[00:22] <GunnarHj> robert_ancell: I updated https://code.launchpad.net/~gunnarhj/lightdm/startup-dialog/+merge/198004 so it includes a simple way to disable the thing.
[00:23] <robert_ancell> cool
[00:24] <GunnarHj> robert_ancell: So now I for one have no remaining doubts about merging it. ;-)'
[00:25] <robert_ancell> GunnarHj, did you work out the translations?
[00:26] <GunnarHj> robert_ancell: Well, I added a line in po/POTFILES.in. Isn't the rest done automatically somehow?
[00:26] <robert_ancell> does it work in bash scripts?
[00:27] <GunnarHj> robert_ancell: Yes. The .sh extension makes xgettext recognize it as a shell script. I did test.
[00:27] <robert_ancell> ok, good
[06:52] <pitti> Good morning
[07:13] <larsu> xnox: that'd be great!
[07:13] <larsu> good morning :)
[08:14] <mlankhorst> morning
[08:21] <pitti> didrocks: bonjour, ça va ?
[08:21] <pitti> hey mlankhorst
[08:21]  * pitti winkt zu larsu
[08:21] <didrocks> pitti: bonjour, ça va bien, et toi? :)
[08:21]  * larsu winkt zurück zu pitti
[08:21] <larsu> didrocks: salut!
[08:21] <pitti> didrocks: oui, merci; nous avons aidé ma sœur avec déplacement à ce week-end
[08:22] <didrocks> hey larsu!
[08:22] <didrocks> pitti: oh, pas trop fatigué?
[08:22] <larsu> bon Monday!
[08:23] <pitti> didrocks: no, we didn't have to carry most of the furniture, they took a company for that; but as always, there are a gazillion small things aside from that (packing dishes, cleaning old flat, moving contents of cellar, etc.)
[08:23] <didrocks> ahah, indeed ;)
[08:24] <didrocks> we went out for the light festivals. We walked 23 kms to see most of the places, drink hot wine, eat churros… it was fun :)
[08:44] <sil2100> alex-abreu: hello! Are you around right now? Would need you to approve a merge
[08:47] <voldyman> hey larsu
[08:48] <voldyman> i have a question, does upstart only launch the indicators or does it do more?
[08:48] <larsu> voldyman: afaik, it only launches them
[08:49] <larsu> they then get a name on the bus, which is how the panel finds them
[08:50] <voldyman> so in theory they can be started by gnome-session ?
[08:51] <larsu> yes (or dbus activation)
[08:51] <larsu> xnox wants to do some work there, it might be worth talking to him
[08:53] <voldyman> thanks, you might be interested in https://lists.launchpad.net/elementary-dev-community/msg02913.html
[09:04] <Laney> hey
[09:06] <seb128> good morning desktopers
[09:06] <seb128> hey Laney
[09:06] <seb128> how are you? had a good w.e?
[09:07] <Laney> yes, was most excellent thanks
[09:07] <Laney> party Saturday and went to see a great brass band yesterday
[09:07] <Laney> https://www.youtube.com/watch?v=USf1zRaDyH4
[09:07] <Laney> you?
[09:08] <xnox> voldyman: it's a bit in flux in trusty at the moment, but yeah i'm planning to get them back to dbus-activation (and optionally upstart)
[09:09] <Laney> guten morgen xnox
[09:09] <seb128> mine was ok, I was feeling a bit weak/cold since friday, fought that most of the w.e (I had a vaccination on thursday, I think that's side effects from it/or I got a small cold/flu at the same time)
[09:09] <xnox> voldyman: it's a bit ahead of time to mandate upstart. Independantly of the indicators, i'd be interested in helping out elementary to adopt upstart for the user sessions =)
[09:09] <seb128> otherwise good, mostly sleeping and playing video games
[09:09] <xnox> Laney: guten morgen! =) ich bin nach office heute ;-)
[09:10] <seb128> xnox, hey, how are you?
[09:10] <xnox> seb128: good good =)
[09:10] <seb128> xnox, what do you plan to get back to dbus-activation? indicators? did you talk to ted about it (he was at the office as well, last week though, no today ;-)
[09:11] <xnox> seb128: damn. well we are talking on the merge proposals.
[09:11] <voldyman> xnox: will the indicators have customizable UI ? our aim is to make something like this http://bassultra.deviantart.com/art/Sync-Indicator-393526929
[09:11] <Laney> aww, hope you're feeling better now!
[09:11] <seb128> Laney, I am, thanks
[09:11] <Laney> xnox: how does the christmas tree look? ;-)
[09:12] <seb128> xnox, if you could teach dbus activation that would make everybody happy
[09:12] <seb128> xnox, larsu would probably hug you for it as well
[09:12] <xnox> seb128: basically two things need to happen: readd dbus activation files, and finally readd upstart-dbus-activation such that dbus activated things are managed by upstart (when upstart is available)
[09:12] <seb128> the hug might even come with ice cream
[09:12]  * larsu confirms: there might be beer involved as well. And ice cream.
[09:12] <seb128> ;-)
[09:13] <seb128> xnox, was that dropped because of the dbus patches we had that didn't get forward ported?
[09:13] <xnox> Laney: it's thin and apart from a single angly birds toy doesn't look geeky enough.
[09:13] <xnox> seb128: correct.
[09:13] <Laney> yeah dbus needs to support it
[09:14] <xnox> seb128: if we find anyone with spare cycles to port them. I don't think the port is that hard, it's just wasn't done, because at the time it wasn't used..... yet touch assumed it will be available once session init lands....
[09:14] <xnox> but first, i need to find coffee =)
[09:15] <seb128> xnox, we have been looking for somebody to port those patches for a while, pushing back on ted for the upstart job changes previous cycle
[09:15] <seb128> but since the upstart side was not happening he finally landed his stuff this cycle
[09:15] <seb128> I didn't know touch was assuming that it would be available
[09:15] <seb128> if we get that working that's going to make everybody happy ;-)
[09:29] <xnox> voldyman: indicators already have customizable UI, one can declare multiple type of interfaces, and request different profiles.
[09:29] <xnox> voldyman: e.g. indicator on the phone, phone greeter, desktop greeter, desktop, installer can already look different
[09:29] <xnox> voldyman: in practice a lot of it is shared (e.g. installer mostly reuses desktop greeter UI for the indicators)
[09:30] <xnox> voldyman: no reason why other / alternative "elementary" UI cannot be exported as well.
[09:32] <xnox> larsu: seb128: is there a way to "inspect" indicators? i think one too many is being loaded on the desktop - I have network indicator and network indicator. (one looks like normal network-manager one looks "phone" like with just wifi on/off, list of APs and settings label)
[09:33] <larsu> xnox: you probably have the indicator-network package installed?
[09:33] <seb128> xnox, one is nm-applet one is indicator-network
[09:33] <seb128> we still use nm-applet on the desktop
[09:33] <seb128> because the proper indicator is feature limited
[09:34] <xnox> larsu: ah, let me check.
[09:34] <voldyman> xnox: ah, nice.
[09:34] <xnox> larsu: =( removing that would remove "account-plugin-ubuntuone indicator-network ubuntu-system-settings ubuntu-system-settings-online-accounts unity-scope-click"
[09:35] <xnox> larsu: can we get a profile up to not load on the desktop at the moment (even though it's installed?)
[09:35] <xnox> (or me blacklisting it locally)
[09:35] <seb128> xnox, edit /usr/share/unity/indicators/com.canonical.indicator.network
[09:35] <seb128> drop the desktop section
[09:35] <seb128> restart it/unity-panel-service
[09:35] <seb128> you might want to dpkg-divert to avoid doing that at every update
[09:35] <chrisccoulson> hmmm, i didn't realize that binutils-gold is an empty package in saucy. how do we tell gcc to use gold to link?
[09:35] <seb128> (I wonder if we should just do that in the vcs)
[09:36] <xnox> chrisccoulson: via ldflags as before. binutils-gold is now provided by binutils themself, so one can't change subvert ld.
[09:36] <larsu> xnox: what seb128 said
[09:36] <xnox> (or some such, haven't actually done that myself, since the switch)
[09:37] <chrisccoulson> xnox, yeah, i used to just install binutils-gold before :)
[09:37] <seb128> grrrr ppas
[09:37] <chrisccoulson> (because it replaced the /usr/bin/ld link)
[09:38] <larsu> seb128: it's Monday. Don't be angry!
[09:38] <seb128> larsu, yeah, and a good week start so far, you are right ;-)
[09:39] <Laney> only two more weeks to go until christmas holidays ;-)
[09:39] <xnox> seb128: diverted + $ stop indicator-network did the trick ;-)
[09:39] <seb128> Laney, the amd64 build of webkit in the desktop ppa got stopped by ld being killed, I wonder if it took longer than the ppa timeout to run ld
[09:39] <seb128> xnox, great
[09:39] <Laney> haha
[09:40] <Laney> how did it get killed?
[09:40] <Laney> could be OOM
[09:40] <seb128> https://launchpad.net/~ubuntu-desktop/+archive/ppa/+build/5308637/+files/buildlog_ubuntu-trusty-amd64.webkitgtk_2.3.2-1ubuntu3%7Ebuild1_FAILEDTOBUILD.txt.gz
[09:40] <seb128> -9
[09:40] <Laney> yeah probably that
[09:40] <seb128> have to love webkit
[09:40] <xnox>  i guess i could of just $ echo manual > ~/.config/upstart/indicator-network.override
[09:40] <Laney> love to hate
[09:40] <Laney> or hate to love, or both
[09:41] <seb128> well, good news is that I'm using i386 and that this one built ;-)
[09:41] <seb128> retrying the amd64
[09:41] <Laney> worth a try
[09:41] <Laney> if not I can upload it to a non-virtual ppa
[09:41] <Laney> can you tell if it fixes the issue?
[09:41] <seb128> Laney, I'm just going to upload to the real archive, we want that fix there anyway ...
[09:42] <seb128> well, let me see if I can reproduce the bug
[09:42] <seb128> last week I got it 3 times in a row
[09:42] <seb128> and then not the 15 next times when I wanted debug info
[09:42] <seb128> well the patch got review/commited upstream, so if it runs fine I'm just going declare it correct and upload
[09:43] <alex-abreu> sil2100, I am
[09:45] <sil2100> alex-abreu: hello! Ping no longer valid ;) We anyway saw that Ken was the only one in content-hub with the power of top-approving
[09:46] <sil2100> alex-abreu: but we managed to work-around that
[09:46] <alex-abreu> ok
[10:10] <seb128> Laney, the webkit fix doesn't work :/
[10:11] <Laney> doh
[10:11] <Laney> no change?
[10:12] <seb128> downloading 343M of dbg to get a stacktrace, to see if it's the same one
[10:12] <seb128> it might be they have another similar issue
[10:35] <Sweetshark> Moin!
[10:35] <Sweetshark> seb128: http://people.canonical.com/~bjoern/libreoffice-4-1/libreoffice_4.1.3-0ubuntu2_source.changes
[10:36] <Laney> should upload harfbuzz first
[10:38] <Laney> but yay, I can upload both if you like
[10:38] <seb128> Sweetshark, \o/
[10:38] <seb128> Laney, we want poppler 0.24.4 as well before
[10:38] <seb128> we can as well tackle all transitions :p
[10:38] <Laney> heh
[10:39] <seb128> https://bugs.launchpad.net/ubuntu/+source/poppler/+bug/1256627
[10:39] <ubot2`> Launchpad bug 1256627 in poppler (Ubuntu) "New upstream release, merge with Debian" [Wishlist,New]
[10:39] <seb128> http://cgit.freedesktop.org/poppler/poppler/commit/?id=ebe49d597a62aa94
[10:40] <seb128> Sweetshark, ^ is libreoffice using one of those functions?
[10:40] <seb128> Laney, the poppler transition might not be a no change rebuilds one, so maybe we can delay it to another round, that's a minor update with only a few commits
[10:42] <Sweetshark> seb128: no idea TBH ...
[10:42]  * Sweetshark digs for the patch we had for the last poppler transistion
[10:43] <xnox> Laney: right, so ideally i'd want that script to set "subscription name" on the subscription (like one can do on the webui) and only delete "script generated subscriptions", well I'm gonna just comment out unsbuscribing ;-)
[10:44] <Laney> xnox: yeah, want to file an LP bug to ask for that over the api?
[10:45] <xnox> Laney: can i file bugs against junk branches? against a person? =)
[10:45] <Laney> the LP API needs to support it first
[10:45] <Sweetshark> seb128: so from a quick look it could well be that we dont use the getObj and shift functions ...
[10:45] <xnox> Laney: ah. I see. sure will do.
[10:46] <Laney> also, if you want, you could patch it to make -u 0 turn off the unsubscribing
[10:46] <Laney> I think you're going to end up subscribed to a lot of stuff :-)
[10:47] <xnox> Laney: well, i'm adding an upstart job to run in debug mode and collect logs for me & will cron "start lp-subscribe-uploads" thus if I spot it doing it naughty things I'll have logs =)
[10:48]  * xnox is thinking 30d unsubscribe timeout (things get stuck in -proposed for a while sometimes)
[10:48] <Laney> mmk
[10:48] <seb128> Laney, do you want to do the new poppler as well or should we just go for harfbuzz/rebuild lo?
[10:49] <seb128> Laney, I'm busy looking at webkit and gtk 3.10 (and dealing with W.E backlog) so I'm not going to step up today, your call
[10:49] <Laney> seb128: I'm worried about getting entangled in other stuff
[10:49] <Laney> let's do them separately
[10:50] <seb128> Laney, +1
[10:50] <seb128> Laney, can you do the few rebuilds for harfbuzz as well? (pango, texsomething, ...)?
[10:50] <Laney> yep
[10:51] <seb128> thanks
[10:54] <Laney> grr, what are these dh-autoreconf changes for?
[10:54] <seb128> the ones from doko?
[10:55] <Laney> yeah
[10:56] <seb128> that's support for new archs I think
[11:26] <Laney> harfbuzz'd
[11:31] <Laney> Sweetshark: you didn't need to build with 4.7?
[11:46] <Sweetshark> Laney: nope.
[11:47] <Sweetshark> Laney: it was caused by the mdds version bump, so it was my own stupidity mostly (I havent had that one on the radar).
[11:47] <Laney> okay
[11:48] <Sweetshark> Laney: so we are builing this one with an internal copy of an (older, compatible) mdds, and already have the mdds in archive that we'll need for LibreOffice 4.2 in the end.
[11:49] <Sweetshark> (which then will again build with a proper --with-system-mdds)
[11:49] <Laney> yep, I got it, thanks
[11:51] <Laney> seb128: harfbuzz is ready for NEWing btw
[11:53] <seb128> Laney, done
[11:53] <Laney> cool, cheers
[12:46] <Sweetshark> seb128: meh, saw the failure on trusty. Likely just that the system ./debian/control rules was run on wasnt updated recently enough. FWIW I _did_ build that locally (but not in a PPA as a full-l10n build wouldnt work in a ppa).
[12:48] <Sweetshark> seb128: oh, hang on -- the broken deps dont seem to come from my control file. So its just a case of SSDD elsewhere it seems.
[12:48] <seb128> Sweetshark, oh, I didn't notice, Laney did sponsor the upload ... can you get a fixed version?
[12:48] <Laney> it was me
[12:48] <Laney> I think it's just pango needing rebuilt
[12:48] <seb128> right
[12:49] <seb128> chrisccoulson, your firefox 26 trusty upload is not happy on !i386/amd64
[12:51] <chrisccoulson> seb128, http://bazaar.launchpad.net/~mozillateam/firefox/firefox-beta.trusty/revision/1032 ;)
[12:51] <seb128> chrisccoulson, good ;-)
[12:52] <seb128> chrisccoulson, I was trying to see if f26 would fix that hunspell issue by any chance, it's ranking high on e.u.c
[12:52] <seb128> but I guess not enough got that version yet to say
[13:04] <Laney> seb128: do you want to upload that webkit patch anyway?
[13:04] <Laney> or wait a while to see if Company comes up with a new one?
[13:04] <Laney> (got to upload it for harfbuzz)
[13:04] <seb128> Laney, Company just ponged, give him some time to look at the bug in case he has an idea there
[13:05] <Laney> okay
[13:05] <seb128> thanks
[14:59]  * Laney eyes the notify-osd autopkgtest failure
[14:59] <Laney> how did that not hold back any migrations?
[15:04] <seb128> yeah for useless logs
[15:04] <seb128> we should make verbose default for tests
[15:06] <Laney> yeah :/
[15:09] <Laney> hm
[15:10] <Laney> pitti: is the working directory preseved when using run-adt-test?
[15:10] <pitti> Laney: yes, for things like -S
[15:10] <Laney> I'm SSHed into the instance, but the directory in ubuntu's HOME is unbuilt
[15:10] <pitti> (not relevant otherwise)
[15:11] <pitti> I usually run it with -S file://`pwd`
[15:11] <Laney> It's has the build-needed restriction but looks like that happened somewhere in /tmp
[15:11] <Laney> which is now gone
[15:11] <Laney> s/It's/It/
[15:12] <pitti> yes, it doesn't touch the current source dir, it copies everything into a temp location of the virt runner
[15:12] <Laney> yeah - I can't even find it inside the instance
[15:12] <Laney> I have the unpacked source, but not the directory that run-adt-test already built and ran the tests in
[15:13] <Laney> so I can't find test-suite.log
[15:13] <Laney> ah well, will run manually
[15:18] <Laney> http://paste.ubuntu.com/6546229/
[15:19] <pitti> ah, that glib change again
[15:24] <desrt> stgraber: hey.  got a sec to chat cgroups?
[15:24] <stgraber> desrt: sure
[15:24] <desrt> stgraber: what is /sys/fs/cgroup/systemd and why do i have it on my ubuntu system?
[15:24] <seb128> Laney, pitti: is the old test running?
[15:24] <stgraber> desrt: it's the cgroup hierarchy setup by logind
[15:25] <Laney> seb128: no, it's eral
[15:25] <Laney> real
[15:25] <seb128> Laney, pitti: http://bazaar.launchpad.net/~indicator-applet-developers/notify-osd/trunk/revision/476 was supposed to fix that issue
[15:25] <seb128> or is the test having the same bug as the runtime was having?
[15:25] <pitti> he desrt
[15:25] <desrt> stgraber: okay.  that makes sense.  is something else in trusty setting up other cgroup hierarchies?
[15:25] <Laney> oh
[15:25] <desrt> pitti: hi :)
[15:25] <Laney> let me see
[15:25] <stgraber> desrt: it's not doing any resource limiting at this point but is used by logind to group all processes of a given seat and user session
[15:25] <pitti> desrt: <gentle prod for shim 6> :-)
[15:25] <desrt> pitti: already released, no?
[15:26] <stgraber> desrt: not at the moment. We have cgroup-lite which comes with LXC and sets up all the others but it's not installed by default
[15:26] <desrt> [ ]systemd-shim-6.tar.xz06-Dec-2013 14:07 75Ktar archive
[15:26] <pitti> desrt: oh, so it is, thanks! should have Ctrl+R'ed..
[15:26] <desrt> stgraber: so http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=707201 started hitting me after the upgrade to trusty
[15:26] <ubot2`> Debian bug 707201 in libvirt-bin "Unable to initialize /machine cgroup: Invalid argument" [Serious,Fixed]
[15:26] <stgraber> desrt: we're also working on our own cgroup manager for LXC which will be in trusty and will be managing /sys/fs/cgroup for us. We'll have some kind of abstraction layer so that logind (and possibly anything else using the systemd cgroup API) will just work and talk to our manager instead of systemd's own
[15:27] <desrt> stgraber: there is a script in there that claims to fix it.... but when i tried to do what the script was doing, i got a lot of warnings about the cgroup being busy
[15:27] <Laney> seb128: pretty sure it's using the right version
[15:27] <stgraber> desrt: ah, can you try installing cgroup-lite and see if that helps? if it does, then we probably should just have libvirt recommend cgroup-lite
[15:28] <Laney> I'm reproducing it locally with the trusty-proposed version
[15:28] <seb128> Laney, so maybe the tests have the same bug as the service had
[15:28] <seb128> Laney, or do you get the issue on the runtime?
[15:28] <Laney> not suer
[15:28] <desrt> stgraber: already installed....
[15:28] <desrt> not sure how i got that by default
[15:28]  * desrt tries removing it =)
[15:29] <stgraber> desrt: it's a depends of libvirt-bin apparently
[15:29] <desrt> recommends it seems, because i just uninstalled it without trouble
[15:29] <desrt> although i tried upgrading to debian's libvirt last night, so that could also be why
[15:30] <desrt> http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=725261 is the bug with the script that fails, rather
[15:30] <ubot2`> Debian bug 725261 in libvirt-bin "libvirt-bin: Cgroup configuration for LXC and Qemu on the same host" [Wishlist,Fixed]
[15:30] <stgraber> it's a cgroup-lite | cgroup-bin depends in Ubuntu apparently
[15:30] <stgraber> do you have cgroup-bin installed?
[15:30] <desrt> no
[15:31] <stgraber> ah that's probably explained by Debian not having cgroup-lite and so not having that depends on their libvirt
[15:31] <desrt> mount: cgroup_cpuset already mounted or /sys/fs/cgroup/cpuset busy
[15:31] <desrt> this is the message that makes me wonder if something else is trying to manage this... maybe logind?
[15:31] <stgraber> can you pastebin "grep cgroup /proc/mounts"?
[15:32] <desrt> sure.  gimme a sec.
[15:33] <desrt> http://paste.ubuntu.com/6546282/
[15:33] <desrt> oh.  neat.  /cgroup
[15:34] <stgraber> ah yeah, that /cgroup really shouldn't be there :)
[15:34] <stgraber> I'd suggest getting rid of that one, then installing cgroup-lite again and check that you get them all setup under /sys/fs/cgroup properly
[15:34]  * desrt wonders who is responsible for this one :(
[15:35] <Laney> okay I got a fix
[15:36] <pitti> stgraber, desrt: anything wrong in /etc/init/systemd-logind.conf ?
[15:36] <desrt> stgraber: working!!
[15:36] <pitti> (just following with half an eye, deep in debugging)
[15:38] <desrt> stgraber: thanks for your help.... looks like something weird wrong during the upgrade
[15:38] <desrt> serves me right for upgrading to an upstable distro, i guess
[15:38] <desrt> *unstable
[15:39] <stgraber> :)
[15:40] <seb128> desrt, what was the buggy/wrong part at this end?
[15:41] <desrt> seb128: had an old entry about cgroupfs in my fstab that the upgrade didn't remove
[15:41] <Laney> seb128: larsu (since you looked at this before): https://bazaar.launchpad.net/~indicator-applet-developers/notify-osd/trunk/view/head:/src/timings.c#L359 ← I think we need to set priv->…_id = 0 there & in other similar places, so that line 176 doesn't try to remove it again later on - agreed?
[15:41] <seb128> we have packages "hacking" fstab?
[15:42] <desrt> seb128: i have no idea
[15:42] <desrt> i also have /cgroup
[15:42] <seb128> no cgroups in fstab here
[15:43] <desrt> hmm https://bugs.launchpad.net/ubuntu/+source/libcgroup/+bug/869364
[15:43] <ubot2`> Launchpad bug 869364 in libcgroup (Ubuntu) "cgconfig gets confused with /cgroups fstab entry" [Low,Triaged]
[15:44] <larsu> Laney: definitely. Also, the assert() is unnecessary now (g_source_remove always returns TRUE)
[15:44] <Laney> ack
[15:44]  * desrt shrugs
[15:44] <desrt> i hate cgroups
[15:44] <seb128> Laney, I'm in hangout, can have a look after that
[15:44] <seb128> (oh seems larsu is on it, good)
[15:44] <Laney> no need, one confirmation is enough for me
[15:44] <Laney> :P
[15:44] <seb128> ;-)
[15:44] <larsu> Laney: are you doing the patch?
[15:45] <Laney> yeah
[15:45] <larsu> cool
[15:45] <Laney> you can review it then if you want :-)
[15:45] <larsu> of course!
[16:03] <seb128> Laney, ok, I doubt Company is going to have a fix out today for that webkit issue, he's doing some refactoring there (he said the issue is created by some optimization webkit is doing and he's pondering dropping those or make them work, he needs to evaluate the win they create before though)
[16:03] <Laney> nod
[16:03] <seb128> Laney, I'm going to upload the patch from the ppa
[16:03] <seb128> it's correct and fixes some issues, even if that's not enough for -sc
[16:03] <Laney> okay
[16:04] <Laney> if a patch comes in the next couple of days then great, otherwise there should be a release next monday
[16:04] <Laney> so we might as well roll it in with that if it's any later
[16:05] <seb128> ok
[16:06] <Laney> hopefully notify-osd can be released too, then that should be harfbuzz providing lo builds
[16:07] <Laney> already did on ppc. sagari ♥
[16:07] <Laney>  (took 1 hour, 17 minutes, 50.8 seconds)
[16:07] <Laney> that box is amazing
[16:16] <larsu> Laney: might as well get rid of the booleans
[16:16] <larsu> and the last one is missing the = 0, no?
[16:17]  * larsu is flying with partial vision. Only looking at the diff
[16:17] <Laney> true
[16:17] <Laney> the last one is always overwritten so I didn't bother there
[16:17] <Laney> can do though if you want
[16:18] <larsu> nah that's fine
[16:18] <Laney> booleans: DEAD
[16:19] <xnox> meh, can't upload ubiquity with webkit/harfbuzz transition in-flight
[16:19] <Laney> is all in hand
[16:20] <larsu> Laney: more red \o/. approved
[16:20] <Laney> excellent, thanks!
[16:36] <Laney> larsu: did you mean to top approve it?
[16:38] <larsu> Laney: I wanted to wait for Jekins
[16:38] <Laney> ah
[16:38] <Laney> doesn't the merger do that itself?
[16:38] <larsu> yeah, I guess. I'll approve it right now
[16:38] <Laney> hope so, I've been approving before it for ages :P
[17:02] <sil2100> kenvandine: hi! I hope you don't mind I re-targetted the lp:content-hub branch to be part of ~phablet-team? :)
[17:02] <sil2100> kenvandine: since sadly ~content-hub-team has only 2 members and we couldn't do any magic
[17:02] <kenvandine> sil2100, again?
[17:03] <kenvandine> well, nobody paid any attention to it before
[17:03] <kenvandine> i was hoping to get a couple people added and merges would get noticed, etc :)
[17:03] <sil2100> kenvandine: now it was rather crucial, since there was an urgent branch we needed top-approving but no one besides you could do that
[17:03] <sil2100> We can re-organize later
[17:23] <desrt> Laney. seb128: with your debian soname-purist hats on, do you know if the libtool versioning stuff is _at all_ useful?
[17:25] <desrt> i'm thinking about dropping our ridiculously complex treatment of it from glib
[17:25] <seb128> I don't see much use to it
[17:25] <seb128> I'm basically happy if the soname change when there is an incompatible change
[17:25] <seb128> the minor, etc, who cares...
[17:26] <desrt> so if we have glib-2.0.so.0.0.0 you're fine
[17:26] <desrt> as long as it's .so.0
[17:26] <walters> desrt: owen designed it, and gtk does the same thing, worth at least running by him
[17:26] <walters> (i think he did anyways)
[17:27] <Laney> yeah I basically don't care how you generate it as long as the SONAME is used to indicate binary compatibility
[17:28] <Laney> We use .symbols files to generate lower bounds on dependencies
[17:29] <desrt> libtool claims to have the idea of multiple versions of an interface being supported
[17:29] <desrt> which interests me quite a lot
[17:29] <desrt> it means that some day we could have a libglib-2.0.so.1 with a link for libglib-2.0.so.0
[17:29] <desrt> and after the whole universe rebuilds, we can drop the backcompat
[17:30] <seb128> desrt, one thing handy about the current minor handling is that it includes the version
[17:30] <desrt> or maybe release .so.2 which is not compat at all and have .so.0 parallel-installable but no longer maintained (old binaries?)
[17:30] <seb128> desrt, so when you get e.g a stacktrace you know what version is being used
[17:30] <desrt> seb128: ya... i find that to be really annoying though
[17:30] <seb128> it came handy a few time, getting non debug bt
[17:30] <desrt> seb128: like when bisecting, if i go back through time, 'make install' installs a .so with an older minor and ldconfig links to the newer one instead
[17:31] <seb128> or "gtk.so.10"... he's running the new gtk!
[17:31] <desrt> ya....
[17:31] <Laney> I don't know that I've done anything with that parallel interface stuff
[17:31] <desrt> the parallel interface stuff actually doesn't work
[17:31] <Laney> do you get n libraries built?
[17:31] <desrt> it's documented as being supposed to work with a note that it works differently depending on the platform
[17:31] <desrt> in reality it doesn't work at all
[17:32] <desrt> no.  only one
[17:32] <desrt> in glib we claim interface 3902 (for 2.39.2, for example) and 'interface age' ==, ie. 3902)
[17:32] <desrt> which, in linux, gives only soname .0
[17:32] <desrt> ie: interface - age
[17:32] <Laney> yep
[17:32] <desrt> it's totally pointless
[17:32] <desrt> on other systems, we get soname == interface
[17:32] <desrt> so .so.3902
[17:33] <desrt> which is .... well, completely fucking awful
[17:33] <desrt> every single micro release is ABI incompat with the one before
[17:33] <desrt> it seems like we could dodge the entire mess by just always having version 0
[17:34] <desrt> until we actually break ABI, then we can manually set it to 1
[17:34] <desrt> ((i make comments like that ^ just to make sure seb128 continues to live in fear))
[17:34] <seb128> lol
[17:34] <seb128> nothing to fear
[17:35] <desrt> in any case, i'm starting to warm to the idea of a rolling ABI
[17:35] <seb128> weI can stop updating glib, easy
[17:35] <seb128> I just need to knock down Laney first
[17:35] <seb128> ;-)
[17:35] <desrt> the only problem is that it doesn't fit neatly into the debian worldview of things
[17:36] <desrt> would be a bit weird for you guys to get libglib2.0-0 and libglib2.0-1 out of the same source package and version, right?
[17:36] <Laney> attente: w00t!
[17:36] <Laney> what do you mean by rolling ABI?
[17:36] <desrt> ...particularly when -0 depends on -1
[17:36] <desrt> Laney: basically, bump the soname of libglib-2.0 to libglib-2.0.so.1
[17:36] <desrt> but install a symlink for libglib-2.0.so.0
[17:37] <desrt> so we now have ABI version 1, but it's claiming to be backcompat with .0
[17:37] <desrt> meanwhile, everyone who rebuilds their stuff gets linked to .1
[17:37] <desrt> after a while, once we think there are no old binaries, we drop support for the old ABI and stop installing the .0 symlink
[17:38] <desrt> you'd end up with a libglib-2.0-0 package that essentially contained only a symlink and a dependency on libglib-2.0-1
[17:39] <desrt> (or a set of symlinks, i suppose)
[17:41] <Laney> it feels weird but I don't know if I can articulate any philosophical objection
[17:41] <desrt> i think it's legit
[17:41] <attente> Laney, i'm not sure why, but setting something like es_CU or de_BE on AS only gets set as es and de respectively on the device, while en_NZ and en_CA still work
[17:41] <desrt> of course, we can still never break API...
[17:42] <Laney> yeah
[17:42] <Laney> see what pochu thinks, he might have some clever thought
[17:43] <Laney> attente: you get them back from locale -a?
[17:43] <attente> Laney, yep, they're present
[17:43] <Laney> hmm
[17:44] <Laney> I guess you could reproduce by poking AS yourself and see what's going on in there
[17:44] <Laney> by 'set' you mean in ~/.pam_environment?
[17:44] <seb128> ok, calling it a day, see you tomorrow
[17:44] <Laney> bye seb128!
[17:44] <attente> no, in the user's AS storage
[17:45] <Laney> ah
[17:45] <Laney> so we can rule out that script, that's good
[17:45] <Laney> doesn't happen on your desktop I guess?
[17:46] <attente> happens for certain locales
[17:46] <Laney> guess you should go see what AS is thinking :P
[17:46] <attente> could be something is not installed..
[17:47] <attente> heh
[17:47]  * attente uses psychic powers
[18:02] <Laney> attente: interesting, Language is wrong and FormatsLocale is right
[18:03] <attente> setting the formats locale does less validation than setting the language
[18:07] <Laney> I'm guessing not everything can be used for display language
[18:09] <Laney> /usr/share/language-tools/language-options is supposed to list those ones AFAIK
[18:09] <Laney> so maybe it is right
[18:09] <Laney> alas, got to go
[18:10] <attente> ok, thanks again Laney
[18:15] <attente> '/usr/share/language-tools/language-validate en_PH.utf8' == 'en_US'
[19:42] <GunnarHj> attente: See that you have questions about the language code i a-s. Anything I can do to clarify?
[19:43] <attente> GunnarHj, thanks, not really, but do you happen to know anything about libicu?
[19:44] <GunnarHj> attente: libicu? No, I don't even know what it is. :(
[19:45] <attente> GunnarHj, oh, ok. no worries. do you know if there's a way to get the likely locale for a language? e.g. for 'en', the likely locale would be 'en_US'
[19:47] <GunnarHj> $ /usr/share/language-tools/language2locale en
[19:47] <GunnarHj> en_US.UTF-8
[19:48] <GunnarHj> attente: ^
[19:48] <attente> GunnarHj, wow, thanks
[19:51] <Laney> 8
[19:51] <Laney> whoops
[20:02] <GunnarHj> attente: <quote> i'm not sure why, but setting something like es_CU or de_BE on AS only gets set as es and de respectively on the device, while en_NZ and en_CA still work</quote>
[20:02] <GunnarHj> That's because there is only one German respective Spanish translation installed, so a-s picks the most reasonable available language. /usr/share/language-tools/language-options lists all the installed translations. And, after all, selecting a language is about deciding which of the translations that is going to be used to display things, right?
[20:06] <attente> GunnarHj, right. i'm rewriting that code in ubuntu-system-settings to use the formats locale field instead
[20:09] <GunnarHj> attente: Given that the u-i sets both language and formats, and that you use a locale -a type list of options, that seems to be a good way to see which option is the currently selected one.
[20:15] <GunnarHj> attente: (All that stuff in a-s assumes that you set language and formats separately, so it does a lot of things that don't really apply to u-s-s.)
[20:18] <attente> GunnarHj, makes sense
[20:28] <GunnarHj> robert_ancell: ping?
[20:28] <robert_ancell> GunnarHj, hello
[20:28] <GunnarHj> robert_ancell: Hi!
[20:28] <GunnarHj> robert_ancell: Did you see sabdfl's comment on bug 435930?
[20:28] <ubot2`> Launchpad bug 435930 in lightdm (Ubuntu) "Guest session does not warn about temporary nature of files" [Wishlist,In progress] https://launchpad.net/bugs/435930
[20:29] <robert_ancell> yes
[20:29] <robert_ancell> I agree shells should ideally implement the dialog, though that could be an enhancement
[20:30] <robert_ancell> GunnarHj, which shells did you propose the feature for?
[20:30] <GunnarHj> robert_ancell: Do you mean enhancement at a later time?
[20:30] <robert_ancell> yeah, I was thinking along those lines
[20:30] <robert_ancell> since it will be reasonably hard to implement this feature in a shell
[20:31] <GunnarHj> robert_ancell: "which shells did you propose the feature for" - not sure what you are asking
[20:31] <robert_ancell> GunnarHj, is this feature intended just for Unity, or LXDE / XFCE / KDE etc as well?
[20:33] <GunnarHj> robert_ancell: Aha. Basically for everyone who seeds zenity. I suppose that KDE might want to add some code to make it work with whatever they use.  Not sure about the other.
[20:34] <robert_ancell> GunnarHj, what do you think about trying to implement a dialog for Unity specifically. I was just concerned we wouldn't get it done in time
[20:34] <robert_ancell> And for other shells should they want it
[20:36] <GunnarHj> robert_ancell: Do you mean it would test for Unity rather than zenity, so it only happens in Ubuntu? But is there a reason to assume that Xubuntu, Kubuntu etc. are less interested as long as they use the guest session feature?
[20:36] <robert_ancell> GunnarHj, no, lightdm would indicate to the Unity session somehow that it was running in guest mode and Unity would show an appropriate UI for the user to inform them of the limitations
[20:36] <robert_ancell> zenity is never going to be the perfect solution for all desktops
[20:37] <robert_ancell> which is hinted at with the delay to show the dialog
[20:38] <GunnarHj> robert_ancell: I'm afraid that such a solution would be over my head. :(
[20:38] <robert_ancell> yeah, me too
[20:39] <robert_ancell> bregma, ^ any spare cycles to implement a dialog in Unity to tell the user they are running as guest?
[20:39] <robert_ancell> or any existing code we can use to do this?
[20:39] <bregma> I think we're already stretched to the limit this cycle
[20:40] <GunnarHj> robert_ancell: So to make at least some progress short time, how about taking the wording suggested by sabdfl and stick to zenity for now?
[20:40] <robert_ancell> GunnarHj, I think that's the best option on the table at the moment
[20:41] <GunnarHj> robert_ancell: Ok, thanks. Then I'll update the branch with the new text.
[21:02] <bregma> robert_ancell, I'm looking for a good way in lightdm to indicate to a session an X server will not be running, any suggestions on the best way?
[21:02] <robert_ancell> bregma, i.e. if the session is a Mir server or similar?
[21:02] <bregma> exactly... like an env var passed to lightdm-session, or something
[21:03] <bregma> I can just check DESKTOP_SESSION, but that seems unextensible and inelegant
[21:03] <robert_ancell> bregma, well, you should be able to tell if there's no X server by the absence of $DISPLAY
[21:04] <robert_ancell> or that you're in a Mir compositor by the Mir env variables
[21:04] <bregma> mm, right, I'll go with that for now
[21:16] <GunnarHj> robert_ancell: sabdfl's suggested wording talks about "when you logout or reboot". But you can't reboot from within a guest session, I then thought, at least if it was launched from within a normal session. But when I tested I could reboot from a guest session, even if there were two normal (locked) sessions open. Isn't that a pretty serious bug?
[21:17] <robert_ancell> GunnarHj, yeah, that seems serious. It will be a bug in logind/policykit
[21:18] <GunnarHj> robert_ancell: Ok, I'm going to check later on if the issue is addressed, and file a bug if not.
[21:18] <GunnarHj> robert_ancell: Dropping "or reboot" now.
[22:47] <GunnarHj> robert_ancell: Done with the branch update. Leaving it to you to answer sabdfl if deemed appropriate.
[22:47] <robert_ancell> GunnarHj, thanks
[22:48] <GunnarHj> robert_ancell: yw ;-)