[00:00] robert_ancell: hey, did seb128 ping you about renaming gnome-calculator.desktop back to gcalctool.desktop? [00:00] jbicha, not recently [00:03] robert_ancell: http://irclogs.ubuntu.com/2013/02/28/%23ubuntu-desktop.html#t17:24 [00:04] desrt: gconf oddities? gnome-themes-standard should be fixed if that's what you meant [00:05] jbicha, yeah, so either distro-patch it or convince the maintainer (I've passed it on) [00:06] robert_ancell: but I thought you were the maintainer... same problem for the games though [00:07] jbicha, I was but no longer. I also don't have a good answer - as an upstream you want the names to change but I see the migration issue [00:10] jbicha: and gnome-shell? [00:12] jbicha: it declared a dependency only (as far as i can tell) because it installed a migration script [00:35] desrt: ok I dropped the gconf dependency for gnome-shell & pushed to raring & the GNOME3 PPA, I guess you don't need to use tweak tool? === cjohnston_ is now known as cjohnston [00:39] jbicha: ya... tweak tool is kinda rough [00:39] not sure what we can do about that [00:44] jbicha: actually, it looks like nothing in here is even using GConf anymore... [00:44] there is an abstraction for GConf, but nothing currently using it... [00:44] desrt: I think it just uses gconf for backwards compatibility, I believe the maintainer uses a stable Ubuntu release instead of trying to run the alpha stuff [00:44] jbicha: i can't find any setting that this thing tweaks that's stored in gconf [00:45] at least in git master version of g-t-t [00:48] yeah, you should open a bug and ask him to kill the gconf support [01:03] Huh. What's caused https://launchpad.net/ubuntu/+source/libgusb to be in the ‘publication pending’ state for 24h? [01:06] RAOF: https://launchpad.net/ubuntu/raring/+queue?queue_state=0 [01:16] jbicha: Ah, of course! Binary NEW. Silly me. [01:17] The reason I uploaded that was to add the new binary package! === VD is now known as Guest18424 [04:20] RAOF: you got a bug and a merge proposal :) bug 1143448 [04:21] Launchpad bug 1143448 in colord (Ubuntu) "colord postinst fails with "27: [: which: unexpected operator"" [Undecided,New] https://launchpad.net/bugs/1143448 [04:22] * RAOF is curious as to how the package installed fine locally [04:31] jbicha: Thanks. === pitti_ is now known as pitti [05:59] Good morning [06:53] good morning [09:07] hey! [09:08] ello [09:08] does anyone else find chroium a bit over sensative for scrolling atm? every time I attempt to scroll I end up on a different application [09:10] czajkowski: is your mouse over the tab list ? [09:10] czajkowski: or perhaps ctrl key down ? [09:10] nope [09:10] it's only happening on chromium as well which makes it a little bit more annoying [09:10] lifeless: morning btw [09:11] czajkowski: morning :) [09:12] hi desktop team! [09:13] hey desktopers [09:13] Laney, chrisccoulson: hey, had a good w.e? [09:13] seb128, yeah, not too bad thanks. and you? [09:13] chrisccoulson, mine was quite ok, thanks ;-) [09:14] yeah, pretty quiet [09:14] ah, crap, chromium failed to link on i386 because the linker hits the virtual address space limit :/ [09:14] so we can have chromium as long as we don't want any debug symbols and crash reports [09:15] chrisccoulson, still having fun with chromium I see [09:15] :> [09:15] seb128, yeah, it's becoming a pain [09:15] chrisccoulson: can't you do a frankenbuild? [09:15] seb128, it's too big ;) [09:15] install 64-bits kernel and gcc with multiarch [09:15] erm linker, probably don't need gcc of amd64 [09:15] mlankhorst, on the archive builders? [09:15] :P [09:16] yeah probably a bad idea [09:16] that's the problem. we can't build it in the archive ;) [09:18] does building all chromium things shared instead of linking into a single blob help? [09:18] oh right probably not, it's just too big [09:19] we may have to start splitting things out of the main binary, but i suspect that's not going to be an insignificant amount of work [09:19] and fairly unsupportable in the long term too [09:19] yeah I just thought you could build some part of it shared [09:19] but even then it's probably too big [09:19] too many c++ debug symbols [09:19] chrisccoulson, just build without debug symbols on i386... [09:19] seb128, that kind of sucks though :( [09:19] it does [09:20] but only so much we can with the resources we have [09:20] chrisccoulson: http://code.google.com/p/chromium/wiki/LinuxFasterBuilds [09:21] you could try the shared thing, see if it works or not [09:21] or remove just the webkit debug symbols [09:22] mlankhorst, oh, i didn't know about the "component-shared_library" option. it looks like that might do what i want (still using bundled libraries but splitting them out) [09:22] i'll give it a try === tkamppeter_ is now known as tkamppeter === vrruiz_ is now known as rvr_ [10:20] xnox! [10:20] !xonx [10:20] Factoid 'xonx' not found [10:21] Laney: hola! =) [10:21] hello! [10:21] so, you know the latest upstart package from stgraber? how it moves its Xsession.d file to 99upstart? [10:22] Laney: yeah.... [10:22] it overwrites all this lovely stuff (except dbus obviously): http://paste.ubuntu.com/5584714/ [10:22] how do we go about getting it back [10:30] Laney: right, so even before we were chewing some of these things. let me check what it is currently doing after the move. [10:31] chewing? [10:31] those things can generally work by setting some environment variables to be available in the session [10:32] but I don't think making gnome-session be start on started somehugelistofthings works [10:39] UDS session? ^o) [10:42] to produce a buglist ? [10:49] Laney: the point is that all of those snippets should be handled as upstart jobs and/or be sensitive to UPSTART environment variable: aka there is UPSTART environment variable and I do ship an upstart job, so let's not polute the STARTUP var, vs I have no idea about upstart and will polute STARTUP var. [10:52] Laney: we have gnome-session & dbus, the bits that are missing are the ssh & gpg agents. [10:52] and im-config [10:52] yes. [10:54] how do you express that in upstart-fu then? so that e.g. $GPG_AGENT_INFO is available to the gnome-session [10:54] Laney: looking at other STARTUP scripts we are also potentially missing consolekit session?! although we will be planning to move to logind at the vUDS. [10:55] Laney: so in user-session upstart mode we support inheriting environment and setting/modifying global environment variables for _all_ jobs from any job. [10:55] specifically for these type of weird jobs ;-) [10:55] but you have to make gnome-session wait on those jobs then don't you? [10:56] you have to make X wait for them [10:57] (thats why it is callled Xsession :P ) [10:57] hmm? [10:57] that's done implicitly by the STARTUP mechanism [10:57] Laney: gnome-session already waits on dbus. [10:57] yes [10:57] but not e.g. gpg-agent [10:57] Xsession is executed on startx as well as on DM startup [10:57] and dbus is on-startup. [10:57] not on login [10:58] ogra_: the new user jobs stuff shortcuts out the Xsession files [10:58] so your gpg-agent job should be "on starting dbus" and ideally a task, thus blocking dbus job, without explicetely modifying the dbus job. [10:58] right [10:59] oh, on starting blocks? [11:00] start on starting foo blocks foo until i've started? [11:00] and thus if I initctl set-env --global some stuff foo will see it? [11:01] it blocks until the depending process emitted the "starting" event [11:03] ah, 'task' was a hint [11:05] Laney: yeah, so in the task job one would want to export all the vars. The question is whether it's fast enough. E.g. I'm not sure if dbus job will notice the new environment vars, but gnome-session will. [11:06] I was thinking they'd be start on starting gnome-session, not dbus [11:06] that would be different to the original order though [11:06] hmm [11:07] maybe not then [11:07] Laney: maybe. Reading cookbook it seems like task is not necessory and normal jobs can block as well. [11:07] http://upstart.ubuntu.com/cookbook/#block-another-job-until-yours-has-started [11:07] Laney: as long as nothing tries to access gpg/ssh-agents over dbus and expect dbus to know about them..... [11:08] which would be weird. And I don't expect order to mater. But hey it's all sequential at the moment. [11:11] let me try it [11:11] * ogra_ still wonders how we plan to do such big transitions in a rolling release [11:12] stuff will break and will stay broken over a certain amount of time until we have shaken out everything [11:13] it's a bit tough considering that we made having users run proposed be an antigoal [11:13] otherwise maybe we could do such things there with migration blocks [11:13] maybe we just have to use PPAs more/better [11:17] well, if your transition touched half of the distro PPAs will get hairy [11:17] C++ ABI bump in libc anyone ? [11:34] so glad I've never been around for one of those [11:37] Laney, we can ask Ryan to break the glib api if you want some fun :p [11:37] ah you know I love a good transition anyway [11:37] * Laney snuggles ghc [11:38] Laney: one x1.14 coming up? :D [11:53] xnox: seems 'task' is required [11:53] 'sleep 5' doesn't block startup without it [11:53] good news is the env variable is set in my session [11:55] Laney: hm. I'd think that pre-start start portion will block, the exec/script will not. [11:55] Laney: as the "starting" event is emitted after pre-start. [11:56] yeah [11:56] I suppose it won't work for daemons [11:56] as task means that it has to transition back to stopped before unblocking === MacSlow is now known as MacSlow|lunch [12:52] qengho, wow, looks like you were really close with your last chromium build [13:01] seb128: ping? [13:02] how can I disable ALT to open the search bar into app menu ? [13:02] seb128: any opinion on making libreoffice-core C/P/R libreoffice-filter-binfilter ? [13:02] just can't get used to it, it always open when I don't want it :( [13:02] Sweetshark, does anything has a versioned depends on libreoffice-filter-binfilter? [13:03] xclaesse, system settings -> keyboard -> keybindings -> launcher -> the bottom line [13:04] seb128, perfect, thx :) [13:04] xclaesse, yw [13:04] btw, when opening settings focus should go to search bar ;-) [13:04] seb128: the 'libreoffice' metapackage used to have a dep on it. Apart from that: no. [13:05] Sweetshark, I would just use C,R,P, we can decide to replace it by a dummy transitional package if update-manager gets confused and that proves needed [13:05] but let's start with the easy solution [13:06] seb128: k. currently its just an unversioned breaks. I guess rene want people to explicitly remove binfilter ... [13:07] well, C,R,P will lead to it being removed as well [13:08] seb128: sure, but the R is a lie and -core doesnt do what binfilter did ;) [13:10] oh, just use the versionned breaks then... [13:10] just make sure you have proted it to surfaceflinger before we switch the desktop :P [13:10] https://soltesza.wordpress.com/2013/03/03/ubuntu-may-switch-to-android-technologies-to-keep-the-linux-desktop-competitive/ === MacSlow|lunch is now known as MacSlow [14:14] i guess we may have to teach localed about how to install packages on ubuntu as well... [14:14] if we want to go that route [14:15] desrt, rodrigo has a branch in git by then for that iirc [14:15] oh nice [14:15] well, maybe it was not localed [14:15] is that new work, or from before? [14:15] but the region capplet [14:15] ah [14:16] lennart said that he'd be willing to accept patches upstream that made my life easier [14:16] well, we have been looking at replacing language-selector by the region capplet for a while [14:16] nice [14:16] so if we need a package install hook or something, probably we can get it [14:17] excellent [14:18] so we have a long-term plan to move ubuntu to the upstream one? [14:18] fwiw, the ibus stuff is starting to get sane by 3.8 [14:18] they realise "oh ya... maybe we actually need per-window keyboard layout after all" [14:18] they started to add lot of options back, that's good [14:18] like the "never blank the screen" is back as well [14:18] Laney, xnox: right, my plan was to update the Xsession script for ssh-agent, gpg-agent, ... to be no-ops if $UPSTART is set and then have those be upstart jobs that start before gnome-session (currently "start on starting gnome-session" is good but we may have to change that if we ever want to support something other than gnome-session in there) [14:18] and some of the options are annoyingly changed [14:19] which we had as a distro patch since GNOME3 [14:19] but in a way, better [14:19] like, that xkb options window is completely gone [14:19] desrt, long term we do plan to move to the upstream region panel yes [14:19] but a lot of the more common choices are moved to a better place [14:19] we just need to sort out the ibus story, which includes having an indicator written on our side [14:19] and we need to plug the langpacks somewhere in there [14:20] i was about to volunteer until you added the part about 'having an indicator written' :) [14:20] ;-) [14:20] anyway [14:20] you can try to push larsu to volunteer for that one :p [14:20] today i'll sort out hostnamed [14:20] then i'll start looking into locale stuff [14:20] attente: maybe you're interested as well? [14:20] great [14:21] seems like we unify a little bit more with upstream this cycle... [14:21] are we still planning for u-c-c? [14:21] desrt, btw, how do you handle the fact that ubuntu-system-services conflicts with systemd-services atm? [14:22] i don't know [14:22] k [14:22] i guess i uninstalled ubuntu-system-services :) [14:22] did we have a hostname-changer in there before? [14:22] I've a deb of u-s-s without the systemd stuff dropped [14:22] no we didn't [14:22] so that was another g-s-d mechanism? [14:22] we didn't even have the prettyname stuff [14:22] no [14:22] we just didn't have an interface to change hostnames [14:22] missing feature [14:23] ahh [14:23] it's not like it was the most important feature ever ;-) [14:23] hi seb128, desrt [14:23] i bet there's just a patch in there [14:23] that needs to be dropped [14:23] thumper: hey! [14:23] thumper, hey [14:23] thumper, what are you doing up at this time? [14:23] nice to see you around these parts :) [14:23] seb128: I'm in atlanta [14:23] desrt, u-c-c ... yes, planned, after we land GNOME 3.8 [14:23] huh [14:23] juju sprint [14:23] thumper, I see ;-) [14:24] my dream come true and suddenly i am unsatisfied :) [14:24] lol [14:24] seb128: i have systemd-services and ubuntu-system-service both installed [14:25] no conflict here, apparently? [14:25] desrt, k, so there is a chance that org.freedesktop.hostname1 is claimed by u-s-s [14:25] that's the case for me [14:25] according to d-feet [14:25] neat [14:26] so we implement the interface [14:26] but the UI doesn't work [14:26] well, it works, it gets the hostname [14:26] we don't implement the edit part [14:26] ah [14:26] so we need also a 5th upload then [14:26] stop having u-s-s do this [14:27] otherwise it's gonna muck up hostnamed [14:28] desrt, yeah, let me get that upload, I did hack on that during the London sprint week [14:28] desktop team ppa gets an early warm-up to the week :) [14:28] ;-) [14:28] desrt, btw I might wait for gtk 3.8 to upload the new serie to raring [14:28] i wonder if there is conflicting policy installed by those two services [14:28] debian guys pointed that DND is broken with the current version [14:28] seb128: i had a feeling this might happen [14:28] that's fine [14:29] desrt, @conflicting policy, systemd-service Replaces u-s-s [14:29] desrt, so it overrides its dbus files [14:29] if it replaces, why do i still have both installed? [14:29] Replaces means "you can overwrite files" [14:29] huh [14:29] til [14:30] desrt, u-s-s is still needed for a few things [14:30] like it has a "set system proxy" interface [14:31] seems weird to be living in a state where one package is replacing files of another, though === m_conley_away is now known as m_conley [14:31] desrt, right, that's buggy and transitional [14:31] desrt, https://launchpad.net/~ubuntu-desktop/+archive/ppa/+sourcepub/3009976/+listing-archive-extra [14:32] nice [14:32] desrt, with that upload they stop conflicting [14:32] hostnamed is weird [14:32] or rather, the panel is weird [14:35] stgraber: Laney: well there is lxde-session starter thing... maybe they all should have a common even they start on and a dummy job that emits "about-to-start-session-starting-thing"..... [14:35] sounds ugly. [14:36] so i guess we somehow have a different polkit setup than fedora does [14:36] because on fedora the hostname is changeable as soon as the dialog opens [14:36] and for ubuntu it's not [14:37] and it seems that nobody thought to add an unlock button to that dialog incase different people have different policy [14:37] desrt, could be that we need to update /var/lib/polkit-1/localauthority/10-vendor.d/com.ubuntu.desktop.pkla [14:38] erm [14:38] this is somehow totally different in 3.8 [14:38] oh right [14:38] the new js based polkit [14:38] https://bugzilla.gnome.org/show_bug.cgi?id=687772 [14:38] Gnome bug 687772 in Sharing "Implement the Sharing panel designs" [Normal,Resolved: fixed] [14:38] the security team is making nightmares over it [14:39] specifically http://bugzilla-attachments.gnome.org/attachment.cgi?id=233431 [14:39] let's make polkit depends on a js interpreter... [14:39] xnox: yeah, I started to think about it a bit. Basically have an xsession-init.conf job that's "start on startup", then have this one emit "xsession SESSION=$NAME" [14:39] xnox: as a result, gnome-session would then be changed to be "start on xsession SESSION=gnome-session and started dbus" [14:40] desrt, http://lists.freedesktop.org/archives/polkit-devel/2012-June/000368.html [14:40] seb128: i know about this [14:40] k [14:40] desrt, well, we didn't do that update yet either [14:40] we discussed it quite a lot at FOSDEM [14:40] the RHEL security guys flipped their lid about this as well [14:40] what was the outcome? [14:40] seems that it's a change that nobody likes [14:40] i don't know [14:41] k [14:41] pitti knows more about it [14:41] someone proposed that we may end up getting the old format back alongside the new format, somehow? [14:41] that's an update I'm not in a hurry to make in any case [14:41] but also: davidz is no longer with red hat [14:41] so uh... [14:41] oh, he's not? didn't know about that [14:42] ya. got sucked into the black hole [14:43] desrt: I don't know much more, just that I don't like it at all either [14:44] desrt: oh, I wasn't aware of that either -- did he go to the big G? [14:44] ya [14:44] seb128: FYI, I updated udisks today, but not gdu yet (as this needs libdvdread in universe) [14:45] pitti, I saw, thanks! [14:45] pitti, will make those users, who are afraid about devices not being powered off on eject, happy [14:46] :) [14:50] desrt, subscribed you to https://blueprints.launchpad.net/ubuntu/+spec/foundations-1303-consolekit-logind-migration btw [14:50] oh. that nightmare :) [14:50] * desrt is starting to notice a theme here [14:51] hehe [14:51] stgraber: Yeah, that's basically where I'm at. My upstart skills are crap enough though that I can't make these jobs work; can't get upstart to track the processes right (in the case of im-launch there's a few processes spawned for example) [14:51] pitti: btw: any plans for having a PID1 systemd drop-in? [14:52] desrt: not from my side; this is pretty much foundations team area [14:53] Laney: right, I'm busy with NM/ofono this morning but I'll poke upstart again this afternoon. 1.7 was now released upstream so I need to get things in shape to land in Ubuntu now. [14:53] desrt, just aftter we spent massive work on getting upstart into user session management ? [14:53] verz verz unlikely [14:53] *very even [14:54] upstart managing the user session is..... annoying [14:55] desrt, how so ? [14:55] desrt, systemd managing them is better ? [14:55] ogra_: it's going to cause us problems going forward [14:55] like when dbus-in-the-kernel lands [14:55] will be a systemd-based feature [14:55] upstart will learn to cope [14:55] and, like usual, we will be playing catch-up [14:55] ogra_: problem is that upstart tends to cope ~2 years after [14:56] look how long logind is taking, as an example [14:56] upstart is far ahead in other areas [14:56] you dont really want systemd on a production server for example [14:56] i don't know of any [14:56] afaik not even RH does that yet [14:56] people keep making hand-wavy arguments around this point, but i don't understand why [14:57] because systemd is far from being done as a generic init [14:57] there are many cases where it cant cope atrm [14:57] and which lennart simply ignores [14:57] i keep hearing lots of statements like 'many cases' [14:58] /usr on a foreign machine or different disk is a very typical server setup in datacenters [14:58] systemd cant handle that [14:58] desrt, ogra_: let's not have another systemd/upstart argument today [14:58] yeah [14:58] * ogra_ stops :) [14:58] thanks ;-) [15:01] ogra_: (also, it's not true) [15:01] pitti: stop :p [15:01] * pitti STFU, sorry [15:02] we'll go all android anyway ... so that discussion is moot :P [15:02] ah right -- time to dust off my ancient java sk1llz from uni? :-) [15:02] hehe [15:16] chrisccoulson, bah, I deleted my firefox's corruption screenshot from the other day ... if you have one please add it to https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/1144558 [15:16] Launchpad bug 1144558 in xserver-xorg-video-intel "Images corruption in firefox when using "sna"" [Undecided,New] [15:17] seb128, sure. 1 second [15:17] chrisccoulson, thanks [15:18] * seb128 is out for some exercice, be back in ~1h [15:40] Hey guys, it turns out we're moving to Android SurfaceFlinger instead of Wayland. [15:40] http://soltesza.wordpress.com/2013/03/03/ubuntu-may-switch-to-android-technologies-to-keep-the-linux-desktop-competitive/ [15:41] It's deductions like this that make me scoff at Sherlock Holmes stories. There are so so many places to go wrong. [15:44] seb128: add a 6th package to your list :) [15:44] policykit-desktop-privileges needs an update [15:46] pitti: where is this package maintained? [15:46] desrt: just apt-get source or UDD [15:46] no magic [15:46] huh [15:46] it's maintained via source package uploads? [15:46] neat version control system :) [15:46] desrt: bzr branch ubuntu:policykit-desktop-privileges :) [15:46] (IOW, UDD) [15:47] iow? udd? [15:47] * desrt is not hip with the lingo [15:47] "in other words", "ubuntu distributed development", sorry [15:50] seb128, FYI: bug 1137907 should keep GTK+ 3.7 in the desktop PPA for now (we were planning to move that to raring eventually right?) [15:50] Launchpad bug 1137907 in gnome-screenshot (Ubuntu) "After pressing my Print Screen button, screen turns white and stays white" [Undecided,Confirmed] https://launchpad.net/bugs/1137907 === Ursinha_ is now known as Ursinha [15:56] why does ubuntuone-client depend on apport? [15:57] larsu: to help you uninstall things easier when you setup a new ubuntu system =) [15:58] desrt: erm, lol?! [15:58] * larsu wants u1, but not apport [15:58] oh ya. you love your u1 :) [15:58] just disable it in /etc/default/apport [15:59] its not like it eats massive amnounts of disk space [15:59] ogra_: ah, didn't know about that one. Thanks! [15:59] still, the dependency is a bit weird... [16:00] file a bug to downgrade it to a recommends [16:00] doing so right now :) [16:02] polkit is smart. [16:02] * desrt is happy [16:04] seb128: do you mind if i do uploads to desktop team ppa? [16:06] oh. he's doing that weird being outside thing [16:07] you won't be able to - you're not in the ubuntu-desktop team [16:07] interesting [16:08] seb128: http://fpaste.org/QK5O/ for when you get back :) [16:08] yeah, and it's a team that grants upload rights [16:08] so ... there's A Procedure [16:09] my Normal Procedure is 'ping seb' :) [16:09] * desrt will keep doing that [16:09] if you make your own PPA then we can copy from that into the desktop one [16:09] let me upload that thing now [16:09] nah... there's no point in that [16:10] we're using the desktop team PPA to stage this changeset before tossing it in the distro already [16:10] no need to add yet another layer of indirection [16:13] ok done [16:13] I gave it a PPA version [16:15] I have some questions to the online UDS. Can I participate without G+ account? Or without webcam (audio-only)? [16:16] i guess to run a session you need G+ [16:16] for partitcipating it should work via IRC and audio (guessing here though) [16:16] there's some thing to get audio out of hangouts? [16:17] oh, that i dont know ... but i would expect the old ways to still work [16:18] worst case through an analog speaker jack with some volume levelling into the input jack i bet [16:18] (to then forward it to icecast) [16:18] Laney: you could "pacat ... >capture". [16:18] yes I am aware there are ways [16:18] someone will need to set this up though [16:19]