=== chriadam|away is now known as chriadam === afkthairus is now known as athairus === Ionic is now known as Guest67057 [01:31] I want to save some data when the application is closed. is there any such an event for me to catch? === Guest67057 is now known as Ionic === chihchun_afk is now known as chihchun [07:17] good morning [07:57] dpm, I'd appreciate your feedback on https://lists.launchpad.net/ubuntu-phone/msg12954.html [08:07] dholbach, I'm on the fence on this one. Logically it would make sense to use u-a-d, but traffic pretty much died in favour of u-p [08:07] and I'm not sure if all experts will move to u-a-d [08:07] not move [08:07] but I think we could twist their arms into signing up for a low-traffic mailing list? [08:08] ... maybe [08:08] :) [08:08] https://strawpoll.me/4482403/r is the feedback I got on the list [08:08] yeah, got it on my screen now [08:11] dholbach, I guess we can just try, as it should not be too much effort updating developer.ubuntu.com/community to point to u-a-d, sending a "revival" e-mail to the list and getting the SDK guys to subscribe [08:15] dpm, dholbach: I think it would make sense, -phone has many details about the OS building part as well, it's likely that the people around atm are interested in that as well, but normal appdevs are probably not going to care about it [08:16] bzoltan, not sure if you saw the conversation above: ^ - the buy-in of you guys would obviously be important as well [08:17] thanks for the feedback seb128 [08:17] * bzoltan reads the backlogs [08:18] dpm, dpitkin's mail on the mailing list gave me the idea to try it again :) [08:18] (I think we talked about this a while ago already) [08:42] bzoltan, so... what do you think? [08:42] bzoltan, the last 6-7 messages of scrollback would have been enough :-P [08:42] LOL [08:43] * bzoltan is already back in 2013. [08:44] dholbach: why do not we merge the two lists?ű [08:45] signal/noise ratio for people who *just* want to write apps and don't care about how the OS is put together? [08:46] and not everyone is used to 100+ mails per week on a list [09:02] bzoltan, ^ :) [09:02] what do you think? [09:03] dholbach: your concern is valid... I do not know the answer. Maybe mailing lists are not the best channels. G+, Facebook, Twitter? [09:06] bzoltan, but who's going to monitor G+/FB/Twitter for questions and reply to them? [09:08] dholbach: wait a sec :) Do we have an army of community engineers to respond to all questions on any channel. As I see we hardly have time and energy to respond to the askubuntu and irc questions. [09:09] bzoltan, it's not that complicated I think [09:09] dholbach: the best would be if we could build a competent community team who could do it [09:09] right now we get a bunch of questions about app development on the u-phone@ list [09:09] dholbach: With all respect... I disagree :) I think it is very complicated. [09:09] which is where we send people right now [09:10] and they get overwhelmed [09:10] u-app-devel@ is low traffic and we have it already [09:10] so we'd just ask app developers with questions to ask there [09:10] (... if we went with this plan) [09:10] dholbach: obviously if somebody is more into app development then a dedicated list is better. [09:11] so it'd be the same questions, the same general amount of mail, but a more dedicated place for app developers [09:11] but it'd require a few people to sign up for 2 lists [09:11] I'd say that the questions on askubuntu/fb/twitter/g+ are a separate thing [09:11] and I don't mean to talk down the amount of questions you and your team generally get [09:13] where can my app write for caching? QStandardPaths cache location is inaccessible (~/.cache probably) [09:13] (and why isnt QStandardPaths updated to reflect this? :) [09:16] kbroulik: you need to append the applicationName to the standard path. [09:16] ~/.cache/project.myname/ [09:16] DanChapman: it does that automatically. on my desktop the path is ~/.cache/companyname/appname [09:17] but QNetworkDiskCache complains it cannot write there [09:18] kbroulik: IIRC the application name needs to be in the format "appname.companyname" or "appname.developer" [09:19] * DanChapman looks for the docs === chriadam is now known as chriadam|away [09:23] kbroulik: it also needs to match the "name" field in you manifest.json. The Runtime Environment section here https://developer.ubuntu.com/en/start/platform/guides/app-confinement/ has a good explanation [09:24] popey, how do you normally go about building the terminal click? [09:24] DanChapman: but shouldnt ubuntu phone provide a qstandardpaths impl that does that? [09:25] dholbach: jenkins builds it [09:25] ah, ok, makes sense of course [09:25] why? [09:25] on every other platform I tried (blackberry 10, ios, android, windows rt, ..) qstandardpath returns sensible values that are accessible [09:26] dholbach: also using this script from DanChapman to build using a local click chroot, to make a fat package http://paste.ubuntu.com/11667958/ [09:26] dholbach: you could change line 11 to make it only do one arch [09:26] ok, I'll take a look at it - thanks! [09:26] dholbach: got a moment for a packaging question? :) [09:28] popey, sure [09:28] dholbach: do you have an armhf 15.04 click chroot handy? [09:28] dholbach: grab lp:ubuntu-filemanager-app and try and build it in an armhf chroot [09:29] it fails because it needs libsmbclient, which if I try to install libsmbclient-dev:armhf it pulls in a bunch of other stuff, which is fair enough, but it still complains when building file manager [09:29] so basically, bzr branch lp:ubuntu-filemanager-app, and then navigate to that directory and run that fat package script. [09:29] popey, yes, updating it right now [09:29] I am wondering of samba has an issue with multi-arch [09:30] I can build file manager on my local pc natively as amd64, but i can't build using the click chroot [09:30] tried all kinds of combinations of libsmbclient and samba and can't get the right voodoo [09:33] * popey gets coffee [09:35] kbroulik: i've never really thought of it as an issue myself. But I can see it would be nice to have an impl that gives correct paths. Might be worth filing a feature request for it [09:49] popey, I looked into it just now, but I can't find the issue [09:49] maybe bring it up on ubuntu-devel-discuss@ or something? [09:49] dholbach: it builds for you? [09:49] no [09:49] oh, ok [09:59] hey sverzegnassi :) [10:00] i was just testing your zoom patch :) [10:00] popey: o/ [10:01] looking forward to your review [10:03] Well, it certainly works. but I have the same issue you noticed, that it jumps sometimes [10:20] humm, where do I change the manifest file in the new QTCreator^WUbutuSDK? :) [10:21] I think I remember it was on "Publish" some day [10:26] popey: does it go like this https://www.youtube.com/watch?v=nOXRBK7EITI [10:27] zbenjamin: ^ [10:28] Anyone knows from the top of their head which policy_version I should use when I'm on 15.04 framework? [10:29] ybon: 1.3 [10:29] thanks :) [10:30] ybon: fyi, when you change the framework to 15.04, the ubuntu sdk automatically changes the policy_version to 1.3 as well [10:30] :-1: error: security_policy_version_matches_framework (OSMTouch.json): 1.3 != 15.04 (ubuntu-core-15.04-dev1) [10:30] nik90: how do you change the framework? :) [10:30] I've made it by hand [10:30] ybon: ubuntu-core-15.04-dev1 is wrong [10:30] ybon: just open the manifest in qtc [10:30] using the open file dialog if not through the project tree [10:30] I can't find it on Ubuntu SDK [10:31] ah [10:31] it doesn't appear on the project tree [10:31] but let me open it :) [10:31] ybon: is it a cmake project? [10:31] no, still OSMTouch, basic QML [10:32] nik90: about the wrong framework, I've followed the link given by QTC with the error http://askubuntu.com/questions/460512/what-framework-should-i-use-in-my-manifest-file [10:32] which then point to an EvilDoc [10:32] ybon: well all projects regardless of whether they are basic qml are recommended to be transitioned to cmake or qmake project types [10:32] maybe we should give this info to someone that this link isn't valida anymore [10:32] nik90: any "how to" around to do so? [10:33] ybon: that link states ubuntu core which != ubuntu touch [10:33] ubuntu core is snappy stuff [10:33] ybon: but either way it should be updated [10:33] yep [10:34] beuno, popey, bzoltan: Can you ensure the doc https://docs.google.com/spreadsheets/d/1t_JGpg4r8BLluzfzmqa-gAbcKUjKUOufSCTSdPpFc5g/edit#gid=0 is updated with the correct frameworks version pls. It seems people use that after looking at https://askubuntu.com/questions/460512/what-framework-should-i-use-in-my-manifest-file [10:34] (which is the link given with the error on QTCreator aka Ubuntu SDK) [10:36] ybon: as for the qmlproject to cmake project transition..I don't think there is a automatic way to do it..you will have to do it manually..but i guess that's a problem for another time [10:36] ok [10:36] I'll add an issue for that for future me :) [10:37] :) [10:41] popey, on terminal, how are we getting on in terms of having a converged version? [10:41] dpm: since we talked yesterday? [10:41] what are your expectations? [10:42] hello guys, I'm not sure this is the correct channel, but I'm having problems overriding default application for applicaiton/sql file type which is set to gedit. Does anyone knows anything about mime handling? [10:42] popey, I don't mean snappyfying it, I mean only, is there anything specific to do (e.g. layout), or should the version on the phone just work as is? I couldn't remember if we discussed this specifically [10:43] dpm: oh, i see. Yeah, I'm making notes of bugs and will file them in a batch. (once I figure out how to take screenshots on Unity8/Mir that actually _works_) [10:44] dpm: we have a multi-arch click and yes it does "just work" but there are bugs of course :) [10:45] popey, ok, so for terminal, in general, it's more about testing it and fixing potential bugs than new designs for the desktop, then? [10:45] yes [10:45] ok, thanks [10:45] it only needs a few tweaks IMO [10:46] seems there's a few toolkit bugs too, but this build is super old, and if I update it, chances are I'll break it [10:46] hence waiting for the "fixes everything" snappy desktop build from the desktop team [10:47] hi David Planella :) [10:52] seb128: do you know if there is a plan to add keyboard options to system settings? Specifically autorepeat delay/speed? [10:55] nik90: which framework from the list do you suggest I use? [10:56] test [10:56] failed [10:56] ybon: use ubuntu-sdk-15.04 with policy version 1.3 This should be supported in the upcoming OTA-4 [10:56] ok thanks [10:56] +1 [10:57] but my phone seems to be still in 14.10 :p [10:57] other phones have switched usually? [10:57] OTA4 hasn't been released yet [10:57] when it is, you'll bump from 14.10 to 15.04 [10:57] * ybon only updates when the phone ask for update [10:57] okay [10:57] but then if I use 15.04, I'm not able to test on my phone right now ;) [10:58] could I switch to OTA-4 yet? [10:58] ybon: what phone is it? BQ? [10:58] nope, Nexus4 [10:58] the same I've broken time ago [10:58] but still working a bit to test [10:58] ybon: well if you switch over to ubuntu-touch/rc-proposed/ubuntu-developer [10:58] that should have OTA-4 [10:58] ok [10:58] well it has vivid really [10:58] let's do that so :) [11:00] ahem http://askubuntu.com/questions/555004/having-problems-trying-to-switch-channels-on-nexus-4-utopic-to-14-09-ubuntu maybe I should wait then :p [11:01] popey: Have you already got an answer regarding your e-mail about problems with ubuntu-filemanager-app and libsmbclient? [11:03] no [11:07] sturmflut2: do you have any ideas? :) [11:07] popey: Hmmm, I can't even get libsmbclient-dev:armhf installed in my ubuntu-sdk-15.04 chroot. It fails because of python2.7, I do an "apt-get -f install", but the latter only *seems* to fix the problem, without actually fixing it. libsmbclient-dev:armhf is still not installed [11:08] :( [11:08] you're right [11:08] looks like some kind of packaging issue [11:08] I am not convinced samba is multi-arch safe [11:09] The problem is with python2.7, the post-install script of the python2.7-minimal package fails [11:13] popey: Oh, I think I see something. The post-install script of the python2.7-minimal package tries to call /usr/bin/python2.7 to pre-process some files, but /usr/bin/python2.7 is an armhf binary and that one will probably never run inside the chroot because the host ist an amd64 machine === MacSlow is now known as MacSlow|lunch [11:47] I've run into these python2.7 issues several times in chroots [11:49] to get some -dev packages installed I've had to apt-get download then dpkg-deb --extract [11:57] popey, no idea, I mentioned those to mpt, up to design [11:58] seb128: okay, thanks! [11:58] popey, yw! [11:58] is anyone really changing those options? [11:58] I never touched that since I've a computer [11:58] well, there's two issues [11:59] 1) autorepeat doesn't work _at_ _all_ [11:59] 2) no way to change those settings [11:59] 1) is more important :) [11:59] right [11:59] just trying to figure out how much the settings are needed [11:59] I never changed them nor know anyone who does afaik [11:59] well i only went looking for them, thinking it might be switched off [12:00] filed bug 1463370 anyway [12:00] bug 1463370 in unity8 (Ubuntu) "Keyboard autorepeat missing" [High,Confirmed] https://launchpad.net/bugs/1463370 [12:02] popey, thanks [12:02] fwiw, I never fiddle with those settings either [12:03] k, so just fixing the default behaviour should be good enough for most of us ;-) [12:03] ya [12:06] that issue might be more a mir one that unity8, but I guess those teams are used to bounce bugs between projects ;-) [12:08] :) [12:09] mzanetti touched it last, out of my hands now :) [12:09] yeah... actually it's probably QtMir, the glue between mir and unity [12:09] I assigned it to daniel, he'll fix or dispatch it further [12:10] when he's back from holiday, that is [12:33] zsombi: please take another look https://code.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/darkBackgroundText/+merge/260647 [12:33] t1mp: would you review this? https://code.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/unitTestApiCheck/+merge/260348 [12:34] * kalikiana feeling like the princess sitting on too many branches === _salem is now known as salem_ [12:37] woot, new phone in the store http://store.bq.com/en/ubuntu-edition-e5 [12:38] :D [12:38] \o/ === pat__ is now known as pmcgowan [12:40] moin. I'm looking to port an erlang-based app to μTouch, is there a way to see what s/w is already "available"? basically is erlang apt-gettable in principle, to start with. [12:40] I've not found this on the FAQ so far, sorry. [12:42] dch: it's not part of the phone images, that means you need to bundle it. much like Python. there's bindings apparently https://github.com/krant/eqml though I haven't used them myself [12:43] kalikiana: thats fine, AFAICT its an armfh arch so I'm confident building it from source should be fine. === chihchun is now known as chihchun_afk [12:44] likely I won't need the QT layer though this is background service stuff [12:44] what kind of app is it? [12:44] dch: background services aren't well supported just now [12:45] kalikiana: its an implementation of a IETF transport protocol called PPSPP for streaming media. typical use case is to provide an additional URL handler to a browser or app. [12:46] we want to be able to send/recv arbitrary udp packets, incl multicast if connected to local wifi or lan (for MDNS style discovery) [12:49] dch: right now you'll require a phone with an exception for the app from the life cycle management [12:49] out of the box no app can run unless the user is looking at it [12:50] (yet) [12:50] not quite true. You can exploit a hole in the lifecycle management, but this hole is likely to be fixed "soon'ish" [12:51] kalikiana: ok, for the moment I can ignore the UI constraint, either wrap it in qt or some html5 bridge I guess. [12:52] thanks, this is enough to go on for the moment [12:56] mcphail: which one are you referring to? something in the browser engine? [12:57] kalikiana: no - apps can run in background (or screen off) if they don't create a GUI. See my syncthing app in the store [12:58] kalikiana: I believe there are plans to kill this behaviour [12:58] (which makes me sad) === MacSlow|lunch is now known as MacSlow [13:36] DanChapman: docs say "Qt applications can find the values of the XDG directories by using the QStandardPaths API as well as QCoreApplication::applicationName" ← so perhaps because I'm manually setting a applicationName in my main breaks it? [13:42] kbroulik: what's breaking exactly? [13:43] when I do QStandardPaths cachelocation, it returns "/home/phablet/.cache/foo with foo being my project name, but the specs say I can only write to ~/.cache/com.foo.bar (with that being the app id) [13:45] kbroulik: would foo match your binary? [13:45] yes [13:45] in that case it might be a race condition [13:45] at what point are you using the path? [13:46] on startup when the qml engine creates a NAM from my factory (I set a QNetworkDiskCache for it) === chihchun_afk is now known as chihchun [13:49] ybon: in case you did not find it yet: you need to open the manifest.json file like any other. Then you get the manifest editor [13:49] zbenjamin: thanks, nik90 told me that too, and it worked :) [13:50] zbenjamin: why not showing it in the files tree then? [13:50] kalikiana: when I do this app.setApplicationName(QString::fromLocal8Bit(qgetenv("APP_ID").constData()).split(QLatin1Char('_')).value(0)); ← then it works xd [13:52] kbroulik: could you set it later? as a response to setting a qml property for example [13:52] or component completion [13:52] no [13:52] not sure what best to suggest as I don't really know your code [13:52] "The factory must be set before executing the engine." [13:53] hmmmm [13:53] kbroulik: and do you have to set the cache at that time? [13:53] could you wait before you store any files? [13:56] kalikiana: no === chihchun is now known as chihchun_afk [17:05] ybon: it should be shown in the files tree [17:05] ybon: what project type do you have [17:09] zbenjamin: where do I see that? [17:16] ybon: he means whether it is qmlproject or cmake or qmake [17:18] ah, so I guess qmlproject === oSoMoN_ is now known as oSoMoN [20:21] https://athoneycutt.wordpress.com/2015/06/09/apn-changes-for-ubuntu-touch/ [20:21] my 2 cents [20:23] ybon: weird the qmlproject should show it ... [20:24] ahoneybun, heh, you caused mpt to reconnect with your design suggestion :) [20:24] ogra_: right [20:43] What's the difference between using "ubuntu-html5-app-launcher" and "webapp-container" for a HTML5 game? The former is the default when you create a new HTML5 project in Qt Creator, and it stutters heavily when running my game. The latter requires a change to the AppArmor profile, but then runs the game perfectly. [20:49] Oh, they're using completely different browser engines, aren't they [21:07] Elleo: I think we should allow people asking questions about podbird on our launchpad. Just saw a bug report which is more a question rather than a bug report. [21:17] rpadovani: hi [21:21] rpadovani: just wanted to let you know I didn't make anything for you today. I had some unexpected shopping journey and was working a bit on other project. Hopefully I'll show you something tomorrow :) === gh is now known as Guest10037 [21:52] SturmFlut: this has some information: https://wiki.ubuntu.com/SecurityTeam/Specifications/WebAppsConfinement [21:53] SturmFlut: webapp-container is the correct one to use with new apps. "In 14.10 and earlier, app deverlopers could use ubuntu-html5-app-launcher with the ubuntu-sdk template. In 15.04 ubuntu-html5-app-launcher has merged with webapp-container and ubuntu-html5-app-launcher is deprecated and users should use the ubuntu-webapp template and specify webapp-container instead of html5-app-container in their .desktop file."