[01:03] Is there a roadmap or preliminary release date for the rewritten unity? (QT/QML) [01:05] Also, will the QML version be free from all gtk* cruft? [01:05] or will it still depend om gtk in any way? [03:21] Not very lively today [03:21] Ill wait a few more hours before i go to bed. [03:24] EgonR|2: Did you have a question that needs answering? [03:25] I asked two and an half hour ago the following: Is there a roadmap publically available or preliminary release date for the rewritten unity? (QT/QML). Also, will the QML version be free from all gtk* cruft? [03:25] perhaps you would know [03:27] EgonR|2: Yes *gtk* should be gone. And a basic roadmap is https://wiki.ubuntu.com/Mir/Spec#Roadmap [03:28] Awesome! I am looking forward to is so much [03:29] Will it be MIR exclusive? Or can you run the new unity on Xorg [03:30] I mean, writing and entire displayserver takes time.. [03:31] EgonR|2: Not totally sure. You should assume it's Mir-exclusive but we'll always aim to make things portable eventually. It's just not a priority right now [03:31] Yeah, i understand that. [03:32] This is where we need some Unity Next people online. I think they're mostly in Europe so won't wake up for a few hours yet [03:32] Oh ok [03:33] Looks like that is very much a pattern for most OSS projects on freenode. [03:34] I think US programmers use EFNet a lot more than the europeans though. [03:36] Maybe... Though it's Canonical policy to have its public discussions on FreeNode [03:36] yup [03:36] Ubuntu policy even [03:37] freenode seems like a better host for a software project. efnet is littered with strange channels as well.. [03:37] hehe [03:39] duflu: Thank you for your time, it was most helpful. I will go back to lurking. [03:45] EgonR|2: No problem. Sorry more people don't pay attention very often [06:10] hey Mirv, how are you? [06:16] didrocks: morning, fine, and you? [06:21] Mirv: I'm good, thanks! :) [06:21] Mirv: I know it's Qt4 and not Qt5, but do you mind having a look at bug #1155327 and tell us what's the best course of action for it? [06:21] bug 1155327 in skype (Ubuntu Raring) "skype crashed with SIGSEGV in malloc@plt()" [High,In progress] https://launchpad.net/bugs/1155327 [06:28] didrocks: ok, looking.. [06:56] commented, although can't reproduce myself as it needs nvidia hardware to happen (+ possibly prelink installed). I don't know what to do other than revert to qtwebkit-source 2.2.1 (or ask for rebuild/fix from Microsoft) [06:57] or finding out skype is the only affected package and accepting the workaround for it [06:57] Mirv: do you mind discussing that with the kubuntu guys? [06:58] Mirv: I think we should, from your comment revert [06:58] Mirv: as you told, we are not sure if skype is the only affected package [06:58] ok, I'll ask around on kubuntu channel [06:58] thanks :) [07:01] riddell has looked into it a week ago, asked for an update [07:01] Mirv: excellent, do you mind keeping me in touch? [07:05] didrocks: ok [07:20] sil2100: good morning, how are you? [07:23] didrocks: morning! Good, how about you? [07:24] sil2100: I'm good, thanks! [07:24] sil2100: do you have time for some touch apps boostrapping help for preparing that for S? [07:25] didrocks: should be no problem, what can I do to help? [07:25] sil2100: should we have a quick hangout? It will maybe be easier [07:27] didrocks: ok [07:27] sil2100: now is fine for you? [07:29] didrocks: yep, ready when you are ;) [07:35] mzanetti, whoa http://s-jenkins:8080/job/unity-phablet-ci/ [07:35] didrocks: just point me to the hangout when ready ;) [07:35] Saviq: yeah... I just cancelled them and am debugging right no [07:35] w [07:36] sil2100: yeah, finishing a discussion and will ping you soon :) [07:36] mzanetti, thanks [07:40] sil2100: https://plus.google.com/hangouts/_/effa6372491bd8d9829359be5de48d6464cfbc83?authuser=0&hl=en [08:11] Saviq, ping [08:14] dandrader, pong [08:18] Saviq, I've been working on removing those FooFake vs. FooWrapper switches and components throughout the qml code (e.g. get rid of ApplicationManagerFake). The plan is to have a fake implementation of Ubuntu.Application module. That way we can get the qml tree pretty much clean of those things. It started out as infrastructure to test RunningApplicationTile but then I though it would be better to just apply it on the whole code base [08:18] Saviq, what do you think? [08:20] Those FooFake vs. FooWrapper switches might be fine for running on the desktop but get in the way of testing [08:24] dandrader, yeah, if we can make without those wrappers, I'm game [08:25] dandrader, only so that the actual Ubuntu.Application is used elsewhere, make sure you put it in a separate module path [08:25] and export QML2_IMPORT_PATH in the ./run script [08:27] Saviq, right [08:38] hey, anyone knows whats going wrong that after recent upgrades in raring the machine doesn't come up any more because of a read only rootfs? [08:39] mzanetti, why would you shut down? it's so '90s ;) [08:40] mzanetti, but seriously, nope, haven't had that [08:40] Saviq: let me rephrase: the machine doesn't come up after a quick reboot because of kernel upgrades [08:40] :D [08:40] veebers has the same in his raring VM [08:41] o/ [08:42] veebers: yeah, if you could find out if its only with raring too or happens after inclusion of phablet ppa's that would be great [08:42] tsdgeos, hi :) [08:42] mzanetti: I'm just fire-ing up the VM now :-) [08:42] Cimi: hi there [08:42] I need a little help here if you have 5 mins [08:42] https://code.launchpad.net/~unity-team/unity/phablet.moving_tests [08:43] I was moving tests under subdirs [08:43] but hud test complains of missing hudclient [08:43] I think is something to do with the makefiles [08:43] can you help me? [08:45] sure [08:45] give me a sec [08:48] mzanetti, we got a PASSED! [08:48] Saviq: huh? [08:48] the tests? [08:48] mzanetti, http://s-jenkins:8080/job/unity-phablet-ci/349/ [08:48] Saviq: yeah... seems one of the VM's is configured different than the other [08:48] uh [08:48] Saviq: and makes the tests stall... I'm on it [08:49] Saviq: right now Mir is building on the "broken" one which takes 1.5 hours :/ [08:49] right [08:49] * mzanetti is happy that unity phablet ci runs in ~10 mins :D [08:52] Saviq: so i got a patch by "the qml guy" [08:52] tsdgeos, nice [08:52] it does indeed fix the crash but i'm not sure his solution is great [08:52] tsdgeos, does it help? [08:52] tsdgeos, :) [08:52] tsdgeos, is it in gerrit? [08:52] problem he is in US west coast, so it's going to take a while to ping-pong between us [08:53] yep [08:53] https://codereview.qt-project.org/53235 [08:54] tsdgeos, can you think of a better solution? [08:54] Saviq: it's not a better solution, it's that this one is only partial [08:55] i.e. accounts for removals to the cache but not additions, not really sure that this is a problem, but it might [08:55] and also the delayed stuff comes from somewhere that does "delete this" thus we effectively end up with a dangling pointer in the cache too [08:55] it "might" not be a problem [08:55] but i'd want to make sure he has thought about these [08:56] tsdgeos, that's what gerrit is for, isn't it :) [08:56] it is [08:56] i'm still perfecting my text :D [08:56] tsdgeos, fortunately it's not awfully pressing [08:56] tsdgeos, so we'll get there when we'll get there [08:57] Cimi: what happened is this "Panel/qml/HudClient/" [08:57] tsdgeos, will you own getting this distro-patched? [08:57] tsdgeos, you have commit access if you can fix :) [08:58] Saviq: once we agree on a solution upstream? sure, it's just messing with zoltan/mirv to get it into the packages, no? [08:58] tsdgeos, yup [08:58] sure, no worries [09:02] mzanetti: after a dist-upgrade with no PPAs and a reboot, machine came up fine. I'll try with a single ppa now see if that breaks [09:02] veebers: cool, thanks [09:02] nic-doffay, you joining the hangout? [09:02] One sec Saviq [09:02] nic-doffay, cheers [09:04] mzanetti: ping [09:04] tsdgeos: yes sir [09:05] mzanetti: you added/wanted to add a make -i/-k to qmluitests, right? [09:05] tsdgeos: yes [09:05] is that a good idea? [09:05] added [09:05] why not? [09:05] i understand CI wise it makes sense [09:05] but if i'm sitting here [09:05] Cimi, ping [09:05] i run "make qmluitests" [09:05] it writes lots of stuff in the shell [09:06] MacSlow, pong [09:06] like two screens [09:06] and most probably i'll only look at the last test and see a "passed" [09:06] tsdgeos: its only when you run "./runtests.sh" [09:06] ah [09:06] ok :-) [09:06] right [09:06] * tsdgeos hits head [09:06] tsdgeos: if you run "make qmluitests" manually there is no -k :D [09:06] yeah, sloooooow [09:06] tsdgeos: hint: run "make alltests" [09:06] Cimi: pushed [09:08] tsdgeos, do panel tests work now? [09:08] Cimi: sure, why wouldn't they? [09:08] tsdgeos, thought thy needed qml subdir [09:08] they do [09:08] it's still there [09:08] the part of it they need [09:09] this is good too [09:09] since we're splitting the "qml" subdir that holded separate stuff for two separate tests [09:09] tsdgeos, because you modified cmakelist of panel [09:09] tsdgeos, the part you left doesn't need it? [09:09] and putting it into the correct places [09:09] Cimi: no [09:09] ok [09:09] tsdgeos, if everything works we're ready for MR [09:10] mardy: ping [09:10] nic-doffay, you with us? ;D [09:15] Saviq, mzanetti tsdgeos https://code.launchpad.net/~unity-team/unity/phablet.moving_tests/+merge/157821 [09:15] Cimi: ack [09:16] mzanetti, better sooner than peter so we have less tests to move afterwards :) [09:16] *later [09:20] mzanetti: after the raring dist-upgrade, I added the phablet-team ppa and again did an update && dist-upgrade. I get this error: https://pastebin.canonical.com/88744/ [09:26] mzanetti: running dist-upgrade again I get this error: http://pastebin.ubuntu.com/5691831/ (the apt-get -f install log: http://pastebin.ubuntu.com/5691833/) [09:27] veebers: hmm... somehow I can't figure how this relates to a read-only filesystem at boot [09:27] mzanetti: the dist-upgrade after that went fine, but after the reboot I get the read only filesystem [09:28] Saviq: any idea? ^ [09:29] mzanetti, Saviq: this is what the phablet-team dist-upgrade wanted to install: http://pastebin.ubuntu.com/5691841/ [09:29] veebers: mountall... that sounds like it could be involved [09:29] veebers: mind reverting your VM and just upgrade that package? [09:30] Cimi: can you merge master into https://code.launchpad.net/~unity-team/unity/phablet.test_responsiveflowview/+merge/157343 ? It's what making the qmluitests fail [09:31] mzanetti: yeah I can do that, I'll upgrade it using the phablet-team ppa [09:31] veebers: thanks [09:32] np [09:37] tsdgeos, ok [09:38] tsdgeos, done [09:42] tx [09:43] luv: pong [09:44] mzanetti, Saviq: After adding phablet-team and installing mountall, rebooting gives me a readonly filesystem [09:44] veebers: \o/ thanks a lot!!! [09:45] mzanetti: no worries, good spotting with the mountall issue [09:45] anyone knows what exactly this does, and expecially why we need a -phablet version for that? [09:46] mardy: Hi! I've been messing around with accounts-sso and implementing the logout functionality to gnome-control-center-signon. I am trying to figure out an API call for libsignon to delete the stored password. What is also confusing me is that there is no signond running on my system. [09:46] http://docs.accounts-sso.googlecode.com/git/libsignon-glib/html/SignonIdentity.html#signon-identity-remove just does not work for me [09:47] mzanetti: do you mind if I leave that with you? I need to finish for the night [09:47] on the other hand, if I delete an account in gnome-control-center-signout the credentials are removed (I guess it communicates with signon via dbus - but no signond running?) [09:48] luv: it should work, because we are using that function call from inside libaccount-plugin to delete the identity when the account creation is aborted [09:48] luv: signond must be running, but it exits after a few seconds of inactivity [09:49] luv: set the debugging level to 2 in /etc/signond.conf, then keep an eye on the syslog [09:49] yes, I was looking at libaccount-plugin to see how it deletes the account - but I didnt see a direct use of that function only a lot of dbus stuff [09:49] luv: you should be able to see if/why your call fails [09:49] umm, I will try that [09:50] luv: http://bazaar.launchpad.net/~online-accounts/gnome-control-center-signon/trunk/view/head:/libaccount-plugin/oauth-plugin.c#L187 === dandrader is now known as dandrader|afk [09:50] thanks a lot, I will have another look tonight :-) [09:51] luv: but I'm a bit worried, why are you deleting identities? :-) [09:51] alright - to implement "logout" functionality [09:52] luv: I don't think you should delete the identity in that case -- doing so will mess things up [09:52] tsdgeos, I've also rebased to the moving tests branch [09:52] oki [09:52] mardy: well sure, but you cant keep passwords and tokens around if you want to implement logout [09:52] luv: if you really want to, you could store an empty password on the identity, but I'm not 100% sure it's a good idea [09:52] that'd be useless [09:53] luv: why? [09:53] keeping passwords and tokens around would be useless [09:53] luv: ah, yes, agreed [09:54] well, I can modify gnome-control-center-signout to deal with a deleted identity properly [09:55] and the apps themselves - they dont have access to identities, do they? [09:55] luv: you could retrieve the current IdentityInfo, clear the password (actually, I think you'll get it empty anyway), delete the record and then store the IdentityInfo as a new one [09:56] luv: they have access -- they use it when authenticating [09:57] luv: oh, a better solution, to make sure that there aren't race conditions, is to create the "cleared" identity record first, then store its ID into the account, and finally remove the previous ID [09:58] luv: maybe that's exactly what you wanted to do :-) === dandrader|afk is now known as dandrader [10:07] mardy: sounds good! [10:07] i will experiment with that for now ;-) and see where it goes [10:09] luv: but I think that it would be much better to implement a logout API directly in signond [10:12] sure, but let me try this first [10:13] Cimi: you know whats weird... [10:13] mzanetti, no :) [10:14] Cimi: if I call "runtests.sh" on current trunk it says 35% coverage [10:14] Cimi: if I run it on your branch its less [10:15] lol [10:15] mzanetti, let's fixthe script then [10:15] Cimi: ... (following code guidelines from mzanetti) ... [10:15] interesting... which ones? [10:15] :D [10:16] mzanetti, when I asked you yesterday [10:16] Cimi: you mean the ordering of testcase and component [10:16] yes [10:16] Cimi: yeah... its fine... I was just wondering... my guidelines? :D [10:16] mzanetti, you now own a guideline, cheers :) [10:19] e [10:23] Cimi: mystery solved... and approved. cheers [10:24] mzanetti, where was it? :) [10:24] untested OnScreenKeyboard.qml was removed in trunk but still present in your branch [10:24] ah ok [10:26] we just reached > 50% coverage! [10:26] thanks everybody [10:30] mzanetti, who told you to add phablet-team ppa? ; [10:30] ;P [10:30] and to upgrade from it! [10:30] mzanetti, some tests are strikeout in the coverage document, means not needed? [10:30] Saviq: still, this is an issue [10:31] Saviq: I think our target should be to enable people to add the paa, no? [10:31] mzanetti, not really, stuff should land in distro instead [10:32] Saviq: yeah... but I guess the packages in our ppa will be merged to the distro... as soon that happens it'll smash your system too [10:32] Cimi: yep [10:33] mzanetti, I don't know what's changed in mountall [10:36] mzanetti, looking at https://code.launchpad.net/~phablet-team/phablet-extras/mountall-quantal I'd say this should not land in distro [10:37] mzanetti, but ping rsalveti or awe about it [10:37] Saviq: yeah, I will! thanks for the research! [10:48] Guys, how can I compile the Hud with qmlscene? [10:48] trying qmlscene Hud.qml [10:55] tsdgeos, ^ [10:55] nic-doffay, you might need to pass the import path where the qml plugins it needs to run are located [10:55] with -I path [10:55] nic-doffay, there's a chance you nees -I and a path [10:55] nic-doffay: the test? [10:58] tsdgeos, not the test, the actual Hud. [11:02] nic-doffay: why do you want to run the actual Hud with qmlscene? [11:02] tsdgeos, just to take a look at it easily. [11:03] i'm not sure it even works [11:03] but sure, you'll have to pass the correct path to the HudClient plugin === MacSlow is now known as MacSlow|lunch [11:13] dednick: what creates unity.qmlproject.user ? QtCreator? [11:13] tsdgeos: yes [11:13] ok [11:14] tsdgeos: yes [11:15] my bad [11:15] ok [11:15] :D === greyback is now known as greyback|lunch [11:26] someone *did* test Panel/IndicatorItem.qml? [11:27] there's a Y in the document, but it's not green nor link on MR [11:31] i'd say it's "part" of the regulra Indicators test [11:31] maybe i'm mistaken [11:33] Cimi, if you don't see a test for it then it's not tested. I would guess that original list was generated by checking which files were read by the tests [11:33] dandrader, ok thx [11:34] anyone ever used the qml's item::mapFromItem ? it's driving me insane. trying to map a repeaters items to global position. i'm getting the correct location for the first item, but each subsiquent item is 2x the width of the previous item away... [11:35] dednick: i used it once, worked without problems [11:35] dednick: mapFromItem doesn't map width, you mean x? [11:38] tsdgeos: i mean the x of items after the first repeaters itms are 2xthe previous items width away. [11:38] so if the first items x's are 0, 28, 56, 84. when i do mapFromItem, i'm getting 0, 56, 112, 168 [11:38] right, you're probably maping it from the wrong item [11:39] but it shouldnt change the x scale if the item is in the parent heirachy surely? [11:39] not sure [11:39] try ignoring the repeater [11:39] and map against the row [11:40] or whatever is holding the repeaetr [11:40] i'm mapping to the indicator [11:40] indicator -> indicator row -> row -> row repeater -> item [11:41] bur i've tried doing it to a few of them with the same resules. [11:41] *results [11:46] What's the deal with the frantic key presses when running qmlscene? [11:46] It's really annoying. [11:46] tsdgeos: found the issue. me being a general idiot. [11:47] i was including the x,y of the item in the mapFromItem call, which was adding width... [11:49] tsdgeos: found the issue. I was including the x,y of the item in the mapFromItem call, which was adding width... [11:49] ouch [11:49] yeah. pretty stupid [11:50] i take solice in the fact that i'm not the first to do it. found answer on the web ;) [11:50] mzanetti, hey, I'm testing the Clock: I have two ways to get the text label: using objectName and grepping the label, or creating a readonly property alias of those labels [11:50] mzanetti, which is the preferred? [11:51] Cimi: if possible, avoid API changes just for tests... prefer findChild() [11:51] ok [11:51] Cimi: if the API already offers that property for other reason, use that because its faster than findChild() [11:56] Saviq, is it so that the ApplicationLauncher plugin is not used anymore? === alan_g is now known as alan_g|ACCU [11:56] dandrader, yes [11:56] Saviq, can we nuke it? [11:56] dandrader, yup [11:56] Saviq, will do [12:02] where do I find qt. javascript functions? [12:03] I want to see if I can get a date from a string [12:03] (testing Greeter's Clock) [12:03] there's Qt.formatTime and Qt.formatDate, how about the opposite? [12:04] found [12:04] ! [12:04] Cimi: interesting... where? [12:04] but it's not documented [12:04] http://qt-project.org/doc/qt-5.0/qtqml/qtqml-javascript-functionlist.html#date-objects [12:05] actually I'm not sure that works [12:05] Cimi: Qt itself doesn't really ship any javascript functions... The javascript engine itself support the ECMA standard [12:05] ah ok [12:06] Cimi: additionally Qt register a object (called "Qt") that offers a little functionality [12:06] Cimi: its mostly a subset of the QDateTime C++ class [12:06] I was speaking of that [12:06] yes [12:06] do you know how can I do the opposite? [12:06] check if a string is a regular date? === MacSlow|lunch is now known as MacSlow [12:08] Cimi: not out of my head, no [12:09] Cimi: but does that make sense? I mean... the string is the direct output of Qt.formatDate()... No need to test if formatDate() works. Qt folks should test that inside Qt [12:12] mzanetti, so I will check if those strings are not empty [12:13] Cimi: yeah, thats good. you can also compare with a static string I guess [12:13] mzanetti, static string of that? [12:14] something linke this in the data() function: [12:14] { input: 123456; outputData: "Mon Feb. 12"; outputTime: "12:34" } [12:15] if you have a predefined input date (e.g. some unix timestamp) you can compare if the label holds that correct time [12:15] just an idea... not sure its really way to go... [12:16] not empty could in theory also be something like "Error converting date" which is not what we want [12:19] Cimi: ^ (not sure if you still saw it) [12:20] mzanetti, do you think qt.formatTime returns that? [12:20] * Cimi tests [12:21] mzanetti, just tested, qt returns empty string [12:21] Cimi: for what= [12:21] ? [12:21] I tested passing "__date" to Qt.formatTime and formatDate [12:21] instead the variable __date === dednick is now known as dednick|lunch === dandrader is now known as dandrader|lunch [12:34] tsdgeos: https://code.launchpad.net/~mzanetti/unity/phablet-copyright-headers/+merge/157858 [12:35] mzanetti, https://code.launchpad.net/~unity-team/unity/phablet.test_greeter-clock/+merge/157859 [12:36] mzanetti, my only question here is if I should change width and height to something else, these are half width and height of greeter/dash tests === greyback|lunch is now known as greyback [12:36] Cimi: I'd say its fine [12:37] ok [12:37] works [12:37] Cimi: if you change it, then I guess childrenRect.height/width would be ok [12:38] I should change to that? [12:38] it does binding loop [12:38] nevermind === _salem is now known as salem_ [12:47] Cimi: reviewed [12:47] hey guys is there an actual package for autopilot if so what is it called or is it installed by default? [12:48] davmor2: autopilot-phablet [12:48] davmor2: if you're working with "the old unity" you can use python-autopilot [12:48] davmor2: both are in ppa:autopilot/ppa [12:48] mzanetti: thanks [12:49] Cimi: could you do a quick check on this please: https://code.launchpad.net/~mzanetti/unity/phablet-copyright-headers/+merge/157858 [12:50] Cimi: we need to merge this soon because I've enabled the check Jenkins and that will put Needs Fixing on all the other MP's until this is merged === dednick|lunch is now known as dednick [12:58] mzanetti: done [12:58] cheers [13:04] we have 33 branches waiting for merge [13:04] instead of down we're going up :D [13:18] mzanetti: you getting https://code.launchpad.net/~paulliu/unity/phablet-add_unit_test/+merge/156859 ? [13:19] mzanetti: want me to do https://code.launchpad.net/~mzanetti/unity/phablet-pageheader-tests/+merge/157711 ? [13:19] tsdgeos: sure [13:20] tsdgeos: hmm... the openeffect thing is indeed a bit strange [13:20] yeah, may an artifact of that the class hasn't really much to test [13:21] "class" [13:21] yeah [13:23] tsdgeos: but it would have potential for the coolest looking test :D [13:24] tsdgeos: well, dunno... I would have tested it quite differently... but not sure if its worth the effort to force paul to invest more time in it [13:24] there isn't much to test indeed. I guess he covered that bits. its not to too obvious [13:31] mzanetti, was sleeping like a baby during lunch break :) [13:32] Cimi: :D [13:32] guys, we'll be late [13:32] Saviq: should we just start or wait? [13:32] mzanetti, go [13:32] ack [13:33] greyback, your turn for notes [13:33] Saviq: ok [13:34] dednick, nic-doffay standup [13:35] Don't have a mic here Cimi . [13:35] ok [13:50] Saviq, are you going to merge the two dirs or shall I? [13:51] Cimi, you [13:51] ok [13:51] Saviq, so to recap: qmluitests -> shell, unittests -> shell [13:51] Is there copyright problem FTBFS? [13:51] Saviq, and two appropriare macro [13:51] *appropriate [13:52] Cimi, yeah, the macro and the correct calls are already in place [13:52] Cimi, you just need to movie it in a single CMakeLists.txt file [13:52]