=== bschaefer_ is now known as bschaefer === _salem is now known as salem_ === salem_ is now known as _salem === duflu_ is now known as duflu [07:06] tsdgeos: good morning [07:07] tsdgeos: I had to merge trunk at this one. mind checking/approving? https://code.launchpad.net/~mzanetti/unity8/quicklist-unset-model-before-closing/+merge/181517 [07:09] mzanetti: checking [07:09] morning [07:09] mzanetti: technically i only need to check you merged right, no? :D [07:09] tsdgeos: yeah :) [07:11] mzanetti is the name of the branch "wrong"? [07:11] it's called unset before closing [07:11] but you close before unsetting, no? [07:12] tsdgeos: ah yeah... I had it the other way round at first [07:12] tsdgeos: caused some issues. [07:13] oka [07:22] mzanetti: so we got a whole lot of autoland fails still :-/ [07:22] tsdgeos: it's not a whole lot any more, but yeah, too many, I agree [07:23] tsdgeos: from what I see this are the main issues: [07:23] autopilot-tests for unlocking sometimes freezes when swiping away the greeter [07:23] one of the builders causes a nullpointerexception somewhere in jenkins [07:23] and at times, the unit tests for lvwph randomly fail [07:24] yeah, we need 5.1 for lvwph [07:24] mzanetti: you set https://code.launchpad.net/~nick-dedekind/unity8/indicators.menu-items-cleaning/+merge/181410 to approved [07:24] but it's been merged already, no? [07:24] isn't it r249? [07:24] there was something weird, yeah... [07:25] ok, i'll edit the MR manually and set it as merged [07:25] done [07:25] sil2100: hi. Has there been any progress on getting mediascanner into main? [07:25] ah... reminds me... [07:25] is didrocks in vacation currently? [07:26] jamesh: working on that ;) [07:26] mzanetti: yes, for 2 weeks [07:26] sil2100: I found a situation where commits get past the releasing process without being tested [07:27] mzanetti: how? What's happening? [07:27] sil2100: if something gets merged by the upstream merger in between the release job starts and ends, it gets merged and then the release job commits the release commit message on top of that [07:27] sil2100: check out unit8's history [07:28] sil2100: there is a commit with a rather long commit message. starting with "tweak launcher folding" [07:28] sil2100: I guess the same goes for unity-scope-mediascanner? :) [07:28] sil2100: it doesn't show up in the changelog because it landed after the release jenkins started, but before it comitted the changelog [07:30] mzanetti: ok, this situation indeed might happen, sadly we don't have a way currently to not make it happen - but the good thing is that it's not really being released without testing [07:30] mzanetti: let me check the lp:unity8 history to make sure [07:30] sil2100: do the tests run again after jenkins posts the changelog? [07:32] mzanetti: from what I see on lp:unity8, there's not really any problem - the process looks like this that the 'changelog' and the revision that the snapshot is being taken is actually created before the testing starts, during the prepare job [07:32] sil2100: yeah. but at that point it also does the checkout to run the tests [07:33] sil2100: so I'm quite positive that this commit was not tested by the release jenkins [07:33] mzanetti: it's tested by the next release, look: [07:33] http://bazaar.launchpad.net/~unity-team/unity8/trunk/revision/253 [07:33] mzanetti: it just didn't appear in the changelog because it wasn't released [07:33] sil2100: sure.. it'll be tested by following released. that's true [07:33] mzanetti: the snapshot was made from 249 [07:34] mzanetti: your commit was 250, so it went to the next release - so it's all safe and nothing goes 'unnoticed' [07:34] sil2100: yeah... it creates a snapshot at 249 and tests this [07:34] then 250 comes in [07:34] then the release jenkins finishes testing and commits its changelog bump to 251 [07:35] 251 gets released because of that, containing 250 - untested [07:35] mzanetti: no no no [07:35] It's not [07:35] mzanetti: look, jenkins says clearly: Releasing 7.81.3+13.10.20130826.4-0ubuntu1 (revision 249 from lp:unity8). [07:36] mzanetti: when releasing it doesn't take lp:unity8 directly, it only uses the commits it prepared during the prepare job [07:36] mzanetti: it creates the source package only once, when doing the snapshot [07:36] mzanetti: 250 didn't land in this release [07:37] mzanetti: what jenkins does is - it prepares the packages in the prepare job (the changelog is then prepared), builds and tests the packages, if all is ok, it commits the 'Releasing ..." commit to trunk and releases the *exact* packages it used for testing [07:38] sil2100: ok... understood it now [07:38] thanks for explaining. [07:39] I was watching it live yesterday (to debug our issues in jenkins) and saw those commits coming in in that order... made me think this must be an issue [07:39] mzanetti: if there were some commits made in the meantime - well, they don't get released, since the source package got created during snapshotting - those commits that got into trunk in the meantime will be picked up by the next release, thanks to the (revision from lp:unity8) <- this tells the build bot which was the last commit it used [07:39] sil2100: I was thinking that the commit to the changelog is the important one [07:39] mzanetti: no problem ;) I guess we can't do anything not to let those things get 'out of order' [07:39] mzanetti: but luckily it's safe [08:19] mzanetti: https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/266/testReport/junit/unity8.shell.tests.test_lock_screen/TestLockscreen/test_can_unlock_passphrase_screen_Native_Device_/ failing again :-/ [08:19] any clue why? [08:19] tsdgeos: I know it will strike back [08:20] tsdgeos: no. I have absolutely no clue and haven't been able to reproduce it locally in hours [08:20] s/know/knew/ ^^ [08:20] weeeeeeeeeird [08:33] stupid tablet ran out of battery while flashing [08:33] hope nothing bad happened [08:34] tsdgeos: seems this phone produced just one failure. the next run on it was a good one again [08:34] tsdgeos: you can't break the device by flashing it with phablet-flash [08:34] good stuff then :D [08:39] jamesh: I need to consult the MIR for mediascanner a bit since mediascanner is using A LOT of typical universe-only packages [08:40] jamesh: since gstreamer1.0-plugins-bad and gstreamer1.0-plugins-ugly we can't get into main [08:41] sil2100: I'm in a standup right now. Can I get back to you in 20 minutes? [08:41] jamesh: are those required? Are those used in the code extensively? [08:41] jamesh: ok [08:41] dednick: ping [08:46] tsdgeos: pong [08:46] dednick: just commented in https://code.launchpad.net/~nick-dedekind/unity8/indicator.multi-icon/+merge/181862 [08:47] sil2100, do we know why are the ap tests for unity failing? [08:48] tsdgeos: hm. maybe i shouldnt be using gu then. [08:49] not sure [08:50] what are you trying to achieve? [08:51] tsdgeos: it's the width between indicator icons. 0.5 is a bit too close, 1 is a bit too far. [08:51] dednick: maybe round the 0.7 gus? [08:51] mhr3: not yet [08:52] the problem is that qml "knows" how to paint in 5.6 pixels [08:52] but that will look mega-crap on some displays [08:52] tsdgeos: maybe should just use dp ? [08:52] and it's also affecting the rendering speed since needs to do "some for this pixel and some for this other pixel" [08:52] tsdgeos, btw sorry for not replying yesterday, we had public holiday here [08:52] dednick: maybe, but then dp is not scaled, no? [08:52] tsdgeos: right. antialiasing. [08:53] mhr3: sure, no worries [08:53] dednick: i mean if you use dp, won't it look "too close" in a tablet? [08:54] tsdgeos: probably on retina it will be super close. [08:54] dednick: so maybe the rounded gu? [08:54] seems a bit lame, but.... [08:54] how about mixing the two: units.gu(0.5) + units.dp(2) [08:55] greyback: is dp scaled? [08:55] i thought it was "always 1 pixel" [08:56] tsdgeos: it is scaled to 2, for high DPI screens [08:56] ah [08:56] dednick: maybe that works then? [08:56] tsdgeos: i'll see how it looks [08:57] galaxy nexus should provide a high enough gu [08:57] greyback: i'm trying to test https://code.launchpad.net/~gerboland/unity8/fix-sidestage/+merge/182212 but my battery keeps running out [08:57] greyback: is that the fix for "starting a side stage app with nothing on the main stage makes the side stage app not be shown"? [08:58] tsdgeos: see attached bug. Side stage apps were completely broken [08:59] so someone ignored my bug report and created a new one [08:59] tsdgeos: as in, when you launched the app, you saw the side-in animation, but then you get black. The app was not allowed to shine through the shell [08:59] yeah [08:59] what i said [08:59] days ago [08:59] https://bugs.launchpad.net/unity8/+bug/1210079 [08:59] Launchpad bug 1210079 in Unity 8 "Side stage app does not correctly paint if there is no main stage app" [Undecided,New] [09:00] tsdgeos: I thought I got it irrespective of if a minstage app open or not [09:00] but thinking about it you're probably right, [09:00] sil2100: hi. As far as the gstreamer dependencies go, the I know -bad and -ugly are not required for mediascanner to run. I'll have to check what is actually needed for the test suite [09:01] greyback: so dupe bugs? or? [09:02] tsdgeos: I've sorted it [09:02] ah you already set the branch [09:02] oka [09:02] jamesh: please do, I'm working on getting the other deps into main in the meantime - we'd probably need to get rid of the gstreamer1.0-libav, gstreamer1.0-plugins-bad, gstreamer1.0-plugins-ugly build-deps, as I'm sure we won't force anyone to let them into main [09:02] greyback: now is when we get some shit for ignoring this since 8 August :D [09:03] tsdgeos: :) === swook` is now known as swook [09:35] greyback: did we got anywhere we can merge https://code.launchpad.net/~aacid/unity-mir/application_manager_tests ? [09:37] tsdgeos: nope, I didn't have chance to look into what's going on. I'm guessing we need to set up those tests to run on actual device [09:37] ok === swook is now known as swook` === swook` is now known as swook [09:55] greyback: ping [10:01] tsdgeos: pong === rachelliu_ is now known as rachelliu [10:04] greyback: should i be able to "move" (i.e. uncollapse) a side stage app if it's the only app open? [10:04] err uncollapse -> collapse [10:04] tsdgeos: no [10:04] it should be locked [10:04] okii [10:04] then it works :D [10:04] yay :) === alan_g is now known as alan_g|vt [10:30] dednick: not sure i understand https://code.launchpad.net/~nick-dedekind/unity8/indicator-location/+merge/182331 [10:31] dednick: so they changed the "default" file for indicators? === alan_g|vt is now known as alan_g [10:32] tsdgeos: there are old backends and new backends. Currently byu default we used the old backend compatible UIs. But now, every indicator we add is going to be compatible with new backend UIs. so i changed the deafult qml file we use. [10:33] tsdgeos: but at the same time, i added the override for the location indicator so that it gets a "friendly" title. [10:33] dednick: where does the title come from otherwsie? [10:33] dednick: and how do we i18n that title? [10:34] tsdgeos: at the moment if just uses the indicator name by default (eg indicator-datetime) if not handled in unity8. It's something that needs sorting out. [10:35] *it just uses [10:35] dednick: so we are not providing i18n facilities for it either? [10:37] tsdgeos: nope [10:38] dednick: because we'll have a better solution so not worth it? or? [10:39] tsdgeos: i'm not sure if we have a solution at hand, but this is not the final one for sure. [10:39] ok [10:39] not sure how important is i18n at this point [10:39] i know paulliu was working on it for a while [10:39] paulliu: you there? was your i18n push for a tradeshow or something? [10:46] tsdgeos: i just respaced those icons to 0.5 [10:46] tsdgeos: with the addition of the location indicator we would be running out of space again anyway, and it wasnt much diff to 0.7 anyway [10:47] ok [10:47] probably going to need a better solution anyway. we've still got loads of icons to go :( [10:50] * greyback_ not fond of 0.7 gu, it means we're not using the pixel grid that the whole shell UI placement should adhere to [10:51] greyback_, ping [10:52] nic-doffay: pong [10:52] greyback_, are you familiar with the Icon class at all in the SDK? Particularly where the assets are stored which it loads by name? [10:53] "image://gicon/%1".arg(icon.name) [10:53] nic-doffay: not off the top of my head. Wellark or larsu probably know more than I about it [11:06] * greyback_ moving to office, back in 30 [11:08] hmmm [11:08] can i link a static lib into another static lib? [11:10] good old google [11:11] so it seems "you can't" === swook is now known as swook` [11:19] tsdgeos, you can :) you just need the right linker magic [11:19] tsdgeos, see the magic whole-archive option approach here: http://bazaar.launchpad.net/~phablet-team/platform-api/trunk/view/head:/src/ubuntu/hybris/CMakeLists.txt#L19 [11:19] tvoss__: yeah well, that's exactly what's failing :D [11:20] tsdgeos, what does it say? [11:20] my ua_url_dispatcher_session symbol ends up in src/ubuntu/hybris/libubuntu_application_api.so.1.0.0 but not in src/ubuntu/mirserver/libubuntu_application_api_mirserver.so.1.0 [11:21] even if both are linking to the .a [11:21] i think it's because in one it's before the whole-archive [11:21] and in another after [11:21] tsdgeos, ah, it's working, but the mirserver part is not pulling it in ... [11:22] my question is why [11:22] since it seems it should :D [11:23] tsdgeos, the issue is here http://bazaar.launchpad.net/~phablet-team/platform-api/trunk/view/head:/src/ubuntu/mircommon/CMakeLists.txt#L41 [11:23] tsdgeos, lacking the whole-archive thingy [11:23] i added it there [11:23] but didn't seem to help [11:23] let me try again [11:35] mzanetti, ping [11:37] Cimi: pong [11:37] mzanetti, was wondering what's the launcherDragArea and the EdgeDragArea [11:38] the difference between the two [11:40] Cimi: one is for dragging in, the other for dragging out [11:40] tvoss__: so, no i did not do it before, i did it on the laptop and then compiled on the phone \o/, but now that i do it on the phone doesn't work either, seems too much stuff is getting inside the whole-archive, let me paste [11:43] tvoss__: http://paste.ubuntu.com/6032225/ [11:44] seems that for some reason ${UBUNTU_APPLICATION_API_LINK_LIBRARIES} is getting expanded to "a lot" ../application/location/libubuntu_application_location.a -lubuntu-location-service -ldbus-1 -lubuntu-location-service -ldbus-1 -lpthread -lboost_system ../application/url_dispatcher/libubuntu_application_url_dispatcher.a -lurl-dispatcher -lgio-2.0 -lgobject-2.0 -lglib-2.0 -lhybris-common [11:44] and thus we get the dupe symbols [11:44] tsdgeos: i18n for indicators? I thought I fix an issue about i18n in the past in the datetime. Yes for a trade show in China. [11:44] tsdgeos: Already got merged I think. [11:45] paulliu: well, it got broken again according to dednick :D [11:45] Cimi: you say it should be fixed with the current code in that branch? [11:46] tsdgeos, it works for me [11:46] Cimi: because it still goes thorugh here, see calculator for example [11:46] tsdgeos: Argh. I'll check... [11:46] tsdgeos, mmm [11:46] tsdgeos, I forgot my ubuntu phone at home, cannot test [11:46] dednick: Do you have a screenshot? [11:47] tsdgeos: nothing was broken that wasnt previously. [11:47] paulliu: ^ [11:47] dednick: sure [11:47] dednick: for some value of "previosuly" [11:47] no? [11:47] i mean it may have been working when paulliu fixed it and then broke again (not in this MR of course) [11:47] tsdgeos: well there was never i18n for indicators as far as i know [11:48] ah [11:48] ok [11:48] paulliu: ^ ? [11:48] tvoss__: any hint? [11:48] dednick: For the datetime there is.. So you can get Chinese date-time. [11:49] paulliu: i c. well that's up to the backend now. if they send a chinese label then that's what we display [11:50] paulliu: we are talking about titles. I dont think they are translated at the moment are they? [11:50] eg "Date and Time", "Battery", "Messaging" [11:51] dednick: yes.. It is from the backend. And datetime strings is based on the locale. for example, LC_TIME.. [11:51] dednick: ah, that is not yet translated. [11:51] dednick: I wonder that is not i18n yet. [11:54] paulliu: dont worry about it. it needs to be fixed anyway [11:54] dednick: yeah. [11:56] nic-doffay: don't use image://gicon (we're in the process of deprecating it) [11:57] larsu, yeah chatting to seb128 about in #sdk [11:57] larsu, not on this server though :P [11:58] nic-doffay: I'm not in there to read the scrollback. I trust seb128 to give you the right advise though ;) [11:59] larsu, I basically pointed them to https://code.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/unity-theme-icon-provider/+merge/180805 and to ubuntu-themes to add the icon [12:00] seb128: thanks. Still not merged, though :( [12:01] Wellark: what's up with that ^^? I thought you needed it urgently? [12:01] larsu, somebody needs to change the top status [12:01] larsu, saviq mhr3 and you approved, why not doing that? [12:01] I guess somebody from the sdk team needs to review/push the trigger? [12:02] seb128: they blocked it on not having tests. Wellark said he'd write some, but I can't see any in the MR [12:02] larsu, try to ping the #sdk guys about it [12:02] seb128, yea, as larsu said, tests missing [12:02] urg, still [12:02] sdk guys wanted that [12:03] Wellark, ^ weren't you supposed to add some test? [12:03] or convince sdk guys that it's not worth it :) [12:03] so the sense of urgency that people communicated to me two weeks ago was greatly exaggerated :P [12:04] Wellark wasn't feeling well yesterday, don't know if he's back yet [12:04] mhr3: well, the tests they'd be happy with are really not worth it :) [12:04] but we want that branch anyway, otherwise there's no support for themed icons with fallbacks [12:05] seb128: I'll wait for Wellark to reply in case he has a half-finished branch on his machine. Otherwise I'll write those tests. [12:05] well unless you count things like "image://gicon/.%20GThemedIcon%20foo%20bar [12:05] larsu, thanks [12:05] mhr3: we don't count that :) [12:06] * larsu certainly doesn't [12:06] mhr3: please tell me that doesn't actually work [12:08] greyback_: for your own sanity, stop thinking about that (and don't try it) [12:08] * greyback_ steps away [12:09] haha === alan_g is now known as alan_g|lunch === hikiko is now known as hikiko|lunch === jhodapp|afk is now known as jhodapp [13:03] greyback_: https://code.launchpad.net/~gerboland/unity8/u8m-osk/+merge/181887 ¿? [13:04] tsdgeos: yeah I needed to land it, else if you used Mir image, once you opened the OSK, you could do nothing else except type [13:04] ok [13:07] dednick: hey [13:08] mzanetti: howdy [13:08] dednick: how does that work in the indicators? [13:08] there is one dbusmenumodel which holds the icons in the panel, right? === hikiko|lunch is now known as hikiko [13:08] and where do the tabs get their information from? not the same model? [13:09] mzanetti: there is the IndicatorsModel, which tells both the panel and the tabs "where to get their content" [13:09] mzanetti: and the UnityMenuModel, which tells each item within the Panel and Tabs "what is the content" [13:10] ie the Indicators model tells us the dbus path etc, and the UnityMenuModels connects to each dbus path [13:11] mzanetti: but we need to connect to the dbus before we can tell if the indicator is visible or not. [13:11] dednick: does that mean that we always have all tabs, regardless if a given system service is not installed? [13:11] mzanetti: yes, at the moment. [13:12] somewhat defeats the purpose of the whole thing, doesn't it? [13:12] mzanetti: somewhat the purpose of this branch :) [13:13] yeah well.. seems to be a workaround for an architectural problem deeper down in the stack [13:13] mzanetti: hang on. no, the service is deemed to be installed if the indicator file exists. [13:14] only the ones with indicator files will appear in the tabs [13:14] or panel [13:14] ok... at least something :) [13:14] the visibility is more than just "being able to connect to dbus" [13:15] hmm... still don't like the fact that we're abusing the model for this... but don't have a better idea right now [13:15] for example, the messaging indicator in future probably wont be visible unless we have unread messages [13:15] but it still needs to be connected to be able to make that determination [13:16] mzanetti: yeah, i didnt really like the fact that i was using setData. [13:16] mzanetti: i have another solution in mind, but it's a bit unqmlish. [13:16] dednick: could the panel have a list property controlling all the visible tabs? [13:17] mzanetti: yeah, that could be a better solution. [13:17] tsdgeos: you happy with this? [13:18] sorry [13:18] wasn't reading [13:18] * tsdgeos reads [13:18] no problem [13:19] mzanetti, tsdgeos: although that wouldnt really work nicely with the filtermodel. [13:19] as it operates on roles... [13:19] mzanetti: yeah, you need a model role [13:19] d'oh [13:19] that's why i suggested an intermediate identitymodel [13:20] that adds the extra role [13:20] but it's a lot of work [13:20] tsdgeos: identityModels dont add roles do they? they just transform current ones [13:20] dednick: you can add more, that's not a problem [13:21] dednick: the other "easy" solution is creating your own proxymodel and reimplementing acceptsrow [13:21] but of course that's also more work [13:22] wait... shouldn't this behavior be controlled by the service? I mean... the service adds and removes all kind of stuff in the UI. shouldn't it also be able to determine when an entry is "active"? [13:22] which is when it is displayed and the tab is there [13:23] dednick: ^ === alan_g|lunch is now known as alan_g [13:26] mzanetti: it is. the visiblity comes from the service. but you need to make the connection (ie create the tab page) to determine if it's visible. Plus we can't change tab visibility. [13:29] dednick: I mean, this setData(isVisible) should happen within the service [13:29] ah... doesn't work again [13:29] crap [13:30] I don't know... if alberts approach works fine for you... otherwise I guess we have to stick with what you have now [13:30] mzanetti: heh. it does. but it has to go through the panel items as they are what is connected to the service. The IndicatorsModel has no connection to the service [13:30] yeah [13:31] dednick: standup [13:31] nic-doffay: you too [13:32] tedg: hi! https://bugs.launchpad.net/ubuntu/+source/indicator-messages/+bug/1217131 ;) [13:32] Launchpad bug 1217131 in indicator-messages (Ubuntu) "Test failure in recent quaternourly build." [Undecided,Confirmed] [13:32] tedg: (just poking) [13:32] sil2100, Yes, I know, looking at it. Unfortunately it happens much more often on the builders than locally :-( [13:33] kgunn: joining us? [13:34] ;/ [13:35] mzanetti: my potential solution was to have an intermediate model which connected to UnityMenuModel for row that was added to get the visibility.. but it's quite a bit of work. [13:35] like tsdgeos solution, but more pimping. [13:36] Cimi: curious, you connecting vga, hdmi or display port ? (mir on projectors) [13:36] kgunn, VGA [13:36] kgunn, but it behaves very weirdly... [13:37] Cimi: gpu ? [13:37] kgunn, intel [13:38] kgunn, test 1.1, 1.2, 1.3 seems to work [13:38] kgunn, cannot do 1.4 and 1.5 [13:38] on scenario 2, 2.1 starts with a weird "mirror mode" with one screen longer than the other [13:38] mirror displays however is off [13:39] greyback_, doh got totally carried away. [13:39] nic-doffay: we're still there [13:39] greyback_, on me way [13:46] scenario 3, I have crashes [13:46] sil2100: can you help me finding an issue in the release jenkins tests setup? [13:47] I can disable primary screen, but then when I switch the projector off, it goes back to lightdm [13:47] mzanetti: what's up? [13:47] sil2100: http://reports.qa.ubuntu.com/smokeng/saucy/image/3770/unity8-autopilot/306767/ [13:48] sil2100: on Galaxy Nexus, all unity8 tests fail to open /dev/uinput [13:48] sil2100: works fine on Nexus 4 [13:49] sil2100: can you point me to the scripts that do the setup before running the tests? [13:49] mzanetti: did you make sure uinput is there and that the user permissions are set-up correctly? [13:50] mzanetti: let me point you to the otto scripts that do the job, since there is some pre-config thing we're doing [13:52] mzanetti: take a look here for instance: http://bazaar.launchpad.net/~otto-dev/otto/testsuite_autopilot-unity/view/head:/target-override/usr/local/share/otto/setup-hooks/10-autopilot [13:53] sil2100: and which is the jenkins job that produces the results for this page? [13:53] http://reports.qa.ubuntu.com/smokeng/saucy/image/3770/unity8-autopilot/306767/ [13:54] mzanetti: hah! This might sound bad, but I have no idea ;) I think you'll have to ask someone from the QA team, gema would probably know best [13:55] sil2100: ok. thanks [13:55] mzanetti: since once it's out of the daily-release process I'm maintaining, I don't really coordinate the path [13:56] sil2100: same here... still we need to figure why unity fails. I just hoped you'd know more than me. But no problem. I'll find someone === dednick is now known as dednick|lunch === om26er is now known as om26er|away [14:12] greyback_: ping [14:12] kgunn: pong [14:18] asac: ping [14:18] kgunn: I've found the issue [14:19] mzanetti: hi [14:19] asac: http://bazaar.launchpad.net/~ubuntu-test-case-dev/ubuntu-test-cases/smoke-touch-apps/revision/80 [14:19] mzanetti: awesome!!! [14:19] this commit broke it [14:19] asac: because it removes the call to unlock_screen [14:19] thich contains the chmod 666 to /dev/uinput [14:19] s/thich/which/ [14:19] mzanetti: cool. btw, we think we fixed the uinput problem for real [14:19] even better [14:20] e.g. adding phablet to autopilot group [14:20] which makes that unnecessary... also fixing the udev rules :) [14:20] +1 [14:20] so lets see what happens in 1-2 hours [14:20] then we should see results coming in [14:25] asac: to be precise, in this one: http://bazaar.launchpad.net/~ubuntu-test-case-dev/ubuntu-test-cases/smoke-touch-apps/revision/80/unity8-autopilot/setup.sh [14:25] the NO_UNLOCK=1 is the culprit [14:35] nic-doffay: time for a review? https://code.launchpad.net/~mzanetti/unity8/fix-1213153/+merge/182406 [14:41] dednick|lunch: apparently qmenumodel is not building on i386 (tests for qmenumodel are failing: https://bugs.launchpad.net/ubuntu/+source/qmenumodel/+bug/1217335). Do you still depend on QMenuModel (the class) or can we remove it? [14:41] Launchpad bug 1217335 in qmenumodel (Ubuntu) "qmenumodel fails to build on i386/amd64 against Qt 5.1.1" [Undecided,New] [14:41] mzanetti, on it [14:42] mterry: I assume because you got dragged away for the html thingie, the demo branch is not fixed/ready yet, right? [14:43] mzanetti, demo should be fixed [14:44] mzanetti, stupid Q_SLOTS/Q_INVOKABLE mixup (I put Q_INVOKABLE after 'public' instead of in front of method prototype) [14:44] mzanetti, there's a qmltest that fails on jenkins but works locally. Just pushed a possible fix for that, but shouldn't block testing [14:44] mterry: hehe... and I've seen in the last version you still have a "pubic:" too much :P [14:45] mzanetti, you kidding me? How does that compile? [14:46] oh, I don't have pubic! [14:46] public [14:47] stupid typos... [14:47] mterry: line 819 for example [14:48] mterry: not a real problem... but not needed either [14:48] 1262 too [14:48] mzanetti, ah, I see what you meaj [14:48] *mean [14:48] mterry: and 1609 there is still the slot thingie [14:49] mzanetti, 1609 needs to be I thought, for the QtDBus auto-binding stuff to work, [14:49] mterry: we really should define that API in lp:unity-api... would fail to compile until all of them are aligned [14:49] ah ok [14:58] larsu: the messaging still using qmenumodel === dednick|lunch is now known as dednick [15:00] dednick: ah okay. So I guess I should have an actual look why the tests fail, then :) [15:02] larsu: fail on amd64 as well :/ works for me... [15:02] dednick: you have qt 5.1? [15:02] larsu: no [15:04] seb128: why am I seeing this? http://paste.ubuntu.com/6032941/ [15:05] that's on todays touch image on my nexus 3 [15:05] I must be crazy or something but it states that the installed version of HUD is 13.10.1+13.10.20130816.1-0ubuntu1 [15:07] instead of that 13.10.1+13.10.20130822.4-0ubuntu1 which we worked so hard to fix critical bugs [15:07] Wellark: did you see our discussion about the theme icon provider from earlier today? [15:07] larsu: nope [15:07] larsu: quick recap please [15:09] Wellark: did you get a chance to work on the tests for that? [15:10] Wellark: I can't see any on the MR and I remember you needed the theme provider desperately two weeks ago [15:10] larsu: doing them today after we get this critical HUD problem sorted out [15:10] larsu: well, we all need it [15:11] Wellark: ah cool. Thanks. Yes, other people have been asking abou this, which is why I'm pinging you right now :) [15:11] in the mean time I fixed upstream QIcon [15:11] ah, very cool [15:11] and HUD [15:11] and HUD [15:11] and HUD [15:11] lol [15:12] larsu: so yeah. unit tests to get that MR forward coming today [15:12] or I will lock my self to bathroom and cry [15:12] haha, please don't :) [15:13] * larsu hears that is not good for mental stability [15:13] mzanetti, you said the launcher jumps in with my launcher branch, how? [15:13] mzanetti, you mean the shadow suddenly appears? [15:13] Cimi: no [15:14] Cimi: the EdgeDragArea has a minimumMovement [15:14] Cimi: which means, you need to drag for half a grid unit or so before the drag gesture is recognized [15:15] Cimi: once the gesture is recognized, the launcher shows up, already having moved for that distance... [15:15] Cimi: I think the easy fix for that is to make it visible onMousePress already [15:15] Cimi: because then the shadow will appear when you put the finger down and the rest happens as before [15:16] mzanetti, but there is no onMousePress [15:17] Cimi: DirectionalDragArea::Status [15:18] mzanetti, recognized? [15:18] recognized is when a drag is recognized [15:18] that's too late [15:18] so what? [15:21] Cimi: how about Undecided? === swook` is now known as swook [15:22] mzanetti, probably better opacity: 0 when is waitingfortuouch or no status? === om26er|away is now known as om26er === swook is now known as swook` [15:30] Cimi: not only... [15:30] Cimi: it also needs opacity = 1 when the launcher is revealed [15:30] Cimi: but you're getting close, yes :) [15:32] mzanetti, implying 1 all other cases [15:36] larsu: just upgraded to 5.1 . still passes :( [15:36] well. :) [15:37] dednick: let's close the bug as WORKSFORME :) [15:38] larsu: lol [15:40] dednick: ah! "QXcbConnection: Could not connect to display" [15:41] I guess we jut need a xvfb-run somewhere in there... [15:41] * larsu wonders why this worked before [15:46] sil2100, So larsu found a fix for the indicator-messages issue. Can we kick off the stack manually to see if it's all green? [15:50] tedg, he's hosting a vUDS track, not sure he has time for IRC [15:50] tedg, let me have a look [15:50] seb128, Ah, okay, thanks! [15:52] tedg, build triggered [15:52] seb128, Thanks /me crosses fingers! [15:57] mzanetti, opacity: (root.state == "" && dragArea.status === DirectionalDragArea.WaitingForTouch) ? 0 : 1 ? [15:57] Cimi: looks reasonable. did you make sure that it doesn't disappear too early? afaik the state changes to "" before the animation is done with hiding it [15:58] mzanetti, could do x == -width instead root.state == "" [15:59] Cimi: yeah, that seems a good one [15:59] mzanetti, you think we need a behaviour? [16:00] hmm... I'd add one. [16:00] but UbuntuAnimation.FastDuration [16:01] mzanetti, do they already work? [16:01] mzanetti, ubuntuAnimation [16:01] Cimi: also, need to check if that doesn't conflict when you drag in the laucnher very fast [16:01] mzanetti, it's why I was asking [16:01] Cimi: yes, UbuntuAnimation works fine [16:04] mzanetti, I pushed, if you could test on your phone... [16:07] Cimi: ok. not sure I'll make it today [16:07] Cimi: but I'll review tomorrow [16:21] larsu: ping [16:22] dednick: yep [16:23] larsu: is the only diff between g_object_unref & g_object_clear that it nulls the pointer after? [16:24] dednick: no, it also checks for NULL (this was more of an FYI then something you have to fix) [16:25] larsu: ok. thanks [16:26] Mirv: I have a patch that makes the qmenumodel tests work for me. Can you please test it? (branch is attached to the bug) [16:27] larsu: hm. did it just fail outright for you then? [16:28] dednick: when sshing into a vm :D [16:28] larsu: ah :) [16:28] I like using vms when adding ppas like that [16:29] and sshing seemed like the easiest way to get a no X situation [16:29] larsu: indeed. [16:35] bregma, hey, is anyone looking at unity failing publication because too many tests fail? [16:37] seb128, bregma is out until tomorrow, and we are trying to look at all the failures...but not a lot of stuff has changed in unity it self the last few days [16:37] seb128, hoping to focus on all the AP failures today... [16:38] bschaefer, ok, great, thanks [16:38] seb128, np! thanks for poking us! [16:40] larsu: tested. added my approve, works now both locally and in the 5.1.1 PPA with your branch! [16:40] Mirv: awesome! Exactly the message I want to hear on my eod :) [16:40] dednick: ^^ can you approve it too, please? [16:49] larsu: sure [16:52] larsu: done. [16:53] larsu: my fixes are in UnityMenuAction branch as well. [17:59] is there anybody around from the unity 8 team who is able to jump into http://summit.ubuntu.com/uds-1308/meeting/21947/client-1308-touch-1404-a11y/ and join the discussion? [18:02] tvoss: ^^ Are you able to join, or could you point others who may be in a better position to discuss the situation? [18:05] Hello, how can I deactivate the animation when the dash opens ? [18:05] TheMuso, gimme5, with you then [18:06] tvoss: Sure, hangout URL is https://plus.google.com/hangouts/_/ff15e632070786351f22a797e5fcf553212c43a8?authuser=0 === olli_ is now known as olli [18:22] mhr3, quick fix will be landing soon! [18:35] greyback_: just wondering when you phlash are you using -b ? [18:35] wonder if that's a diff [18:35] kgunn: not usually, unless things get very confused [18:36] greyback_: specifically what image are on atm [18:36] ive gone back and forth several times...pending+mir got lots of rendering glitches on dash [18:38] kgunn: nexus4, yes I've heard of rendering glitches unfortunately [18:38] ;) [18:38] greyback_: but these aren't just in the panel area....this is why i thot i had no assetts....they are there, just weird [18:39] kgunn: can you send me a screenshot/photo of the phone? [18:40] greyback_: holy crap...it just recovered...what the what [18:40] looks awsome now [18:40] hmmm.....i was measuring framerate...killed that...then it got better.... [18:42] very odd [18:42] * greyback_ hopes that doesn't happen again [18:43] greyback_: ah...weird..."stop unity" unity failed to stop...but, i ran unity again on top.... [18:43] so i had like 2 unitys ? [18:43] kgunn: yep, it's on my todo list to prevent that [18:44] greyback_: just checking...password for su phablet is "phablet" right ? [18:44] kgunn: yep [18:44] greyback_: this definitely worked earlier...but now unity is being stubborn [18:44] kgunn: any strange output in your shell? [18:44] hey [18:45] would someone be so kind as to review this very simple merge request? https://code.launchpad.net/~mathieu-tl/unity-lens-applications/libcolumbus/+merge/182467 [18:45] I'd be so very grateful :) [18:45] phablet@ubuntu-phablet:/$ su phablet [18:45] Password: [18:45] phablet@ubuntu-phablet:/$ stop unity [18:45] stop: Unknown job: unity [18:45] phablet@ubuntu-phablet:/$ [18:45] cyphermox, can do [18:45] seb128: thanks [18:45] greyback_: unknown job unity? [18:46] kgunn: "unity8" [18:46] greyback_: you know...i was just thinking...did i accidently delete the 8 in my notes? [18:46] kgunn: :D [18:49] greyback_: Whats your launchpad user ID? [18:50] TheMuso: "gerboland" [18:50] greyback_: Thanks. [18:50] TheMuso: thank you [18:58] kgunn: everything ok? [18:59] * greyback_ had to go [18:59] has [18:59] o/ [19:02] \o [19:15] mzanetti, did the icon size in the launcher change recently? they look a bit smaller to me with today' image [19:43] om26er: yes. they are a bit smaller [19:44] mzanetti, they look to small in comparison to the dash icons, probably something for design to consider [19:44] *too [19:44] om26er: this is what I got from design [19:45] om26er: actually it was a bug on my side that they were bigger. all the design specs had the smaller ones, right from the beginning [19:45] mzanetti, it could possibly be that my eyes got used to the old size afterall. [19:45] the continued use of the next few days will tell if they look weird still [19:46] mzanetti, offtopic: the quicklists appear on the launcher, shall i raise a bug for that ? [19:46] om26er: yes, if it turns out to cause issues I'm sure the size will be increased again [19:48] om26er: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1204466 [19:48] Launchpad bug 1204466 in Ubuntu UI Toolkit "[Popover] Functionality to control positioning needed" [High,New] [19:48] mzanetti, ah - cool [19:54] om26er: did you see this before? http://s-jenkins:8080/job/generic-mediumtests-runner-saucy/2287/console [19:56] mzanetti, no, never saw that one. try a restart for that job ? [19:56] om26er: there are 4 of this in a row now :/ [19:58] something is broken with that VM maybe ? === seb128_ is now known as seb128 === bschaefer_ is now known as bschaefer [22:25] Dumb question: I want to try unity. How do I get it? === Eric is now known as Guest50120 [22:26] Guest50120, well what are you on right now? [22:28] Ubuntu 10.04. [22:29] Guest50120, hmm well you'll either have to upgrade to 12.04, or partition your current hard drive with 12.04 or 13.04 to avoid messing up your 10.04 version [22:30] Ah. [22:31] theres always this fun online tour: http://www.ubuntu.com/tour [22:31] but, it wont be like the real thing :)