[07:29] no sil today? [07:29] no didrocks today? [07:32] seb128, i have a branch where tests fail on jenkinks, but if i build the pkg locally everything is fine, any ideas what's up with that? [07:32] talking about [07:32] https://code.launchpad.net/~mhr3/libunity/scope-result-in-python/+merge/180158 [07:32] mhr3, today is a national holiday in France [07:32] seb128, but you're in germany! :) [07:33] mhr3, yeah, still not going to work :p [07:33] mhr3, how did you try? in a pbuilder? [07:34] does bzr bd use pbuilder? :) [07:34] if so then yes :) [07:34] not by default [07:34] then no :P [07:34] well, build env are different from your env [07:35] they don't have a dbus session, nor xorg, nor internet access [07:35] do you rely on session stuff or on an internet access? [07:35] i tried it also in a chroot which doesn't have any of that either [07:35] succeeded there [07:36] mhr3, is that lp:libunity ? [07:36] well, it does have internet, but we don't need that [07:36] seb128, link above ^^ [07:36] ha [07:36] let me try [07:41] mhr3, http://paste.ubuntu.com/5988032/ [07:41] mhr3, that's what I get building on a porter box chroot [07:41] seb128, exactly what i needed, and now what is a porter box? [07:42] mhr3, ssh porter-amd64.canonical.com [07:42] mhr3, if you have access/your ssh key configured [07:42] * mhr3 checks [07:43] nope, it doesn't like me [07:43] * apw is trying to find a good example (preferably in python) of something which is a client of the messaging menu. i had thought it would be libindicate, but none of the example with that do a thing [08:00] mzanetti, hey, a small one that will hopefully help with quaternourly release [08:01] :D [08:01] https://code.launchpad.net/~saviq/unity8/ap-passphrase-wait-for-focus/+merge/180293 [08:01] * Saviq forgot to paste... [08:02] Mirv, ping [08:02] Saviq: looks sane to me. but did we actually have issues with this? [08:02] mzanetti, http://10.97.0.1:8080/job/autopilot-saucy-daily_release/1010/label=autopilot-ati/artifact/results/autopilot/videos/unity8.shell.tests.test_lock_screen.TestLockscreen.test_can_unlock_pin_screen%20%28Desktop%20Nexus%2010%29.ogv [08:03] mzanetti, log excerpt http://pastebin.ubuntu.com/5988076/ [08:03] Saviq: I wonder why there is a unity7 running in this video [08:03] mzanetti, because that's how the machines are set up there [08:03] ah... right... they are not using our vms [08:04] mzanetti, they're real hardware [08:04] and still this sort of performance issues? [08:04] that's the other way around [08:04] well, maybe not [08:04] not really, no [08:04] you know, recordmydesktop has its toll [08:05] hmm... [08:09] Saviq: pong [08:10] Saviq: just filed https://bugs.launchpad.net/unity8/+bug/1212580 [08:10] Launchpad bug 1212580 in Unity 8 "Failing unity8 tests since evening Aug 14th" [Critical,New] [08:10] it seems you're on to it already? :) [08:10] Mirv, yeah, somewhat [08:10] Mirv, I hope I have a solution for the unstable test [08:10] Mirv, but then there's unity8 crashing [08:11] Saviq: there's another one in addition to the unlock pin screen, test_hide_hud_click [08:11] Mirv, you think we could get /var/crash out of the machines? [08:11] Mirv, yup, that's a crash, and we can't seem to reproduce [08:12] Saviq: I'm on the nvidia machine, where that test consistently fails, but the /var/crash is empty [08:13] Mirv, can you `ulimit -c unlimited` and see if core is dumped? [08:14] Saviq: I guess that'd need setting globally somehow, for the next run. trying. [08:15] Mirv, /etc/security/limits.conf [08:15] Mirv, can you unhook the machine for some manual debugging? [08:18] Saviq: ok added, is there some sort of need to reload something? [08:18] Saviq: sorry, no idea, I haven't ssh:d to the machines before and I don't know what didrocks has used to do [08:18] Mirv, no, if it crashes it should dump a "core" file in $CWD [08:19] Mirv, that we can then use to see the crash locally [08:20] Saviq: after the current test run on normal unity finishes, I can rerun the unity8 tests before 10 UTC when the next run starts [08:20] Saviq: meanwhile, you don't have any nvidia machines that could possibly bring up the issue? [08:20] Mirv, I'm running one [08:20] Mirv, no crashes here [08:21] ok [08:24] * Saviq runs the test 20 times to see if anything comes up [08:28] meanwhile the intel machine looks to have gone down [08:29] Saviq: do you happen to know which people have access to fix something like that? I think jibel was perhaps one, but he is also on holiday today [08:30] trying IS.. [08:32] nope, need QA team people [08:33] Mirv, got the crash [08:33] or at least *some* crash [08:34] with an unusable core ;| [08:42] * Mirv progresses on finding people to get intel machine back up [08:44] * greyback_ moving to office [09:44] Mirv: hi. sil2100 was helping out with getting mediascanner into the archive, but I think he is on holiday today. Are you able to help, or do you know anyone who could? [09:52] jamesh: it's starting to look impossible for today for me, but possibly kenvandine / robru when they wake up. sil2100 is also back tomorrow. [09:52] thanks. [09:52] jamesh: furthermore, the QA lab is currently broken so we wouldn't progress on getting to saucy archives anyhow. [09:55] Mirv, can we re-run unity8 to see if the fix we merged has that one issue sorted? [09:55] thostr_: ^^^ so we might have trouble getting mediascanner in the archive today [09:56] Saviq: I tried, I got some other error and was just about to try again. I'm not sure if the intel machine being broken could prevent the other machines from being used as well. [09:58] Saviq: now it runs on the nvidia machine, let's see [09:59] what a nice day to everything get broken when half of the people is on holiday. the usual :) [10:09] Saviq: at least the ati machine succeeded without that one error it had earlier (although not everytime). as expected nvidia machine still getting the other error, but I don't see coredump on the machine. but I gathered your hunt for that problem is also progressing? [10:17] Mirv, :) [10:51] mzanetti, did you have a look at the issue with autopilot? [10:51] mzanetti, or can you tell me how to see the bug? (make test what?) [10:54] Cimi: I couldn't reproduce yesterday. today I merged trunk and jenkins failed again. [10:54] didn't investigate yet [11:10] Saviq: I hope this will do https://code.launchpad.net/~mzanetti/unity8/more-robust-lockscreen-tests/+merge/180316 [11:12] who can help me in imports? [11:12] I'm working on the settings app, I have issues in importing the plugins === MacSlow is now known as MacSlow|lunch [11:22] Cimi: I think I fixed the autopilot test... lets see if it works on jenkins [11:22] Cimi: whats the issue with the imports? [11:23] mzanetti, I am in settings app [11:23] mzanetti, from that branch [11:23] actually, I'll be back in 10 [11:28] Mirv, ok so there's a fix for the crasher https://code.launchpad.net/~mhr3/unity/only-valid-filters [11:29] Mirv, already approved and autolanding now [11:30] Mirv, and https://code.launchpad.net/~mzanetti/unity8/more-robust-lockscreen-tests/+merge/180316 should be even more robust for the other issue (even though it passed I saw in the logs in other tests that the issue is still there) [11:31] Mirv, I assume we need to wait for the intel machine to come up first? [11:32] greyback_, ping, got a minute? [11:32] Saviq: yep [11:32] greyback_, mumble? [11:32] Saviq: am there [11:36] Saviq: awesome! yeah, the -intel issue is now breaking up the rest. [11:36] Saviq: after it hopefully gets fixed kenvandine will be the next one from our team to wake up [11:36] Mirv, got i [11:36] t [11:37] Mirv, did you start working early to cover the gap or something? [11:38] Mirv, you're just +1 from me, but you sound like you're EOD'ing soon :) [11:38] mzanetti, so from this https://code.launchpad.net/~cimi/ubuntu-system-settings/welcome-wizard [11:38] mzanetti, I tried adding [11:38] import Ubuntu.SystemSettings.Language 1.0 [11:39] for example in Pages/WelcomePage.qml [11:39] and run qmlscene -I /path/to/plugin/language MeetUbuntu.qml [11:39] no luck :-\ [11:41] Cimi, for plugins to load [11:41] Cimi, they have to be on the same path [11:42] Cimi, that you provide with the "import" statement [11:42] Cimi, so import Ubuntu.SystemSettings.Language 1.0 [11:42] Cimi, equals "Ubuntu/SystemSettings/Language" [11:42] or, if versioned [11:42] "Ubuntu/SystemSettings/Language.1" [11:42] "Ubuntu/SystemSettings/Language.1.0" [11:42] mmm valid [11:42] in reverse priority order [11:42] Cimi, so without installation, you won't get that path [11:43] Saviq, but there's the qmldir, no? [11:43] Cimi, doesn't matetr [11:43] ok [11:43] Cimi, it needs to be on the path [11:43] if this is located in /home/cimi/foobar/Ubuntu/SystemSettings/Language you'd need to do -I /home/cimi/foobar [11:43] thought was responsible of registering the name [11:43] Cimi, the "module" in qmldir is just for protection [11:44] Cimi, not related to where QML looks for the module [11:45] Cimi, because .po files don't seem to respect the install prefix [11:45] Cimi, I'm afraid it's easiest if you just go `bzr bd` [11:45] build the packages and install whichever ones you need [11:46] Saviq, but the plugin is private [11:46] Cimi, so? [11:46] Saviq, at least that's what web told me [11:46] *seb [11:46] Cimi, still has to be installed [11:46] Cimi, you just need to point QML2_IMPORT_PATH at it [11:46] or -I if using qmlscene [11:46] at the private path, that is [11:46] ah ok [11:46] Cimi, and you should read http://qt-project.org/doc/qt-5.0/qtqml/qtqml-syntax-imports.html [11:47] Cimi, if you're not doing anything with the plugins, just install settings-app [11:47] I need to use them [11:47] but not changing for the moment [11:48] ubuntu-system-settings is the package name [11:49] Saviq, perfect works with /usr/lib/x86_64-linux-gnu/ubuntu-system-settings/private [11:50] Saviq, wondering how I get this to be dynamically loaded [11:50] Cimi, you'll need a c++ main() around it, just as system-settings has [11:50] Saviq, maybe not using qmlscene but a main.cpp? [11:50] ok [11:50] yup [11:50] Cimi, qmlscene isn't meant to be used "in production" [11:51] Saviq, thought it was for us [11:51] Cimi, it's only there for development/debugging purposes [11:51] Saviq, thought zoltan wanted [11:51] use qmlscene as much as possible [11:51] unless c++ is really a requirement [11:51] Cimi, qmlscene-*like* [11:51] Cimi, qmlrunner, whatever [11:51] Cimi, just one that didn't have all the dev/dbg options qmlscene has [11:51] Cimi, not sure what's the status of it btw === MacSlow|lunch is now known as MacSlow === alan_g is now known as alan_g|lunch [12:05] Cimi, see #sdk for an explanation on qmlscene [12:42] Saviq: could you please look: https://code.launchpad.net/~gerboland/unity-mir/fix-platform-api-pkgconfig-change/+merge/180329 [12:42] it would unblock me [12:43] greyback_, on it [12:43] ta [12:44] greyback_, approved, although not sure why that change in platform-api [12:45] Saviq: thanks! === _salem is now known as salem_ [12:56] Saviq, did something happen to scrolling? seems jerky now [12:57] mhr3, apps running? [12:57] Saviq, on desktop even [12:57] then again, seeing what my screen is doing it might just as well be the intel driver... [13:01] mhr3, nope, not seeing that on the desktop [13:01] mhr3, can you confirm another thing for me, though [13:01] mhr3, launch unity8 [13:01] mhr3, go to home scope, drag down [13:02] mhr3, does the first category get stuck 1/2 down the screen for you? [13:03] Saviq, i've seen it do that a few times... but doesn't right now [13:04] mhr3, how about in ./run -- -fullscreen ? [13:04] mhr3, or actually, GRID_UNIT_PX=12 ./run [13:06] the search bar moved in the middle of the screen, then fixed itself after scrolling [13:06] but still doesn't get stuck === alan_g|lunch is now known as alan_g [13:08] mhr3, bad news, it's category overrides causing that... [13:08] Saviq, how, what now? [13:09] mhr3, not sure yet, but reverting it does get rid of the issue [13:09] no way [13:10] told you i've seen it long before i even thought about category overrides [13:11] mhr3, I agree it might just be triggering something [13:11] but i'm not denying that using actual scopes that give us non-mocked data might trigger it [13:13] mhr3: i think this should get rid of any duplicates. https://code.launchpad.net/~nick-dedekind/unity/scopeview.model.resync/+merge/180335 [13:14] a sec, the screen corruption is getting worse [13:15] * mhr3 wants stable intel driver === olli_ is now known as olli [13:24] mhr3, sorry, I'm afraid it really is category overrides :/ it must be tricking the proxy model somehow [13:24] mhr3, it only affects apps and home, too - they're the only ones with overrides [13:27] Saviq, so you're unable to rep the issue if you comment out the overriding? [13:27] mhr3, applying http://paste.ubuntu.com/5989012/ [13:27] mhr3, is enough to fix [13:27] mhr3, let me try commenting in qml [13:31] plus they're the only ones which don't have static data [13:31] mhr3, ok, so commenting out in QML isn't enough [13:32] dednick, standup [13:32] Saviq, there you go, then it can't be the root cause [13:32] mhr3, it can [13:32] mhr3, if the categories themselves are doing tricks [13:32] mhr3, doing http://paste.ubuntu.com/5989012/ does help [13:33] mhr3, I'll drill down on it after the standup [13:33] s/the/our/ [13:33] if nothing calls the method, the code paths you removed are never triggered [13:33] yea, i have a call now too :) [13:56] mhr3, http://paste.ubuntu.com/5989111/ fixes [14:01] mhr3, not sure why, yet, but it does [14:01] dednick: your packages didnt help [14:01] fwiw [14:02] asac: hm. ok, i'll need to look into the issue some more. [14:02] asac: thanks for testing [14:02] i hope i didnt make a mistake [14:03] asac: is there anything specific you were doing to reprocude? [14:03] dpkg -i failed first [14:03] i ran apt-get install -f to resolve the conflicts after [14:03] and then dpkg -i worked, but i got it after reboot again [14:03] dednick: it must be pretty easy to reproduce... what i did (without beying systematic) was: [14:03] 1. use look at indicators [14:03] 2. open two apps without closing each [14:03] and then kill them [14:04] let me see :) [14:04] if i can be better [14:08] mhr3, don't get me wrong - you might as well be triggering an issue in the dash ListView [14:08] Saviq, hmm, could you try to change the signal connections to QueuedConnection? [14:09] mhr3, http://paste.ubuntu.com/5989165/ didn't help [14:12] thostr_1: Saviq: when will the Click scope be available? Now that the Click app store is going online, it'd be nice to complete the story === mzanetti is now known as mzanetti|otp [14:14] mhall119: mhr3 is working on it, should be any day now [14:14] Saviq, i still think that the count being present in the model is just wrong [14:14] awesome [14:18] mhall119, or you can enable it already [14:19] mhall119, gsettings get com.canonical.Unity.Dash scopes [14:19] mhall119, add 'click/click.scope' and `gsettings set com.... scopes "[...]"` [14:20] mhall119, but yeah, it will show up in the apps scope asap, too [14:20] cool, let me grab my USB cable and try [14:20] mhr3, which model would that be? Categories? [14:20] Saviq, yea [14:21] Saviq, do you have a moment for hangout? [14:21] mhr3, sure [14:21] i want to see what you see [14:22] mhr3, ah then let me make a video [14:22] we can also chat about it in the hangout :) === boiko_ is now known as boiko [14:28] Saviq: I'm getting a dbus-launch error when trying to set gsettings from adb shell [14:28] mhall119, as phablet [14:29] I was [14:31] that's okay, the terminal-app on my phone + bash history meant I could easily re-run it from a session with dbus [14:31] restarting [14:32] Saviq: mhr3 does click/click.scope pull from the online store, or just already installed click packages? [14:32] mhall119, search for "*" [14:33] Saviq: nothing [14:33] mhall119, works here [14:33] do I need to login to something? [14:34] IIRC, anonymous access to the click store was still in discussion [14:35] mhall119, on staging it's enabled [14:35] mhall119, alecu should be able to help [14:36] on a different note, I *love* that indicators are tabs now, makes so much more sens [14:36] sense [14:36] mhall119: click scope gets results from the click webservice, yes [14:37] alecu: ok, so I got my xda-developers-app published in the click store yesterday, but searching for it doesn't show any results [14:37] and, in fact, search only shows the "Installed" category header [14:37] mhr3, here's videos of the two issues I noticed: http://ubuntuone.com/0y3o0lPEaWsVBEP6gdiJdd http://ubuntuone.com/4AGFEchm7D1QqVvDtrbmZb [14:37] mhall119: if you don't get results from searching for "*", may I ask you for the click scope logs? they are in ~/.cache/unity-scope-click.log [14:37] Saviq, yep, downloading them, it's super slow [14:37] mhr3, right.. u1 [14:38] yea... [14:38] mhall119: did you publish it in the production or staging web service? the scope is still using the staging one. [14:39] alecu: http://paste.ubuntu.com/5989261/ [14:39] alecu: ah, I'm not sure to be perfectly honest, let me ask [14:40] mhall119: it should show up for you in https://developer.staging.ubuntu.com/dev/click-apps/ [14:41] alecu: it's on https://myapps.developer.ubuntu.com/dev/click-apps/3/ === alan_g is now known as alan_g|tea [14:41] so I guess that means production [14:42] right [14:42] any easy way for me to point it at production? [14:43] mhall119: not yet, it means recompilation on the device [14:43] mhall119: I plan to point it to production soon [14:44] ok [14:44] oh looks like intel is back [15:09] mhall119: here's the bug to have the scope returning results from production: https://bugs.launchpad.net/unity-scope-click/+bug/1212709 [15:09] Launchpad bug 1212709 in Unity Click Scope "Uses staging webservices" [High,Triaged] [15:10] mhall119: and we'll probably have it fixed in a few hours === alan_g|tea is now known as alan_g [15:13] alecu: that would be great, thanks! [15:13] alecu: let me know when it's ready for me to test [15:15] sure [15:17] tedg: ping [15:18] alecu, Howdy [15:18] tedg: I see that you've mentioned a few times in the app-devel list the click hooks. May I ask you some question about them? [15:18] I can't find cjwatson today, so that's why I'd like to ask you :-) [15:20] Saviq, it's interesting that you always get apps after files, it's vice versa for me [15:20] alecu, Ah, he's at debconf [15:20] alecu, Sure [15:20] tedg: so, I need to tell the .desktop files that are generated by the click installer from other .desktop files created by other means [15:20] mhr3, less files? ;) [15:21] tedg: I was thinking of having the click desktophook add a new X-something field to the .desktop file when it's created on .local/share/applications [15:22] tedg: (all of this may be needed so the click scope can show the right previews for apps that were installed from .clicks) [15:22] tedg: do you think that it makes sense to do it in the click desktophook? [15:23] alecu, talked with thomas, he's inclined to do the de-dup where both scopes will provide the same results [15:23] alecu, You can create a new one, but that might end up in a race condition with the desktop file creation hook (there's not guarantee on ordering). Probably best to do it in the hook that creates it. === mzanetti|otp is now known as mzanetti [15:23] alecu, That hook is in upstart-app-launch, it's pretty simple. [15:25] alecu, Though, if you made them with constant names, and didn't modify the desktop file (which makes sense I think for your use-case) you could do an independent hook. [15:26] mhr3: ack [15:28] alecu, one problem is that it could be racy [15:28] like if the click scope takes a while to return results [15:28] mhr3: it surely will take a while, since it always needs to hit a webservice [15:29] alecu, even for the installed apps?? [15:29] mhr3: you are right, not for installed apps [15:30] mhr3: right now the slow part will be the spawning of "click list --manifest" [15:31] yea, i hope we don't need to do that on each keypress :) [15:32] mhr3: yes, I surely cache that :-) [15:33] mterry, hello [15:33] mterry, you joining our hangout? [15:34] katie, omg, yes [15:34] tedg: looking at desktop-hook.c, I see that X-Ubuntu-Application-ID is being added to .desktops already [15:35] tedg: that should only be present for click packages, right? [15:37] alecu, Yup [16:09] mzanetti, I have issues with the import here lp:~cimi/ubuntu-system-settings/welcome-wizard_app [16:09] it's qmake things... [16:18] anyone could help me on qmake? [16:18] Cimi: [16:18] re [16:19] re [16:19] Cimi: so. what's the issue? [16:19] mzanetti, Error: Error while loading page: qrc:/qml/Pages/WelcomePage.qml:19 module "Ubuntu.SystemSettings.LanguagePlugin" is not installed [16:21] * mzanetti compiles and installs deps [16:24] Cimi: ok... I compiled it and ran system-settings [16:24] its a bit empty [16:25] what do I need to do to reproduce the issue? [16:30] mzanetti, not system settings [16:30] mzanetti, you need to run wizard [16:31] Cimi: ok. can see the issue [16:31] it must be something with the import [16:31] and the PLUGIN DIR definitions [16:32] I use in main.cpp [16:43] Cimi: so... [16:43] If you want to run it with qmlscene, you need to create a directory structure [16:43] I created for example: plugins/Ubuntu/SystemSettings/LanguagePlugin/ [16:43] and copied the lib and the qmldir file into that [16:43] mzanetti, I want to run with main.cpp [16:44] then running qmlscene wizard/qml/main.qml -I plugins/ works [16:44] mzanetti, because the plugin is private [16:44] what you mean with that? [16:44] mzanetti, that the directory changes [16:44] it's somewhere in the system [16:44] you must have such a directory structure [16:45] doesn't matter if its somewhere in the system. it needs to be Ubuntu/SystemSettings/LanguagePlugin [16:45] as a directory structure [16:45] otherwise you need to link stuff statically and use qmlRegisterType in your main.cpp instead of the plugin.cpp [16:47] mzanetti, it has this structure [16:47] mzanetti, but I want to set the path from the main file [16:47] mzanetti, on my system is /usr/lib/x86_64-linux-gnu/ubuntu-system-settings/private === alan_g is now known as alan_g|EOD [16:52] Cimi: you'd need to export QML2_PLUGIN_DIR=/usr/x86_64-linux/gnu/ubuntu-system-settings/private/ [16:52] Cimi: then, if it has this directory structure in there it should work [16:52] mzanetti, I want to set it in main.cpp [16:52] mzanetti, look at how system settings is doing [16:52] you can't. I might differ from system to system [16:52] mzanetti, no [16:53] mzanetti, it works for system settings app [17:14] Cimi: ok. found it [17:15] Cimi: they add a define in qmake [17:15] Cimi: DEFINES += -DPLUGIN_PRIVATE_MODULES=/ubuntu-system-settings/private [17:15] and then view.engine()->addImportPath(PLUGIN_PRIVATE_MODULE_DIR); in the main.cpp [17:16] which I think is quite ugly [17:16] but ok. it works [17:23] Cimi: I see you're doing the same. whats the problem actually? [19:42] mhall119: I've just installed the XDA Developers App with the very latest unity-click-scope [19:43] mhall119: but it fails when starting the app, like this: [19:43] [unity-scope-click] - DEBUG: click-scope.vala:171: cannot execute application: No manifest found for app_id: com.ubuntu.developer.mhall119.xda-developers-app [19:43] mhall119: that means that the app id in the webservice does not match what's on the click manifest [19:44] the manifest says "com.ubuntu.developer.mhall119.xda-developers" [19:45] but the webservice says "com.ubuntu.developer.mhall119.xda-developers-app" [19:46] I think changing the manifest and rebuilding the .click is the faster way to fix it [19:58] ok [20:43] mzanetti, it is not working for me... [20:43] mzanetti, indeed the warning [20:43] Cimi: works here [20:43] °_° [20:43] Cimi: did you install it? [20:43] * Cimi rebuilds [20:43] mzanetti, installed what? [20:43] sudo make install [20:43] but what? [20:44] mzanetti, wizard or the app? [20:44] everything in that repo [20:44] plugins at least [20:44] they are installed [20:44] anyway I'll have a look tomo morning [20:44] to /ubuntu-system-settings ? [20:44] yes [20:44] hmm... works for me [20:45] Cimi: did you run sudo make install? [20:45]