=== Trevinho is now known as Trevinho|holiday === _salem is now known as salem_ === salem_ is now known as _salem [06:58] mo'nin [07:38] morning [07:43] Saviq, remove 0.1.16 from trunk or I'll do :) [07:43] Cimi, check again [07:43] Saviq, I just merged and I got it [07:43] weird [07:44] maybe was last night merge [07:44] Cimi, well, you got it because you have it [07:44] Cimi, not because it's in trunk ;P [07:44] I am infected! [08:06] @unity is canonical irc down for you? [08:06] Cimi: seems like it [08:06] ok [08:07] Cimi: unless they've done something with dns again [08:37] o/ [08:41] RIP Robin Williams :| [08:53] Saviq, do we have scope settings in the mock or I have to c++ them? [08:53] Cimi, in "the mock"? [08:53] Cimi, you mean in unity-api's mock? [08:54] Saviq, in Unity 0.2 mock [08:54] Cimi, right, no [08:56] Cimi, btw, found a bug with your recent changes [08:56] Cimi, there's no right margin in overlay either [08:56] Cimi, where it should be [08:58] Saviq, added to my todo [09:00] Saviq: flashing mako. bar goes from right to left :) [09:00] Saviq: logo rotating clockwise [09:01] mzanetti, so it's just wrapped in the middle or something ;) [09:01] so its either mirrored vertically [09:01] or yeah, mirrored most probably [09:01] or yeah, shifted [09:01] well, no, if it was mirrored [09:01] wouldn't it rotate counter clockwise? [09:01] true :D [09:01] so its a different piece of animation. [09:02] I don't think its shifted either [09:02] because spacing between bar and logo wouldn't match [09:02] so what probably really happens [09:02] "probably really" yes. :) [09:02] is coordinates are topleft vs. bottomright [09:03] or no [09:03] probably really not [09:03] :) [09:03] because that'd have failed in a more weird way [09:03] I'd say its just implemented the other way round. that's it [09:03] mhm [09:05] Saviq, so in fake_scope.cpp it returns a null when you look for settings, is there any other mock code I could use there? [09:05] maybe from unity api [09:05] Cimi, you need to write a mock settings model [09:05] Saviq, yup, but maybe they wrote it already no? [09:05] otherwise how did they test it? [09:05] Cimi, not likely [09:06] Cimi, with the real thing [09:06] Cimi, pete-woods will know [09:06] pete-woods, I need to write a scope settings mock, where is the real thing? :) [09:08] Cimi: the settings model in unity-api gives a reasonably complete mock [09:08] it has bool, string, number and list in it [09:10] pete-woods, found it [09:10] :) === vrruiz_ is now known as rvr [10:22] Saviq: when you have some time, can you please verify this grooveshark thingie? [10:22] https://bugs.launchpad.net/unity8/+bug/1355356 [10:22] Ubuntu bug 1355356 in unity-webapps-grooveshark (Ubuntu) "App spread should not kill media" [Undecided,New] [10:22] Saviq: works fine with youtube and others here [10:22] and I'm not really in the position to test with grooveshark [10:23] mzanetti, will do in a mo [10:34] mzanetti, critical one for you bug #1355729 [10:34] bug 1355729 in Unity 8 "Network caching doesn't work anymore after introducing Dash as an app" [Undecided,New] https://launchpad.net/bugs/1355729 [10:35] mzanetti, we didn't bring over caching from shell to dash [10:35] like if that ever worked :D [10:36] mzanetti, yes it did [10:36] mzanetti, check out your ~/.cache/unity8/network [10:36] Saviq: then the cache is the same speed as the network connection :D [10:37] mzanetti, we're not caching for speed, we're caching for lower data usage [10:38] Saviq: we should cache for speed too :D [10:38] Saviq: jokes aside... how come then the cached content still takes seconds to load? [10:38] mzanetti, that assumes that network is the bottleneck, which we know it is not [10:38] mzanetti, bug #1338430 [10:38] bug 1338430 in Unity 8 "Slow loading/showing of images in the dash" [Critical,Triaged] https://launchpad.net/bugs/1338430 [10:59] Saviq: hi. Have a bit of a conundrum about qml singletons. Can't really use the objects given to qml safely outside of qml, since they're managed by the engine (can be deleted before c++ objects which may use them on destruction). You ever come across that problem before? [11:00] dednick, nope, but sounds legit [11:00] dednick, why would you need to use the singleton in some other object's d'tor? [11:01] Saviq: yeah, but the qml documentation says not to use singlton factory. [11:01] Saviq: because there's some regegistration when objects are destroyed [11:01] Saviq, spot any mistake? http://paste.ubuntu.com/8025733/ [11:01] de-registration [11:02] Cimi, is that a review or debugging? [11:02] Saviq, QMetaProperty::read: Unable to handle unregistered datatype 'unity::shell::scopes::SettingsModelInterface*' for property 'Scope::settings' in tryDash [11:02] Saviq: http://bazaar.launchpad.net/~unity-team/unity8/trunk/view/head:/tests/mocks/Unity/Application/ApplicationInfo.cpp#L55 [11:03] Saviq, my C++ still lacks... [11:03] Saviq: although i'm not saying that's very good design [11:03] Cimi, and where are you registering that type in your mock plugin that you expect it to work? [11:04] Saviq, right [11:04] Saviq, adding it [11:04] Cimi, that message was so clear it couldn't be more so ;) [11:05] Saviq, it is clear if you understand qml and c++ [11:05] Cimi, afternoon read http://qt-project.org/doc/qt-5/qtqml-cppintegration-topic.html [11:05] dednick, so... AppMan is being destroyed at the same time SurfMan is, correct? [11:06] Saviq: surface manager has already been destroyed [11:06] dednick, well, yeah, that's what I mean [11:07] dednick, why doesn't SurfMan connect to the surface's destroy signal? does AppMan really need to tell SurfMan that's happening? === pete-woods is now known as pete-woods|away [11:11] Saviq: well, i guess there are ways to mitigate the problem, but still need to use SurfMan from App/Surface at some point. Just wondering if you knew a better way. [11:11] dednick, nothing that would save us this issue, engine can destroy the singleton at will [11:12] dednick, I don't think there's a way to "get" it from the engine [11:14] Saviq: right. I've done a bit of a crazy solution decorating a QSharedPointer and passing that as the singleton, then storing the shared pointer where-ever it's used. Works, but it's pretty messy and quite a pain to maintain. [11:14] Holas [11:14] dednick, and then mixing Qt's QML/parent-child ownership and smart pointers is somewhat messy, too [11:15] Saviq, ok of this http://pastebin.com/8zbBKHjC how do I correctly set implicitHeight? [11:15] Saviq: the qml singleton isn't a smart pointer, the object it uses just holds a ref to one. but it's pretty crap [11:16] Cimi, spam detection, please use paste.ubuntu.comn [11:16] i'll just try sort it out with connections [11:16] Saviq, I used qtcreator plugin for the first time in my life :D [11:16] Saviq, http://paste.ubuntu.com/8025836/ [11:16] Saviq, ScopeSetting is Item with list of props [11:17] Cimi, why would you set implicitHeight when in this case this widget only really makes sense in the height of the ItemSelector's? [11:17] Saviq, because they appear all on top of each other [11:17] Cimi, aaand? [11:17] Saviq, this and others [11:17] Cimi, you need to bind ScopeSettings's height to ItemSelector's [11:17] Cimi, but you should be using Ubuntu{ListView,Column} here and an Expandable [11:17] Saviq, http://paste.ubuntu.com/8025853/ [11:18] Cimi, in which case there's more that needs to happen [11:18] Saviq, that one as well, no implicitHeight they are all on top as delegates [11:18] Cimi, your component is 0-height [11:18] yeah I need to set implicitHeight of the children [11:18] NO [11:18] you need to set height [11:19] not implicit height, because that component doesn't deal well with height other than the one of its children [11:20] Cimi, if you have a ListItem.Empty in every ScopeSetting [11:20] Cimi, why isn't ScopeSetting a ListItem.Empty already [11:20] Saviq, because List is not using ListItem.Empty [11:20] Saviq, I did that originally, but then I needed to change ScopeSetting [11:20] Cimi, what's List? [11:20] Saviq, list of words [11:20] hold on [11:21] Cimi, I really don't understand your problem [11:21] Saviq, something like the language setting I think [11:21] ScopeSetting has no height [11:21] ListItem.Empty has height [11:21] Saviq, yes [11:21] ListItem.Empty is in ScopeSetting [11:21] so what height is ScopeSetting supposed to have? [11:21] String, Number, Boolean use Empty [11:22] List might use ItemSelector [11:22] AAND? [11:22] doesn't ItemSelector have a height? [11:22] yes, so I set height of that ScopeSetting as height of the children? [11:23] nope [11:23] because it expands.. [11:23] this is not correct http://paste.ubuntu.com/8025908/ [11:26] :/ [11:27] Cimi, why isn't that correct? [11:27] Saviq, btw someone linked me this http://theburgerlist.tumblr.com/post/94443927060/le-bourguinon-bun-le-bun-riverside-feast-battersea [11:27] Saviq, guess could be a treat next week :) [11:28] Cimi, only there until Wed afternoon, though [11:28] ok [11:28] Saviq, so if I set that, when I click on the item it should expand, but is clipped instead to that height [11:31] Cimi, sounds like a bug in the component [11:32] Cimi, except there's http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/trunk/view/head:/modules/Ubuntu/Components/ListItems/ItemSelector.qml#L153 [11:33] Cimi, I'm afraid you can't do ScopeSetting if you want to use ExpandablesColumn (which you should) [11:34] ah damn with a *Factory you can't even do that either [11:34] mzanetti, do you have an idea about using Expandables and a Loader as delegate? [11:35] Saviq: hmm... you don't know the loader's content size in advance, do you? [11:35] Saviq: dunno. would need to try [11:35] might work, might not... [11:36] Saviq: so you can put inside an Expandable whatever you want, also loaders [11:36] mzanetti, no, I was wondering if we can reparent Loader.item to the Expandable{ListView,Column} [11:36] no [11:36] mzanetti, right, so delegate Expandable { Loader { } } [11:36] yes [11:36] Saviq: and: [11:36] we'd have to fwd everything things from item to the parent expandable [11:37] Expandable { expandedHeight: loader.height ... } [11:42] Saviq, shall I use expanded ones? [11:46] Saviq, I'll use expanded: true in the meanwhile [11:47] Cimi, what do you consider a "meanwhile"? [11:48] Saviq, hah, until we realise what's wrong with it [11:48] with the height [11:48] Cimi, we did already [11:51] Saviq, so shall I always use expanded? [11:52] Cimi, no, why? [11:52] so how can I make it work? [11:52] I am confused [11:53] Cimi, http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/trunk/view/head:/examples/ubuntu-ui-toolkit-gallery/ListItems.qml#L106 [11:53] Cimi, this works in the gallery [11:53] Saviq, this is a column [11:53] Saviq, I already checked [11:53] Cimi, so? a column takes the item's dimensions [11:54] Saviq, maybe because is not embedded in one item? [11:54] wrapped [11:55] nope [11:56] just tried removing the item on top of listitem.itemselector [11:56] still issues [11:56] Cimi, http://paste.ubuntu.com/8026115/ [11:56] Cimi, btw, ItemSelector has its own header, use it [11:57] Saviq, I am already using it [11:57] http://paste.ubuntu.com/8025836/ [11:59] Saviq, you might want to try this http://paste.ubuntu.com/8026141/ [11:59] Cimi, yeah, and why would that ever work? [12:00] Cimi, when you set the Item's height to 100? [12:00] Saviq, it doesn't [12:00] Cimi, well, how can it? [12:00] Saviq, I need to set the item height to be exactly of the container [12:00] Saviq, because if the item is collapsed it should not eat more height [12:00] than it needs [12:01] Cimi, does not compute [12:01] not container [12:01] sorry children [12:01] Item height, where you set 600, should be the height of the itemselector [12:01] and the added rectangle [12:02] because when you use the item in a listview, you want to allocate width and height [12:02] Cimi, http://paste.ubuntu.com/8026161/ [12:03] thius works? [12:03] it is basically http://paste.ubuntu.com/8025908/ [12:04] Cimi, why would it *not* work? [12:04] that code I pasted doesn't work [12:04] what do you mean "doesn't work"? [12:05] Saviq, when I click on the itemselector to expand it [12:05] Saviq, the height doesn't grow [12:05] Cimi, but you have a Loader in the middle, how does the Loader's height look? [12:08] Saviq, the height doesn't change [12:08] neither the inner one [12:08] of the item I mean [12:08] Cimi, you need to push the code somewhere if you still want to do IRC debugging, which you know I resent [12:09] Saviq, call it help... [12:09] Cimi, what I say happens: you're overriding the height in the delegate [12:09] Cimi, which means whatever you do *inside* the component doesn't matter [12:10] Saviq, pushed [12:10] ~cimi/unity8/scope-settings [12:10] make tryDash === MacSlow is now known as MacSlow|lunch [12:33] Cimi, ok confirmed, this looks real weird indeed [12:34] Saviq, I am working on a workaround [12:34] Cimi, talk to SDK peeps, too [12:35] Cimi, TBH the ListItem.Empty is complex enough that something there must be getting awry [12:37] Saviq, http://paste.ubuntu.com/8026405/ [12:37] Saviq, ETOOUGLY [12:37] Cimi, indeed [12:37] works btw [12:38] Cimi, the same seems to happen with OptionSelector [12:38] HUH [12:40] Cimi, height: listItem.height + 1 [12:40] Saviq, can I commit that workaround with a comment? [12:40] Cimi, talk to SDK first [12:40] Saviq, that works too [12:40] but without animation [12:41] Cimi, well, there is animation if you look at the console output [12:42] Cimi, it just doesn't get "up" correctly [12:42] Cimi, but in any case that suggests something's awfully broken in the ListItems' height [12:43] Saviq, can you join me in #sdk and probably explain better the issue? [12:48] Saviq, also wondering how containerHeight is not causing binding loops [12:49] Saviq, having headache trying to understand how the properties are nested === _salem is now known as salem_ === MacSlow|lunch is now known as MacSlow [12:56] Saviq, can I also do http://paste.ubuntu.com/8026544/ to make tryGenericScopeView work? [12:57] Saviq: FYI: https://bugs.launchpad.net/unity8/+bug/1318533 [12:57] Ubuntu bug 1318533 in indicator-network (Ubuntu) "provide unity-notifications-service virtual package" [Wishlist,Triaged] [13:08] greyback: Hey, I have a question about the transition to qtmir wrt to https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1353041. Whenever you have some time, could we chat about it for a bit? [13:09] Ubuntu bug 1353041 in unity8-desktop-session (Ubuntu) "Logging out of the Unity8 desktop preview session hangs" [Undecided,Confirmed] [13:10] ChrisTownsend: hey. Sure, happy to talk. [13:11] greyback: Cool. So I've done some investigating and made some notes in the bug, but in a nutshell the QGuiApplication dtor is blocking waiting on some pthread_cond_wait. I have no clue what it's waiting on though. [13:12] greyback: So teardown gets stuck at this point and the log out hangs. I'm thinking it might have to do with the transition to qtmir, but wanted to see your opinion. [13:13] has the category header links changed? ours in our aggregating scopes aren't working atm [13:14] ChrisTownsend: hmm yeah, I've skimmed your comments so far. It possible qtmir waiting for something and thus blocking shutdown, but I see nothing that obvious aside from one QMutexLocker, but I'd be surprised if that the culprit [13:14] ChrisTownsend: I can certainly investigate it further [13:14] greyback: Awesome, that would be great! [13:14] ChrisTownsend: please add qtmir as an affected project, and assign to me [13:15] I'll give it a look today [13:15] Saviq: ok... I guess this is it now: https://code.launchpad.net/~mzanetti/unity8/new-lockscreen-design/+merge/230478 [13:15] greyback: Ok, will do. And thanks! [13:15] np [13:15] mzanetti, kk [13:27] Saviq: any flaws in my reasoning at bug #1318533 ? [13:27] bug 1318533 in indicator-network (Ubuntu) "provide unity-notifications-service virtual package" [Wishlist,Triaged] https://launchpad.net/bugs/1318533 [13:30] Wellark, nope [13:30] Wellark, all good [13:31] Saviq: hate to reject a community contribution, though.. [13:37] Wellark, well, it might be better at some point soon [13:37] Wellark, when we drop the whole UnityMenuModel concept from notifications [13:37] Wellark, at least in the shape they're there now [13:38] Wellark, one idea was to use trusted sessions for this (so i-n would actually *draw* the UI it wants, not only direct it) [13:39] Saviq: hmmm... do we still need the caching in unity8? [13:39] mzanetti, we don't *need* it, but it can't hurt [13:40] I guess it doesn't hurt... [13:40] ack [13:40] mzanetti, we can get remote images for avatars in notifications, for example [13:40] and in messaging menu [13:40] I guess it *can* hurt... but lets hope all goes well :D [13:40] ;) [13:52] pstolowski: https://code.launchpad.net/~mzanetti/unity8/dash-caching/+merge/230487 [13:53] mzanetti, awesome, thanks! [13:54] Saviq, did we change dash so that it no longer goes back to the apps scope when you switch back to it? [13:54] dandrader, depends on how you switch to it [13:54] dandrader, it should still happen when you left-swipe or press the button [13:55] mzanetti, is it enough to register it in main? no qml changes? [13:55] Saviq: except when you're already at the dash [13:55] mzanetti, well, button still works [13:55] Saviq, right [13:55] mzanetti, left swipe doesn't [13:55] and overview broke it [13:55] pstolowski: yep. it registers a QNetworkAccessManagerFactory [13:55] dandrader, but yeah, the bug you were looking at [13:55] dandrader, swipe dash to a scope [13:56] (different than apps) [13:56] pstolowski: so if QML wants something it'll create a new QNAM and will get it from that factory [13:56] dandrader, launch an app from the launcher [13:56] dandrader, long-left-swipe to dash [13:56] dandrader, you see the old scope, until you touch it [13:56] dandrader, I wonder, though, ig the dash getting suspended is causing this [13:57] if [13:57] Saviq: yes, I think so [13:57] but unlikely [13:57] mzanetti, nah, happens without even waiting for the other app to launch [13:57] mzanetti, k, thanks for explaining. +1 [13:57] mzanetti, so within the 5 or so seconds before it gets suspended [13:58] Saviq: not here [13:59] oh. yes... this time yes [13:59] Saviq: not if you're really quick [13:59] Saviq: its 3 secs btw [14:00] strage... it does not happen all the time anyways [14:00] what (additional) branch is needed to get current lp:unity8 trunk to to work with the run-script on the desktop? [14:00] but not related to the 3 secs thing [14:00] MacSlow: forget about the run script :D [14:01] Saviq, when you reassign bugs to settings, please do it against the package, we don't use the project for bugs [14:01] mzanetti, but I should now use instead.... ? [14:01] MacSlow: make tryShell for instance [14:01] MacSlow: depends a bit on what you need [14:01] MacSlow, lp:~saviq/unity8/tweak-runscript [14:01] it's a start [14:01] or that :) [14:05] Saviq, Is there a way to know when the dash has settled? [14:06] Saviq, i.e. when it's done loading icons, etc. [14:06] Saviq, Trying to make the app startup results more reliable, avoiding having the dash stealing CPU. [14:06] tedg, right now it's never "done loading icons", because it won't start loading them until you look at the dash [14:07] Saviq, So I look at the dash by unlocking Unity before starting the app. [14:07] tedg, just start the app, it should unlock itself [14:07] I *think* [14:07] mzanetti, Saviq: with lp:~saviq/unity8/tweak-runscript I still only get a tiny black square window... and I don't have any Makefile (and therefore no make-targets) [14:08] MacSlow: not following... [14:08] mzanetti, you said something about "make tryShell" a few lines back [14:08] Saviq, Hmm, that seems like a bad thing, no? [14:08] MacSlow: anyways... IMO you're better off using make tryShell [14:08] MacSlow, the makefile's in builddir [14:09] MacSlow: unless you really require unity to be started by upstart [14:10] tedg, or even better [14:10] tedg, just stop unity8-dash [14:10] Saviq, Ah, that's a good idea. [14:12] Saviq, Though, if that makes apps start faster you might get pressure to do that on all devices ;-) [14:13] Saviq, mzanetti, btw, unity8-dash is never suspended because "ApplicationController::primaryPidForAppId FAILED to get PID for appId= "unity8-dash"" [14:13] tedg, is FINE [14:13] dandrader, right, it doesn't have an appid [14:13] Saviq, which actually helps me with debugging :D [14:14] mzanetti, Saviq: how do I get unity8 (via "make tryShell") to use/load the notifiation-plugin? [14:15] MacSlow, you don't, tryShell will load the mock one [14:16] MacSlow, ./run.sh should still launch the correct shell (and with my branch the correct dash, too) [14:16] MacSlow, if it doesn't work for you, do you get any interesting output on the console? [14:16] MacSlow, does it launch if you just go ./builddir/src/unity8 / [14:16] ? [14:17] Saviq, "./builddir/src/unity8 /" works [14:18] Saviq, what's the pin/password? :) [14:18] oh... mterry collecting abstain votes again :D [14:18] mzanetti, gotta have 'em all! [14:19] Saviq, my local one?! [14:21] MacSlow, yes [14:21] MacSlow, with ./run.sh you won't need it [14:21] MacSlow, yeah the / was meant to be ?, no need for that ;) [14:22] Saviq, yeah... working now [14:23] mterry, oh I missed your query about wrong-pass review [14:23] mterry, and I actually thought I asked mzanetti for it already https://code.launchpad.net/~mterry/unity8/wrong-password-handling/+merge/229184 [14:23] * Saviq greps logs [14:23] yeah... might well be [14:24] mzanetti, well regardless. If you have time today, it would be swell [14:24] mzanetti, I did, back on the 8th! ;) [14:24] yeah... I had a feeling [14:24] THAT TONIGHT'S GONNA BE A GOOD NIGHT [14:25] hmm... not really, no [14:27] mterry: btw... https://code.launchpad.net/~mzanetti/unity8/new-lockscreen-design/+merge/230478 [14:28] mzanetti, bro... [14:28] Saviq: chances horizontal-list gets landed for next week? [14:28] mterry: ? [14:29] cwayne, *for* next week for sure [14:29] mzanetti, I was expressing frustration at all the points of conflict [14:29] :) [14:29] mzanetti, but yours is only an hour old! [14:29] mterry: yeah... [14:29] cool beans [14:32] what's horizontal-list? [14:40] mzanetti: a new renderer for unity that joe wanted [14:40] ah. a card layout thingie [14:41] ok. thanks [14:43] hey, can I get a volunteer to review these branches: [14:43] https://code.launchpad.net/~gerboland/unity-api/surfaceSizerCallback/+merge/230270 [14:43] https://code.launchpad.net/~gerboland/unity8/initialSurfaceGeometry/+merge/230490 [14:43] https://code.launchpad.net/~gerboland/qtmir/initialSurfaceGeometry/+merge/230491 [14:44] greyback, don't touch UNITY_API_MICRO [14:44] Saviq: why not? [14:44] greyback, because you're not touching the library [14:44] greyback, *do* touch VERSION in include/unity/applications/CMakeLists.txt [14:44] Saviq: understood [14:45] greyback, also add -0ubuntu1 so that dch doesn't complain about native vs. non-native versions [14:45] greyback, doesn't *really* matter as train will overwrite anyway, but just for correctness :) [14:45] yeah I just expect train to fix that [14:46] greyback, wonder, could this not ba Q_PROPERTY? [14:46] hmm probably not [14:46]