=== duflu_ is now known as duflu [02:06] robert_ancell, your unity-session-compositor LightDM branch is working really well for me, but is there any way to configure environment variables to pass to unity-system-compositor before starting it? [02:07] bregma, not that I can think of, what sort of variables? [02:07] ANDROID_ROOT in particular, used to find device configuration files (instead of /system) [02:08] bregma, you can make a shell script for the compositor and set the name of that in lightdm.conf [02:08] if that's something more permanent, we could look into options [02:08] yeah, that was my next plan, assuming even using device files will solve my problem [02:09] I'm not sure off hand if the global env variables make it through to the compositor or we have a clean environment [02:10] well, it's probably a sort of last-resort solution anyway, since going that route woukld require shipping device configuration for pretty much every known device [02:10] but if that's what it takes.... [02:12] yeah, shell script time for you :) === tim is now known as thumper === tim is now known as thumper [03:26] cyphermox, hey i installed urfkill... how will i know it's working? what steps should I take o test it? [05:49] Good morning [06:16] pitti: Thanks for chacing down that espeak issue. That really was a weird one. [06:17] Well... Working around the dpkg issue that is. [06:25] TheMuso: well, it's not dpkg, it's overlayfs/aufs [06:25] TheMuso: and I'm far from proud about that workaround, but I didn't find a better one [08:02] Laney: do you know what holds glib 2.38 back from unstable? [09:02] morning [09:02] pitti: the ftbfs [09:02] hey laney [09:02] good morning desktopers [09:03] hey seb128 [09:03] how's it going? [09:04] bonjour seb128 [09:04] hey Laney [09:04] hola [09:04] pitti, guten tag! wie gehts? [09:04] Laney, good! you? [09:04] Laney: ah, of course [09:04] bigon was looking at the build failures a bit but I don't think he got anywhere [09:04] seb128: gut, danke! [09:05] good thanks seb128! [09:05] pub quiz last night ;-) [09:07] seb128: seen https://bugs.launchpad.net/ubuntu/+source/gtk+3.0/+bug/1260473 ? [09:07] Launchpad bug 1260473 in Ubuntu Kylin "The slide bar that allow to rezise the partition does not appear" [Medium,Triaged] [09:11] Laney, now I did ... that seems similar to https://code.launchpad.net/~larsu/ubiquity/lp1260396 [09:11] that fix landed right? (I didn't follow up, that was during my holidays, just saw it mentioned when looking to what happened while I was not there) [09:12] Laney, the mp got approved only on monday and we didn't have an ubiquity upload since [09:12] not sure that's the same [09:13] yeah, me neither [09:14] it would be good to land an update with the fix from Lars though [09:15] would it only be to avoid debugging something that's already fixed (if that turns out to be the same issue) [09:15] xnox, hey, do you plan to do an ubiquity update in trusty? [09:22] didrocks, hey, do you have any opinion on what integration tests we should have before merging https://code.launchpad.net/~diegosarmentero/ubuntu-system-settings/click-updates/+merge/195729 in? [09:23] didrocks, there is some pushback in recent comment saying it makes the review longer and should be in another merge later on, I'm just unsure that they are going to care once that one land so I'm tempted to ask for the branch with tests to be ready before we merge that one [09:23] seb128: for system image update, it would be nice to use barry's mock and that the mock so that we ensure we receive the right signals [09:23] seb128: well, if I were you, I would really bind the 2 merges (tests and code) together, otherwise, you know how it goes [09:24] didrocks, I guess they are going to argue that the system image is your code and that they didn't change/don't own it :p [09:24] right [09:24] seb128: they did change it, right? [09:24] they changed the UI and refactored [09:24] yeah, both the UI and backend [09:25] didrocks, well, it looks like they kept most of your updates.cpp, just did s/updates/systemupdate [09:26] in any case, let's me try to get them to add some more testing there [09:26] seb128: there are some new objects retaining all states, not sure the signaling is still working [09:26] ok [09:26] seb128: and yeah, you know how QML can be, when changing it :p [09:36] didrocks, so, can I try to land the u-s-s update we put in CI train yesterday or is the system down due to the canonistack issues? [09:36] seb128: you can try, there are a lot of network issues [09:37] so be prepare to run and rerun [09:37] didrocks, is #is looking at those? [09:37] seb128: not sure, asked twice yesteday [09:37] nothing happened [09:37] did you get anyone responding? [09:37] in meeting… [09:37] well, everyone using canonistack [09:37] so yeah [10:08] seb128: after bug #1260473 is fixed, ubiquity will be uploaded. [10:08] Launchpad bug 1260473 in Ubuntu Kylin "The slide bar that allow to rezise the partition does not appear" [Medium,Triaged] https://launchpad.net/bugs/1260473 [10:08] xnox, so it's a different issue than the one larsu fixed? [10:08] which also appears to be a gtk+ / themeing bug. [10:08] xnox, do you need help debugging it? is it easy to reproduce? [10:09] seb128: yeah, larsu's fixed is good. the installer is no 3X width the monitor screen. [10:09] seb128: well... we have a custom widget that is using deprecated gtkwidget themeing.... what could go wrong?! =) [10:10] * xnox needs to do a minimal example. As going from start to "dual-boot install" takes some time. [10:13] xnox, yeah, having a small example code would be nice [10:23] seb128: ok, seems we are going to redeploy another instance in lcy01 [10:23] seb128: as I can't get any instance up on lcy02, it means I can't rsync the filesystem :/ [10:23] so, I'll have to rerun your build job with the same parameters [10:24] (and reconfigure the silos) [10:24] mind if I do that? So that you are unblocked ASAP [10:25] Laney: which train are you on to FOSDEM ? [10:25] 1532 or so [10:25] :( [10:25] didrocks, whatever unblocks the upload wfm [10:25] seb128: ok ;) [10:25] y u sad [10:25] asac: FYI, applying the band-aid version (I'll rerun the stuff for seb128 to not have to redo it himself) ^ [10:25] didrocks, I've other changes to land today btw, not sure if we want to batch those or just redo a landing [10:26] Laney: I'm on the 12 [10:26] seb128: thanks for checking my merge-request on indicator-power and spotting the obvious whoopsie. anyway, i was also wondering whether this patch doesn't introduce problems for ubuntu-gnome, or are they not using indicators at all? [10:26] seb128: as you prefer, we can do that, sure [10:26] czajkowski: the other uk-ers too? [10:26] yp about 10 going from our LUG [10:26] :D [10:27] Laney: stop by the NoSQL room on sunday will be there all day for a cupcake! [10:27] ^_^ [10:27] ;D [10:29] ochosi, yw, larsu said he's going to tweak it and land it I think [10:29] didrocks: what is the bandaid version? [10:30] is that the bandaid of the bandaid :P [10:30] ochosi, well, Ubuntu GNOME doesn't use indicators so they shouldn't be impacted by whatever the indicator code is doing [10:30] asac: redeploying somewhere else and rerunning myself the job so that seb128 is in the same state [10:30] seb128: ah, nice! [10:30] seb128: ok, that's good to hear, cause i was afraid that the gnome-control-center call was for them... [10:30] didrocks: yeah. unfortunate, but this exercise of redeploying etc. is probably worth it [10:31] didrocks: how would the system behave if we just deploy it 2-3 times (e.g. on different regions) [10:31] and the switch to whereever we get what we want? [10:31] ochosi, I think the g-c-c call was more "we didn't think much about the different usecase and let the old command as default" [10:31] asac: not possible accross regions [10:31] asac: the rules for having something quick were clear: one instance, one filesystem [10:32] didrocks: sure, but we could have one spare? [10:32] asac: well, you have to rsync the filesystem between the 2 [10:32] * asac guesses that causs races etc. [10:32] which isn't possible here [10:32] as we even can't brought up an instance on lcy02 to attach to filesystem [10:32] seb128: i see. well i'm sure ubuntu-studio will also be greatful for this getting in, as it'll make it very easy for them to also use -power if they want to, so thanks for the quick review! [10:32] ochosi, yw ;-) [10:36] larsu: let me know if you want me to update the MR or if you're just going to throw the "gchar *path;" and "g_free (path);" into the Unity-if yourself [10:37] didrocks: that is only if you dont want to risk loosing progress on active landings/silos? [10:37] e.g. what will we loose if we redo the filesystem? [10:37] asac: right [10:37] ok thats fine [10:37] all progress [10:37] all silo attributions [10:37] for a moment i was scared that we need to backup filesystem [10:37] ochosi: I'm doing it right now. Thanks! [10:38] asac: and changing instance means changing some metadata on the spreadsheet, meaning if people keep IP directly, they won't work [10:38] didrocks: the IP is recorded in the spreadsheet for the silo? what else? [10:38] well, here, we need to have new public IP [10:38] e.g. what else is instance specific [10:39] and new firewall rules [10:39] but that's only one time, we'll keep both opened then [10:39] ok. we cant reassign the IP:) [10:39] like elastic ips? [10:39] maybe we have something like that? [10:39] note: this is just theory talk so i understand better [10:39] I don't think they have this notion accross instances [10:42] larsu: ok, thank *you* in fact :) [11:13] bah [11:13] webkit ooming on the porterbox [11:16] Laney, well, does it mean that it went far enough that it should build? [11:16] not quite ... [11:16] :-( [11:18] -Wl,--reduce-memory-overheads HEAR MY CALL [11:18] Didn't Debian fix the FTBFS on i386 recently? [11:20] I'm not looking at that [11:21] (ppc) [11:21] Ah, yes, ppc [11:22] Then I am waiting for your fix to apply it to qtwebkit :) [11:22] does that track webkit 2.3.x? [11:23] 2.3 is webkitgtk numeration, not upstream one [11:23] you know what I mean [11:24] I'm trying something quite similar to http://pkgs.fedoraproject.org/cgit/webkitgtk3.git/tree/webkitgtk-1.11.5-libatomic.patch?id=2d4347bc2e76a07a8a0d6e8260730dfc91b0410e [11:24] qtwebkit 5.2.0 is webkit 538.1, don't know about the gtk one [11:25] trying one more time with --reduce-memory-overheads [11:25] otherwise I'll just take the chance probably [11:29] seb128: sil2100: can you check that you can login in and have enough rights on http://162.213.34.102/ ? (to run some jobs) [11:31] didrocks, what job can I try on? [11:32] seb128: just check if you have the "build" button [11:32] didrocks, I don't [11:32] didrocks, on what job should I have one? [11:32] hum [11:32] everyone but deploy [11:32] I tried http://162.213.34.102/job/landing-001-1-build/ [11:32] and configure silo [11:32] ok, let me check [11:33] seb128: when you logged in, you have tick the teams you are in, right? [11:33] hmm, I have some problems logging in [11:33] didrocks, yes, u-c-d [11:33] didrocks, but even without that, before I had a build button even when not logged in [11:33] clicking on the button would send me to the auth page [11:33] didrocks, http://ubuntuone.com/3X3SNx74hOelkX9J5kuZDE [11:34] seb128: I don't see a build button here, you meant, you had one when not logged in? [11:34] didrocks: do I have to create a new ubuntu one account here? Since it seems I cannot get my existing credentials to work [11:35] didrocks, yes, remember the discussion the other time? I had the "build" button even when not logged in before [11:35] sil2100: yeah, it's using staging sso [11:35] didrocks, I first though it was a bug, but clicking on it was sending to a login page [11:35] seb128: it's when you go to the build page directly [11:35] (this isn't the build page) [11:35] oh [11:35] I got confused by jenkins UI again [11:36] well, anyway, I'm logged in and checked the u-c-d team (which was the only listed) [11:36] ok [11:36] let me check why [11:36] didrocks, is the canonical team supposed to be in the list? [11:36] seb128: no, is it? [11:36] (it's not) [11:36] ok ;) [11:36] we can't [11:36] no, but I though it was before [11:36] as we are not whitelisted [11:36] ok [11:37] there was a canonica-ci [11:37] canonical-ci* [11:37] * didrocks finished to resetup the rsync [11:37] ok, now let's see those creds [11:43] seb128: better now? [11:44] didrocks, yes [11:44] didrocks, http://ubuntuone.com/3X3SNx74hOelkX9J5kuZDE [11:44] ups [11:44] didrocks, http://162.213.34.102/job/landing-001-1-build/1/ [11:44] "/tmp/hudson1182320129477473757.sh: line 5: cd: /var/lib/jenkins/silos/landing-001: No such file or directory" [11:44] didrocks, :-( [11:44] seb128: yeah, I didn't assign you the silo yet [11:45] and the firewall for rsync isn't opened by IS [11:45] (as the IP changed) [11:45] seb128: want me to attribute the silo so that you can build? [11:45] or you want to add more components first? [11:53] sil2100: in addition to seb128's right, you should be able to see the build button for deploy-jenkins === alan_g is now known as alan_g|afk [11:54] didrocks, let's attribute the silo, we can do another landing later [11:56] ok :) [11:56] launchpad authorization done [11:56] let's attribute [11:57] seb128: ok, you can try to run a build [11:58] seb128: sil2100: I'm reverting to login.ubuntu instead of staging [11:59] \o/ [12:00] sil2100: mind testing and so answering if the creds are fine? [12:00] * didrocks pushed his changes to prod [12:00] didrocks: ok, so I'm logged in now and I had ubuntu-unity team in it [12:00] sil2100: so, do you see the build button to the deploy-jenkibs job? [12:01] didrocks: and I see build with parameters on landing-001-1-build [12:01] didrocks: yes [12:01] great! [12:01] :) [12:03] sil2100: can you try to reassign the second silo? would be a good exercise :) === alan_g|afk is now known as alan_g [12:04] sil2100: only the jenkins job is needed, not the spreadsheet part [12:04] asac: so, we're back on shape, I just rebuilt an instance and improved the install rules [12:05] asac: just the status isn't synced back to the spreadsheet and we can't copy packages to distro until rsync from the new IP address is allowed [12:05] asac: ev will work on that once he's back [12:05] didrocks, build running (sorry, was have lunch, I'm back now) [12:05] seb128: let's look if it's starting fine [12:05] didrocks: how can I do a reassign by using only jenkins? Which job to use? [12:05] didrocks, it is, it's installing packages [12:06] seb128: sweet! [12:06] sil2100: you have the link in the assign silo menu [12:07] * didrocks now praises that there is just one call to change to update all links :p [12:07] seb128: great, uploaded to the ppa it seems [12:08] nice [12:08] let's hope the gpg is rightly setup [12:08] we'll see it soon :) [12:08] seems it signed successfully at least [12:08] didrocks: destination is Ubuntu archive, yes? ;) [12:08] launchpadlib creds matches [12:08] sil2100: yeah, so keep it empty [12:09] seb128: https://launchpad.net/~ci-train-ppa-service/+archive/landing-001/+builds?build_state=building [12:09] didrocks: reconfigure_silo set to landing-002 (just confirming if doing it irght) [12:09] ok, now, I can work on what I planned at 10am :p [12:09] didrocks, great [12:09] sil2100: you can try this ;) (it's a special case, as we don't really have a config file) [12:09] sil2100: let's see ;) [12:10] sil2100: if you don't set anything, it will take next available silo anyway [12:10] which is landing-002 :p [12:10] didrocks: oh noes... sil2100 is missing the Job/Build permission [12:10] ;_; [12:10] sil2100: hum… you checked ubuntu-unity, right? [12:10] oh [12:10] I redeployed all jobs [12:10] Yes [12:10] not that one [12:10] uno momento [12:11] sil2100: retry [12:12] didrocks: yay, seems to work! [12:12] great! [12:12] Finished: SUCCESS [12:12] sweet, all good then :) [12:12] didrocks: ok thanks. i hope it takes not too long [12:12] Assigned to landing-002 [12:12] asac: let's see… [12:12] SO all is cool [12:12] sil2100: thanks :) [12:12] sil2100: back to landing -> did yu publish the keyboard? [12:12] didrocks: thank *you* ;) [12:13] (I don't see it in -changes) [12:13] didrocks: yes, I pressed the button at least [12:13] as I'll need that to update the seeds [12:13] hum, and the artefacts were created? [12:13] Let me check [12:14] I have an idea what can happen [12:14] http://q-jenkins.ubuntu-ci:8080/view/cu2d/view/Head/view/Services/job/cu2d-services-head-3.0publish/127/console <- hum [12:14] So not good... [12:14] ok [12:14] I guess the rsync on snakefruit doesn't happen [12:14] as the script is set -e [12:14] and now that the jenkins other node isn't reachable… [12:16] sil2100: yeah, that was it, unblocked [12:16] and added || true to first request [12:16] didrocks: thanks ;) [12:16] sil2100: yw! think about looking at -changes please :) [12:16] didrocks: there is this as well -> https://code.launchpad.net/~sil2100/ubuntu-keyboard/some_packaging_changes/+merge/202654 (not top priority) [12:17] didrocks: it's still not lintian-free, but well... [12:17] didrocks: will do :)! [12:17] sil2100: approved [12:22] sil2100: I don't see hub and libcolumbus though [12:48] seb128: hum, do you see deja-dup, wasn't it in g-c-c? (or visible in the dash) [12:48] I may have remove the g-c-c plugin part, just checking [12:49] shrug [12:49] hum, deja-dup binary package has the plugin normally [12:49] 3 person to ask the question this week :p [12:49] ahah :) [12:49] didrocks, it's showing in unity-control-center and not gnome-control-center [12:49] mterry knows about it [12:50] (it's supposed to show in both) [12:50] we didn't switch defaults yet because everything didn't get transitionned for u-c-c [12:50] but that's ongoing work [12:50] seb128: ok, as we are still shipping g-c-c by default… :p [12:50] yeah [12:50] well, the intend was to have it in both [12:50] there is a bug, we discussed with mterry on monday [12:50] I can see that as /usr/lib/x86_64-linux-gnu/control-center-1/panels/libdeja-dup.so [12:50] I guess he's going to fix it [12:50] ok ;) [12:51] * didrocks install u-c-c [12:51] that's likely a key in the .desktop issue [12:51] you can run the standalone deja-dup UI otherwise [12:51] probably, yeah [12:51] I never know how to get the main panel [12:51] would give you an idea of how client side decoration don't work on Unity ;-) [12:51] it's only restore or save [12:52] didrocks, run "deja-dup-preferences"? [12:52] ah -preferences [12:52] seb128: come on, there is a beautiful decoration there! [12:53] invisible, but I'm sure, beautiful :) [12:54] ;-) === alan_g is now known as alan_g|lunch [13:33] happyaron: are kylin doing some alpha 2 testing? === alan_g|lunch is now known as alan_g === Sweetsha1k is now known as Sweetshark [14:04] Laney: I think they are mainly busy working on those third party cooperation projects [14:21] seb128: thanks for the new glade! so much better. [14:22] seb128: i'm however, loading ubiquity custom python widgets catalog, and it appears to be marking one of my widgets as "deprecated" (but no other), not sure why [14:26] xnox, yw! [14:27] I don't know about the deprecation either === m_conley_away is now known as m_conley [14:29] happyaron: would be good to have some if there's to be a release [14:45] Laney, hum, would it make sense to override the fileselector setting in ubuntu-settings rather than in the gtk source? [14:46] hmm [14:47] Laney, that's where we moved the Ubuntu overrides for other sources, it makes easier for other flavors to make different decisions than us [14:48] to take* [14:48] on that note, I'm away for some exercice, be back in ~1h [14:48] I was thinking of it as a "gtk" choice rather than an "ubuntu desktop" one [14:48] don't really mind [14:50] Laney, that works for me, especially if we patch gtk2 we sort of enforce that as a gtk choice [14:50] or we would just give others the flexibility to make gtk2/3 softwares inconsistant [15:02] seb128, Looking at your comment: https://code.launchpad.net/~robert-ancell/indicator-power/ucc-only-in-unity/+merge/202559/comments/472287 [15:03] seb128, What do you think we should be checking? I thought we only wanted u-c-c for Unity so that GNOME could be "vanilla"? [15:03] seb128: for some reason, I'm getting a race at boot with trusty where gnome-keyring doesn't end up as the ssh agent...have you seen that? [15:03] about half the time [15:10] Laney: I see, will check with our PM how to talk about it with them. [15:10] mdeslaur: I think it races with the upstart job [15:10] happyaron: cool, thanks [15:11] :) [15:14] Laney: hrm, ok, I'll file a bug. thanks === om26er is now known as om26e === om26e is now known as om26er === jhernand1z is now known as jhernandez [15:35] mdeslaur, what Laney said, and yes sometime the agent is not working for me either (wrong env, because if I run gnome-keyring-daemon and export the env it prints things works) [15:36] seb128: I filed 1271591 [15:36] tedg, right, GNOME doesn't use indicators (there is no way to use them under gnome-shell and the UI wouldn't fit the shell anyway) [15:37] seb128, So what's the issue of checking for the Unity desktop? [15:37] tedg, GNOME is going to be vanilla by not using indicator and using g-c-c [15:37] tedg, xfce use indicators, gnome-panel session use indicators [15:37] I'm more worried about folks like Xubuntu [15:38] Don't they want g-c-c ? [15:38] tedg, did you see https://code.launchpad.net/~larsu/indicator-power/xfce-control-center/+merge/202641 [15:38] tedg, I think those desktops want, in priority order, their-own-tool, u-c-c, g-c-c [15:39] seb128, Oh, okay, I thought they'd want g-c-c before u-c-c. [15:39] tedg, g-c-c is not going to include indicator settings, so better to call u-c-c which does [15:39] tedg, I'm also unsure that stuff like CSD is going to work well on those desktops [15:39] Seems weird that the behavior could change based on what is installed though. [15:39] tedg, I think they are closer from Unity than from shell [15:40] i.e. if you had Xubuntu and Ubuntu Desktop you'd get one, but not if you didn't. [15:40] you mean? [15:40] xfce would always have their own tool called [15:40] Ubuntu Desktop installed or not [15:40] I guess Xubuntu is a bad example as they have their own. But just saying because we're checking for the existence of the utility, we could behave differently. === gatox is now known as gatox_lunch [15:41] Seems we should decide and error. [15:41] I would be fine with that [15:41] I don't care enough to spend time on it though [15:41] if somebody files a bug or some other desktops have concerns they can file a bug/send a patch [15:41] Is there any reason an indicator would want to call g-c-c? [15:42] better than erroring out I guess? [15:42] but yeah, at this point g-c-c seems a random choice [15:42] we could as well call kdeconfig [15:42] Yeah [15:42] we should maybe just hide the entry if u-c-c is not installed === alan_g is now known as alan_g|tea [15:43] though having the g-c-c case makes it easier for e.g Debian to keep packaging indicators [15:43] or at least having them working as designed [15:44] It also makes the migration story easier. [15:44] tedg, I think we should claim the current code as good enough and not spend more effort on it until somebody comes with a valid concern/bug/reason to change [15:44] So my thought is: keep the patches like they are today. But when we get to a stable point we can start dropping g-c-c. [15:44] +1 [15:45] K, will review as such. [15:45] thanks [15:47] seb128: my custom do_draw() is borked. All it does is request the current split between pack1 & pack2 of a Gtk.Paned widget and updates labels appropriately with the split. [15:47] seb128: upon resizing, does Gtk.Paned emit any singnal that I can key on? [15:48] xnox, do you have code to share/a small testcase showing what you do/the issue? [15:48] seb128: yeah, one sec. === alan_g|tea is now known as alan_g [16:04] seb128: awwwwww! [16:04] xnox, works now or what? [16:04] seb128: so my do_draw() method ( a derived / subclassed Gtk.Paned) [16:05] seb128: did not have "Gtk.Paned.do_draw(self, cr)" as first thing [16:05] seb128: so nothing was drawn! [16:05] seb128: now it all works! [16:05] great! [16:15] seb128: do you know if I can increase the corner radius of the ShadowType.ETCHED_IN ? [16:15] of a Gtk.Frame [16:16] seb128: \o/ OMG, migrating away from HPaned to Paned, made that custom widget render in the new glade! #firstinhistory [16:20] haha [16:20] dunno for the corner [16:20] but larsu might [16:20] or Cimi [16:21] xnox, it's static IIRC [16:21]