/srv/irclogs.ubuntu.com/2013/09/11/#ubuntu-unity.txt

Saviqdidrocks, we got a release! awesome man, thanks!01:54
didrocksSaviq: yw ;)01:57
didrocksyou should go to bed as well!01:57
didrocksSaviq: you missed some fun though01:57
didrockslike launchpad ppa not uploading the binary package01:57
didrocks(beeing in some kind of freeze to add arm64…)01:57
=== salem_ is now known as _salem
Mirvbregma: https://bugs.launchpad.net/unity/+bug/1223671 seeing FTBFS now that cu2d starts to be about operational again05:58
ubot5Launchpad bug 1223671 in Unity "Unity FTBFS week 37, failing tests" [Critical,New]05:58
=== duflu_ is now known as duflu
mzanettiveebers: hii06:29
mzanettilol... I like how Saviq puts "Needs Information" comments on his own merge requests06:58
mzanettigreyback: good morning08:48
greybackmzanetti: hye08:49
mzanettigreyback: I messed up with the review (once again) :D08:49
mzanettiapp.desktopFile doesn't exist any more08:49
mzanettiI'll fix it08:49
greybackmzanetti: oh dear08:49
mzanettigreyback: but I guess now it's the time to switch to the real model stuff anyways08:50
mzanettigreyback: also, I need to align the items in the launcher to the dash.08:50
mzanettigreyback: is it true that as long as an app is contained in the appmanager model, it's shown in the dash too?08:51
greybackmzanetti: yes08:51
mzanettiso basically I just display everything in there08:51
mzanettiok. /me is on it08:51
greybackack08:52
Saviqmzanetti, :PPP08:56
mzanettiSaviq: ?08:56
Saviqmzanetti, "Needs information"08:56
mzanettihaha08:56
mzanettiSaviq: yesterday evening, after flashing my device with -b it took like half an hour to compile unity and it went really hot...08:58
mzanettiSaviq: so I decided to give cross compiling a go. But seems I'm hitting a bug in Qt's cmake files. Do you know anything about it?08:59
Saviqmzanetti, oh yes!08:59
mzanettiSaviq: I thought we could quite easily set this up... just extracting a pbuilder tarball in /opt. I've prepared a toolchain.cmake file already... so run_on_device could quite easily be changed to cross-compile, given we manage to get around that bug09:00
Saviqmzanetti, there's people working no this09:01
Saviqmzanetti, you'09:01
Saviqve got mail09:01
mzanettiSaviq: hmm... my approach would have been to go with just an armhf chroot for the build root and using the cross gcc from the host system.09:04
mzanettiSaviq: from what I understand in that mail they want to do it for real in the desktop system by supporting all multiarch stuff in there09:04
mzanettiSaviq: which might be quite nice for us. But might be an additional issue when trying to get this stuff running on e.g. windows09:04
mzanettior other distros09:04
Saviqmzanetti, well, yeah, it's a more complete solution, where you can easily build a package etc.09:05
mzanettiSaviq: you can do that in the chroot too09:05
Saviqmzanetti, I know, that's what's happening there, although a more "proper" way maybe - obviously the net result would be to be able to cross-build in buildd / PPAs09:06
Saviqmzanetti, either way, it worked for me after having tweaked some things09:06
mzanettiSaviq: so you patched the Qt cmake files to actually regard CMAKE_PREFIX_PATH?09:07
Saviqmzanetti, https://code.launchpad.net/~unity-team/unity8/cross-building/+merge/179248 shows what I changed in unity8's code, and an email from me in that thread describes what else needed to happen09:08
Saviqmzanetti, the MultiArchCross.cmake fixes are in distro09:09
mzanetticool... I'll give it a try in some lazy evening09:09
Saviqmzanetti, I had to hardcode CMAKE_INSTALL_LIBDIR09:09
sil2100greyback: hi!09:37
sil2100greyback: I see yesterday unity8 got published - all was ok?09:37
greybacksil2100: there are still some branches to go. But yeah getting there09:41
* greyback back in 4010:10
Saviqnic-doffay, here?10:14
SaviqMirv, ping10:15
MirvSaviq: pong10:37
SaviqMirv, hey, just wanted to let you know if you want to save time on builders - https://lists.ubuntu.com/archives/ubuntu-devel/2013-June/037380.html10:37
SaviqMirv, when I built qtdeclarative yesterday, it took some 20 mins on my laptop to just optipng them...10:38
mzanettiSaviq: as gerry is away, https://code.launchpad.net/~mzanetti/qtubuntu/register-application-interface/+merge/18501010:38
SaviqI can only imagine how long that takes on armhf builders...10:38
MirvSaviq: ah yeah sorry, I noted the link but didn't reply. indeed that export NO_PNG_PKG_MANGLE=1 is worthwhile in some cases.10:38
Mirvoptipng is really, really slow10:39
SaviqMirv, obviously if there's a separate package with assets (that's Architecture: all) the gain is smaller10:39
Mirvfor either doing more rapid Qt builds or alternatively permanently for our upstream projects10:39
=== Guest72454 is now known as karni
SaviqMirv, but still worth thinking of10:40
Saviqmzanetti, happrovd10:40
mzanetticheers10:40
MirvSaviq: yes, it could be helpful for the daily release of some packages (or all)10:40
=== alan_g is now known as alan_g|afk
=== MacSlow is now known as MacSlow|lunch
MirvSaviq: btw we shouldn't have panda builds anymore, which is great :) but the optipng takes ages on even high-end x86 in some cases10:48
SaviqMirv, yeah10:48
SaviqMirv, that was sent back when we didn't have the calxeda setup yet ;)10:48
=== alan_g|afk is now known as alan_g
mzanettiSaviq: and another one: https://code.launchpad.net/~mzanetti/unity-mir/register-app-interfaces/+merge/18501711:01
Saviqmzanetti, happroved11:02
Saviqseb128, I forgot so many times... you know that settings app is laid out wrong on maguro? it seems to use a preset width for the items in the grids11:29
Saviqseb128, and they're not horizontal-aligned evenly11:29
Saviqseb128, also, "Language & Te..." is cropped here11:29
seb128Saviq, no, I don't know about that ... do you have a screenshot?11:29
Saviqseb128, sure, sec11:29
seb128Saviq, thanks11:30
seb128Saviq, bonus point if you file a bug (https://bugs.launchpad.net/ubuntu-system-settings/+filebug) and include it11:30
Saviqseb128, bug #122382711:34
ubot5bug 1223827 in ubuntu-system-settings (Ubuntu) "Home page misaligned on Galaxy Nexus" [Undecided,New] https://launchpad.net/bugs/122382711:34
Saviqseb128, we actually have a ResponsiveGridView for such use cases in unity811:35
Saviqseb128, I always thought it could earn its place in the SDK11:35
seb128Saviq, I wonder what is wrong with the standard grid :/11:36
Saviqseb128, it's not dynamic11:36
Saviqseb128, you need to set the item width / height manually11:36
Saviqand the item count per row11:37
seb128Saviq, well we have "columns: width / unitys.gu(<width_item>)" and width: <width_item> for the elements11:40
seb128Saviq, it does work/wrap correctly on resize on the desktop11:40
seb128Saviq, but indeed you have cases where it eats a marging11:40
Saviqseb128, then maybe there's just not enough space so that it misaligns?11:41
seb128Saviq, right, that logic works fine where the width a multiple of the item width11:41
seb128is a multiple*11:42
Saviqseb128, yeah, you need to adapt the item width size11:44
seb128Saviq, you need to fix a variable somewhere...11:44
Saviqseb128, and work with a minimum / maximum11:44
Saviqseb128, sure, have a look at ResponsiveGridView in unity811:45
seb128Saviq, so ranges?11:45
seb128if resolution is in [..] 4 items11:45
seb128in [...] 5 items11:45
seb128etc?11:45
Saviqseb128, not resolution, but width11:45
seb128then have width_item = width/number11:45
seb128Saviq, right11:45
Saviqseb128, minimum_width <= item_width < maximum_width11:45
seb128Saviq, I start really disliking qml btw :p there is so much hackish stuff to do/things the toolkit doesn't do for you11:46
Saviqseb128, and you minimize the number of columns having that condition in mind11:46
seb128it's fun for nice demos11:46
seb128but as soon as you need things to behave in a solid way it's piling of hacks11:46
Saviqseb128, lol11:46
Saviqseb128, sure it's not complete (hence our own SDK)11:47
Saviqseb128, did you see 5.1 layouts, though?11:47
Saviqseb128, it's just not complete yet :)11:47
seb128Saviq, thanks for the hint, I'm going to have a look to ResponsiveGridView11:47
seb128Saviq, yeah, I saw that, I wish we would be on 5.1 already :p11:47
seb128Saviq, sorry for the ranting, I just had a very frustrating day yesterday trying to parse simple json to display items in a listview11:49
seb128Saviq, it's like you need to go to cpp, do manual parsing, do abstract model ... it's all crazyness and too hard for a simple usecase11:49
seb128where XmlListModel is great, it just doesn't exist for json :/11:50
Saviqseb128, did you write a JsonListModel, then? ;)11:51
Saviqseb128, http://qt-project.org/wiki/JSONListModel11:51
seb128Saviq, there is one in https://github.com/kromain/qml-utils/tree/master/JSONListModel I pondered just copying the qml insource11:52
seb128Saviq, but then I've hundred of lines of code copied from somewhere and virtual unmaintained11:52
Saviqseb128, right11:53
seb128Saviq, I ended up doing a List<QObject *> and turn that to a QVariant, which works but doesn't give me easy sorting of the model or dynamic update ... Laney is going to turn that into a QAbstractListModel that should do the job, it's just a non trivial api and I tried to go for the easy stuff at first :p11:54
seb128Saviq, https://code.launchpad.net/~seb128/ubuntu-system-settings/storage-backend-list/+merge/184967 btw if you are curious/ have comments ;-)11:55
Saviqseb128, QAbstractListModel is only scary at first11:55
Saviqseb128, but for general use it's not *that* big of a deal11:56
seb128Saviq, well, google let me land on http://doc.qt.digia.com/4.7/qdeclarativemodels.html#c-data-models11:56
seb128Saviq, which recommends the QList<QObject *> before the qabstractmodel, saying to use that one only if the easy case doesn't work11:56
seb128Saviq, I feel like I got misleaded to spend time on the wrong solution :p11:57
seb128Saviq, that's part of learning the toolkit though, so no big deal11:57
Saviqseb128, indeed11:57
seb128Saviq, it just that people try to sell you qt/qml as an easy way to do stuff, but at this end it's as hard with any other toolkit11:57
=== alan_g is now known as alan_g|lunch
Saviqseb128, sure, the backends need to be there11:57
Saviqseb128, and it obviously depends on the UI you need11:58
=== MacSlow|lunch is now known as MacSlow
Saviqseb128, in your case there's just more work on the backend than the UI11:58
seb128right11:58
Saviqseb128, so in that sense it's just as any other indeed11:58
seb128though my backend need at not crazy ones11:58
seb128I just need to display a list of items from a json model11:58
seb128at->are11:58
seb128anyway, moving on11:59
Saviqseb128, yeah, it's not Python (yet) ;)11:59
seb128Saviq, thanks for listening to my semi-ranting and still providing constructive hints ;-)11:59
Saviqwhere there's barely anything you're the first one doing11:59
seb128Saviq, right, it has potential, as said the xmlmodel magic is awesome11:59
Saviqwith python someone somewhere did what you need already in 99% of the cases, with QML we're the ones ploughing through12:00
Saviqseb128, and yeah, well... as soon as you have the second place where you need to parse json... maybe looking at the Json model is not that bad...12:01
seb128yeah, I was somewhat hopping that more of the heavy lifting would have been done by the other people who have been using it before us12:01
Saviqwe might even try to push it into Qt or our SDK at least12:01
Saviqseb128, that's the problem - there really isn't that much use of QML out there still12:01
Saviqseb128, with unity-2d we were one of the biggest projects with it at that time12:01
seb128right12:01
Saviqmaybe not so much nowadays with Blackberrys and such12:02
seb128which is why I though more would be done12:02
seb128I though that with the BB and Jolia out there we would have a bit more of the "standard usecase" covered12:02
seb128but they didn't use it for that long either12:03
seb128it's going to get there, eventually ;-)12:03
seb128Saviq, https://bugreports.qt-project.org/browse/QTBUG-12117 btw12:03
seb128doesn't seem to be actively worked on atm though12:04
mzanettiSaviq: just saw this stop by in the debug output: file:///usr/share/gallery-app/rc/qml/Utility/UbuntuApplicationWrapper.qml:19:1: module "Ubuntu.Application" is not installed12:05
mzanettiI guess we have a problem :/12:06
greybackmzanetti: yep, saw it too. It's on my list12:06
Saviqmzanetti, we renamed to Unity.Application?12:06
Saviqdang12:06
Saviqlet's push to fix bug #1186556 then12:07
ubot5bug 1186556 in qtubuntu "Browser does not respond to Qt.openUrlExternally(link)" [Critical,In progress] https://launchpad.net/bugs/118655612:07
Saviqso that we can get rid of those completely12:07
mzanettigreyback: another issue (not high priority tho) is that the Mocked ApplicationManager doesn't use icon path's but only the appId12:08
mzanettigreyback: so right now, running on the desktop breaks icons in the launcher12:08
greybackmzanetti: ack. Could you log a quick bug for me please?12:08
mzanettiok12:08
greybacktnx12:12
=== _salem is now known as salem_
Saviqmzanetti, you up for a review of https://code.launchpad.net/~saviq/ubuntu/saucy/qtdeclarative-opensource-src/add-qtquick-delegate-range/+merge/184774 ?12:19
Saviqmzanetti, obviously the .patch file does not need a review, as it's applied to the code already12:20
=== hikiko is now known as hikiko|lunch
greybackSaviq: I hold off on this, to encourage gallery folks to use Qt.openUrlExternally:  https://code.launchpad.net/~gerboland/gallery-app/gallery-app-fix-appman/+merge/18503012:30
Saviqgreyback, thing is... it's not ready yet12:31
Saviqgreyback, https://code.launchpad.net/~aacid/platform-api/papi.rules.typo/+merge/18235412:31
Saviqhttps://code.launchpad.net/~aacid/qtubuntu/qtubunturl/+merge/18175212:31
Saviqgreyback, and with tvoss away, ricmm on a sprint12:31
greybackSaviq: ah, I thought it was. Ok12:32
Saviqgreyback, won't get done before next week, so let's just fix it12:32
greybackSaviq: ack12:32
Saviqgreyback, thanks12:32
mzanettiSaviq: I hoped not having to review this :D12:35
=== alan_g|lunch is now known as alan_g
Saviqmzanetti, ;D12:42
Saviqmzanetti, why? it's actually pretty small12:43
Saviqmzanetti, but if you refuse, I'll find someone else ;)12:43
* mzanetti never refuses12:43
* Saviq notes that down12:43
mzanettihaha12:43
mzanettiSaviq: ok... I'll try to get that done soonish.12:44
Saviqmzanetti, thanks12:44
veebersmzanetti: hey how are you doing?13:03
mzanettiveebers: dude... isn't it like 3am for you?13:03
veebersmzanetti: heh, I'm in Boston on a sprint ^_^13:03
mzanettiah... you're at the sprint13:04
mzanettiright13:04
mzanettiyeah... doing good now... didn't do so well in our hangout this morning. luckily noone showed up :D13:04
veebersmzanetti: oh the sync meeting? I thought that I deleted the ones for this week?13:05
mzanettiveebers: yes you did...13:05
mzanettiveebers: obviously I noticed that after getting up and ready for the hangout :D13:05
mzanetticoding till 1:30 am and hangouts at 8:30 am don't go well together13:05
veebersmzanetti: oh I see now :-P Sorry I should have emailed when I did it]13:05
mzanettiveebers: no... really not your fault13:06
mzanettianyways... how is Otto?13:06
veebersmzanetti: it's annoying me because I almost did send an email but didn't :-P Also, I've deleted the next one on Monday as I get back that day and will probably be in bed13:06
mzanettiack13:07
veebersmzanetti: ah right so I'll ping fginther about that today (now with the qa/ci split) I talked to him about it yesterday13:08
kgunngreyback: hey...so, i am assuming the pending image has unity-mir landed in it ?13:10
kgunnat least from the scrollback on #ubuntu-touch13:11
kgunni got on late...but misses asac13:11
kgunni assume there is a script to switch to mir ? from sf...13:11
kgunnor i should ask...how does one switch ?13:11
greybackkgunn: I've only landed the last piece there under an hour ago. The next pending image will have it in, for sure. I need to find out when it is generated13:12
kgunngreyback: ah..dang13:13
Saviqgreyback, we can always ask to trigger one13:21
Saviqgreyback, as there's only going to be one later today13:21
greybackSaviq: Good to know. When we're ready, we'll have one spun up then13:22
kgunnSaviq: curious...why only one later today ?13:23
=== hikiko|lunch is now known as hikiko
Saviqkgunn, there's only two images daily13:24
Saviqkgunn, we daily release 4 times a day13:24
Saviqkgunn, but only build an image twice13:25
kgunnSaviq: consider me educated :)....i didn't realize that...thot release lead to image13:26
kgunnrealizing now image means release plus load of tests13:26
kgunnand duh...i've only ever seen 2 images a day13:26
mzanettiveebers: fyi. we have the biweekly shell hangout now. if you wanna join13:27
sil2100bregma: hello!13:28
bregmasil2100, howdy!!13:28
veebersmzanetti: I'm keen, just in a meeting. Can you link me and I'll try pop in13:29
mzanettiack13:29
sil2100bregma: sorry to ping again with bad news, but it seems unity FTBFS for amd64 and i386 because of failing unit tests13:29
sil2100bregma: https://launchpadlibrarian.net/149963225/buildlog_ubuntu-saucy-amd64.unity_7.1.0%2B13.10.20130911.2-0ubuntu1_FAILEDTOBUILD.txt.gz13:29
bregmasil2100, https://bugs.launchpad.net/unity/+bug/122356113:30
ubot5Launchpad bug 1223561 in Unity "Some TestLauncherEntryRemote unit tests are failing with new libdbusmenu changes" [Undecided,Confirmed]13:30
sil2100Right! How can we get this fixed?13:30
bregmawe have to figure out a way to rewrite the tests so it works with the new libdbusmenu13:31
bregmait's always a problem when folks justthrow their changes over the fence and run13:31
sil2100heh, right13:32
Saviquh oh13:32
SaviqI broke hangout :D13:33
Saviqmzanetti, can you drop out and back in, otherwise it won't let me in :D13:33
greybacklol13:33
Saviq"You can't join this hangout because you blocked one of the users. You can make a new one instead."13:33
Saviqd'oh13:33
Saviqmzanetti, well, it's a change to how the right edge behaves13:54
Saviqmzanetti, as now you just pull a single app in in one gesture13:54
mzanettiSaviq: in that prototype I can't see any difference to that really13:55
Saviqmzanetti, the first one was broken13:55
mzanettiSaviq: might be because mouse handling is somewhat broken tho13:55
Saviqmzanetti, lp:~willow-team/willow/RightEdgeNavigation13:55
mzanettiah. was the second link any different?13:55
Saviqyes, no "Phone"13:55
mzanettiah...13:55
mzanettiwow13:56
mzanettilooks amazing13:56
mzanettino clue how usable it will be... but it looks good :D13:56
=== dandrader is now known as dandrader|afk
greybackmzanetti: do you know: what listens for a bottom edge swipe to show the application toolbar?14:16
greybackmzanetti: I thought it used to be the bottomBarVisibilityCommunicatorShell14:16
mzanettigreyback: it's the toolbar itself14:16
mzanettigreyback: wait14:16
mzanettigreyback: you talking about apps? or the black one in the shell?14:16
greybackmzanetti: I open a web browser, the toolbar hides. I want to show that toolbar again14:17
mzanettigreyback: yeah... it's the Panel itself... it creates a 2 gu area at the bottom14:17
mzanettigreyback: timp might be able to provide details14:18
=== dandrader|afk is now known as dandrader
=== alan_g is now known as alan_g|tea
=== 3JTAAEL6B is now known as tvoss
=== jono is now known as Guest91193
Cimihey tedg14:54
tedgHowdy Cimi14:54
veebersmzanetti, Saviq: you have a moment? I'm having issues running a test script using qmlscene when unity8 is using mocks.15:00
mzanettisure15:00
veebers I have a desktop file and I'm using the desktop_file_hint. It works when I run the command [1] under a `start unity8` session, but when I export the mock paths etc. (i.e. what the autopilot tests do) I don't see the application pop up15:01
veebersdoes that sounds weird to you?15:01
=== alan_g|tea is now known as alan_g
Saviqveebers, on the device?15:01
veebersSaviq: yes15:02
Saviqveebers, or regardless - yeah, we're still using mocked applications15:02
Saviqveebers, i.e. placeholder images15:02
veebersSaviq: oh right, does that mean that I can't run an actual application while under those conditions?15:02
Saviqveebers, obviously we need to move away from that15:02
Saviqveebers, yes, if unity8 is using the Unity.Application mock15:03
mzanettiwell... we need to keep that in the mock15:03
mzanettibut autopilot is NOT supposed to use the mock15:03
Saviqyeah, ultimately it should not15:03
Saviqbut we just never got around to switching to the real backend for it15:03
veebersmzanetti: oh, so are the autopilot tests doing something wrong then? (with the exports for QML2_import_path and ld_library_path)15:04
mzanettiyes15:04
veebersmzanetti: rats, right so I need to figure out what it should actually be doing15:04
mzanettiveebers: well... for the lightdm stuff we don't have the real thing in place yet. so that's why we introduce the mocks to be able to start writing tests15:04
mzanettiveebers: but the mocks are supposed to be used for unit tests (qmltestrunner)15:04
mzanettiveebers: autopilot supposed to do integration tests => test the real thing15:05
Cimitedg, hay, was wondering if there was some update on the wifi plugin15:05
veebersmzanetti: aye, correct. I guess stuff was pulled over with the revamp of the autopilot suite15:05
veebersmzanetti: http://pastebin.ubuntu.com/6092858/15:05
mzanettiyep15:06
tedgCimi, Uhm, nothing really news worthy.  Working on it right now infact :-)15:06
veebersthat is in effect what we're doing15:06
veebersmzanetti: so is any of that needed?15:06
mzanettiveebers: so in the mid-term that should go away for autopilot15:06
Cimitedg, good15:06
tedgCimi, Connecting models to delegates to factories to.... and wondering when this gets easy :-)15:06
Cimitedg, second question, dednick as well… I did some work on system components, to become indicators. It's missing backends15:06
veebersmzanetti: alright, so no env exports needed then? or do we still need the lightDM stuff?15:06
mzanettiveebers: the qml2_import_path should not be there... for the others you'd need ask mterry on the current state and plans15:07
Cimi(date time, users, media player etc etc...)15:07
Cimiwhat do we do for them?15:07
veebersmzanetti: awesome will do. Thanks15:07
tedgCimi, ?  I'm confused.  You did the QML widgets?15:07
mzanettiveebers: I'm not entirely sure how to autopilot the real thing with lightdm...15:07
mzanettiveebers: so that might stay an exception (altough I'd prefer not)15:07
tedgCimi, My guess would be that you need to put them in dednick's MenuItemFactory class to instantiate them.15:07
Cimitedg, https://code.launchpad.net/ubuntu-settings-components15:08
mterrymzanetti, veebers:  I think I 'm missing something in the scrollback due to disconnects.  What's up?15:08
veebersmzanetti: ack. thanks15:08
mzanettimterry: the fact that our autopilot tests use mocks is bad15:08
mzanettimterry: and we should steer away from it instead of using more of them15:08
tedgCimi, Yeah, so they need to get used by dednick's code in Unity8, no?15:08
mzanettimterry: so the question is what the plans are in regard to autopilot the real thing with lightdm15:09
tedgCimi, By backend you mean plugged into Unity815:09
mzanettimterry: which I agree can be somewhere between extremely tricky and impossible15:09
mterrymzanetti, I always figured lightdm was a reasonable thing to mock, because we don't want to set up machines with real accounts that we'd need, eh?15:09
mzanettimterry: actually we should... in theory15:09
Cimitedg, yep I think15:10
mterrymzanetti, is our QA machinery even remotely able to?15:10
dednickCimi: components for indicators?15:10
mzanettiveebers: ^15:10
Cimidednick, yes15:10
dednickthey need to go in the Indicators.MenuItemFactory to be instantiated by the backends15:10
dednickCimi: it has a string to component mapping15:11
mzanettimterry: there's not muck point of integration tests when mocking the integrational parts :) when just testing the greeter code itsel, qmltests would be way better15:11
mzanettimterry: that's the main thing behind the discussion15:11
veebersmzanetti, mterry thats a good question15:11
Cimidednick, are you planning to come to the office in next days? would be better to chat live15:11
mterrymzanetti, well...  you can still test other bits :)15:11
tedgdednick, Is it possible we could put that factory in the library as well?  I've copied it for system settings, but it seems like it'd be better to include.15:11
dednickCimi: didnt have any plans to15:11
mterrymzanetti, I think we tend towards qmltests for the greeter itself15:12
mzanettibut as I said... I have no clue yet how to autopilot something with logging out the user15:12
veebersmzanetti, mterry are we able to setup the accounts stuff programatically?15:12
mterrymzanetti, we need a robot arm driving the machine  :)15:12
veebers^_^15:12
mzanettidon't we have students?15:12
dednicktedg: mmm. it's very unitymenumodel speficific in the factory.15:12
mterryveebers, uh, I mean, I could probably write a script to do so...  I just don't know if we have capabilities to run such things sensibly with our QA setup15:12
tedgdednick, Sure, but do we expect anyone to use System Components without Unity Menu Model?15:13
dednicktedg: not all the setttings have indicator backends15:14
tedgdednick, The good ones do ;-)15:15
tedgdednick, But what I'm saying is that they're still going to be pulling in that lib, they'd just not use the factory in that case.15:15
dednicktedg: ;) i guess the plugins for settings which use indicators will use the same thing15:15
tedgdednick, So we're not adding a dependency or anything they don't already have.15:15
dednicktedg: yeah i guess so15:15
dednickCimi: i was working on getting the ubuntu-settings-components package into a workable form so I could put all the indicator items in there.15:16
Cimidednick, great, can I help?15:16
veebersmterry: well we can incorporate it into the autopilot tests itself. What specific capabilities are you concerned about?15:17
veebersi.e. what capabilities are there (that I'm not aware of)15:17
dednickCimi: i've been pulled onto something else in the intrim, and it's not really in a fully working form yet. Are you wanting to do this immediately?15:18
Saviqmzanetti, veebers but the fact that we *are* using the application mock is on purpose15:18
Saviqmzanetti, veebers as no one looked yet at moving away from it15:18
Cimidednick, it was one of my tasks so I am happy to help15:18
Saviqi.e. making sure that no apps are left running when you exit etc.15:18
Saviqas AFAIK there's no way for autopilot on touch to handle that15:19
dednickCimi: ok, well i'll try get it into shape quickly and you can take over my branch maybe?15:19
veebersSaviq: understood. Perhaps for the specific tests I'm looking at I can do something different with the mocks (or perhaps this isn't the best place for the tests)15:19
Cimidednick, yes15:19
veebersSaviq: if you launch the app through autopilot it ensures that apps are killed at test end, but not if you 'click' on an application to launch it15:20
dednickCimi: ok, give me 30 minutes or so.15:20
Cimidednick, no problem15:20
Saviqveebers, I believe the blocker for us to move away from the application mock is autopilot's ability to kill the processes on touch15:20
Saviqveebers, yeah, exactly15:20
Saviqveebers, AFAIK there was some work towards that, but not completed yet15:20
Saviqor waiting for the new app manager under Mir, at least15:20
SaviqI *think*15:21
Saviqveebers, but obviously thomi knows more15:21
veebersSaviq: ack, right I'll touch base with him and see further. At this point I might see about changing the mocks for this test specifically. Thanks15:22
Saviqveebers, yeah, that might work, too15:22
mzanettiSaviq: greyback: https://code.launchpad.net/~mzanetti/unity-api/launcher-and-appmanager/+merge/18508515:32
Saviqmzanetti, 107+ /// @cond15:33
Saviq108+ void applicationManagerChanged();15:33
Saviq109+ /// @cond15:33
Saviqmzanetti, should be @endcond?15:33
mzanettiSaviq: yes15:33
Saviqmzanetti, hmm, didn't we establish this morning that this won't work?15:34
Saviqmzanetti, i.e. we need a common library with ApplicationManagerInterface for this to work?15:34
mzanettiSaviq: it works for the mock in this case15:34
mzanettiSaviq: as it doesn't really load the real appmanager plugin15:35
mzanettibut compiles the application mock itself15:35
Saviqmzanetti, yeah15:35
mzanettiSaviq: won't register this in the real launcherplugin ofc15:35
Saviqmzanetti, yeah, but that essentially means that we need the common library *now*15:35
Saviqmzanetti, for Unity.Application and Unity.Launcher to work together15:36
Saviqmzanetti, as we need to pass the ApplicationManager from Unity.Application to Unity.Launcher15:36
mzanettiSaviq: no... we don't really *need* it now. but it would be cleaner, yes15:36
greybackyeah it would be cleaner15:36
mzanettiSaviq: it works for now if the application plugin registers stuff15:36
Saviqmzanetti, even if both plugins compile their own copy of ApplicationManagerInterface?15:37
mhr3_pstolowski, pls https://code.launchpad.net/~mhr3/unity-scope-home/fix-1223940/+merge/18508815:37
Saviqinteresting...15:37
mhr3_eh, wrong channel15:37
mzanettiSaviq: yes... seems to do15:37
mzanettiSaviq: that's why I was a bit surprised that it didn't work when registering the type in main.cpp15:37
mzanettiI would have expected that to work15:38
Saviqmzanetti, btw, I asked thostr to invite you to a dash preview-related mtg tomorrow, I'll try to be there, but if not, you have my "power and trust" as thostr called it15:38
mzanettiack15:38
Saviqmzanetti, if I'm not there and you don't conclude, we'll revisit next week15:38
greybackwhy did I read that as "power and thrust" ?15:38
* mzanetti opens dict.leo.org15:39
mzanettihehe15:39
mzanettiSaviq: greyback: my favorite lines from that merge are lines 173 and 17415:40
greybackmzanetti: yeah, I had just noticed that15:40
Saviqlol15:40
mzanettistupid test... you got your chance to check the property... now its too late15:41
Saviqhehe15:42
greybackmzanetti: hey, you mentioned to me this morning I broke application icons in unity8 on the desktop. Which icons are broken, in the launcher?15:55
mzanettigreyback: your mixing up 2 things15:55
greybackso I broke 2 things? Oh yay :D15:55
mzanettigreyback: the one in the morning is that newly launched apps are not added any more to the launcher (on the current phone image that is)15:55
mzanettigreyback: the other isn't visible yet. That's only in the mock for the appmanager. icon() returns the appId instead of the "url to the icon"15:56
mzanettigreyback: and the latter one will only be visible when using the fake appmanager. so no biggie15:56
mzanettigreyback: the former one will be fixed by me in my upcoming merge that integrates the appmanager for real15:57
greybackmzanetti: oh ok, I'll not bother then15:57
mzanettiack15:57
greybackmzanetti: fake appmanager got running "./run --fake" right?15:58
mzanettigreyback: on the desktop you always get the fake app manager15:58
mzanettiso far15:58
mzanettibut yes, with --fake too15:58
greybackok, so when I just use "./run" I have no icons in the launcher (except dash)15:58
dednickSaviq: ping16:01
=== dandrader is now known as dandrader|lunch
* greyback afk for a few hours16:41
Saviqdednick, pong16:54
dednickSaviq: sorry, unping16:54
Saviqdednick, no worries :)16:55
* mzanetti is fed up with mocking the appmanager :D16:59
mzanettiSaviq: here it is... I hope you don't freak out when you see what I did in the launchermodeltest.17:05
mzanettihttps://code.launchpad.net/~mzanetti/unity8/launcher-and-appmanager/+merge/18510717:05
mzanettiSaviq: this will align the recent apps in the launcher with the ones on the dash \o/17:06
=== alan_g is now known as alan_g|EOD
dednickCimi: hm. ok, I'm not actually going to do anything with the components at the moment. It's just missing a qmldir file i think.17:19
dednickCimi: and qmltypes would help as well.17:19
bregmasil2100, we have a fix for #1223561 in, would be be possible to kick off an extraordinary head/unity daily build job?17:33
sil2100bregma: did it get merged in already?17:42
bregmasil2100, I got the merge mail17:44
dednickmzanetti: just locked my simcard. entering pin incorrectly too many times locks out puk as well...doh.18:06
=== tiagohillebrandt is now known as tiagoscd
=== tiagoscd is now known as Guest67365
mzanettidednick: really?18:11
mzanettidednick: that's news to me18:11
dednickmzanetti: apparently.18:11
mzanettidednick: what is "too many times" ?18:11
mzanettiapprox18:11
dednickmzanetti: well you're allowed incorrect puk 10 times i think18:12
mzanettiah... you entered the puk wrong 10 times in a row18:12
dednicki think it's just counting an incorrect "pin" after the 3 normal retries as puks18:12
mzanettioh...18:12
dednicki dont think i ever sent "puk", although i may be wrong18:12
dednicki dont think i want to try again :)18:13
mzanettithat might be a bug tho...18:13
mzanettiwell... don't know... could be that the specification for sim cards doesn't even support sending choosing what you enter18:14
dednicknow i need a new sim card... sigh18:15
dednickshops are closed18:15
dednickputting in an invalid pin makes ofono go crazy...18:15
dednickinvalid sim.18:15
=== dandrader|lunch is now known as dandrader
veebersmzanetti: hey are you still around perchance?18:44
mzanettiveebers: sort of18:47
mzanetti:)18:47
veebersmzanetti: ah wait it's late for you sorry :-\18:51
mzanettiveebers: no worries18:51
mzanettiveebers: what's the issue?18:51
veebersmzanetti, perhaps this is a question for Saviq maybe too. I'm wanting to tell if an application is in focus or not18:51
mzanettiveebers: qmltests? or autopilot?18:52
veebersI'm told that the application managment interface (via dbus) will be available when MIR lands, and not before18:52
veebersmzanetti: autopilot18:52
mzanettiveebers: that should be doable with latest trunk18:52
mzanettiveebers: give the ApplicationManagerWrapper.qml an objectName and get it with autopilot18:53
veebersi.e. the qml introspected properties will say visible == True, I assume because as much as the qml app is aware it is visible , it's just not the one on top18:53
veebersoh really? Cool, I'll give that a poke and try18:53
mzanettiveebers: the ApplicationManager has a property "focusedApplicationId"18:53
veebersmzanetti: nice, that should be what I want, thanks. Good that you're up late :-)18:54
mzanettiwell, its not that late18:54
veebersmzanetti: FYI I have a branch that removes the mocking stuff (except lightDM for now) from the autopilot tests18:59
mzanettiveebers: nice18:59
veebersmzanetti: also I had meant to ask you or Saviq (I think it was Saviq that added it) the UNITYSHELL_GSETTINGS_SCHEMA / Gio et al.19:00
mzanettiveebers: no ida about that19:08
mzanettiveebers: btw: https://code.launchpad.net/~mzanetti/autopilot-qt/fix-1218971/+merge/18475319:09
mzanettiveebers: if you have time... it's not urgent at all19:09
Saviqveebers, I didn't add it, but I might know what it's about :)19:17
veebersmzanetti: can do, might be a little later this afternoon19:19
veebersSaviq: ah ok, I assume that it's used to hide the launcher in a unity7 Desktop?19:19
Saviqveebers, aah I know now19:19
veebersi.e. on the desktop that a test is being run on (as opposed to the unity that 's actually under test)19:19
Saviqyes exactly19:19
Saviqveebers, in daily release, the tests are run in normal unity7 session19:20
Saviqveebers, and because of bug #1204480 we needed to disable it19:20
ubot5bug 1204480 in Autopilot "Display should include strut information" [Wishlist,New] https://launchpad.net/bugs/120448019:20
Saviqveebers, 'cause we couldn't rely on the available geometry19:20
veebersSaviq: oh ok odd :-\ Right I might just shuffle it around a little bit, add comments stating why it's there etc. Perhaps it needs to exist outside the tests themselves19:21
Saviqveebers, yeah, well, it's probably gonna go away soon enough, too19:22
veebersSaviq: ah nice, that's even better :-) Thanks19:22
Saviqveebers, as in when we start running real Unity8 with Mir on PCs19:22
Saviqveebers, when we can't have unity7 running of course19:22
veebersah true19:23
veebersSaviq: is there a nice way to say how many tests a directory of qml tests contain? i.e. how many tests are there in tests/qmltests/Dash/?19:26
Saviqveebers, I'm afraid not19:27
veebersmzanetti, Saviq: FYI I'm going to fire up some autopilot tests covering the Dash so we have qml and autopilot coverage of the Dash19:27
Saviqveebers, although we could improve that in the run_tests script19:27
Saviqveebers, awesome19:27
veebersI'm looking at a document that shows design/expectations etc. and there is a deficiency19:28
veebersSaviq: perhaps a grep <lines starting with function test_*> | wc -l should be a quick and dirty count right?19:28
Saviqveebers, better yet, look at the results http://s-jenkins:8080/job/unity-phablet-qmluitests-saucy/1661/testReport/%28root%29/qmltestrunner/19:29
veebersSaviq: ah nice, thanks ^_^19:29
=== dandrader is now known as dandrader|afk
=== dandrader|afk is now known as dandrader
=== salem_ is now known as _salem
Saviqkgunn, it might be difficult for me to join, dunno if I find a place good enough to talk20:37
kgunnSaviq: good grief man....20:39
kgunnoh yeah...forgot your at the convention20:39
kgunngo to bed20:39
Saviqkgunn, can it wait until next week?20:39
kgunnyeah20:39
Saviqkgunn, let's just talk it over in person20:39
kgunnyep20:39
kgunnhave a good convention Saviq20:39
Saviqkgunn, will do, soon, gotta catch a train 6am...20:39
kgunnmterry: ping20:43
kgunnmterry: so...pre mir-on-mir...assuming the current greeter is phone v1 greeter, will it/can it support custom background/wallpaper ?20:52
mterrykgunn, sorry20:54
mterrykgunn, back20:54
kgunnnp20:54
mterrykgunn, yeah, it supports it today20:54
kgunnsweet20:54
mterrykgunn, the system-settings doesn't let you change it yet, but I can give you commandline if you want20:54
mterrykgunn, also, we're just one branch-land away from using lightdm (albeit in just autologin mode)20:55
kgunnmterry: got it, jasoncwarner was thinking they were blocked by integration of lightdm for this...20:55
kgunnmterry: sweet20:55
mterrystill will need mir-on-mir for actual greeter stuff20:55
* kgunn single handedly bringing back the 80's w/ overuse of the word "sweet"20:56
kgunnmterry: sure...20:56
mterry:)20:56
* mterry goes afk for a few minutes20:57
* mterry is back21:10
jasoncwarnerhey mterry , thanks for that. Seb is on holiday for a bit, though so maybe sync with laney tomorrow on lightdm progress?22:07
=== _salem is now known as salem_
mterryjasoncwarner, lightdm progress?  OK22:17

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!