[06:23] <pitti> Good morning
[07:39] <seb128> wgrant, hey, still no vivid translations?
[07:45] <wgrant> seb128: Ah, I forgot to unhide them. Fixed.
[07:46] <seb128> wgrant, which means it's going to start importing?
[07:46] <wgrant> seb128: Yes.
[07:46] <seb128> wgrant, excellent, thanks!
[07:48] <pitti> Riddell: are you looking at the kate file conflict? (cf. the autopkgtest regressions)
[07:53] <pitti> mvo: do you have an idea about https://jenkins.qa.ubuntu.com/job/vivid-adt-system-image/3/ARCH=i386,label=adt/console ?
[07:55] <mvo> pitti: no, that is something that barry will know about
[07:55] <pitti> mvo: ack, thanks
[08:17] <highvoltage> n/win 22
[09:40] <themusicgod1> http://pastebin.com/XufeN6b8 tl; dr #ubunt+1 suggests I ask here.  Rogue firefox thread caused system to no longer be able to initialize video card on boot(acer white screen of death).  Suggestions online suggest to not suspend on laptop-close.  Would it be worth suggesting ubuntu use this setting for laptops of this kind?
[09:46] <themusicgod1> (firefox confirms that web content is a firefox process, but it was a non-ubuntu build of firefox/nightly)
[10:04] <darkxst> slangasek, I commented on bug 1391102
[10:05] <Laney> yofel: you aware of https://jenkins.qa.ubuntu.com/job/vivid-adt-kate/lastBuild/ARCH=amd64,label=adt/console and others?
[10:26] <Laney> Riddell: ^ or you
[10:52] <sil2100> @pilot in
[11:22] <LocutusOfBorg1> hi developers, can anybody please explain me why armhf is so different from debian?
[11:22] <LocutusOfBorg1> https://launchpadlibrarian.net/190819459/buildlog_ubuntu-vivid-armhf.openimageio_1.4.14~dfsg0-1~ubuntu1_FAILEDTOBUILD.txt.gz
[11:23] <LocutusOfBorg1> the same program builds correctly on debian, whithout that /usr/include/qt4/QtOpenGL/qgl.h:85:17: error: conflicting declaration 'typedef GLfloat GLdouble'
[11:23] <LocutusOfBorg1>  typedef GLfloat GLdouble;
[11:23] <LocutusOfBorg1> stuff
[11:23] <cjwatson> https://wiki.ubuntu.com/ARM/FTBFS#OpenGL_and_Qt_combination
[11:24] <LocutusOfBorg1> cjwatson, thanks
[11:46] <caribou> Q: When fixing an old package (python-pywbem in that case) for SRU, should I worry about lint complaining about obsolete build-depends or old standards ?
[11:47] <caribou> or I should just adapt the package to the minimum required to fix the issue (add on new depends in my case) ?
[11:48] <cjwatson> caribou: Don't change things you don't need to change.
[11:48] <cjwatson> (for SRUs)
[11:48] <cjwatson> caribou: Those things are in the "don't need to change" category.
[11:48] <caribou> cjwatson: perfect, that was what I was inclined to do
[12:39] <sil2100> @pilot out
[12:57] <ScottK> didrocks: It'd be nice if there was some way to respond to your blog post that wasn't registration required.
[12:58] <didrocks> ScottK: I did have my own system for this, however, the amount of spam, even with some spam filters was just crazy and I ended up giving up on hosting those comments myself
[12:58] <ScottK> didrocks: Maybe App Developer Tools For Ubuntu.
[12:59] <ScottK> Ubuntu Developer Tools of course made me think of ubuntu-dev-tools, which is completely different.
[12:59] <didrocks> ScottK: right…
[12:59] <didrocks> ScottK: sounds good, I like it! what would be the command line in your opinion?
[12:59] <barry> pitti: those are just the typical timeout errors.  you have to retry the build
[13:00] <ScottK> It also made me think of http://qa.ubuntuwire.org/multidistrotools/
[13:00] <ScottK> How about just adtu.
[13:00] <ScottK> Short, sweet, and unlikely to have collisions (I didn't check though)
[13:00] <pitti> barry: hm, it failed three times in a row, but I'll retry
[13:01] <didrocks> ScottK: well, like udtc, but people couldn't type it ;) but I'm adding that one to the pool myself, thanks for the suggestion!
[13:01] <barry> pitti: this is vivid?
[13:01] <pitti> barry: yes
[13:02] <barry> pitti: i'll try locally, but i've seen this plenty of times before.  si tells udm to download a bunch of files and then no 'finished' signal is ever received from udm
[13:02] <pitti> barry: ah, ok
[13:02] <mdeslaur> @pilot in
[13:05]  * dholbach hugs mdeslaur
[13:05]  * mdeslaur hugs dholbach
[14:17] <didrocks> Riddell: ScottK: stupid question, but I don't find kdm in vivid, kde-workspace doesn't exist anymore as replaced by plasma-workspace, but the binary package doesn't seem to me to provide a DM, where is it?
[14:20] <pitti> hm, nothing relevant in Contents.gz either
[14:20] <cjwatson> It's called sddm now IIRC?
[14:21] <cjwatson> (No specific knowledge here apart from having dealt with a number of MPs that added sddm support ...)
[14:21] <didrocks> ah, seems so, thanks cjwatson :)
[14:25] <didrocks> as it's sysvinit only, I guess that's why it's not on the list of services to transition to systemd (where it actually is if we want it to work accordingly to /etc/X11/default-display-manager choice and not having 2 stepping on each other)
[14:28] <ScottK> didrocks: Yes, it's sddm.
[14:28] <pitti> stgraber: working on bug 1346734 ATM (I have a first working PoC); do we really want to put logind cgroups into *all* controllers? including "devices"? don't we want to use "devices" in some more coarse-grained/different way?
[14:28] <didrocks> ScottK: thanks for confirming
[14:28] <ScottK> We ought to have that work with systemd.
[14:29] <ScottK> It just recently actually made it into the archive.
[14:31] <didrocks> ScottK: I'll push a .service file in the coming days, I don't see any Vcs-Bzr though, do you have a link?
[14:32] <didrocks> (I'm handling debian first, but they need to have a not-that-nice snippet for jessie, we'll get what we'll push in jessie+1 to vivid directly)
[14:33] <ScottK> didrocks: It should be at lp:~kubuntu-packagers/kubuntu-packaging/sddm but it does appear that someone neglected to push it.
[14:34] <ScottK> Riddell: ^^^ can you push sddm packaging to LP.
[14:34] <didrocks> ScottK: ok, thanks! If it's not pushed yet, I'll just dput the change at worst
[14:36] <ScottK> didrocks: I forgot we're doing the KF5/Plasma 5 stuff in git.  It's at http://anonscm.debian.org/cgit/pkg-kde/kde-std/sddm.git/
[14:36] <didrocks> ScottK: hum, so it means people having upload rights in ubuntu can't push it there?
[14:36] <ScottK> In some cases.
[14:36] <ScottK> Feel free to just dput.
[14:37] <didrocks> ScottK: I'll dput and propose a git-format patch then
[14:37] <ScottK> Thanks.
[14:37] <didrocks> yw
[14:43] <Riddell> didrocks, ScottK: sddm is in debian git http://anonscm.debian.org/gitweb/?p=pkg-kde/kde-std/sddm.git
[14:43] <Riddell> in the kubuntu_vivid_archive branch
[14:43] <ScottK> Right.  We eventually figured that out.
[14:43] <Riddell> didrocks: turning on systemd is just a matter of editing debian/rules and installing the service file
[14:45] <didrocks> Riddell: it's a little bit more complicated than that if you want to handle multiple dms on the same machine and not having multiple ones starting at the same time
[15:03] <stgraber> pitti: putting the user in all cgroups is fine. The user will never be allowed to grant themselves more access than they already have anyway.
[15:04] <pitti> stgraber: ok, thanks; I now have a patch which I'm reasonably happy with as a first iteration (one bit should go upstream, though)
[15:05] <pitti> stgraber: for now I just tested it in nspawn and checked /proc/$$/cgroup; this now looks like http://paste.ubuntu.com/9233771/
[15:05] <pitti> stgraber: I'll now create a VM with user containers to test it there, but does that look good to you at first sight?
[15:06] <stgraber> pitti: well, cpuset is clearly wrong
[15:07] <pitti> stgraber: right, that one is missing; I mean the idea is that the user session processes are in all controllers in a cgroup that's writable to the user, right?
[15:08] <stgraber> correct
[15:08] <pitti> (I'm not sure why cpuset is missing, it's in my list of allowed cgroups; but I'll debug that
[15:13] <pitti> stgraber: argh, actually this is wrong -- the .slice cgroups are root owned, the only thing that's owned by the user is user@1000.service
[15:13] <pitti> stgraber: anyway, I found the various places in the code which deal with that :)
[15:13] <pitti> and they should be root -- a user shouldn't be able to transfer processes into different sessions, that'd be a potential privilege escalation
[15:37] <pitti> stgraber: so I added the veth config to /etc/lxc/lxc-usernet and the two lxc.id_map to /etc/lxc/default.conf, but I get http://paste.ubuntu.com/9234222/
[15:37] <pitti> stgraber: this is vivid host (qemu), trusty guest
[15:37] <stgraber> pitti: read the error message :)
[15:38] <stgraber> pitti: specifically, the maps should be in ~/.config/lxc/default.conf not in /etc/lxc/default.conf
[15:38] <stgraber> unless you're doing unprivileged containers started by root
[15:38] <pitti> oh, ok; /me creates that dir then
[15:39] <pitti> that works, thanks!
[15:39] <stgraber> right, the procedure is, create ~/.config/lxc, copy /etc/lxc/default.conf to ~/.config/lxc/default.conf and add the id maps in there
[16:38] <sabdfl> sladen, mpt, very glad to see you two driving DM ;)
[16:44] <sladen> sabdfl: have you been panning the camera around...
[16:45] <sabdfl> no, just watching my bug email, sladen :)
[16:50] <LocutusOfBorg1> pitti our internship student wants to thank you for this link http://www.piware.de/2013/09/how-to-watch-system-d-bus-method-calls/
[16:50] <LocutusOfBorg1> :)
[16:55] <pitti> LocutusOfBorg1: heh, yw :)
[16:59] <kees> infinity, pitti, stgraber: TB meeting in 2 minutes :)
[16:59] <pitti> kees: I sent apologies, another meeting now
[17:00] <kees> pitti: cool, no worries.
[17:03] <mpt> sabdfl, we are nothing if not versatile ;-)
[17:30] <tkamppeter> pitti, mvo, doko, around?
[17:32] <tkamppeter> pitti, mvo, doko, it is about apt-daemon, package repository setup and package installation via /usr/share/system-config-printer/install-printerdriver.py does not work any more, bug 1340830.
[17:52] <kees> on vivid, what is spewing "calling: info" during apt-get dist-upgrade?
[18:17] <didrocks> jamespage: hey, I saw your docker.io upload, is there any reason you reenable starting docker unconditionnaly at boot (like debian does) on systemd?
[18:18]  * didrocks did remove it some days ago (with adding socket activation) and didn't notice any issue with this
[18:26] <mdeslaur> @pilot out
[21:42]  * xnox thinks my ideapad is broken
[21:42] <xnox> lsusb hangs =(
[23:07] <tych0> so suppose i want to rebuild all of debian/ubuntu with a custom gcc
[23:07] <tych0> is there some infrastructure set up to do that?
[23:30] <bdmurray> pitti: the ddeb for libqt5widgets5 for vivid and armhf is missing
[23:30] <bdmurray> pitti: it was uploaded yesterday
[23:31] <bdmurray> pitti: also libqt5xml5