=== shuduo is now known as shuduo_afk [07:59] Saviq, fixed issue [08:01] Cimi, bug #1355046 btw [08:01] bug 1355046 in ubuntu-ui-toolkit (Ubuntu) "Colourized icon with invalid source is black" [Undecided,New] https://launchpad.net/bugs/1355046 [08:01] Saviq, cool [08:01] Saviq, for now I rely on source != "" [08:01] Cimi, yup, sounds good [08:03] Saviq, I think we want to branch and fix overview [08:04] Saviq, or maybe merge and branch later on [08:04] Cimi, please comment on the MP with any issues you have [08:04] Saviq, most of them are UX issues [08:04] Saviq, only UI issue is the white indicator on black frame [08:05] Cimi, white indicator? [08:05] Cimi, ah activity [08:05] Saviq, your dash activity indicator has a white box AROUND [08:06] Saviq, let's call it processing indicator [08:06] or activity bar [08:06] Cimi, yeah, this is a slightly weird issue, 'cause I'd have to pass opacity back to all the respective views so that they draw the background on which I overlay the orange bar... [08:07] Cimi, not a problem of the overlay branch in any case [08:07] Cimi, is a problem in trunk [08:08] but we just don't have any other scope that exposes it [08:11] Cimi, you're done with settings are you? [08:18] Saviq, nope [08:18] ah ok [08:18] Saviq, doing tests today then is done [08:25] Saviq, when we flick the dash and change scope, which is the component in charge of that? [08:25] Cimi, DashContent [08:25] dashcontentlist? [08:25] ok [08:26] Saviq, I have to stop it to accept flicks when the previews are on screen [08:27] Cimi, previews have their own horizontal flickable, why would those propagate up to dashcontent? [08:27] (since they have not until now) [08:28] Saviq, settings don't :) [08:28] Saviq, a quick interactive false [08:28] Cimi, so settings, not previews :P [08:28] Cimi, you said that you have to disable it if previews are up [08:28] pff :P [08:28] monday morning :P [08:29] Cimi, if you'd just put it all in a page stack already you'd be done with it by now ;) [08:30] Saviq, there's all the work for the header to be done, isn't it? [08:30] Cimi, I don't think we'd do anything special with the header, it'd be part of a page in our case [08:31] anyway currently my dash doesn't run === vrruiz_ is now known as rvr [08:37] Saviq, does dash run for you? [08:38] today? [08:38] only thing I did was rebooting from friday I think [08:38] forget [08:39] done [09:09] Saviq, how do I mock something like data.property = value in js? [09:09] Saviq, i remember how do do it for arrays, but not .property [09:10] (disctionary I mean) [09:10] Cimi, data = { "property": "value" } you mean? [09:10] Saviq, data.property will work? [09:11] Cimi, actually scrap the first quotes, but yes [09:11] Cimi, FWIW both data.property and data["property"] willl [09:11] JS is weird like that [09:11] Saviq, I thought that was data["property"] [09:11] ah ok [09:11] that was my concern [09:11] Cimi, { } is an object [09:11] Cimi, so it has properties [09:12] (something completely different, that just seems to fit here: https://www.destroyallsoftware.com/talks/wat ) [09:19] Saviq, settings should behave like listItems? [09:19] Saviq, touching everywhere to have focus? [09:20] or action [09:20] mmm maybe not [09:20] because they can be very mixed [09:59] Cimi, small one for you https://code.launchpad.net/~saviq/unity8/fix-preview-anchor/+merge/230268 [10:02] Saviq, the label setting [10:03] Saviq, horizontal space is limited [10:03] Saviq, I don't see how I can have both the description and the entry [10:03] Cimi, ENOCONTEXT [10:03] HAH [10:03] Saviq, scope settings, what else am I doing now? :) [10:03] Saviq, we have no design for them right? [10:04] Cimi, no [10:04] so the description label is usually is [10:04] Cimi, so just add one on top of another? [10:04] instead of next to each other? [10:04] "display grooveshark result in dash" or something long like that [10:05] it is already ellipsised (cannot write that word) [10:05] with the switch [10:05] (that was for boolean) [10:05] then we have settings with a text field [10:05] they might have a long label with the description [10:05] then a text field [10:06] I was wondering if we want to consider using two rows for settings [10:06] description on top, action/control on bottom [10:07] Saviq, ~cimi/unity8/scope-settings [10:08] Saviq, you can see boolean settings (real) in Music and Video scopes [10:08] Saviq, then test the switch with make try(test)ScopeSettingSwitch [10:08] Saviq, label with make try(test)ScopeSettingLabel [10:09] Cimi, yeah, just put it on top [10:09] Cimi, also, calling it 'label' feels wrong [10:10] Cimi, shouldn't it be ScopeSettingText? [10:10] Saviq, ok [10:10] Cimi, label implies static text [10:10] ah good pppoint [10:14] looks really ugly in two lines [10:14] better to use checkbox [10:40] greyback, if you didn't know, you can uncheck "Needs review" in MP submission page to make it WiP to start with [10:40] Saviq: yep, I just forgot [10:40] greyback, sure [10:40] sorry for the noise === MacSlow is now known as MacSlow|lunch [11:14] freaking ctrl q [11:15] so close to ctrl tab [11:24] cwayne, you'll be happy with this: https://code.launchpad.net/~saviq/unity8/dash-horizontal-list/+merge/230279 [11:24] Saviq: :D thank you! [11:31] Holas! [11:32] elo [11:38] /food [12:02] so if I add text in a textfield with insert(0, "new text) [12:02] the text property does not seem to update [12:13] Saviq, hey, the flag we talked about for disabling rotation sensors, any docs on that yet? :) === _salem is now known as salem_ [12:42] Saviq, do we want to darken and make setting touchable? [12:42] scope settings [12:43] because putting textfield in a listitem might be weird - if I subclass listitem.empty I get the triggered signal when I press the content, but this might not be ready [12:44] or not valid [12:44] when I press the listitem I mean [12:44] so I can create a new signal and such... [12:44] Cimi, it probably depends, for a checkbox I'd actually trigger the checkbox as you press on the item [12:44] and makes sense [12:44] and that's why I used it [12:45] Cimi, Empty has a property to disable the highlight [12:45] but now I am confused on the string and number setting [12:45] what shall I do when user taps the listitem? [12:45] Cimi, you could focus the text entry [12:46] Cimi, for numbers obviously nothing [12:46] highlightWhenPressed seems visual only [12:46] Saviq, triggered is always emitted though [12:46] Cimi, yeah, but what would ever listen to that? [12:46] Saviq, so when I tap the string listitem it will emit trigger [12:46] Cimi, so? [12:46] I was thinking of listening to trigger in the scopesettings [12:47] I will have to create another signal [12:47] settingUpdated [12:47] that's fine I just wanted to keep code simple [12:47] Cimi, is that for the model.value = foo part? [12:47] Cimi, triggered wasn't a good name for that anyway [12:47] Saviq, it is the signal I want to emit from each single setting component [12:48] when the editing was finished [12:48] ok [12:48] settingUpdated maybe [12:48] Cimi, yeah, and "triggered" doesn't sound like a good name for it at all [12:48] Cimi, no need for the "setting" part there [12:48] Saviq, is the same name we use in PreviewWidget (triggered) [12:49] Cimi, yeah, but those are, actually, triggered [12:49] Cimi, actions are triggered on those, there's no value associated with the triggering [12:59] Saviq, how do I focus the textfield to a specific char? [12:59] cannot find in APIS [13:00] Cimi, I don't think you do, why? [13:01] Saviq, we said would be good to focus the textfield when clicking the listitem? [13:01] Cimi, yeah, but why on a particular char? [13:01] Saviq, last one [13:02] Cimi, I think we should go for http://qt-project.org/doc/qt-5/qml-qtquick-controls-textfield.html#selectAll-method [13:02] ok [13:02] Saviq, I'm using ubuntu components textfield btw [13:03] (which has same method) [13:03] Cimi, which is just a wrapper around the upstream one [13:03] Saviq, nope [13:03] Saviq, it doesn't have editingFinished() [13:04] Cimi, right, it's not, another waste of API :| [13:05] * Saviq really doesn't get why we're replacing all the upstream components ;( [13:05] :/ [13:05] editingFinished is also quite handy [13:06] otherwise you have to do onActiveFocusChanged: if (!activeFocus && acceptableInput) [13:07] and listen to other keys [13:07] Cimi, onAccepted doesn't help does it? [13:08] Saviq, is only on enter [13:08] return key [13:08] so I have both [13:08] Cimi, yeah thought so [13:09] Cimi, sounds like a bug against UITK is in order [13:15] karni, I believe you can mark https://bugs.launchpad.net/unity8/+bug/1352940 fixed for Savilerow? [13:15] Ubuntu bug 1352940 in Unity 8 "Thumbnails nolonger appear in carousel for videos scope" [High,Triaged] [13:16] karni, the fix was in UITK after all [13:16] Saviq: done [13:16] karni, tx [13:31] Saviq, any way to fake pressing return? [13:31] Cimi, why? [13:31] Saviq, testing that I emit the signal when I press enter [13:32] Cimi, just go .accepted() [13:32] Saviq, is not exposed [13:32] Saviq, and I cannot test it... [13:32] Cimi, wdym it's not exposed :D [13:32] I have the widget and I want to fake an user interaction [13:32] I want to fake the user pressing enter on the keyboard [13:32] Cimi, but this way you'd start testing the TextField widget [13:33] Cimi, that it does indeed emit accepted when enter is pressed [13:33] ah ok [13:33] so I do textField.accepted() ? [13:33] Cimi, which is the responsibility of the UITK test suite [13:33] Cimi, yes [13:44] dednick_, hey, any ETA on dash-overview review? [13:53] mzanetti, ping [13:54] dandrader: hi [13:55] mzanetti, hi! when would you have time to review https://code.launchpad.net/~dandrader/unity8/lifecycle/+merge/230090 ? [13:55] just rebased it on top of latest trunk [13:56] right... hmm... dunno. will try to get to it asap === MacSlow|lunch is now known as MacSlow [14:16] Saviq, if I use Clipboard in the test, the test segfaults at the end [14:16] Saviq, uitk bug? [14:17] Cimi, you gave rather little info to determine that [14:17] Saviq, to check I am selecting text on clicking the listitem, I copy the content of the selection and check if Clipboard.data.text is equal to what I expect [14:18] Saviq, this pass, but after all tests are running, I see qmltestrunner segmentation fault [14:18] Cimi, more testing of TextField, eh? [14:18] Cimi, just check if it's focused ;) [14:18] Saviq, yes, did that, then wanted to test if we were selecting all the text [14:19] Cimi, which is testing TextField [14:19] I can avoid using clipboard by calling cut and checking displayed text [14:19] or just testing that it's focused and relying on UITK to have tested their own component's selectAll === alan_g is now known as alan_g|tea [14:32] Saviq: i'll get back on it shortly. i've just been re-installing ubuntu on new ssd. [14:39] dednick_, tx === alan_g|tea is now known as alan_g [14:42] MacSlow, https://www.youtube.com/watch?v=-HoHkFgslJc [14:44] MacSlow: recommendation to get started: https://developer.oculusvr.com/forums/viewtopic.php?t=1539 [14:44] MacSlow: the planetarium is also amazing [14:45] Saviq, mzanetti: I assume I won't get much sleep starting tomorrow :) [14:45] heh [14:45] MacSlow: well, first time I tried I got sick after 5 mins [14:45] MacSlow: takes a bit to get used to... [14:46] mzanetti, I'm used to some really bad VR back from my university-days... [14:46] MacSlow: yeah... but back then the brain was able to figure that its not real [14:46] that changed a bit [14:46] now it seems real without gravity lining up with what you see [14:48] mzanetti, I'm willing to play guinea pig for you all :) [14:48] MacSlow: as I said... I tried them in June last year... [14:49] MacSlow: was sharing a room with the guy that developed that "alone in the rift" demo I posted before [14:49] mzanetti, I assume that was still the DK1 version?! [14:49] yeah... it was the low-res version still [14:50] but already impressive enough [14:50] Saviq, do you know if anyone has cycles to review https://code.launchpad.net/~mterry/unity8/wrong-password-handling/+merge/229184 ? That has the delay-on-failed-login feature that security requested and kgunn wanted to get in by Thursday [14:50] mzanetti, they've done much work to improve upon the usual motion-sickness factors (higher res, less latency, motion "prediction"...) [14:51] MacSlow: they still can't simulate gravity, which is the main reason [14:51] MacSlow: well, they can, but I doubt you've got that piece of hardware in your mailbox :) [14:51] mzanetti, true... [14:51] if you do, I'm on my way [14:52] :) [14:52] mzanetti, hehe [14:53] Saviq: is the dash overview handle always supposed to show? (the little tag you are supposed to pull at the bottom) [14:53] dednick_, no, and it doesn't [14:53] Saviq: it's always there for me. [14:53] dednick_, it only shows on the dash, and only if the header is all on screen [14:53] oh right. yes, thats what i mean [14:53] dednick_, if you scroll down, it'll go away [14:53] dednick_, then yes [14:54] dednick_, https://sites.google.com/a/canonical.com/unity8dash/scopes/dash-overview [14:54] dednick_, it actually even says "The hint is always visible - never hidden or smaller (unlike hints used in apps)" [14:54] dednick_, but we alleviated that slightly [14:54] (with design ACK) [14:59] Saviq: any idea why the CI failed for overview? can't see any qmltest failures. [14:59] dednick_, looking [15:04] dednick_, yes, that's what I mentioned [15:05] dednick_, it fails because gcovr hangs for 2h in qmluitests [15:05] Saviq: ok. cool [15:05] dednick_, but if the test report says no failures, then it's passed [15:05] dednick_, there are two failures there though https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-utopic/845/testReport/? [15:06] Saviq: ah. i wasn't looking at the full log [15:07] or the report [15:07] dednick_, yeah, it doesn't show them if the job failed [15:07] dednick_, testDashContent segfaulted there [15:08] dednick_, and looking at https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-utopic/845/testReport/%28root%29/qmltestrunner/Dash__test_dash_overview_show_select_different_favorite/ it's a timing issue (didn't wait for something) [15:08] Saviq: hello hello; i have a hung unity/shell (password screen for unlocking does not accept input etc.); anything i shall extract? [15:08] (before rebooting) [15:08] i can adb shell into device luckily [15:08] asac, power button shows the greeter? [15:09] Saviq: yeah, that thing i can swipe away [15:09] but then the pin screen doesnt take any input [15:09] asac, ok, ~known issue, touch vs. mouse events (don't ask, Qt limitation, we need to dig into it) [15:09] indicators still work etc. [15:09] ok so you have all info? [15:10] or er s/limitation/weird choice/ [15:10] the pin screen is greyed out though [15:10] asac, only info we'd be interested in is steps to reproduce... [15:10] i had similar feeling hangs regularly [15:10] ok, guess its something racy, so not easy to repro [15:10] Saviq: no gdb dump or anything that could help? [15:11] e.g. force a coredump and submit that? [15:11] asac, not really, as the whole thing works [15:11] asac, so it doesn't hang anywhere [15:11] asac, it's just part of the even logic gets confused [15:11] asac, so we need to get it under gdb in that state and step it [15:11] *event [15:12] asac, so steps to reproduce are what we're most interested in [15:12] Saviq: remote access could help? [15:12] e.g. ssh into my machine? === alan_g is now known as alan_g|afk [15:13] anyway rebooted now... see that often enough so we can still do that [15:13] asac, it's happening to us somewhat frequently, too, so once we focus on it we should be able to find out what's going on [15:13] kk [15:13] thanks [15:14] asac, bug #1295623 btw [15:14] bug 1295623 in unity8 (Ubuntu) "Sometimes input breaks and only edges are responsive" [High,Triaged] https://launchpad.net/bugs/1295623 [15:14] Saviq: I don't see it often any more... its still there, but I get it like once in 2 weeks nowadays [15:15] thanks! [15:15] asac, if you find anything that suggests triggers it, please write down there [15:16] is it normal getting segfaults after some qmltests have run? [15:16] i.e. the tests pass, but then maybe xvfb crashes?? [15:16] no [15:17] ok, I musta broke something === dandrader is now known as dandrader|lunch [15:17] greyback, DashContent is the only one crashing sometimes that I know of [15:18] dednick_, in any case, if you can verify those tests passing locally, I wouldn't block on those failures (unless you can reproduce the overview related one) [15:18] Saviq: i'm running it now [15:21] * Saviq loves qmltests with xvfb, can run them on all my 8 cores ;) [15:21] Saviq: what's an xfail? [15:21] dednick_, expected fail, meaning this is a negative tests [15:21] *test [15:21] ah [15:22] dednick_, hence the PASS after that === alan_g|afk is now known as alan_g [15:23] dednick_, the issue you reported, sounds like not a dash overview one, you can get there by tapping on the search entry and dismissing the keyboard by swiping it down [15:23] Saviq: ok, passed for me. i approved branch [15:25] Saviq: hm. i think the switch to overview needs to cancel input focus [15:26] Saviq: you can also reproduce putting keyboard into layout mode (orientation). you can then use the bottom (now left) edge to swipe in overview. [15:27] not exactly intended, but the popup still remaing [15:27] dednick_, you mean landscape mode? ;) [15:27] :) yes [15:27] bit edge case though! [15:27] dednick_, hmm interesting, I can't get the OSK to rotate when in dash [15:28] dednick_, yeah, and for my usecase the field should unfocus on keyboard dismissal [15:28] Saviq: rotate, then click search [15:28] dednick_, yeah so... we can't really unfocus on bottom swipe as we're getting a snapshot just then [15:29] Saviq: popup not included in snapshot it would seem [15:29] dednick_, yeah, popups are parented to root object unfortunately [15:29] dednick_, bug in UITK IMO [15:29] or well, something [15:30] but yeah, there's a bit of animation and shit happening when you unfocus the search. [15:31] dednick_, yeah, that would get snapshotted in the middle [15:31] Saviq: why do we snapshot? shouldn't it just be same as app switch? considering dash is an app? [15:31] dednick_, overview is the same app [15:31] Saviq: ah [15:31] right [15:31] dednick_, we *could* just scale the actual object down [15:31] dednick_, but that wouldn't solve this issue [15:32] dednick_, I think both are edge cases, and not really bugs with the overview branch, so not blockers IMO [15:32] dednick_, keyboard should never rotate when on dash ('cause it doesn't rotate) [15:32] Saviq: yep. that's why i approved. [15:32] dednick_, and field should get unfocused on keyboard dismissal [15:33] dednick_, sure, just justifying myself here ;D [15:33] well, you can always just blame me for approving! [15:33] ;) [15:35] greyback, do we have a bug# for "launch two apps quickly → non-foreground app does not suspend"? [15:36] Saviq: we do, hunting... [15:38] Saviq: just tried unity-scope-tool after updating today, I just get a blank results area now. I think this is because we are now dependent on the "scopes" (overview) scope being loaded [15:38] is that something you're aware of [15:38] pete-woods, overview not landed yet [15:39] pete-woods, and scope tool works fine here? [15:39] * Saviq checks again [15:39] Saviq: it only breaks for an individual scope [15:39] i.e. if you do scope-tool PATH_TO_INI [15:39] pete-woods, oh hmm [15:40] when it generates a new scope config, starts a registry, etc [15:40] pete-woods, file:///usr/share/unity8/ScopeTool.qml:139:23: Unable to assign null to QString [15:40] ERROR! Caught unity::scopes::TimeoutException: Request timed out after 5000 milliseconds [15:40] Unable to add overview scope, can't find with ID: "scopes" [15:40] Saviq: yeah, that's the error I get [15:40] I think the config generator in the tool needs to include the scopes scope somehow [15:40] or something (perhaps the shell plugin) needs to stop being totally dependent on that scope being present [15:42] pete-woods, should be somewhat easy http://bazaar.launchpad.net/~unity-team/unity8/trunk/view/head:/tools/scopetool.cpp#L73 [15:42] pete-woods, to just include scopes in the extraScopes list [15:42] Saviq: yeah, that makes sense [15:43] pete-woods, will you make a branch or will I? [15:43] Saviq: I don't mind doing it, but obviously I'll need your review [15:43] pete-woods, that's easier than me doing it and getting yours ;D [15:44] pete-woods, only concern is getting the right path [15:44] :) [15:44] yes [15:44] pete-woods, but perhaps you can get it from the registry somehow? [15:44] pete-woods, or even just get it from .pc build time [15:44] pete-woods, but you'll manage I'm sure [15:44] Saviq: hmm, the difficultly [15:45] Saviq: can't find the damn bug, must be imagining it. Will log one [15:45] is that you want the dir: /usr/lib/ARCH/unity-scopes/ (which will contain the scopes scope) [15:45] Saviq: but that will also pull in all the other scopes :( [15:47] greyback, thanks, I couldn't find it either [15:47] might need to do something fun with symlinks in the temporary directory that the tool creates [15:47] pete-woods, why can't we just pass the .ini for the scopes scope alone? [15:48] pete-woods, I thought we could just push it into the extraScopes set [15:48] Saviq: oh, is that how that works? [15:48] * pete-woods actually reads docs [15:49] I'm (fairly) confident you have to give directory paths to the registry ini [15:49] and those dirs are expected to contain scope directories [15:49] pete-woods, no, you pass .ini files, read --help [15:50] 'passing a path to a scope config file' [15:50] pete-woods, but anyway that doesn't work too well [15:50] pete-woods, I end up with just the scopes scope in the UI [15:50] pete-woods, which we probably don't want to show [15:51] pete-woods, ah no wait, wrong .ini [15:51] Saviq: hmmmm. just tried starting the user session version of the registry [15:51] and it magically owrks [15:51] *works [15:52] Saviq: https://bugs.launchpad.net/unity8/+bug/1355275 [15:52] Ubuntu bug 1355275 in Unity 8 "Launch 2 apps quickly, the non-foreground app evades lifecycle suspend" [Undecided,New] [15:52] unity-scope-tool /usr/lib/x86_64-linux-gnu/unity-scopes/scopes/scopes.ini /usr/lib/x86_64-linux-gnu/unity-scopes/musicaggregator/musicaggregator.ini [15:52] pete-woods, ↑ [15:52] pete-woods, only we might want to do something special to UNITY_SCOPES_LIST [15:53] pete-woods, or maybe that's fine... [15:53] I have difficulties with implicitHeight of expandable listitem http://paste.ubuntu.com/8018151/ [15:53] implicitHeight here doesn't grow when I expand the listitem [15:53] Saviq: I think something as simple as that works well enough for me [15:54] pete-woods, so yeah, we might append the scopes .ini to extraScopes in the tool [15:54] pete-woods, whether we want to tweak UNITY_SCOPES_LIST or not... [15:54] pete-woods, you tell me ;) === dandrader|lunch is now known as dandrader [15:55] if this is only affecting the scope tool, I don't want to play with anything else [15:55] I'm also more relieved now I know that it currently does work if the main registry is running [15:56] Cimi, does ListItem.ItemSelector change its height? [15:56] Cimi, do you have height set explicitly on ScopeSetting maybe? [15:56] Cimi, which of those items is Expandable, for that matter? [15:57] Saviq, height changes if I don't set implicitHeight on the ScopeSetting item [15:58] Cimi, does it really? where does ScopeSetting take its height from? [15:58]