cron job should be in… upstart?
[15:39] <awe_> there *are* known bugs, and I'm working on them
[15:39] <tedg> Not sure of a package there.
[15:40] <awe_> however your bug is still incomplete...
[15:40] <Wellark> awe_: you said that you saw the switch in the UI not to match what rfkill reports, right
[15:40] <ogra_> tedg, upstart ships a job for logrotate already in the user session
[15:40] <ogra_> tedg, but it will need adjustment
[15:40] <tedg> ogra_, Yeah, it only runs on session start, which is part of the problem. People don't reboot their phones enough. :-)
[15:40] <ogra_> right
[15:41] <tedg> We could just reboot their phone for them. That'd solve the problem ;-)
[15:41] <awe_> Wellark, please review the bug.  Your last comment indicated that you would re-review and provide more data
[15:41] <ogra_> tedg, great, go for it
[15:41] <ogra_> :P
[15:41] <tedg> No more memory leak issues either!
[15:42] <ogra_> yeah, just reboot every hour
[15:44] <Wellark> awe_: sure. will do.
[15:45] <tedg> ogra_, Oh, it looks like the session job will run on the system event "rotate-logs". So if you start issuing that, the session may just follow suit.
[15:46] <ogra_> ah thats from cron.hourly or some such
[15:47] <tedg> It's on cron.daily, "upstart"
[15:47] <ogra_> yeah
[15:48] <tedg> Seems like someone solved this already :-)
[15:48] <ogra_> well test it :P
[15:48] <tedg> ogra_, Can you ping me when you get anacron on the image?
[15:48] <ogra_> tedg, yup, should happen this week
[15:48] <tedg> Cool
[15:49] <mterry> fginther, heyo!  Any objections to landing https://code.launchpad.net/~mterry/ubuntu-test-cases/no-wizard/+merge/223768 ?
[15:49] <ogra_> tedg, so emitting that event seems to gzip the logs and rotate them but i fear we still end up with five gzipped logs per app or so
[15:50] <fginther> mterry, I need to make sure the new phablet-tools is installed in CI first. Will get back to you
[15:50] <ogra_> tedg, that should be at most one
[15:50] <ogra_> imho
[15:50] <seb128> mhr3, you should be on #ubuntu-devel
 sil2100: FYI, libunity-scopes2 now fails to install in -proposed due to "GError: Can not find a single database provider in /etc/click/databases"; this fails the tests for ubuntu-system-settings-online-accounts and unity-scope-click and thus holds back stuff
[15:50] <seb128>  ubuntu-system-settings-online-accounts and unity8 in particular
[15:51] <seb128> Saviq, ^ you might be interested by it, as a fyi, since it blocks unity8
[15:51] <pitti> seb128: ah, thanks for forwarding
[15:51] <seb128> pitti, yw
[15:51] <tedg> ogra_, Yeah, perhaps we adjust the log to differ between apps and session jobs. But let's make sure we're not overflowing first :-)
[15:51] <ogra_> it already differs
[15:51] <ogra_> upstart ships a session job
[15:52] <ogra_> which differs from logrotate's system job
[15:52] <tedg> I'm saying different between application*.log and indicator*.log
[15:52] <mhr3> seb128, i'd need to somehow convince my xchat to react to the auto-join checkbox... cause it knows better :/
[15:52] <mhr3> seb128, what causes that error?
[15:53] <mhr3> i have no idea tbh
[15:53] <ogra_> tedg, oh, i misunderstood :)
[15:53] <tedg> We're reasonably certain that our jobs aren't overly verbose. Apps we have less control over.
[15:53] <ogra_> right
[15:53] <seb128> mhr3, https://jenkins.qa.ubuntu.com/view/Utopic/view/AutoPkgTest/job/utopic-adt-unity-scope-click/lastBuild/ARCH=amd64,label=adt/console
[15:53] <seb128> mhr3, seems to happen at package installing
[15:54] <seb128> mhr3, http://paste.ubuntu.com/7690980/
[15:55] <mhr3> pete-woods, ping?
[15:55] <mhr3> you're the click expert
[15:56] <seb128> hum
[15:56] <seb128> could be a click issue
[15:56] <seb128> mvo, cjwatson: ^
[15:56] <seb128> https://launchpadlibrarian.net/178217792/click_0.4.26.1_0.4.27.diff.gz added an error case for that it seems
[15:56] <seb128> +		if (db.size == 0)
[15:56] <seb128> +			throw new FileError.FAILED(
[15:56] <seb128> +				"Can not find a single database provider in %s", real_db_dir);
[15:56] <mvo> seb128: that is my change, sorry for the trouble
[15:57] <cjwatson> OK, I think we should revert that bit?
[15:57] <mvo> cjwatson: yes
[15:57] <cjwatson> So that libclick is still usable even if click isn't installed
[15:57]  * mvo nods
[15:58] <mvo> I need to go to dinner, I can look at it later tonight or tomorrow early morning
[15:58] <cjwatson> mvo: I should have time for a quick revert
[16:00] <mhr3> seb128, uuuh, and you're blaming me... guess you owe me a beer now?! :P
[16:00] <seb128> mhr3, don't drink on the tap of the messenger!
[16:01] <seb128> mhr3, the days where you could redirect the blame on others in a consistent way is over, even if you sometime still manage to escape blames ;-)
[16:02] <mhr3> seb128, nope i still do it consistently, and you owe me another beer for doubting that :P
[16:02] <seb128> mhr3, lol
[16:02] <seb128> mhr3, alright, you won a beer for being funny ;-)
[16:02] <mhr3> \o/
[16:03] <pete-woods> mhr3: hi
[16:03] <pete-woods> click export??
[16:03] <mhr3> pete-woods, nvm, solved
[16:03] <pete-woods> mhall119: okay
[16:03] <pete-woods> mhr3: okay
[16:03] <pete-woods> autocomplete fail
[16:06] <Saviq> seb128, :/
[16:06] <K1773R> ogra_: thanks
[16:07] <cjwatson> Partial click revert building in silo 12
[16:09] <seb128> cjwatson, thanks
[16:18] <charles> ricmm, ping
[16:18] <charles> ricmm, earlier this year there was an indicator-datetime ticket for it to use the in-house audio stack to play the alarm sound, rather than using libcanberra
[16:19] <charles> ricmm, the discussion at the time was that you were working on a media API that would be the right fit for this
[16:20] <charles> this was quite some time ago, did the media API land s.t. datetime could use it for playing the alarm sounds?
[16:20] <charles> ricmm, https://bugs.launchpad.net/ubuntu/+source/indicator-datetime/+bug/1283065/comments/4
[16:21] <ricmm> charles: so in the end that dropped out of the scope for RTM, im sorry if it wasnt well communicated, we need to figure out a way to reach the media-api without much pain from desktop, or disable it altogether in there for now
[16:22] <ricmm> charles: would you care to start an email thread with jhodapp salveti and me for this matter?
[16:22] <ricmm> charles: or a call, tomorrow
[16:22] <charles> ricmm, from my pov with indicator-datetime, we could punt the "Use the existing audio stack to play event sounds" ticket to post-RTM
[16:23] <jhodapp> ricmm: charles: yeah we need add this...the plan right now is to work on this right after RTM for 14.10
[16:24] <charles> ricmm, ok. I'll start an email thread so that the topic moves forward, but let's not spend much pre-RTM time on it
[16:24] <charles> s/much/any/
[16:39] <ogra_> stgraber, i just dropped SWAP file creation from livecd-rootfs ... coudl you do the same with system-image (not sure where that bit of code lives)
[16:39] <stgraber> ogra_: as in, we don't want any swap at all or we just want it created at flashing time?
[16:39] <ogra_> none at all to make tvoss happy :)
[16:40] <stgraber> hmm, may I ask why? because that's pretty universaly considered as a very bad idea...
[16:40] <stgraber> the kernel expects you to have some swap or some things will get confused
[16:40] <stgraber> so before doing that kind of thing, you may want to talk to the kernel team first :)
[16:41] <ogra_> stgraber, OOM and the app lifecycle get confused by having it apparently
[16:42] <ogra_> it breaks the app lifecycle
[16:42] <ogra_> it was discussed several times
[16:42] <ogra_> OOM might need adjustment ... we'll take care for that after we have dropped SWAP
[16:42] <ogra_> (and have some real world data for it)
[16:43] <ogra_> kernel team was involved in discussions btw
[16:44] <stgraber> still seems horribly wrong, but whatever. The creation bit you want to change is in the android-bootable-recovery git branch
[16:44] <stgraber> and there may also be some initrd thing you need to change as well
[16:44] <ogra_> hmpf
[16:44] <ogra_> thats more than i thought
[16:44] <ogra_> (especially since it take me two days (with blocked internet) to sync the android tree here)
[16:45] <stgraber> just clone that one branch, it's pretty small
[16:45] <stgraber> actually, it may be that if you replace that mkswap by an rm in there, the initrd will just DTRT
[16:46] <stgraber> (and you should do an rm regardless, you don't want to waste that space if it's not used at all)
[17:03] <mpt> cyphermox, hi, does anything record how long a paired Bluetooth device has been paired for? (e.g. 1 hour 14 minutes)
[17:20] <mterry> fginther, thanks for merging!
[17:20] <fginther> mterry, np, it should be all ready to work now.
[17:20] <mterry> sweet
[17:21] <mterry> ogra_, so we look like we're capable of landing welcome wizard if we want.  When would be the best time to do so?
[17:23] <ogra_> mterry, any time :)
[17:24] <ogra_> mterry, i will need to bug you this week (not today anymore) abput some info we need from the greeter for the developer mode
[17:24] <mterry> OK, I'll do a final smoketest and change seeds
[17:24] <ogra_> *about
[17:24] <mterry> ogra_, OK
[17:42] <sil2100> boiko: hello!
[17:42] <sil2100> :)
[17:46] <sil2100> boiko: so... I forgot to mention that before, but we seem to be having new autopilot failures in dialer-app and messaging-app
[17:46] <sil2100> boiko: could you maybe take a look on those?
[17:48] <boiko> sil2100: hi, yes, do you by chance have a link to the output easy there?
[17:48] <sil2100> boiko: sure! Give me a moment
[17:48] <sil2100> boiko: http://ci.ubuntu.com/smokeng/utopic/touch/mako/94:20140623:20140530/8685/messaging_app/ and http://ci.ubuntu.com/smokeng/utopic/touch/mako/94:20140623:20140530/8685/dialer_app/
[17:49] <sil2100> boiko: thanks :)
[17:53] <sil2100> ogra_: https://plus.google.com/109159869108744115904/posts/jihdsBrHgYx
[17:53] <ogra_> sil2100, already shared :P
[17:53] <dobey> wtf
[17:54] <dobey> if "init" has a variable set in its environ, why would a child process of that init, not have that variable?
[19:30] <cwayne> anyone successfully make a go+qml click package for armhf?
[19:35] <mazzchri> hi
[19:35] <mazzchri> mtk development
[19:35] <mazzchri> ??
[20:02] <dobey> mazzchri: huh?
[20:45] <dobey> QQuickWindow: possible QQuickItem::polish() loop
[20:46] <dobey> ^^ anyone have any idea what that means?
[20:46] <ogra_> i guess Saviq does ... (it says "polish" :P )
[20:47] <dobey> heh
[20:58] <dobey> well i managed to get that message to disappear, though i've no idea why
[20:59] <dobey> it was complaining about the ListModel i had
[21:05] <cwayne> agh, im getting an "Invalid policy version" error from apparmor, but the policy version is just 1.1...
[21:05] <jjohansen1> cwayne: invalid policy version, likely means the cached binary policy is incompatible with your kernel
[21:06] <jjohansen1> cwayne: try recompiling the policy with /etc/init.d/apparmor restart
[21:07] <cwayne> jjohansen1: same error
[21:08] <jjohansen1> cwayne: what is the exact error message and where are you seeing it, console/terminal, syslog, dmesg?
[21:08] <cwayne> jjohansen1: the exact error is:ERROR: Invalid policy version for 'com.ubuntu.developer.cwayne18.aroscontrol_aroscontrol_0.2.json'. Skipping
[21:09] <cwayne> which i see either when trying to run the click system hooks, and i also see it in /var/log/upstart/click-system-hooks.log
[21:09] <jjohansen1> alright that narrows it down to the click wrappers
[21:11] <jjohansen1> cwayne: what is the policy version in the other click files?
[21:11] <jjohansen1> cwayne: /var/lib/apparmor/clicks/
[21:11] <cwayne> jjohansen1: 1.1
[21:12] <jjohansen1> jdstrand: ^ any ideas
[21:13] <jdstrand> 1.1 is the apparmor-easyprof-ubuntu policy version
[21:14] <jdstrand> cwayne: can you paste /var/lib/apparmpor/clicks/<path to affected file>.json
[21:15] <jdstrand> /var/lib/apparmpor/clicks/com.ubuntu.developer.cwayne18.aroscontrol_aroscontrol_0.2.json
[21:15] <cwayne> jdstrand: http://paste.ubuntu.com/7692251/
[21:15] <jdstrand> cwayne: can you paste the manifest too?
[21:16] <jdstrand> cwayne: ie, the click manifest
[21:16] <cwayne> jdstrand: http://paste.ubuntu.com/7692254/
[21:16] <jdstrand> "framework": "ubuntu-sdk-14.10-dev1"
[21:17] <jdstrand> you can't use 1.1 policy with 14.10
[21:17] <jdstrand> use 1.2
[21:17] <jdstrand> did you use the click-reviewers-tools on it? (it should have caught that)
[21:17] <jdstrand> well, either use 1.2 policy or keep 1.1 and use 14.04 framework
[21:22] <cwayne> ah, i didn't even know those existed!
[21:25] <cwayne> jdstrand: worked like a charm, thanks!
[21:25] <jdstrand> np
[21:25] <jdstrand> we have the concept of vendor and version for easyprof policy
[21:25]  * sarnold awaits the juju team saying "the charm works like aa-easyprof!" hehe
[21:26] <jdstrand> we use version with our policy cause if we move stuff around it might not be valid on other frameworks
[21:26] <jdstrand> hehe