/srv/irclogs.ubuntu.com/2013/05/29/#ubuntu-app-devel.txt

Mirvmorning04:07
Mirvcould someone check/approve https://code.launchpad.net/~timo-jyrinki/ubuntu-ui-toolkit/qmlrunner_use_dpkg_architecture/+merge/166172 ? sergiusens reported a problem and it was indeed there.05:04
=== chriadam is now known as chriadam|away
mzanettihttp://wstaw.org/m/2013/05/29/plasma-desktopSR2564.png08:40
mzanettihttp://wstaw.org/m/2013/05/29/plasma-desktopMy2564.png08:40
mzanettifirst draft ^08:40
mzanettiThe Dialog would need to remove the space for the "text" if it is empty...08:41
mzanettishould I report a bug for that?08:41
t1mpmzanetti: which app is that?08:49
mzanettit1mp: fahrplan. lets you find trains/busses.08:50
t1mpmzanetti: we don't have a date/time picker yet in the UI toolkit so when we work on that we can re-use yours perhaps :)08:50
mzanettit1mp: I'm porting it from MeeGo08:50
t1mpmzanetti: ah cool. for which countries does it work?08:50
guschstill pending for review: https://code.launchpad.net/~schwann/gallery-app/gallery-modularity/+merge/16605408:50
mzanettit1mp: yeah, yesterday I had a discussion with zsombi about this. right now I'm talking with design and once they're happy with it I'll try to upstream it to the SDK08:51
mzanettit1mp: http://wstaw.org/m/2013/05/29/plasma-desktoptX2564.png08:51
DanChapmanHi folks, does anyone know how to delete/remove a tarball in python. I'm doing an autopilot tests for fileroller and need to delete the created archive in teardown, but os.remove claims no such file looked through the tar modules and cant find a way to remove an archive just to create/extract/addfiles etc.08:52
t1mpmzanetti: ns.nl would be useful for me :)08:52
t1mpmzanetti: currently the dialog always assumes that you have text and one or more buttons, so if you need something else (check with design), then file a bug/feature request on bugs.launchpad.net/ubuntu-ui-toolkit please08:53
mzanettit1mp: hehe... shouldn't be too hard... there seems to be a standard called HafasXml. If your provider uses that adding support would be quite straigt forward08:53
t1mpI wasn't aware that different companies in different countries are using a standard08:56
t1mpthat is surprising actually, especially since they are big companies08:56
mzanettit1mp: dunno... just looking at the implementation, half of the sites use the same parser (called ParserHafasXml) while the other half has custom parsers.08:56
netcurlihttp://www.hacon.de/unternehmen/referenzen/hafas/hafas-referenzen08:57
mzanettinetcurli: awesome :)08:58
mzanettitrenitalia suports it... interesting08:58
mzanettit1mp: seems you're lucky... ns.nl seems to support it09:00
=== gusch is now known as gusch|away
t1mpmzanetti: yes, its on that hafas-referenzen list :)09:03
mzanettit1mp: so in theory it should be just a matter of adding the website to some list. I'll check it out when I work on it next time. (spare time project)09:04
t1mpcool, thanks09:06
dpmmzanetti, that's pretty awesome. When you think the app is in a working state, please add it to https://wiki.ubuntu.com/Touch/Collection - looks like a cool app we'd like to showcase :)09:07
t1mpI'm planning to switch to using ubuntu on a phone that I really use (not just development)09:07
mzanettidpm: its already in polishing phase :) I expect it to be done by the end of the week09:07
dpmexcellent!09:07
t1mpindeed, excellent :)09:08
mzanettit1mp: +1. thats why I started all that porting mania. can't use the phone without the 5 apps I use regualry09:08
t1mpmzanetti: at the moment I am abroad and I'm not sure whether I can reliably disable data (except wifi) on the phone. if that doesn't work well it could cost me some unneccesary money09:09
mzanettioh... that's indeed a reason to be careful...09:09
mzanettit1mp: but... can you ENable data already?09:10
mzanetti:D09:10
* mzanetti still runs an image from end of last week09:10
t1mpmzanetti: I'm not sure. So far I only used ubuntu in a phone without a sim-card09:10
dpmt1mp, I've got a question on C++ plugins for QML if you've got a minute: I'm working on a personal project to decode QR codes, and I decided to use a C++ plugin for decoding. There is some existing library for Qt 4.x and I decided to port it to 5.0 and package it up as a plugin,09:11
t1mpdpm: yes I have a minute :)09:11
dpmgreat, so here's the question :)09:11
dpmt1mp, I've done the port and the packaging, but my C++ is very rusty, not to talk of my Qt skills. Does the change in http://bazaar.launchpad.net/~dpm/+junk/qzxing-plugin/revision/2 look sensible to you?09:12
dpmfor the .c and .h files09:12
dpmthose were the main changes I did, but as my Qt experience is limited, it felt a bit like stabbing in the dark at times :)09:13
mzanettidpm: you should keep the QObject *parent in the ctor09:14
t1mpdpm: I'll have a look09:14
t1mpQT += widgets in your pro file is needed if you use the "old" qt widgets if I'm not mistaken09:14
t1mpso if your gui will be qtquick you don't need that09:14
dpmthanks mzanetti. I think that gave me a warning IIRC, but I can try again09:15
mzanettirest looks sensible to me09:15
dpmgreat, thanks mzanetti09:15
t1mpI don't see a reason not to make the constructor call QObject constructor09:16
dpmt1mp, good point, I'm not using widgets, so I'll remove that include.09:17
mzanettidpm: for reference, I did exactly the same yesterday for this: https://gitorious.org/qml-process-element/qml-process-element/merge_requests/109:17
dpmah, interesting :)09:18
dpmt1mp, mzanetti so on line 13 I should basically do + QZXing::QZXing(QObject *parent) right?09:19
mzanettidpm: QZXing::QZXing(QObject *parent) : QObject(parent)09:19
dpmah, so keep it as it was before. Let me try09:20
mzanettidpm: oh... you changed it to inherit QQmlExtensionPlugin...09:20
mzanettidpm: that seems hackish...09:20
t1mpI just noticed that too09:21
mzanettidpm: better have another file/class plugin.h that does only the registering and the QZXing then only inheriting from QObject and doing the logic09:21
dpmI warned I wasn't experienced in QT ;)09:21
mzanettidpm: not saying that there isn't a way to make it work this way, but gets weird... especially once you want to register an additional class...09:22
t1mpdpm: if you have a look at the UITK code, we have there classes like UbuntuI18n in i18n.h/cpp, and we register all the classes that we need exposed to qtquick in plugin.cpp09:22
dpmmzanetti, I agree, I prefer doing it the right way. Just I didn't know what the right way was. This feedback is really useful09:23
dpmok, let me see if I can create those plugin cpp/h files now...09:45
nerochiarogrr, hate it when i start working in the morning and forgot to turn on irc09:48
dpmmzanetti, t1mp, does this look more sensible now? -> http://bazaar.launchpad.net/~dpm/+junk/qzxing-plugin/revision/510:12
mzanettidpm: purrfect10:13
dpmexcellent, thanks for the help!10:13
nerochiarotmoenicke: some updates: i'm using your latest library (the one with set Property) and the simple test we used yesterday, but with a TextArea doesn't cause the keyboard to bounce. so it's something in the notes-app, and i'm trying to figure out what it is10:33
tmoenickenerochiaro: ok cool10:34
nerochiarotmoenicke: in the notes app, when i hide the keyboard manually i see this debug print: QMaliitPlatformInputContext::imInitiatedHide() step 2, setProperty("focus", true)  ... << why is it set to true if i'm hiding the keyboard ?10:34
tmoenickenerochiaro: right, recompiling10:35
nerochiarotmoenicke: that's weird that even with that mistake it was working in the simple case. wtf ?10:36
tmoenickeindeed10:36
nerochiarotmoenicke: hmm, actually no, in the simple case i just checked and the focus wasn't being removed anymore. yesterday i think i had just time to verify that it wasn't bouncing back and that it programmatically setting focus to false and then true was doing the right thing regarding focus on the parents10:37
nerochiarotmoenicke: so it's all right. let's check with this new version you're recompiling that sets focus to false on imInitiateHide and let's hope it's the last thing we need :)10:38
tmoenickenerochiaro: I just sent you an updated version10:39
tmoenickehope that works now10:40
nerochiarotmoenicke: unfortunately no. if i manually hide the keyboard the blinking cursor is removed but the textarea still has focus. but let me try with a textinput and also if i can simply work around this in the notes app10:54
nerochiarotmoenicke: though knowing why focus isn't removed is probably important10:54
=== MacSlow is now known as MacSlow|lunch
nerochiarotmoenicke: ok, we have something. TextInput gets the focus removed correctly. TextArea doesn't11:01
nerochiarotmoenicke: when hiding the keyboard manually11:01
nerochiarotmoenicke: let's talk to zsombi ?11:01
kalikianaKaleo, standup11:02
tmoenickenerochiaro: yeah, after lunch?11:02
Kaleosorry, conputer broke; hd total faikurr11:04
nerochiarotmoenicke: ideal11:04
Kaleogoing to look if cn buy one now11:04
nerochiarotmoenicke: i'll see if i can give him a peek now anyway, i'm short on time11:04
kalikianaKaleo, ouch, I hope you didn't lose much11:04
nerochiarozsombi: can you please poke me when you have some minutes ?11:04
t1mpKaleo: good luck :s11:04
tmoenickeok11:09
=== gusch|away is now known as gusch
om26ernerochiaro, removed the nexus 7 stuff as you said: https://code.launchpad.net/~om26er/camera-app/ap_tests_improvements/+merge/16573311:13
nerochiaroom26er: added a few comments11:25
nerochiaroom26er: then i will need to test it and if htey all pass approve11:25
nerochiaroom26er: (on desktop)11:25
guschstill pending for review: https://code.launchpad.net/~schwann/gallery-app/gallery-modularity/+merge/16605411:25
om26ernerochiaro, thanks fixing11:26
om26ernerochiaro, the python docstring convention says that if a doctring is multi-line it should have an empty line before ending it.11:26
nerochiarozsombi: the thing me and tmoenicke would like to discuss with you is https://bugs.launchpad.net/notes-app/+bug/1182108 and specifically the last few comments about focus in TextArea11:26
ubot2Ubuntu bug 1182108 in notes-app "keyboard re-appears after swiping it away" [Critical,In progress]11:26
nerochiaroom26er: there are a bunch of other multiline docstrings in that code that don't have that. either fix them all if it's required by python or leave them all more compact. I honestly prefer them more compact, the extra lines are a waste of screen space and make readability worse11:28
nerochiaroom26er: for example11:28
nerochiaro159+    """Test that the shoot button gets disabled for a while then re-enabled11:28
nerochiaro160+    after shooting"""11:28
nerochiaroom26er: i think this is good and readable11:28
om26ernerochiaro, ok then i'll change that back11:29
oSoMoNzsombi: hey, it looks like the Panel is broken in today’s image, tapping outside it doesn’t hide it anymore, can you confirm?11:30
nerochiaroom26er: thanks11:30
oSoMoNtmoenicke: hey, regarding inputMethodHints, do you know if there’s already a bug report to track its addition to the TextField component in the SDK?11:38
nerochiarotmoenicke: zsombi: I think i know what's happening. the OSK is removing the focus from the inner TextInput but the SDK TextArea doesn't track that and remove focus on the entire component11:39
nerochiarotmoenicke: zsombi: if we fix that I think we can solve our problems11:40
oSoMoNtmoenicke: nevermind, I filed bug #118539211:42
ubot2Launchpad bug 1185392 in Ubuntu UI Toolkit "Need to expose 'inputMethodHints' in the TextField component" [Undecided,New] https://launchpad.net/bugs/118539211:42
oSoMoNt1mp: hey, it looks like the Panel is broken in today’s image, tapping outside it doesn’t hide it anymore, can you confirm?11:49
dpmt1mp, is it ok to reopen bug 1182577? I've just found what I think is a valid use case11:51
ubot2Launchpad bug 1182577 in Ubuntu UI Toolkit "Cannot use the i18n plugin in C++" [Wishlist,Won't fix] https://launchpad.net/bugs/118257711:51
t1mpoSoMoN: ok. I'll install today's image to test11:53
oSoMoNt1mp: thx11:54
oSoMoNt1mp: I can observe that in all applications that have a toolbar11:56
t1mpdpm: okay. I re-opened it.11:57
dpmthanks t1mp11:57
=== t1mp is now known as timp
tmoenickeoSoMoN: no i dont think so, thx11:57
tmoenickenerochiaro: cool11:58
nerochiarotmoenicke: let's see what zsombi says when he comes back12:03
tmoenickeok12:03
tmoenickenerochiaro: also, using the property does not add the dependency to declarative, so the patch should be fine12:05
tmoenickenerochiaro: though, should be done for widget-based code too at some point12:06
nerochiarotmoenicke: sweet, let's worry about what we use now ATM and put a note there in the code about widgets12:07
timpoSoMoN: I confirmed. Did you create/re-open a bug for that?12:08
om26ergusch, please approve this one, its just some style changes https://code.launchpad.net/~om26er/gallery-app/reduce_pep8_complaints/+merge/16585412:08
oSoMoNtimp: I’m filing a bug right now12:08
guschom26er: ok12:09
om26errenato_, Hey! this one is now ready to go https://code.launchpad.net/~om26er/mediaplayer-app/skip_if_phone/+merge/16588012:10
guschom26er: approved12:11
oSoMoNtimp: bug #118539712:12
ubot2Launchpad bug 1185397 in Ubuntu UI Toolkit "[regression] The panel doesn’t hide when tapping outside it" [Undecided,New] https://launchpad.net/bugs/118539712:12
om26ergusch, thanks12:12
oSoMoNtimp: can you please confirm the bug, and can we please revert the faulty revision asap? (I’m suspecting rev 517, needs confirmation)12:13
om26eroSoMoN, when you get some time: calendar branch is now ready.12:15
timpzsombi: png12:16
timp+i12:16
renato_om26er, done12:30
om26errenato_, thanks12:30
guschstill pending for review: https://code.launchpad.net/~schwann/gallery-app/gallery-modularity/+merge/16605412:31
guschnerochiaro oSoMoN renato_ tmoenicke ^12:31
zsombitimp pong12:32
nerochiarozsombi: ping12:44
zsombinerochiaro: pong12:44
renato_gusch, I will take it12:44
kalikianaif anyone's up for a review of documentation/ui toolkit https://code.launchpad.net/~kalikiana/ubuntu-ui-toolkit/fataldocs/+merge/16624012:44
nerochiarozsombi: can you please check the messages that i directed to you in the scrollback, or I can repeat if you like12:45
zsombi(02:01:31 PM) nerochiaro: tmoenicke: ok, we have something. TextInput gets the focus removed correctly. TextArea doesn't12:48
zsombinerochiaro: this is the issue, right?12:48
zsombinerochiaro: when hiding the keyboard manually12:48
zsombinerochiaro: manually means by swiping down the OSK, right?12:49
nerochiarozsombi: yes12:50
zsombinerochiaro: is this happening with some special build?12:50
zsombior is it in the latest image?12:50
guschrenato_: cool thx12:50
nerochiarozsombi: let's hop on mumble when you have a minute and i'll explain to you12:53
timpzsombi: did you read up about the panel regression?12:54
timpzsombi: https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1185397 do you think it could be related to r517 with IMA changes?12:55
ubot2Ubuntu bug 1185397 in Ubuntu UI Toolkit "[regression] The panel doesn’t hide when tapping outside it" [Undecided,Confirmed]12:55
nerochiarozsombi: it's happening with yesterday image in any case, plus a library from tmoenicke12:56
=== MacSlow|lunch is now known as MacSlow
nerochiarozsombi: it's explained in the bug. but let's mumble so i can give you more context and we can talk about what it can be done to fix it13:01
renato_gusch, I added some comments13:01
guschrenato_: thx - I pushed the fixes ;)13:03
zsombinerochiaro: I am supposed to have a meeting with ckpringle now, so eventually after that13:07
zsombitimp: weird... perhaps... I tested the apps back then and haven't noticed such an issue...13:09
zsombitimp: easy to find out: print the IMA's sensingArea in Panel to see whether it has a valid value13:10
timpoSoMoN: ^ will you have a look? I'm working on something else now13:11
oSoMoNzsombi, timp: will do13:13
zsombinerochiaro: what is the bug id?13:48
nerochiarozsombi: https://bugs.launchpad.net/notes-app/+bug/118210813:50
ubot2Ubuntu bug 1182108 in notes-app "keyboard re-appears after swiping it away" [Critical,In progress]13:50
nerochiarozsombi: but i really need to explain to you what's going on13:50
nerochiarozsombi: on voice13:50
nerochiarozsombi: i have a standup in 10 minutes13:51
zsombinerochiaro: ehh :)13:51
zsombinerochiaro: I just finished my meeting13:51
nerochiarozsombi: let's hop on mumble now then13:51
zsombiconnecting...13:52
nerochiarozsombi: i'm in the PS apps channel13:52
=== mzanetti is now known as mzanetti|food
zsombiconnecting takes ages...13:52
nerochiarozsombi: confirm that it works fine with TextArea14:08
nerochiarozsombi: TextEdit, sorry14:08
zsombinerochiaro: ok14:08
zsombithx14:08
=== mzanetti|food is now known as mzanetti
* kalikiana wonders if jppiiroinen would like to claim https://code.launchpad.net/~kalikiana/ubuntu-ui-toolkit/fataldocs/+merge/16624014:19
Mirvdpm: alert, alert, the 'gomobile' page no redirects to http://developer.ubuntu.com/get-started/ which has completely bollocks instructions, like an antique version with wrong package names / copy-paste errors updated with some Ubuntu version numbers..14:25
dpmMirv, on it, thanks for the heads up14:25
Mirvdpm: thanks. it refers to a PPA from half a year back et cetera, just received an e-mail from a precise user, noticed the redirect and then noticed the instructions14:27
oSoMoNom26er: re- the calendar app and objectName, are you using autopilot vis to visualize the hierarchy of items?14:31
dpmMirv, phew, well spotted. This was due to the replacement of the old get-started page (i.e. Quickly) by the new gomobile page, which unnoticedly set the fields with the PPA instructions to defaults (i.e. as they were at launch date) - the deployment happened a couple of hours ago, so hopefully the confusion didn't last for long. You can now refresh http://developer.ubuntu.com/get-started/ and see the proper instructions14:32
Mirvdpm: much better now! :) it's good that it was caught early.14:33
dpmnicely spotted :)14:34
om26eroSoMoN, yes14:35
oSoMoNom26er: then it must be a problem in autopilot…14:35
om26eroSoMoN, but I also tried a different technique with select_many() and gave only the objectName but there are no objects found14:36
oSoMoNom26er: how do you use autopilot vis on an app started with qmlscene? I’m not getting the hierarchy of items, only the root one14:38
om26eroSoMoN, qmlscene calendar.qml -testability14:39
om26eroSoMoN, if that's not working you might have an old version of libautopilot-qt14:39
oSoMoNom26er: that’s what I’m doing, and I’m only getting the QQuickView root object14:39
om26eroSoMoN, that was a bug which have been fixed in autopilot-qt14:39
om26eroSoMoN, make sure to have libautopilot-qt 1.3daily13.05.24ubuntu.unity.next-0ubuntu114:40
oSoMoNom26er: where do I get this one from? I have 1.3daily13.05.09ubuntu.unity.nextbzr58raring1 from the autopilot PPA14:42
om26eroSoMoN, ppa:ubuntu-unity/daily-build-next14:43
oSoMoNok, thanks14:46
oSoMoNom26er: ok, I understand the issue, you’re setting the objectName on the Action items, right?14:50
om26eroSoMoN, yes14:50
oSoMoNom26er: so in fact it’s not a bug but a missing feature in the UITK, the objectName of the action should be transferred to the corresponding button14:52
om26eroSoMoN, right, makes sense14:53
om26eroSoMoN, so currently what we are doing is fine ?14:53
oSoMoNom26er: I wouldn’t say that, let’s try and find a cleaner way of doing it14:55
oSoMoNom26er: I’m pretty sure the same problem was already solved by other apps’s autopilot tests, have a look at how it’s done in the gallery for example14:55
om26eroSoMoN, will look there now.14:56
om26eroSoMoN, about the issue of my test not saving the event. I have not found a way to delete the event in the app so running the test each time will add a new event14:57
zsombinerochiaro: so this happens in every TextArea, not just in Notes?14:58
oSoMoNom26er: the application is using the localStorage API, so it’s just a matter of deleting the DB before starting the app, the notes-app does something similar in its tests14:59
oSoMoNom26er: but in a first step I wouldn’t bother with such a complex set up, I’d just test that clicking the timeline button switches to the timeline view, and then back to the main view, for example15:00
oSoMoNom26er: a test for creating a new event can be written later on, the point of this MR is to bootstrap the tests with something simple and illustrative15:00
zsombijppiiroinen: weird thing: I had enabled the developer mode on my tablet yesterday, and now after I connected it, it sais I should enable the dev mode again...15:01
zsombijppiiroinen: In QtC...15:01
om26eroSoMoN, ok I can do that15:01
om26eroSoMoN, about the gallery-app, its also using the index number technique it seems15:01
oSoMoNom26er: the technique used by the gallery is slightly safer though (see get_toolbar_button(…)), because it ensures that it gets buttons only inside the panel, not every single button in the application15:04
om26eroSoMoN, yeah, I am going with that. Infact that's what I was using back in the days before I was told to use select_single()15:06
om26eroSoMoN, as now I am only targeting a single test to show/hide timeline, what happens to my already written emulators? shall I just keep them or remove them? They are definitely going to be used in future.15:25
oSoMoNom26er: keep them around15:26
zsombiKaleo: greyback: new stuff needed in layouting: http://studio.sketchpad.cc/MnCl8Uj9Qa15:27
zsombiKaleo: greyback: please comment15:27
greybackzsombi: use-case please15:28
greybackbecause right now I totally don't understand it15:28
greybackzsombi: did you see the other portion of my Layout API proposal: "LayoutItem"15:29
zsombigreyback: as said, I saw it, and it may make sense to use that, however in there you need a LayoutItem per each re-parented item, right? This would give the parameters on a specific fragment15:35
oSoMoNnerochiaro: would you be available for a review today?15:37
zsombigreyback: I'll come back with an example tomorrow... I should really step off now...15:39
greybackzsombi: ok, have a good evening15:43
nerochiarooSoMoN: to be honest no, i am behind with a task ATM15:56
oSoMoNnerochiaro: ok, nm16:05
guschrenato_: can you please have a look at this again? https://code.launchpad.net/~schwann/gallery-app/gallery-modularity/+merge/16605416:09
nerochiarooSoMoN: om26er: do you guys know why this wasn't reviewed yet ? I think it's pending from last week, for some reason i had assumed i gave it to one of you for review16:22
oSoMoNnerochiaro: link?16:22
om26ero_O16:26
oSoMoNom26er: I approved your MR, is there anything specific to do on jenkins to activate autopilot tests for the calendar app?16:55
om26eroSoMoN, I am not sure about that, will have to ask about that from mmrazik16:55
om26eroSoMoN, he was looking into that a few days ago16:56
oSoMoNom26er: ok, anyway I’ll let you handle it16:56
=== chriadam|away is now known as chriadam

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