[01:59] HI, I am using Ubuntu 12.04. The version of Unity installed on my computer is 5.18 but I can see the latest version in 5.0 branch is 5.20 on launchpad. Anyone have any idea why is it so? [04:13] (5.20 has not yet veen uploaded) [06:56] didrocks, ping [06:56] Saviq: pong [06:57] didrocks, hey, when you have the time, could you please look over https://code.launchpad.net/~saviq/unity-api/fix-packaging-multiarch/+merge/163564 [06:57] didrocks, and potentially explain why we only use macro version in SONAMEs [06:57] s/macro/major/ [06:58] Saviq: sure, I'm catching up on hundreds of emails though, so maybe tomorrow morning, is it fine? [06:58] didrocks, sure [06:58] will do :) [07:17] good morning [07:18] hey there [07:20] greyback: jono contacted me and asked for a video of unity-next running on mir [07:20] greyback: I tried to create that but I fail unlocking the screen because of the missing mouse cursor :D [07:20] greyback: any chance you have the cursor visible already? [07:21] mzanetti: lp:~unity-team/unity/phablet-integrate-mir/ [07:22] mzanetti: from your terminal, do "./run -m" and it should open a new VT, and run Mir + Unity inside that [07:22] the usual Ctrl+C to quit [07:22] didrocks, hi - do you have 3 min. ? :) [07:22] hey mzanetti, greyback! [07:22] didrocks: hey there, hope you had a good holiday! [07:22] MCR_: still on the catchup, but sure, what's up? [07:23] hey didrocks [07:23] it is about EZoom [07:23] I want to disable it by default [07:23] greyback: yeah, nice weather! Helped to change my mind :-) [07:23] MCR_: the plugin you mean? [07:23] because of these arguments: [07:23] yes [07:23] https://code.launchpad.net/~mc-return/compiz/compiz.merge-ezoom-cleanup/+merge/159991/comments/360604 [07:24] didrocks: sunshine is good for the soul. [07:24] greyback: hmm... explains why I haven't been able to unlock it... with your branch I see the cursor, but I can't move it [07:24] greyback: its stuck in the top left corner [07:24] mzanetti: did you chmod the /dev/input ? [07:25] lol [07:25] thanks man [07:25] MCR_: you didn't address my main concern [07:25] which is ? [07:25] MCR_: we have no way for people to enable the plugin easily [07:25] mzanetti: there are times when we truly use our own language :) [07:25] MCR_: for people needing that a11y support [07:25] in g-c-c though, you can setup the shortcuts [07:25] didrocks, but there is no way without CCSM either [07:25] which is what the current situation enabled [07:26] that is possible ? [07:26] yep [07:26] not for plugin [07:26] but for the zoom shortcut [07:26] I did not find those shortcuts when I looked [07:26] so the current situation is still better [07:26] strange [07:26] it's there IIRC [07:26] I searched [07:27] I am quite sure it is not in All Settings->Keyboard->Shortcuts... [07:28] MCR_: if not, we should expose it [07:28] there is just some "Zoom in/Zoom out" and "Turn zoom on or off", but this is not for EZoom [07:29] MCR_: anyway, I'm not thrilled to change the default configuration as long as we ship unity 7 by default [07:29] I just want to optimize things [07:29] but now I know where you would want to go... [07:30] thanks [07:30] didrocks, ok - I won't change the config then, although I really have to say it is "not the most efficient solution" [07:31] MCR_: thanks :) [07:31] I'll look into exposing the shortcuts in gcc... but it won't be main priority [07:33] MCR_: sure, making sense [07:33] ok, ack - thx [07:33] MCR_: I really wish we had a better a11y story in Unity 7 tbh :( [07:35] didrocks, Compiz would be capable of delivering that of course... [07:35] MCR_: yeah, it's all on nux/unity side, I know [07:35] we are moving really fast forward with Compiz now [07:36] the best solution in the long run would be to give users CCSM back, but I know you do not want that [07:36] gcc's configuration is kindergarten compared to CCSM in terms of functionality [07:37] CCSM coders were really intelligent guys [07:37] it is the perfect configuration utility [07:39] hem, no time to argue on this "perfect", a perfect configuration utility should prevent you shooting on your feet IMHO :) [07:40] didrocks, exactly that is CCSM's strength [07:40] just Compiz could not cope with some issues in the past (and still WIP) [07:40] we are working on making it rock-solid [07:40] didrocks, have you seen: https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/745159/+attachment/3660185/+files/GridAllFunctionalityFullyFixed.mp4 [07:41] Launchpad bug 745159 in compiz (Ubuntu) "Segfault in compiz when unchecking "Snap windows back to original size" in the Grid options" [Medium,Confirmed] [07:41] ? [07:41] MCR_, if modern industry success show anything is that "exposing a zillion options users don't understand" is everything but perfect... [07:41] seb128, that is why we have tooltips explaining things to users [07:41] MCR_, what users want is something that just work, with only simple options for things they need [07:41] MCR_, you miss the point [07:41] seb128, which I cannpt find in gcc for example [07:41] I don't want to go through 15 panels reading tooltip for half an hour [07:42] if you need tooltips you loose [07:42] seb128, I still do not know what "Zoom in/Zoom out" and "Turn zoom on or off" means [07:42] the tooltip concept doesn't even work on a touch screen [07:42] in gcc [07:42] seb128, sorry I do not care about touchscreens too much... [07:43] well, that's only one limitation of the tooltips [07:43] but let's not argue about that [07:43] sure [07:43] ccsm is a fine power user tool agree [07:43] normal users just don't want that much complexity [07:43] and "zoom in/zoom out" does what it says [07:43] it zooms in or out [07:44] and turn zoom on or off is pretty self explanatory as well [07:44] it turns zoom on or off... [07:44] seb128, which zoom ? [07:45] that settings is coming from GNOME [07:45] I know [07:45] and zooming is built in for gnome-shell [07:45] but it is not ideal to configure Compiz [07:45] how the zoom is implemented doesn't matter to users [07:45] and is far less powerful [07:46] well, it's not the UI that is buggy there, it's our integration [07:46] we should ping a plugin that can do zoom [07:46] and bind those actions to the increase/decrease zoom effects [07:46] yes [07:46] I'll look into that (see above) [07:48] cool [08:31] Anyone available for a QML review? [08:32] hi, i have my friend's laptop and the keyboard works in the start and the resposnse in a funny way; like the CTRL(control) button was pressed hard [08:32] can anyone help [08:32] pls === alan_g|life is now known as alan_g [09:00] mardy: Hi, I would like to add GUI for signon_identity_signout to gnome-control-center-signon - I understand I will need signond from branch lp:~mardy/signon-keyring-extension/lp1172710 ... that's all fine, what I don't know how to do is how to build .deb packages from that branch. [09:01] (or anyone else who can help, is there a easy way how to build .deb from a "custom branch" (for an already packaged software) on launchpad?) [09:01] Thanks! [09:01] luv, `bzr bd` [09:02] luv, and if you want it to build in a PPA [09:02] luv, `bzr bd -S` and then push to the PPA as usual - `dput ppa:something package-file.dsc` [09:02] where the dsc file was created by bzr bd -S === mmrazik is now known as mmrazik|afk [09:32] mzanetti: hey, do you install desktop-autopilot or the touch one when you run the medium tests to merge unity8? (and maybe you know about apps as well?) [09:32] sil2100: ^ [09:33] didrocks: desktop for now. touch too as soon as we have the hardware set up to run them [09:33] mzanetti: so the apps and media medium tests should run just fine with desktop-autopilot for now? [09:34] autopilot-desktop* (sorry) [09:34] didrocks: yes. om26er fixed them all and is currently re-enabling the jobs... [09:34] didrocks: btw: quite some broke already... [09:35] mzanetti: excellent, finally a clear info! we saw some failures as well. Thanks :) [09:35] Saviq, when I run the greeter via ./run touch events aren't being picked up. Any idea as to why? [09:35] didrocks: actually introducing bugs that would have been catched if tests would have been enabled [09:35] mzanetti: yes, we saw some AP failures finally and were wondering what was happening [09:35] Sorry not ./run [09:35] sil2100: yeah... they were disabled too long [09:35] on the device. [09:35] mzanetti: I asked fghinter to enabling them back during the sprint, that didn't happen? [09:35] didrocks: no [09:36] shrug :( [09:36] didrocks: yesterday I decided to take care about it myself [09:36] mzanetti: yeah, I was away for a week. I thought that asking it was enough… [09:36] mzanetti: thanks for taking care of that [09:37] didrocks: np... it has been a lession for us all... [09:50] didrocks: https://code.launchpad.net/~sil2100/cupstream2distro-config/oif_fix_dest_typo/+merge/163670 [09:57] nic-doffay, not really, are you sure there's nothing on top of what you're looking at that eats the events? [09:58] Saviq, most likely. Any idea who is best to ask about this. [09:58] nic-doffay, you mean it works fine on desktop, broken on device? [09:59] nic-doffay, just put a MouseArea at the bottom of Shell.qml and see if inClicked gets fired [09:59] *onClicked [09:59] Saviq, will try. [09:59] nic-doffay, we never had any such issues, anything I could try on one of mine devices? [10:00] Saviq, https://code.launchpad.net/~unity-team/unity/infographics-with-lightdm [10:00] If you have time could you do the review too? [10:01] nic-doffay, I'll try and get to it soon, yeah [10:01] ta Saviq [10:23] Saviq, who wrote the shell? [10:23] There's a comment I'd like to enquire about. [10:25] nic-doffay: lots of people contributed. Which comment, or in what part of the shell? [10:26] greyback, Shell.qml line 517 [10:27] nic-doffay: "his should be handled in the input stack, keyboard shouldnt propagate" ? [10:27] greyback, yep [10:27] greyback, basically trying to figure out why the touches are being swallowed on the infographics. [10:29] nic-doffay: it's there because at the moment, the shell receives _all_ touch events, even those destined for the application. For the case when the OSK was open on top of the shell, we wanted to stop the shell reacting to a keyboard press. Hence that mousearea, which is the same size as the OSK. [10:30] nic-doffay: I doubt that mousearea is to blame for your issue anyway [10:39] mzanetti, do you remember what to do for cobertura to find the source files? https://jenkins.qa.ubuntu.com/job/unity-api-raring-i386-ci/13/cobertura/src_unity/Exception_cpp/ [10:40] Saviq: use the latest build [10:40] mzanetti, that is the latest build, no? [10:42] Saviq: hmm... it seems it is.. [10:42] greyback, any idea how to propagate the touch events to the infographics? [10:42] Saviq: ah... I guess the workspace is not exported to the public jenkins [10:43] mzanetti, ah, might be [10:43] nic-doffay, does it work on your desktop and not on the device? [10:43] mzanetti, yeah [10:44] Saviq, yeah [10:44] nic-doffay, I will try in 5 mins, bear with me [10:44] Saviq, I'm used onPressed. [10:44] The branch a linked you to has onClicked. [10:44] *using [10:45] nic-doffay, both should work [11:13] nic-doffay, it works just fine here on a Nexus 10 [11:13] Saviq, O_o [11:15] nic-doffay, only issue I can see straight away is that it changes the data source twice for some reason [11:16] Saviq, I'd ask pete-woods about that. [11:16] I need to find out why it's not working on my device too. [11:16] nic-doffay, i.e. for each user switch in tablet mode I get two "Data sets for..." one "old" and one "new" [11:17] and there's an error, too: [11:17] Greeter/Infographics.qml:75: TypeError: Cannot read property 'presentUnlockAnim' of null [11:17] Saviq, I'm using a Galaxy Nexus. [11:17] nic-doffay, mine's charging, will try as soon as its usable again [11:17] Saviq, that error is also related to the data sets from what I recall. [11:18] there's some blocking, too, when switching users [11:18] greyback, any idea why touch events wouldn't be working for the infographics on a Galaxy Nexus, but working on Saviq's Nexus 10? [11:18] i.e. the user doesn't switch smoothly, it jumps once or twice === MacSlow is now known as MacSlow|lunch [11:18] nic-doffay, does everything work when you run shell from the image? [11:18] nic-doffay, when did you last flash? [11:19] Saviq, now. [11:19] nic-doffay, Saviq: the present unlockAnim error is one nic will have to sort out, it happens when the data set is empty - I'll investigate the double data set change issue [11:19] pete-woods, right. I'll handle that when I'm writing the tests then. [11:20] nic-doffay: is the MouseArea where it should be on the phone - check by putting a Rectangle inside it. I'd need to see code to investigate further [11:20] in terms of the non-smoothness, I know that changing user was very slow after mterry's phablet-greeter-lightdm branch for a while, then he added some performance-related change that helped a lot [11:20] greyback, I see no reason why not if it's in the correct place when running on the desktop. [11:20] it might still be in tat area [11:21] pete-woods, yeah we fixed that [11:21] pete-woods, and I'll merge that branch in today if everything's fine [11:21] cool [11:21] pete-woods, nic-doffay, so you'll be fine working on top of trunk [11:21] our's depends on that branch (being more lightdm stuff) [11:22] Saviq: woo! hopefully I can get to the bottom of the double change thing, then [11:22] nic-doffay: nor do I. Phone and desktop should behave the same. But there are differences, the GU definition for instance [11:22] Saviq, if it is fine then yes. I'd like to get to the bottom of this Galaxy Nexus issue though so I can check the frames of the circles. [11:23] If everything is ok then I'll begin writing tests. [11:23] And make amendments. [11:23] But I'd prefer to do that after potentially re-writing code for any GL related changes. [11:24] nic-doffay: see line 64 "if(presentCircles.count === 0 || pastCircles.count === 0) {" of Infographic.qml [11:24] that's where I "fixed" one of the empty data set issues [11:24] you might need to robust that up a bit [11:24] me not knowing QML at all [11:31] strange that the diff in launchpad is nonsense - it seems to be including the changes from mterry's branch [11:31] even though that's down as a prerquesite [11:32] pete-woods, yeah that happens sometimes, I'm about to approve mterry's branch so that should be gone soon [11:32] just checking if it works on the galaxy [11:32] okay, cool [11:39] katie, ping [11:40] saviq pong [11:40] katie, hey, can we ask you for a comment on https://code.launchpad.net/~nick-dedekind/unity/phablet-greeter-indicators/+merge/157330 [11:40] katie, re: tap on indicators in greeter opens device menu [11:40] sure [11:42] Saviq, is this in the latest build? [11:42] katie, no, it's not merged yet [11:43] katie, I wasn't sure we should merge it before that issue was raised [11:43] pete-woods, I'm doing some subtle changes to the way the animations are done. [11:43] I'll let you know when I've pushed it, mind checking if it's broken anything I'm unaware of? [11:44] Saviq, what would you like me to comment on in particular? [11:45] pete-woods, will there always be past circles? [11:45] I mean surely not? [11:45] Regarding your check on line 64... [11:46] nic-doffay: no - not if a particular user has no data yet, or you're on the guest account [11:47] Saviq, just realised what you're talking about :) will comment now [11:47] pete-woods, mind pushing a branch for me where there are no past circles? [11:47] nic-doffay: that was just a quick hack to stop it from crashing all the time [11:47] nic-doffay: just select the guest account [11:48] pete-woods, ok great! [11:48] I'll change some of the user's data so that they have some present and no past, though, just to test that situation [11:49] pete-woods, awesome cheers. I'd like to cater for every situation. [11:50] Even before testing. [11:51] nic-doffay: just to be clear about the API - it will always try and give you "complete" months, just the data values might be null - this way it carries the concept of the length of the month, and for days without any data [11:51] pete-woods, I see. I assuming 0 would be passed. I guess that's the reason so many things are breaking. [11:51] *assumed [11:54] nic-doffay, works fine on a Galaxy Nexus, too [11:54] Saviq, this is weird. [11:54] Any ideas? [11:54] Saviq, what's the frame rate like? === _salem is now known as salem_ [11:58] nic-doffay, seems fine [12:03] pete-woods, one thing I can see is bad for sure: nextDataSource() takes 250ms [12:04] nic-doffay: I'm getting two calls to setUesrname from the GUI in the API [12:04] pete-woods, it needs to be asynchronous [12:05] pete-woods, that might be the hiccup when switching users [12:05] Saviq: I'll have a look at it === MacSlow|lunch is now known as MacSlow [12:05] wonder what on Earth I could be doing [12:06] I'm only copying two lists with 30 floats in [12:09] pete-woods, generateFakeData() needs to be called as a slot [12:10] pete-woods, nextDataSource should be a signal [12:10] Saviq: okay, will make that change! :) [12:10] pete-woods, to which generateFakeData() is connected [12:10] pete-woods, this way nextDataSource() will return straight away [12:11] Saviq: I'm still concerned about why it's so slow, though [12:11] pete-woods, generateFakeData looks quite intensive [12:11] a lot of rand()s [12:12] Saviq: that's only called once [12:12] ah right [12:12] I get 50 ms on a desktop machine, which is still not great [12:12] pete-woods, /me can't read [12:13] pete-woods, yeah [12:16] nic-doffay, here's the dump from a single data switch http://pastebin.ubuntu.com/5664285/ [12:16] nic-doffay, at first it's fine, but as the count of circles grows we're getting past the 16ms mark [12:18] nic-doffay, pete-woods remember you can connect to the device with the QML profiler to see a lot of profiling data [12:19] nic-doffay, pete-woods, ./run_on_device; Analyze > QML Profiler (External) and give it the IP of your device === alan_g is now known as alan_g|lunch [12:19] Saviq, brilliant thanks. [12:20] Saviq, still stumped as to why the infographics don't have any touch events on my phone though. [12:21] nic-doffay, does the shell at all? i.e. put a huge MouseArea at the bottom of Shell.qml and console.log() onClicked [12:21] Saviq, how will I be able to test that one the phone? I won't see any output... [12:22] nic-doffay, with ./run_on_device you get all the output [12:22] nic-doffay, it's just like you'd run it locally [12:22] but it goes over adb+ssh to the device [12:23] Saviq, ah ok. [12:23] brb [12:24] Saviq: that doesn't work for me [12:26] maybe I'm doing something wrong [12:26] bash: cd: shell: No such file or directory [12:26] maybe that's the right error message [12:33] pete-woods, run_on_device doesn't? did you run_on_device -s first? [12:33] nope! [12:33] rtm :$ === mmrazik|afk is now known as mmrazik [12:45] Saviq: just trying to wrap my head round this asynchronous stuff [12:45] currently nextDataSource is a slot [12:45] you were saying it should be a signal [12:46] the UI reacts to the signals directly from the data i.e. the QAIM and label properties [12:47] pete-woods, yeah, but it's calling it as any other method, not as a slot [12:47] Saviq: so how do I call it as a slot? (completely have no knowledge of QML here) [12:47] pete-woods, so it's blocking until it return [12:48] pete-woods, if nextDataSource would be a signal [12:48] pete-woods, infographic.model.nextDataSource() would emit that signal [12:49] so I listen to my own signal? [12:49] pete-woods, yes [12:49] Saviq: right, okay, that makes sense, thanks! [12:50] pete-woods, but regardless, it shouldn't take so long indeed ;) [12:53] Saviq: I found the crunch point, it's the call to data.first_month.setVariantList(new_data->first_month.variantList()); [12:53] clearly I must be doing something dumb in there [12:54] I think I'm inadvertently copying the list about 5 times [12:54] something like that must be happening, yeah [12:57] Saviq: lightdm is merged! \o/ [12:57] mzanetti, indeed === Guest61231 is now known as mfisch === mfisch is now known as Guest83719 === dbarth__ is now known as dbarth [13:07] mterry: approved [13:08] thanks! [13:08] mzanetti, Saviq: thanks, sorry that took so long! [13:08] mterry, right back at you :) [13:08] mterry: no worries. it was a quite complex merge [13:09] mterry: do'h! now that its merged I've found a small issue :D [13:09] mterry: I think we don't need qtdeclarative5-xmllistmodel-plugin in the deps any more [13:09] mzanetti, hah [13:09] mzanetti, I believe we used it elsewhere [13:10] mzanetti, at least we did back when I started it [13:10] maybe things have changed in the last months ;) [13:10] mterry: hmm... I remember adding that dep for the greeter... but I'll check again and will clean it up if not needed any more [13:13] alesage, Howdy, it seems we have a branch not autolanding: https://code.launchpad.net/~ted/hud/context-api/+merge/158939 [13:15] alesage: hi [13:18] greyback: any joy ? (segfault, efence) [13:19] kgunn: some, that old segv has been cured anyway. I'm hitting the issue racarr mentioned in his mail, re who gets the Surface first. [13:20] kgunn: I'm looking more into that, and learning the code on the way [13:20] pete-woods, all the dots are filled for the first infographic too. Any idea what's going on? === olli__ is now known as olli [13:21] Cimi, tsdgeos I'm getting this when running tests: http://paste.ubuntu.com/5664445/ [13:22] greyback: cool, feel free to pester alf & alan_g in #mir [13:22] dandrader: your sdk is old [13:22] oops/#mir/#ubuntu-mir [13:22] tsdgeos, ok. will dist-upgrade [13:22] kgunn: yep understood [13:22] dandrader: make sure you have the sdk ppa [13:23] tsdgeos, http://ppa.launchpad.net/phablet-team/desktop-deps/ubuntu not enough? [13:23] no clue [13:23] not what ./build adds [13:23] tsdgeos, you don't use that ppa? [13:24] i do [13:24] but the sdk doesn't seem to be there [13:24] is there? [13:25] tsdgeos, seems not. I must be getting it from raring itself [13:26] * dandrader checks the ppas added by the build script === alan_g|lunch is now known as alan_g === sergiuse1s is now known as sergiusens [13:29] mzanetti, we're using it for the video preview [13:29] mzanetti, xmllistmodel, that is [13:30] Saviq: ah. thanks [13:30] dandrader: that's not good enough anymore, get the ppa [13:31] tsdgeos, yeah, I'm running "./build --setup" now. lot's of packages being installed or upgraded (I also didn't have qt-edgers ppa) [13:36] Saviq, I'm not getting any output with ./run_on_device ... [13:36] Adding a MouseArea test to shell works on the desktop too. [13:40] nic-doffay: I don't know what would cause all the dots to be filled [13:40] it's not something I've been able to reproduce [13:40] nic-doffay, what do you mean no output? none at all? or just nothing you'd expect? [13:40] Saviq, none at all. [13:41] nic-doffay, did you ./run_on_device -s first? [13:41] Saviq, no, what does -s do? [13:41] setup [13:41] nic-doffay, read CODING [13:42] nic-doffay, or at least ./run_on_device -h [13:42] Will do ta Saviq === mardy_ is now known as mardy [13:48] Saviq, as requested: https://code.launchpad.net/~dandrader/unity/phablet_move_test_targets/+merge/163715 [13:49] dandrader, aactually... do we actually still have the "check" target? [13:49] dandrader, I think that might be why alltests failed, 'cause we dropped the "check" target that was make-specific [13:50] s/failed/didn't run everything/ [13:52] Saviq, right, there's no "check" target now. [13:53] dandrader, problem is "test" is not really a target [13:54] dandrader, that's why we had a custom "make test" target [13:54] dandrader, if you can find a way to do "$generator test", that would work then [13:55] Saviq, I'll add a commit to this MP that removes "check" and "alltests" [13:55] dandrader, k [13:55] Saviq, after that, what else would be missing or needed [13:55] ? [13:56] dandrader, nothing, if we can live without the "alltests" target [13:56] Saviq, without the "check" target it doesn't make sense anymore [13:57] dandrader, yeah I agree [13:59] MacSlow: i totally reproduce the problem in my local chroot [13:59] chroot pow3r [13:59] tsdgeos, hm [14:00] i'm going to randomly guess the need for a waitForRender call [14:00] tsdgeos, before the mouseClick()?! [14:01] Saviq, done [14:02] MacSlow: yep [14:03] tsdgeos, waitForRendering(object) ?! [14:03] MacSlow: see ./tests/qmltests/Dash/tst_PeoplePreview.qml [14:03] guys, I'm still having trouble running the run_on_device script [14:03] tsdgeos, ./Dash/tst_PeoplePreview.qml yes.... [14:03] Saviq, actually, without alltests tests/plugins/Utils won't be run by any "top-level" test target. should it be added to qmlunittests? [14:03] I deployed networking and ssh [14:03] and ran the run_on_device -s stage [14:03] but running the shell I get a strange error [14:03] dandrader, won't it run with "test"? [14:04] MacSlow: i'm randomly guessing that what happens is that at the second 000000 you are getting both buttons at the same position since they haven't been layouted yet [14:04] if that's the case, it makes qmltest a bit too flacky [14:04] but oh well [14:04] saying debug port option not recognised, but split over like 30 lines, one character each [14:04] tsdgeos, going to try now... thanks for the hint [14:04]