[06:06] <didrocks> robru: excellent explanation on the changelog, thanks for doing that :)
[08:04] <Laney> gooooooooooood morning desktoppers
[08:04] <Laney> happy friday
[08:09] <didrocks> happy frrrrrrrriday Laney!
[08:09] <Sweet5hark> Happy Friday everyone!
[08:10]  * Sweet5hark packs stuff for german LibreOffice community weekend.
[08:10] <didrocks> hey Sweet5hark!
[08:11] <Sweet5hark> (... and is depressend by LibreOffice4.3/utopic failing on the buildds in tests, but working fine in a local pbuilder)
[08:11] <Sweet5hark> well, actually LibreOffice 4.2 is failing on the buildds too.
[08:11] <Laney> hey didrocks et Sweet5hark
[08:12] <Sweet5hark> (which used to work fine)
[08:12] <Laney> there's some kind of law here
[08:13] <Laney> distro developers spend x% of time fixing tests of which y% is actually not the fault of the tests but the environment
[08:13] <Laney> where x and y are large
[08:13] <seb128> good morning desktopers, happy friday!
[08:13] <Sweet5hark> seb128: morning!
[08:13] <seb128> hey didrocks Laney Sweet5hark
[08:14] <Laney> hey seb128
[08:14] <Laney> happy friday!
[08:14] <seb128> Laney, well, at least pitti&co work on improvement the infra which should reduce y
[08:14] <Sweet5hark> Laney: The law is actually not new. We just have more controlled environments now to pinpoint issues IMHO.
[08:14] <Laney> it's things like no network in buildds
[08:15] <Laney> or different kernel versions
[08:15] <seb128> right :/
[08:15] <seb128> what's the solution? stop running tests?
[08:15] <seb128> (if you reply "yes" I'm going for some slapping!)
[08:16] <Sweet5hark> seb128: yes
[08:16] <Laney> USERS = QA!
[08:16] <seb128> Sweet5hark, lol
[08:16]  * seb128 slaps Sweet5hark
[08:18] <Sweet5hark> seb128: IMHO, we should move ~all tests to autopkgtests instead of during the build. Throwing away 24 hours of building because a tests ran in the wrong moon phase is bad. Test should obviously still be run and watched closely etc.
[08:19] <seb128> you are free to do that on your packages
[08:19] <seb128> I'm not sure I've much of an opinion on the topic
[08:19] <seb128> out of that tests shouldn't be flacky
[08:25] <Sweet5hark> seb128: well, this case is a bit different: the tests _are_ stable in my local pbuilder/upstream/everywhere else. Just not in the buildd which is fun to debug ...
[08:28] <Sweet5hark> speaking of which: does anybody know where the "Cannot create pipe for readline" in https://launchpadlibrarian.net/175001708/buildlog_ubuntu-utopic-i386.libreoffice_1%3A4.2.4~rc2-0ubuntu1~utopic1_FAILEDTOBUILD.txt.gz comes from?
[08:30] <Sweet5hark> what should happen in the case of a test crash is that it creates a core dump and parses the dump with gdb (all inside the LibreOffice build system) -- that would be rather helpful here.
[08:30] <Sweet5hark> I need to forced a test to crash in my local pbuilder to see if I get a crash here ...
[08:30] <Sweet5hark> s/get a crash/get a stack trace/
[09:04] <seb128> marga, hey, reopening https://bugs.launchpad.net/ubuntu/+source/evince/+bug/1016103 doesn't seem a good idea, if it's an apparmor issue neither the title nor the description, nor the tags, nor the subscribed people are correct, it's just going to sit ignored
[09:05] <marga> seb128, yeah, I had the feeling it was not the best course of action, but I'm kind of in a tight spot right now and don't have the time to report a new bug properly :-/
[09:05] <seb128> marga, ok, I'm just pointing it, don't expect much
[09:05] <marga> Sorry
[09:05] <seb128> marga, I don't know apparmor well, but to me it still seems you are in the case of https://wiki.ubuntu.com/DebuggingApparmor#Adjusting_Tunables
[09:06] <marga> It could be
[09:06] <marga> But why would adding the rule I said fix the issue, then?
[09:06] <seb128> no idea
[09:06] <seb128> is the issue specific to GNOME3?
[09:07] <seb128> or does it happen under Unity?
[09:07] <marga> I couldn't repro under Unity
[09:07] <seb128> even weirder
[09:07] <marga> Yeah, well.
[09:07] <marga> You could just close the bug if you think it will not get anywher.
[09:07] <marga> I won't get offended :)
[09:08] <seb128> well, if it's a real issue we should keep it open
[09:08] <seb128> it just that it's not defined well enough atm I think
[09:08] <seb128> so it's likely to get anywhere
[09:08] <marga> As I said, I'm currently under a lot of pressure for other things and I didn't have a lot of time, I just wanted to report the magic line
[09:08] <seb128> k
[09:09] <seb128> bug #1016103
[09:09] <seb128> marga, ^ I updated the title at least
[09:12] <marga> seb128, thanks
[09:14] <seb128> yw!
[09:33] <maxb> Is there any nicer way to configure autostart apps (the ones in /etc/xdg/autostart/) other than first editing the files there to remove NoDisplay=true, so that they show up in gnome-session-properties ?
[09:35] <seb128> not really no
[09:35] <Laney> You can copy them to ~/.config/autostart/ and remove NoDisplay=true in there ...
[09:35] <maxb> Oh. Hiding these appears to be an unhelpful Ubuntu diff :-/
[09:35] <seb128> Laney, "I've got a followup because the same/different toggle is buggy" ... mean you are going to propose a branch?
[09:35] <Laney> yes
[09:36] <seb128> cool, waiting to organize a landing then
[09:36] <seb128> "unhelpful" ... it's helpful, it prevents users to shoot themself in the foot by disabling an essential session service without understand why it's there
[09:36] <Laney> there's the mp
[09:37] <Laney> oh I forgot to prereq it on his one, hang on
[09:37] <maxb> seb128: Well.... the situation is such that so much has been hidden to protect the user from themselves that gnome-session-properties is now entirely useless
[09:38] <seb128> it's not
[09:38] <seb128> it's there to add things
[09:38] <seb128> not to remove system components
[09:38] <maxb> On my fresh trusty install, ironically the only displayed item is indicator-application
[09:38] <seb128> it's meant to be an empty list where users add e.g tomboy
[09:38] <seb128> indicator-application shouldn't be listed, that's a bug
[09:39] <maxb> I suppose the current situation is an uneasy compromise because there wasn't time to add a GUI selection to opt out of the hiding
[09:42] <seb128> why would you want to display the things that are hidden?
[09:42] <maxb> To turn off the ones I don't want (like gnome-keyring-ssh)
[09:43] <maxb> So that traditional openssh ssh-agent can take over
[09:55] <maxb> Hm, apparently gnome-session-properties has been removed upstream anyway
[09:58] <seb128> your ssh think seems like a bug
[09:59] <seb128> not something you should have to fiddle with autostarts to make work
[09:59] <seb128> think->thing
[09:59] <seb128> especially that some of those services are upstart jobs now
[10:01] <maxb> It is hard to see where this is configured, yes.
[10:01] <maxb> The ssh-agent I actually use is the one launched by /etc/X11/Xsession.d/90x11-common_ssh-agent
[10:01] <xnox> maxb: what's wrong? we run user sessions with upstart, so you can use upstart jobs in ~/.config/upstart/
[10:01] <seb128> what's the issue with the gnome-keyring one?
[10:02] <xnox> maxb: ssh-agent is started by upstart. and in utopic upstart also starts the gnome-keyring agents (well makes the environment variables exported the right way)
[10:02] <xnox> maxb: we didn't SRU it into trusty yet.
[10:03] <xnox> what does "$ status ssh-agent" say?
[10:04] <maxb> Oh, hmm, it does say start/running there. So perhaps what I mean is that I *used* to use the ssh-agent started by /etc/X11/Xsession.d/ but this has been taken over by upstart without me noticing in trusty
[10:05] <maxb> As to why I don't use the gnome-keyring agent, it's partly because I want something I can interact with in all cases when SSHed into my workstation remotely, and partially because it has a tendency to fail to sign auth challenges if you initiate lots of SSH sessions at the same time
[10:06] <xnox> maxb: default in pre-upstart sessions was ssh-agent starts, then gnome-keyring starts and takes over the agent vars. What are the symptoms? you do login with password and you never used to be asked for ssh passwords, but now you get a GTK+ dialog for ssh password?
[10:07] <xnox> maxb: to disable gnome-keyring, you need to copy it's autostart file, into ~/.config/xdg/autostart (i think) and set an extra var hidden=true, if that's what you think the problem is.
[10:07] <xnox> we only have upstart user-session in "normal" desktop logins, not when login in via tty/ssh/etc.
[10:08] <maxb> Right - that's what I've been doing for the past few releases. It just occurred to me to see if there was a better way than copying .desktop files around
[10:09] <xnox> not that i know off, per spec it can be a two-liner file with the matching name. But gnome-session rejects files if they lack e.g. Name arg and a few others, hence the copy & paste
[10:10] <maxb> It looks like gnome-tweak-tool would be upstream's designated place for this sort of config these days
[10:10] <xnox> *facepalm*
[10:11] <xnox> in the future of running user sessions under systemd, you'd be able to symlink a unit file to /dev/null to disable it.
[12:50] <om26er> mvo, Hi! I am getting this error http://paste.ubuntu.com/7601417/ when i try to add a ppa on my phone. help ?
[13:07] <mvo> hey om26er - does it happen for all ppas or just for a specific one?
[13:07] <mvo> om26er: hm, actually - that looks a lot like some sort of corruption, it fails to import
[13:08] <mvo> om26er: what does "python3 -c 'from xml.sax.saxutils import escape'" say?
[13:08] <mvo> (or maybe just python instead of python3)
[13:08] <om26er> mvo, works fine, no exceptions
[13:08] <om26er> and phone only have python3
[13:08] <mvo> om26er: both? py3 and py?
[13:08] <mvo> ok
[13:09] <ogra_> did you make the image writable ?
[13:09] <om26er> ogra_, its writable, yes
[13:12] <seb128> kenvandine, mterry: hey, could you review https://code.launchpad.net/~laney/ubuntu-system-settings/background-toggle/+merge/222298 ?
[13:13] <mvo> om26er: hm, badmarshal usually indicates a corrupted .pyc file
[13:14] <mterry> seb128, I'm sort-of off today, but could review this weekend if ken doens't
[13:14] <seb128> mterry, no worry, can be next week
[13:14] <seb128> mterry, don't connect to IRC on your days off! and enjoy the w.e ;-)
[13:43] <kenvandine> seb128, i'll look at it
[13:44] <kenvandine> mterry, stay away from IRC :)
[13:44] <seb128> kenvandine, thansk
[13:44]  * kenvandine shouldn't have highlighted him again :)
[13:44] <mterry> :)
[13:55] <popey> Has anyone tested ubiquity on a 14.10 daily live cd recently? I can't get ubiquity to start - gives http://paste.ubuntu.com/7601879/ "pkexec must be setuid root"
[14:00] <xnox> popey: hm, so latest desktop image that was promoted was in may.
[14:00] <xnox> http://cdimage.ubuntu.com/daily-live/current/
[14:00] <xnox> (20 - 22 May)
[14:01] <xnox> thus pending are borked in automated testing, let me check.
[14:01] <popey> thanks
[14:01] <kenvandine> seb128, approved
[14:02] <kenvandine> seb128, can you arrange landing of that gsettings-qt fix?
[14:04] <popey> xnox: pretty sure I tried the ones from 20/22 may and had the same problem
[14:05] <xnox> popey: the image itself, has pkexec with setuid bit set in the squashfs. Botting to see what's happening in the live system.
[14:09] <popey> xnox: okay, double checked, the amd64 image from 2014-05-20 loads ubiquity
[14:13] <seb128> kenvandine, sure
[14:15] <chrisccoulson> seb128, i think i've figured out bug 1307709
[14:15] <xnox> X starts, panel loads, unity-settings-daemon does not come up.
[14:15] <chrisccoulson> good news, it's not in oxide \o/
[14:16] <chrisccoulson> bad news, it's because the mir egl backend in mesa is incomplete
[14:16] <seb128> chrisccoulson, good that we have a team working on mir ;-)
[14:22] <chrisccoulson> seb128, yeah
[14:23] <chrisccoulson> seb128, https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/1307709/comments/13
[15:10] <Laney> seb128: you uploading u-s-s?
[15:11] <seb128> Laney, yes, is that ok?
[15:11] <Laney> yep
[15:11] <Laney> just checking
[15:12] <seb128> Laney, https://launchpad.net/~ci-train-ppa-service/+archive/landing-020/+packages if you want to check/test it
[15:12] <Laney> oh it's already there
[15:12] <seb128> kenvandine, can you test https://launchpad.net/~ci-train-ppa-service/+archive/landing-018/+packages ?
[15:12] <kenvandine> seb128, sure
[15:12] <seb128> thanks
[15:18] <popey> xnox: would you like me to file a bug for the ubiquity issue?
[15:19] <xnox> popey: yeah. It does launch, and then idle's in perl process and i don't know why yet.
[15:19] <popey> ok
[15:21] <popey> xnox: the boot splash menu is broken too
[15:22] <xnox> popey: we have a bug about boot splash menu already
[15:23] <kenvandine> seb128, works great!
[15:23] <seb128> kenvandine, thanks for confirming
[15:24] <kenvandine> np, i'll look forward to that landing and making my click hook reliable :)
[15:26] <popey> xnox: ok
[15:40]  * dro|desrt eyes bugzilla
[15:50] <didrocks> ok, time to run (was trapped in work, didn't see the time), and EOW! :)
[15:50] <didrocks> Monday is be off in France, see you on Tuesday!
[15:51] <Laney> bye, happy weekend!
[15:51] <didrocks> thanks, you too Laney!
[15:54]  * dro|desrt digs into the codepath that keeps giving back -- signal handler closure invalidation
[16:11] <chrisccoulson> seb128, i just had oxide running in desktop unity8 :P
[16:12] <chrisccoulson> (although, I had to disable the form factor detection, as it still crashes there)
[16:12] <seb128> chrisccoulson, nice!
[16:19] <dro|desrt> larsu: quick signoff?  https://bugzilla.gnome.org/show_bug.cgi?id=731339
[16:21] <dro|desrt> larsu: thoughts also welcome on https://bugzilla.gnome.org/show_bug.cgi?id=730296 if you're looking for a larger time investment :)
[23:27] <Noskcaj> I'm making a MIR bug for uhttpmock, which libgdata now build-depends on. Would this team be willing to be the "owning" team
[23:27] <Noskcaj> ?
[23:44] <Noskcaj> bug t
[23:44] <Noskcaj> bug 1327458