[07:55] mzanetti, ping [07:56] morning Saviq [07:56] mzanetti, hey, any idea about https://jenkins.qa.ubuntu.com/job/generic-mediumtests-builder-saucy/850/console [07:56] mzanetti, ERROR:pbuilderjenkins:H05set_package_version not found in hooks [07:57] Saviq: not seen yet... let me check existing hooks [07:57] mzanetti, btw dh_install: usr/share/unity8/plugins/mocks/Unity/Launcher/libMockLauncherPlugin.so exists in debian/tmp but is not installed to anywhere [07:57] dh_install: usr/share/unity8/plugins/mocks/Unity/Launcher/qmldir exists in debian/tmp but is not installed to anywhere [07:57] ok. I'll fix that [07:58] mzanetti, I think it might've changed with fginther adding our local repo yesterday [07:58] mzanetti, I mean the hook [07:58] Saviq: yeah... seems so [08:03] Saviq: oh crap... thats a mess [08:03] mzanetti, :/ [08:04] Saviq: the jobs pull hooks from 3 different sources [08:04] eek [08:04] * mzanetti is not going to clean this up now... [08:05] I'll try to just remove the set_version hook from the mediumtests job [08:06] not possible :/ [08:09] mzanetti, who do we have in our timezone that could help us fix? [08:10] Saviq: yeah... thats actually a problem... there is no QE in europe timezone any more [08:10] right :/ [08:10] well, vrruiz is, but he's mostly busy with webapp stuff [08:10] yeah, we're hiring for that, probably, then [08:11] Saviq: anyways, that missing hook comes from a branch owned by fginther [08:11] Saviq: and that branch is pulled in the builder jobs but not in the others [08:11] Saviq: I'm trying to pull the hook from there and add it to where it belongs. the real hooks package [08:11] mzanetti, thanks [08:11] mzanetti, why can't you remove it btw? [08:12] because they are inherited from upstream jobs [08:12] mzanetti, which are configured autmagically [08:12] and if I remove it upstream it won't bump the version number any more [08:12] which would cause the local repository to break I'd assume [08:13] mzanetti, yeah, it's probably for that (it needs to mangle the changelog to actually do a release) [08:13] yep [08:13] mzanetti, btw, did qmluitests get the local repo added, too? [08:13] * mzanetti checks, but I would assume no [08:15] Saviq: qmltests does not use the inherited hook list [08:15] Saviq: because that's a specific job for unity8 so it has the required unity8 hooks hardcoded [08:16] mzanetti, yup, we should add the one that adds the local repo, right? [08:16] Saviq: guess so, yes... let me first fix the mediumtests one. [08:17] btw. why is our code coverage dropping like crazy? [08:17] Saviq: ^ [08:18] mzanetti, are you looking at -ci or at -autolanding? [08:18] both... right now autolanding [08:18] mzanetti, anyways, indicators-client isn't tested well yet [08:19] oh right... that was my problem child already when I was in the apps team [08:19] mzanetti, yeah, we're taking it over and we'll be improving that [08:19] mzanetti, at least the QML part [08:20] mzanetti, how about this? http://paste.ubuntu.com/5857753/ [08:22] mzanetti, also, somehow we started gathering coverage for tests again, need to fix that to get some real-ish numbers [08:24] Cimi: what's this? [08:24] mzanetti, instead of clearing the model [08:24] mzanetti, I push/append [08:24] Cimi: ok. give me a minute for my current task, then I'll re-review [08:25] mzanetti, it's a WIP [08:25] mzanetti, like, this works but I believe there might be flaws [08:25] mzanetti, like if you clear also the buffer and remain with one or two items, will it scroll? [08:26] maybe since onCurrentIndex should populate the sides... [08:26] might have to test [08:26] Cimi: ah ok... yeah, I can give it a closer look in a few. Not sure either how the thing behaves in that situations [08:26] mzanetti, it's the corner cases indeed [08:26] mzanetti, when something changes currentDate or maximum/minimum [08:27] Saviq: https://code.launchpad.net/~mzanetti/pbuilderjenkins/add-missing-hooks/+merge/173668 === 36DAAZP3C is now known as tvoss [08:28] Saviq: I'm not entirely sure if fginther will be happy about this... but the current state doesn't work either... So I'd say we go for it and if fginther doesn't like it, I'll help him finding a proper solution [08:28] mzanetti, yeah, we need to unblock [08:28] (I think this is the proper solution though) [08:29] mzanetti, approved [08:29] Saviq: ok... once this landed, I need to update the pbuilderjenkins package in the mediumtests vm's. Then it should work again [08:30] mzanetti, k, thanks [08:45] Saviq: as usual, I forgot to release. Could you approve this one too: https://code.launchpad.net/~mzanetti/pbuilderjenkins/relase-new-hooks/+merge/173670 [08:46] mzanetti, UNRELEASED ok [08:46] ? [08:46] huh? didn't I release [08:46] one sec [08:47] Saviq: now [08:47] mzanetti, saucy ok? [08:47] mzanetti, the previous ones are raring [08:47] Saviq: yeah... there hasn't been a release in a while... [08:47] k [08:51] Cimi: this seems broken: [08:51] + while (i < calendarModel.count && calendarModel.get(i).monthStart < minimumDate) { [08:51] + calendarModel.remove(0) [08:51] + i++ [08:51] + } [08:51] Cimi: you do a get(i++) and a remove(0). [08:51] Cimi: seems like you're skipping every second one [08:52] mzanetti, indeed [08:52] mmm [09:07] tsdgeos: I can reproduce it now [09:07] mzanetti: shoot [09:07] tsdgeos: go to the music lens [09:07] tsdgeos: drag it a little bit up, so that most of the header is hidden [09:08] tsdgeos: but not all of it. the thick sepearator should still be visible [09:08] mzanetti: on the phone, right? [09:08] tsdgeos: yes [09:08] tsdgeos: then drag it to the right (moving content to the right, which would get you to the lens on the left, if there would be one) [09:08] tsdgeos: the header will come back [09:08] tsdgeos: release, and drag again [09:09] tsdgeos: the header will come down a bit more [09:09] tsdgeos: and with every drag it goes further down. once it reaches the bottom of the screen the whole lens is messed up and ever interaction breaks it more [09:09] blame qreal :D [09:09] haha [09:09] * mzanetti blames qreal [09:10] i'll have a look [09:10] i had this before [09:10] thought it was fixed [09:10] tsdgeos: its an image from yesterday morning [09:12] *boom* [09:14] yowza [09:14] tsdgeos, yeah, my header just went far down, it's just GONE! ;) [09:15] tsdgeos, and the content went up, it's GONE, too :D [09:16] Saviq: yep. thats the one :D [09:16] Saviq: reproducable with the left/rightmost lens [09:17] mzanetti, made it with Home, too [09:17] Saviq: I've seen it once there. but way harder to reproduce there [09:17] Saviq: also, slide the header a little out the screen and tap on search. sometimes it'll come in without moving/clipping the listview [09:18] it == the search field [09:18] yeah can repro the thing [09:18] :/ [09:18] will work on it [09:19] tsdgeos: well, if you can reproduce it, isn't that half of the win? [09:19] does pphablet-flash give me a flipped image already? [09:19] mzanetti, yeah, it's the same code - showHeader() [09:19] mzanetti: :D [09:19] tsdgeos, yes [09:19] tsdgeos: imagine you couldn't reproduce it and it happens randomly... that would suck [09:19] Saviq: do we need any adjustment on run_on_Device or just works? [09:19] Saviq, the osk doesn't seem to work with latest image, known? [09:19] mzanetti: well one could say it happens randomly since on the PC it works :D [09:20] :D [09:20] Saviq, well, in the scope views [09:20] mhr3_, yeah, OSK is broken [09:21] tsdgeos: run_on_device is already adusted for flipped images [09:21] Saviq, any workaround? [09:21] mhr3_, 'fraid not [09:22] mzanetti: greatz [09:22] * tsdgeos goes to pay some taxes while the newer image is being downloaded [09:22] mhr3_, bug #1199150 [09:22] bug 1199150 in touch-preview-images "search lens disappeared when hitting any key on OSK" [High,Confirmed] https://launchpad.net/bugs/1199150 === larsu is now known as Guest60692 [09:26] thx [09:32] Saviq, for the container of the expandable, should the expanding call set a property on the container which is a child? This is specifically for the animations and all. [09:32] expanding signal sorry [09:32] Am I making sense? [09:33] nic-doffay, you lost me at "container which is a child" :D [09:33] Let me pastebin a quick example. [09:34] Saviq, here's an example minus everything else of what I meant: http://pastebin.ubuntu.com/5857906/ [09:34] nic-doffay, sure [09:34] nic-doffay, only you don't need (Item expandable) [09:35] nic-doffay, you can use "sender" [09:35] Saviq, cool. [09:35] I _think_? [09:35] I also obviously have to set the Container in the expandable. [09:35] mzanetti, with get(0) should work http://paste.ubuntu.com/5857909/ [09:36] nic-doffay, yeah, you need to pass it, I thought sender is available, but apparently it isn't [09:36] Saviq, so basically the OptionSelector now works whether there's a scrollable container as a parent or not. [09:36] Since this Container is another optional component. [09:37] nic-doffay, yeah, and ideally it still should work without the Container [09:37] Saviq, that's what it does atm. [09:38] Saviq, however what class should the expandable be passed into that signal as? [09:38] Item is a bit vague. [09:38] nic-doffay, ListItem.Expandable, probably, when you introduce it :) [09:39] nic-doffay, don't optimize prematurely [09:39] Saviq, what base feature would ListItem.Expandable have that the OptionSelector doesn't though? [09:39] I'm assuming you mean for the ListItem.Expandable to be a base class for all of these? [09:39] nic-doffay, it can be inherited by other components [09:40] nic-doffay, hmm [09:40] Saviq, just the expanding logic? [09:41] nic-doffay, it should have `real collapsedHeight`, `real expandedHeight`, `bool expanded` at least [09:42] nic-doffay, just think which would be common between the OptionSelector and the expanding Message items you did before [09:43] nic-doffay, it's not like I can tell you from the top of my head what would that be :) [09:49] paulliu, ping [09:54] Cimi: haven't tested it, but looks better, yes === Guest60692 is now known as larsu [10:07] Saviq, what's the best way to get more output from a qmlscene? [10:07] Something (nothing) unexplained is happening. [10:07] nic-doffay, onPropertyChanged: console.log("something") usually [10:07] nic-doffay, you can also run the debugger in QtCreator [10:09] Saviq, I mean the qmlscene isn't even coming up for weird reasons. [10:09] Need to find out why. [10:09] nic-doffay, does your top-level object have dimensions? [10:09] Saviq, yeah. [10:10] Saviq, basically what I did was change that Item in the pastebin I sent you to an OptionSelector, now nothing runs. [10:10] Obviously an item won't have an expanded property. [10:11] nic-doffay, sorry, quite busy now, we've failed miserably at a review yesterday [10:11] Saviq, how so? [10:11] nic-doffay, not you, /me and mzanetti [10:12] but mostly me [10:12] :( [10:12] huh? [10:12] the launcher-backend ones? === dandrader is now known as dandrader|afk [10:18] mzanetti, no, the dee filtermodel [10:18] mzanetti, it broke _everything_ in the dash [10:19] on the device, that is [10:19] Saviq: hmm... weird. I ran it on the device tho... seemed to work [10:25] mzanetti, it "works", but there's no data in Home other than apps, and no Installed in Apps [10:25] mzanetti, not to mention icons and titles and such are broken everywhere [10:26] mzanetti, have a moment to check something out? [10:26] nic-doffay: hit me [10:27] Weird problem where I'm passing something to a signal but as soon as I change the class from Item to the class nothing works. [10:28] mzanetti, lp:~nicolas-doffay/+junk/option-selector [10:29] mzanetti, if you take a look at ScrollableExpansionContainer.qml in modules/Ubuntu/Components/ListItems [10:29] You'll see an OptionSelector is passed from a signal. [10:30] When this is an Item everything works properly. [10:30] However obviously the class needs to be passed in to properly access the variable. [10:30] Running test.qml in modules/Ubuntu/Components/ListItems you'll see what I mean. [10:30] No output, nothing. [10:33] nic-doffay: there is no ScrollableExpansionContainer [10:34] mzanetti, whoops one sec. [10:38] mzanetti, pushed again. [10:40] Saviq: hi [10:40] paulliu, hey, thanks for spotting the video breakage... there's actually more of that elsewhere... am fixing now [10:40] paulliu, it was an epic review fail on my part [10:42] Saviq: ok.. I just fix the video lens to test my branch. :) [10:42] nic-doffay: this looks really scary [10:43] nic-doffay: what do you want to achieve with this? [10:43] nic-doffay: and why is this a signal? [10:43] to me this doesn't make any sense at all [10:49] mzanetti, this container will handle scrolling for expandable SDK components. [10:49] The signal is sent from the expandable component. [10:49] paulliu, can you please include this diff in your MP http://pastebin.ubuntu.com/5858043/ [10:49] nic-doffay: yeah... signals are supposed to be used when a component wants to emit something to the outside. not getting something from the outside [10:50] nic-doffay: also, why don't you just use a flickable then where you embed the other component that needs to be scrolled? [10:51] nic-doffay: do you have a spec how things should look like in the end? I'm having a hard time following what you want to do with this [10:54] Saviq: ok === hikiko is now known as hikiko|lunch [10:58] mzanetti, there aren't any specs I'm aware of. [10:58] mzanetti, let me forward you something though... [10:58] mzanetti, you have mail. [10:59] whattt [10:59] now i can repro the header problem on the pc too [10:59] :S [11:04] btw do you guys get the images of "Music" in either pc or desktop? [11:04] i'm just getting empty ubuntu shapes [11:04] Saviq: mzanetti: Cimi: greyback: MacSlow: dednick: all :D↑↑↑ [11:04] tsdgeos, yes, we broke it yesterday [11:04] tsdgeos, fixing now [11:04] tsdgeos, ? [11:04] ah [11:04] ok [11:04] tsdgeos, not seen that issue yet === mardy_ is now known as mardy [11:05] mzanetti, Cimi, greyback, MacSlow, dednick, all: unping [11:05] Saviq, :) [11:05] :D [11:05] "unping" [11:05] * MacSlow notes that term in his memory [11:05] E_POP_EMPTY_STACK [11:09] nic-doffay: still don't get it... what is the use case for this? if I have an expanding component, why wouldn't it be expanded to what it needs to be? [11:09] mzanetti, the idea is that this container can be used for any expanding component instead of coding it separately for each. [11:10] nic-doffay: ah. so the expanding component itself would always be expanded, but just clipped by the container? [11:11] mzanetti, no. [11:12] mzanetti, the expanding item would call the container which would in turn set the expansion on the expanding item and handle the animation and movement [11:12] does that make any sense? [11:13] nic-doffay: not for me... but according to this mail it does make sense for Saviq. So I must be missing something here. [11:13] mzanetti, lol [11:14] :) [11:14] what should I say [11:14] thats how it is :D [11:14] mzanetti, haha [11:14] nic-doffay, the height animation could happen in the Expandable base component itself [11:14] mzanetti, do you know what the whole deal is about? [11:15] Saviq: not really, no [11:15] mzanetti, an item in a list [11:15] mzanetti, the item is expandable, i.e. it can change its height [11:15] yeah. so far I'm good [11:15] mzanetti, but when it expands, it should be limited to the list's height [11:15] mzanetti, and it should be scrolled into view [11:15] what's the "list" ? === MacSlow is now known as MacSlow|lunch [11:16] mzanetti, anything, a ListView, say [11:16] oh... I see... [11:16] mzanetti, and an item in that ListView needs to be able to expand [11:16] the scrolling thing is interesting. [11:16] mzanetti, and when it does, it needs to be constrained to the ListView's height [11:16] mzanetti, and then scroll it into view [11:16] mzanetti, yup [11:16] right... [11:17] mzanetti, when the item is smaller than the ListView [11:17] mzanetti, it should be minimal travel [11:17] to show that whole item [11:17] ok... I start to understand... [11:17] thanks Saviq [11:17] mzanetti, so the ListView needs to know the item's expanded and collapsed heights [11:17] mzanetti, and change them accordingly [11:18] mzanetti, I might've overcomplicated things in the email, too :) [11:19] why would the listview need to know both sizes? [11:19] ah no... the listview is the one to scroll [11:19] ok... [11:21] nic-doffay: ok... so the naming ScrollableExpansionContainer is the first misleading thing. I thought the content in there needs to scroll, not the listview around it [11:22] mzanetti, you're right the content within does. [11:22] mzanetti, maybe ScrollableContainer is more reasonable? [11:22] mzanetti, how are we re: jenkins? [11:22] * mzanetti checks (should work by now) [11:23] Saviq: I ran a mediumtests job manually and it worked [11:23] mzanetti, k thanks === hikiko|lunch is now known as hikiko| [11:24] Saviq: but wouldn't we need to have our own ListView implementation to support this kind of stuff? [11:26] mzanetti, no, why? [11:27] mzanetti, we just need to have a channel of communication between the Flickable / Scrollable and the Expandables [11:27] but where is the code that scrolls the flickable? [11:28] Saviq: the expandable itself? [11:28] mzanetti, no, in the Flickable [11:29] mzanetti, it knows which Expandable it needs to scroll to [11:29] mzanetti, because the Expandable will signal Scrollable.RequestExpansion(self) (pseudo code) [11:30] mzanetti, that's what it currently does. === dandrader|afk is now known as dandrader [11:30] nic-doffay: ah... so we're not talking about a flickable that uses those components... but another flickable inside the list [11:31] mzanetti, then, the Scrollable will map to its own coordinates and scroll accordingly, animating the height in parallel (that can happen either in Scrollable directly or the Expandable could have a Behavior, but that could be wasteful when you only ever animate one Expandable at a time) [11:31] mzanetti, hum? [11:32] where would you place such a component? do you have an example? [11:32] I really dont see the usefulness of such a thing (except maybe the reviews in scopes) [11:32] previews [11:32] mzanetti, this is the thing. [11:33] Everyone is debating this. [11:33] Design doesn't even have a proper mock up yet. [11:35] nic-doffay: anyways... ignoring that... let me write some pseudocode how I would try to approach the problem [11:36] (besides asking design for clear guidance) [11:36] mzanetti, done that bit already. [11:36] Any idea about that original issue out of interest? [11:44] mzanetti, mhr3_, https://code.launchpad.net/~saviq/unity8/fix-mock-scopes/+merge/173698 [11:44] nic-doffay: http://paste.ubuntu.com/5858161/ [11:44] nic-doffay: hope that makes somehow sense [11:44] Saviq: on it [11:44] mzanetti, re ↑↑↑ http://design.ubuntu.com/apps/building-blocks/option-selector [11:45] http://design.ubuntu.com/apps/building-blocks/expansion [11:45] Saviq: for that I don't see why we would need such a construct [11:45] Saviq: instead of just placing a normal expanding component into a ListView [11:46] Saviq: and perhaps making the ListView "interactive: contentHeight > height" [11:46] mzanetti, it's not always a ListView, mind you, it can be as simple as a Column [11:46] mzanetti, a Column in a Flickable, that is [11:46] mzanetti, it's actually going to be the more common use, I'd say [11:46] Saviq: s/ListView/Flickable/ [11:47] Saviq, why is the Qt.binding necessary? [11:47] mhr3_, otherwise it only sets the value once [11:47] mhr3_, whereas with Qt.binding it will update when changed [11:48] mhr3_, so when you send dataChanged(i, i, RoleResults), it will update automagically [11:48] clever [11:48] mhr3_, we should be able to test by killing the backends, right? [11:48] mhr3_, as the Scope object itself won't change, nor the Categories will [11:49] mzanetti, yeah that makes sense. Cleaner imo [11:49] ok, looks fine to me, but perhaps it'd be better to add also the s/column_/.../ changes? [11:49] Saviq, yep [11:49] mhr3_, prereq [11:51] Saviq, i'd suggest merging it in, one won't work without the other, will it? [11:52] mhr3_, works fine, it fixes two different issues [11:52] mhr3_, with the prerequisite apps and videos will get fixed [11:52] mhr3_, with my MP only the "peeking" into other scopes is being fixed [11:54] ah, ok, i thought that the raw model is still used by some of the renderers [11:55] mhr3_, we need to re-search after the scope is reconnected after it's killed, right? [11:55] Saviq, yea [11:55] mhr3_, 'cause now my model is updated, but the results are not [11:55] mhr3_, k, separate MP then === greyback is now known as greyback|food === alan_g is now known as alan_g|lunch === dandrader is now known as dandrader|lunch [12:06] mhr3_, "m_unityScope->Search(m_searchQuery.toStdString());" is missing the callback, right? [12:08] Saviq, i'm just looking at that, and that should make sure that a re-search happens [12:08] mhr3_, yeah, but there's no callback? [12:08] not sure why it doesn't [12:08] Saviq, right, but no biggie [12:09] mhr3_, should work anyway? [12:09] yep [12:09] there'll be one missed searchFinished signal, don't think shell uses it anyway yet [12:09] right [12:10] Saviq: oh noes! the hooks are there now, but one of them fails to execute [12:10] mzanetti, ;( [12:10] Saviq: I fear we need fginther for this [12:10] only he know what this stuff is supposed to do [12:11] it's a fail day if I know one [12:14] mzanetti, nic-doffay now that I think of it again... I'm not sure the "constrain height to Scrollable's height" I described should actually happen [12:15] nic-doffay, can you please confirm with ckpringle what should happen with "overflow" [12:15] nic-doffay, because if we enable scrolling *inside* the Expandable in case of overflow, it kind of conflicts with the fact that you should be able to scroll to the other items [12:15] nic-doffay, and the "expand in place" pattern [12:18] Saviq: I might have found the issue on jenkins... next test running... [12:18] mzanetti, thanks [12:19] Saviq, I'm busy waiting on confirmation about the actual expandable item too. [12:19] nic-doffay, get busy prodding, then! :D [12:19] Saviq, that's what I've been doing. [12:19] nic-doffay, k :) [12:22] Saviq: yeah! thats exactly what I meant [12:23] mzanetti, you mean flicking within another Flickable? [12:23] I mean... there might be use cases for such a thing. as I said, the DashPreview would be one. [12:23] but really not for such components as this is intended for [12:24] only if the content will always expand to fullscreen === MacSlow|lunch is now known as MacSlow [12:31] mhr3_: ping! [12:31] sil2100, pong [12:31] mhr3_: who's the guy I should pester related to Unity dash Previews? [12:31] mhr3_: you guys or dednick? (or someone else even?) [12:32] sil2100, what is the pestering about? [12:32] mhr3_: failing AP tests of course! [12:32] What else should I pester about? :) [12:34] sil2100, the same thing as on friday? [12:34] it wasn't really failing preview but failing dbus [12:34] probably cause of udev [12:35] Not sure if the same, since I was out on Friday, but there's a lot of preview tests failing [12:36] I see most failures look hm, sane: MismatchError: After 10.0 seconds test on DashView.preview_displaying failed: True != dbus.Boolean(False, variant_level=1) === alan_g|lunch is now known as alan_g === _salem is now known as salem_ === dandrader|lunch is now known as dandrader [13:04] mzanetti, do you know what this error means? https://jenkins.qa.ubuntu.com/job/generic-mediumtests-builder-saucy/872/console [13:04] random jenkins failure or do we have to do something? [13:04] dandrader: yes. I'm working on it. I just had a successful build again [13:04] dandrader: with a bit of luck you should be able to autoland again in a few minutes [13:19] Saviq: hey, nic-doffay and i were just chatting....wondering if we might want to pick [13:19] one of the scopes to actually tie up to the [13:19] option expander as an example to try to drive some feedback [13:20] i thought you might have one in mind ? [13:20] kgunn, there's no backends hooked up yet [13:20] kgunn, nic-doffay so just work with any of them [13:20] off of dummy data [13:21] mzanetti, ↑ see who joined? ; [13:21] ;) [13:21] fginther: \o/ [13:21] Saviq: but I think I just solved it :D [13:21] nic-doffay: can you just go for it for say "music" [13:21] mzanetti, fginther probably saw that you did :) [13:21] fginther, that's why he joined ;) [13:22] uh oh [13:22]