[08:47] ltinkl: do you have a sec? [09:18] tsdgeos, ye sure [09:18] hmmm [09:18] will come to you in a bit [09:18] sorry :D [09:24] ltinkl: you've got mail [09:24] tsdgeos, got it [09:24] ltinkl: please uncompress [09:25] run qmake && make in the ManglingImport folder [09:25] and then from the root run [09:25] qmlscene -I . main.qml [09:26] both with the Image { } from MySingleton commented and not commented [09:26] i think both show bugs :D [09:26] but i'd like you to have a look at all the code and tell me what you think before i convince you [09:27] tsdgeos, convince of what? :) [09:27] that it's a bug in both cases (two different bugs) [09:27] ah [09:27] I get "QML Image: Cannot open ..." in both cases [09:28] that's not the bug [09:28] read the code please [09:28] though you're doing it wrong [09:29] if you comment the Image {} from MySingleton you should not be getting that [09:29] but [09:29] $ qmlscene -I . main.qml [09:29] Ignoring the url you set to QUrl( "file:///home/tsdgeos_work/extendedTypeBug/RegularImage" ) [09:29] Ignoring the url you set to QUrl( "file:///home/tsdgeos_work/extendedTypeBug/thisShouldBeIgnoredByImport" ) [09:29] qml: MoMoMoMo MoMoMoMo [09:30] ltinkl: do you get that? [09:31] tsdgeos, yup I'm getting qml: MoMoMoMo MoMoMoMo [09:31] and if you uncomment the image [09:32] you get the "cannot open" variant [09:32] right? [09:32] yup [09:33] do you agree that both behaviours are wrong? [09:34] not sure, it's either getting the image from the C++ plugin or from the Image QML class right? [09:37] the problem is [09:37] that if the singleton loads an image [09:37] the qmlRegisterExtendedType that gets loaded later is ignored [09:37] that's problem A [09:37] problem B is [09:37] if the singleton is not loading the image [09:38] the loading of ManglingImport in MyItem.qmll viralizes to the Image in main.qml that knows nothing about such import [09:39] problem A - perhaps it's designed that way, so that QML imports get precedence over C++ imports [09:39] problem B - ye, I get that, this is eird [09:39] w [09:40] but understandable imho, you're extending the QQuickimageBase, so basically Image is no longer the one shipped by Qt but your provided by the plugin [09:40] no matter where you import it [09:41] so once you import it, there's no going back to the original Image imo [09:42] how would QML know then which one to use? [09:42] because it knows [09:42] it has different metaobjects [09:42] for different imports [09:43] i.e. the image of 2.0 is different from the image of 2.4 [09:43] it has a different metaobject * [09:43] they just need to also extend that for extended stuff [09:43] i agree it's hard [09:43] but not undoable [09:43] and A it can't be designed that way [09:44] it's a oversight i'd say that they forgot about it [09:44] anyway will file two bugs and see what happens [09:44] BTW A+QMLcache is what is making us crash in unity8 some cases [09:45] aha I see [09:45] ye definitely ask those 2 basic questions (bug reports) and see what they have to say [11:26] pstolowski: greyback landed something so you need to rebuild the silo [11:28] and hopefully there's no conflicts [11:29] greyback: can you do https://code.launchpad.net/~aacid/unity8/workaround_qt47709/+merge/267647 ? [11:32] tsdgeos: lgtm [11:43] greyback: i think the last landing makes a test crash [11:43] let me confirm [11:43] tsdgeos: which test? [11:43] LauncherModelTest [11:44] http://paste.ubuntu.com/12055032/ [11:44] not sure if it's new though [11:44] how nice would be to have cI back ... [11:44] hmm, that's an old one [11:44] somehow has resurfaced [11:44] I recall there were times "app" was null [11:45] which didn't make sense [11:47] greyback: do i open a bug? assign it to who? [11:47] tsdgeos: open bug anyway, will give it to dandrader [11:48] isn't he off for like the next two weeks? [11:48] I *think* that's from tomorrow [11:48] oki [11:49] tsdgeos, rebuilding [11:50] tsdgeos, have you found anything about the error i had? [11:51] pstolowski: hadn't had time to test, promise first thing after lunch [11:51] lunch now! [11:51] greyback: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1483675 [11:51] Ubuntu bug 1483675 in unity8 (Ubuntu) "testLauncherModel crashes" [Undecided,New] [11:52] ta === alan_g is now known as alan_g|lunch === om26er_ is now known as om26er [12:48] dandrader: vivid or wily? [12:49] tsdgeos, vivid [12:49] dandrader: weird, dist-upgraded? [12:49] tsdgeos, yes [12:50] * tsdgeos scratches head [12:54] pstolowski: you mentioned some issues with the silos not installing properly [12:54] pstolowski: what do i do to fix it? or should work? [12:54] tsdgeos, yeah, you need to make sure you get versions from the silos [12:55] oki [12:55] tsdgeos, apt-cache policy for every package [12:55] tsdgeos, then apt-get install foo= using versions from ppa if needed [12:56] tsdgeos, but i'm still battling with silo 4. rebuilt failed [12:56] tsdgeos, looks like something landed also in unity api, so rebuilding [12:56] k === alan_g|lunch is now known as alan_g [13:08] pstolowski: i guesswe need that rebuild [13:08] i'm getting [13:08] [1439298272.585212] mircommon: Caught exception at Mir/EGL driver boundary (in queueBuffer): /build/mir-VroDxW/mir-0.14.0+15.04.20150722/src/client/rpc/stream_socket_transport.cpp(168): Throw in function virtual void mir::client::rpc::StreamSocketTransport::send_m [13:08] essage(const std::vector&, const std::vector&)^M [13:08] and unity8 restarts/crashes [13:10] let me try to see on the desktop [13:26] pstolowski: yeah can reproduce this Playlist thing on the desktop, investigating [13:26] tsdgeos, ok [13:30] pstolowski: pushed a change, sorry [13:31] ok [13:31] tsdgeos, only unity8? [13:31] yes [13:32] how do i test it, the music scope was empty [13:32] am i supposed to get something? [13:32] tsdgeos, do you have local music? [13:32] pfff [13:32] probably not [13:32] phew [13:32] i mean i do have it [13:32] but not where media-scanner wants it i guess [13:32] i'll copy some stuff to ~/Music [13:32] is it? [13:33] tsdgeos, yes [13:33] tsdgeos, then go to Tracks department in My Music [13:34] ok [13:35] hmm wait [13:35] i do actually have stuff in ~/Music [13:35] Caught an error from create_query(): unity::scopes::MiddlewareException: unity::ResourceException: RegistryObject::ScopeProcess::exec(): exec aborted. Scope: "mediascanner-music" took longer than 4000 ms to start. [13:36] pstolowski: ↑ ? [13:36] tsdgeos, do you have all the versions from ppa? including mediascanner2.0 and unity-scope-mediascanner2? [13:36] yeah [13:36] http://paste.ubuntu.com/12055524/ [13:38] tsdgeos, hmm. need to wait for rebuild. what phone image do you have? [13:38] pstolowski: this is on the pc [13:38] pstolowski: on the phone can't test because the mir crash/restart [13:38] tsdgeos, ah... i can't test on the PC ;) [13:39] tsdgeos, waiting for the build, will check on the phone [13:39] ok [13:39] i'll uninstall the stuff from the PC [13:39] and do something else while waiting for the rebuild :) [13:41] tsdgeos, keep fingers crossed for the build if your hands are free ;) [13:43] he he [14:07] tsdgeos, can you check if https://code.launchpad.net/~aacid/unity8/dash_activation_no_special_casing/+merge/264024 merges ok with trunk? [14:08] hmm [14:08] it should [14:08] let me check [14:08] pstolowski: yeah merges fine here [14:08] tsdgeos, ok thanks [14:18] tsdgeos, silo 4 built [14:19] pstolowski: cool === dandrader_ is now known as dandrader|afk [14:44] pstolowski: seems to "do things" :D [14:44] pstolowski: if i play the first track is it supposed to play the next? [14:44] tsdgeos, only if from same album [14:44] ok [14:44] * tsdgeos waits 2:20 [14:45] tsdgeos, does it integrate with indicator? [14:47] pstolowski: not yet, jhodapp said he wanted to wait for something that used it to enable it [14:47] or that's what i understood [14:50] doesn't seem the play next song worked [14:50] let me add some console.log [14:52] tsdgeos, correct [15:00] jhodapp: so silo 4 [15:02] tsdgeos, silo 4? [15:03] jhodapp: if you join it with silo 38 has a somewhat working audio card support on the dash [15:03] tsdgeos, oh very nice [15:03] tsdgeos, thanks [15:03] for some reason playing the first time fails [15:03] you have to play twice [15:03] tsdgeos, so I assume you can't land silo 4 until the indicator controls work? [15:03] and the playlists don't work either [15:04] but not sure whose fault it is [15:04] debugging a bit [15:04] jhodapp: correct [15:04] jhodapp, also we need to wait for silo 27 to land and then rebuild [15:04] tsdgeos, let me give you a Test.qml file that you can use to compare things against [15:05] tsdgeos, last time i checked with scopes-client it was exposing playlists as expected. perhaps we have a mismatch in names or something trivial like that somewhere [15:05] tsdgeos, i mean between dash and scope [15:06] tsdgeos, alright, shared with you on google drive [15:06] pstolowski: sure, also debugging that [15:07] tsdgeos, just use qmlscene Test.qml --desktop_file_hint=/usr/share/applications/mediaplayer-app.desktop [15:07] tsdgeos, and make sure to change the sources listed in there to ones you have on your test device [15:07] k tx [15:07] *music sources [15:10] pstolowski: yeah looks good, seems i'm losing it somehow at some point [15:16] booo me [15:16] you can't obviously store an array in a url :D [15:16] boo [15:16] if the sdk wasn't overflowing us with warnings [15:16] maybe even the qml interpreter told me [15:16] at some stage [15:20] also the song i'm trying is longer that it says it is? [15:20] pstolowski: 2:47 vs 2:45 ? [15:20] wooo, works :) [15:20] \o/ [15:20] tsdgeos, afaik the duration is read from mp3 tags or some such, may not be precise? [15:21] may be [15:21] alecu, ^^^ [15:22] the file doesn't seem to have that in the tags [15:22] http://paste.ubuntu.com/12056055/ [15:22] pstolowski: is that inline playback? [15:22] tsdgeos, hmm in any case this is what mediascanner backend tells me. probably it gets it from gstreamer [15:23] alecu, yes [15:23] pstolowski: what's the silo number? [15:23] pstolowski: wonderful [15:23] alecu, 4 + 38. but indicator intergation is not enabled yet [15:26] it's pretty nifty going back and forth [15:26] and still works :D [15:27] even on the album preview [15:27] even if you played it from the tracks list [15:27] cool [15:35] jhodapp: the debug is weird [15:35] the first time i add things to the playlist i have [15:36] http://paste.ubuntu.com/12056136/ [15:36] while the second i have [15:36] http://paste.ubuntu.com/12056137/ [15:36] but i'm passing the urls the same way [15:36] let me console.log them [15:40] jhodapp: can it be that addSources to a playlist are ignored if it's not part of a player? [15:40] that may explain why the second time works [15:40] * tsdgeos tries [15:41] tsdgeos, yes there's a possible bug in the QDeclarativePlaylist impl [15:41] tsdgeos, let me get back to you in a little bit, on a hangout === dandrader|afk is now known as dandrader [15:46] yeah i sthat [15:46] i guess i can workaround it in my code easily [15:46] i mean i just did [15:46] but would make sense fixing it anyhow [15:49] pstolowski: can you rebuild the silo so that it includes the two last fixes i made? [15:49] pstolowski: unity8 [15:49] tsdgeos, ok [15:53] tsdgeos, yes indeed, there will be several bugs to fix [15:53] tsdgeos, can you file that bug against qtmultimedia-opensource-src package? [15:54] tsdgeos, rebuilding [15:58] jhodapp: https://bugs.launchpad.net/ubuntu/+source/qtmultimedia-opensource-src/+bug/1483806 [15:58] Ubuntu bug 1483806 in qtmultimedia-opensource-src (Ubuntu) "addSource to a Playlist are ignored if it's not part of a player" [Undecided,New] [15:58] pstolowski: cool [15:58] tsdgeos, thanks! [15:59] tsdgeos, oh right, yes it clears the playlist when it attaches to a player [15:59] tsdgeos, I didn't realize that's what you were trying to do...this is on purpose [15:59] ok [15:59] it's not a big thing i just did it the other way around [16:00] well actually no this is a valid bug, nm [16:00] but if that's the behaviour, make sure it's documented :D [16:00] * tsdgeos eods [16:00] it will clear any existing playlist attached to the player [16:07] davmor2, does silo 27 work for you now? [16:15] pstolowski: I'm on another silo currently will return to that after [16:16] davmor2, ok [16:16] anyway, eod. cu === alan_g is now known as alan_g|EOD === dandrader is now known as dandrader|afk === jhodapp is now known as jhodapp|errand === jhodapp|errand is now known as jhodapp === dandrader|afk is now known as dandrader [19:02] josharenson, ping [19:18] dandrader: pong [19:20] When I run "make tryWideView" all users have a "Tap do unlock" button. Some of them should display a text entry for instance, for typing the password [19:21] * josharenson runs that test [19:21] josharenson, do you know what's wrong there (with the mocks most likely) [19:21] ? [19:24] sorry didn't have a good build, compiling now [19:28] dandrader: ah I see. Nothing happens when you "tap to unlock" but there should be a text box, yes? [19:29] does it hang for you? [19:31] josharenson, yeah, nothing happens. [19:31] looking into it, either a problem with the mock, or the wrong lib is being loaded in the test I would imagine [19:31] josharenson, The "Has Password" user should naturally display a message box for typing the password, for instance [19:33] josharenson, I wanted to debug that login message box and because of this bug in the test there's no easy way to do so :/ [19:33] dandrader: ok, give me a few, I'll ping you when I find something [20:46] josharenson, hey [20:46] dandrader: hi [20:46] josharenson, does PhysicalKeysMapper have some logic to ensure you don't change the volume while your' [20:47] josharenson, you're pressing the volume keys to get a screenshot? [20:47] dandrader: it should, saviq/I fixed that a while ago I thought [20:47] dandrader: I honestly don't remember what the exact solution ended up being, I'd have to look at it [20:53] josharenson, we just don't have a test guarding this feature [20:54] dandrader: I could swear I wrote some... let me look [20:54] josharenson, I bet https://code.launchpad.net/~lukas-kde/unity8/globalshortcuts/+merge/267188 will inadvertently break it [20:58] dandrader: I see. Either that code needs logic to handle not showing volume notifications, or it needs to let volume key events through, which is a bad idea [20:59] josharenson, just commented there that it might break this behavior [20:59] dandrader: I'll try and do a review if I have some time later [21:00] josharenson, if you want to. I already did my second pass over there [21:00] ok [21:00] or third, actually === dandrader is now known as dandrader|afk [21:02] josharenson, but it would be great if you could confirm me whether we have a test for this screenshot vs. volume case. I didn't find any in tst_PhysicalKeysMapper.qml [21:03] dandrader|afk: ok ill look for it for sure. Also, before you go away, I can't find anything wrong w/ the greeter wide view [21:03] oh wait [21:03] hang on [21:12] dandrader|afk: ah, there was a negative test when the screenshot sequence was Power+VolDown but it looks like it never changed when we reverted to VolUp+VolDown [21:20] ok. see you next week === g4mby is now known as PaulW2U