2015/06/29/#ubuntu-app-devel.txt

=== chihchun_afk is now known as chihchun
[06:40] <dholbach> good morning
[07:06] <davidcalle> Morning all o/
[08:13] <mihir> popey: ping
[08:13] <popey> hi mihir
=== greyback is now known as greyback|doc
[10:04] <dpm> popey, when you've got a minute, could you have a look at https://code.launchpad.net/~dpm/ubuntu-filemanager-app/rename-pam-package/+merge/263162 ? With this and calculator fixed, then only terminal and clock PPA builds will remain to be fixed
[10:10] <popey> ok
[10:22] <mihir> hey popey
[10:22] <mihir> could you comment and if this is right could you review  , https://code.launchpad.net/~mihirsoni/ubuntu-calendar-app/1464457/+merge/263185
[10:22] <mihir> ?
[10:22] <mihir> popey: read the bug comments first , before review.
[10:31] <dpm> thanks!
[10:42] <popey> mihir: will do
[11:02] <mihir> popey: thanks :)
=== greyback|doc is now known as greyback
[11:56] <karni> I have inputMethodHints: Qt.ImhNone on a TextArea, and I still get input suggestions. Anyone experienced that as well?
[11:57] <nik90> karni: is that TextArea in a Dialog?
[11:57] <karni> no, just a regular one
[11:58] <nik90> karni: I have noticed that in uReadIt where the username and password textarea is shown in a dialog and the inputMethodHints are totally ignored
[11:58] <nik90> but yes I have seen the issue before
[11:58] <karni> noted. however, this is not the situation here, so it seems it's also ignored in other cases :|
[11:58] <karni> nik90: cheers
[12:08] <Elleo> karni: Qt.ImhNone doesn't disable suggestions, that says that you have no hints to provide, whereas to disable suggestions you need to explicitly pass the hint "Qt.ImhNoPredictiveText"
[12:09] <karni> :O \o/
[12:10] <karni> Elleo: thanks! :)
[12:10] <karni> nik90: ^^^
[12:10] <Elleo> karni: no problem :)
[12:21] <alecu> popey, @all: hi! I'm trying to find what are the community apps that use the image thumbnailer, to ask their owners to try out silo 10.
[12:22] <alecu> We have a rewrite of the thumbnailer in that silo, and it needs some testing of the apps that use it.
[12:30] <karni> alecu: any API changes? Telegram uses thumbnailer
[12:30] <alecu> karni: there are no API changes.
[12:31] <alecu> karni: well, you are not supposed to pass -1 as width or height any longer
[12:31] <alecu> karni: if you don't do that, then it's fine
[12:31] <karni> I very much doubt we do that
[12:31] <alecu> karni: apps passing -1 to the thumbnailer will get a warning in their app log
[12:32] <karni> cool
[12:32] <alecu> karni: so, how do you use the thumbnailer in telegram? do you use it for remote images, or local ones?
[12:33] <karni> local ones, to show thumbs instead of showing full sized pictures, and also to generate thumbs of sent videos
[12:33] <alecu> karni: sounds good
[12:52] <popey> alecu: only music app I think?
=== _salem is now known as salem_
[13:31] <alecu> popey: great. Who works on music app?
[13:32] <popey> alecu: https://launchpad.net/~vthompson and https://launchpad.net/~ahayzen
[13:56] <kalikiana> brrrrr it's so annoying when you do certain things in qml and the result is a white canvas and no errors whatsoever...
[13:57]  * ogra_ recommends edding 
[13:57] <ogra_> :P
[13:57] <kalikiana> :-D
[13:57] <ogra_> just draw something on it ... be creative
[13:58] <kenvandine> lol
[14:04] <alecu> popey: thanks, I'll ping them so they can check the new thumbnailer silo.
[14:21] <sverzegnassi> popey, ping
[14:22] <popey> hey sverzegnassi
[14:23] <sverzegnassi> popey, I'm having a second look at bug 1432394, which is about the documents imported twice through content-hub
[14:23] <ubot5> bug 1432394 in Ubuntu Document Viewer App "[content-hub] Do not create new copy of a document, if it already exists in $HOME/Documents" [Low,Triaged] https://launchpad.net/bugs/1432394
[14:23] <popey> sverzegnassi: I tagged it as bitesize hoping we could find someone else to do it ㋛
[14:23] <sverzegnassi> popey, yeah, ok. i'm wondering if it's still valid, and if is it worth to go for that solution
[14:24] <popey> sverzegnassi: not sure you want to be md5summing a potentially 4GB pdf?
[14:24] <popey> on a phone
[14:24] <sverzegnassi> calculating a MD5 for multiple imports (which now we support) could be a long task for workaround a missing string
[14:25] <popey> be faster to check file name and size first
[14:25] <sverzegnassi> there's also another similar bug report for music-app (bug 1413821)
[14:25] <ubot5> bug 1413821 in Ubuntu File Manager App "Music files are imported into the Music app, rather than simply opened" [Undecided,New] https://launchpad.net/bugs/1413821
[14:25] <sverzegnassi> yes, but content-hub does not export the path of the original file :/
[14:25] <popey> uhg, upstream bug.
[14:25]  * popey looks at kenvandine 
[14:25] <mcphail> why would you need to use something as slow as MD5?
[14:26] <popey> mcphail: i think that was just a suggestion
[14:26] <popey> mcphail: better suggestions are most welcome :)
[14:26] <sverzegnassi> yep
[14:26] <mcphail> popey: you you look at the hashing which rsync uses - more risk of collisions but much faster
[14:26] <mcphail> (iirc)
[14:27] <mihir> hey popey
[14:27] <popey> hey
[14:27] <mihir> that was pretty fast
[14:27] <mihir> :P
[14:27] <sverzegnassi> would be nice if content-hub allows to trusted apps to read the source path of the file. everything would be fix by a string comparison
[14:27] <mihir> did you get chance to look at that bug thred ?
[14:28] <sverzegnassi> (english, my dear... s/fix/fixed)
[14:28] <kenvandine> we wouldn't want to expose the path to the original file
[14:30] <sverzegnassi> kenvandine: huh, ok. so should we check for file name and size instead?
[14:30] <popey> mihir: will do now.
[14:30] <kenvandine> that's an option
[14:31] <karni> I know it's called "stack", but it's a shame you can't remove a page from other than top. if you first put a new one on top, but want to remove one from under it, if you do it first - you see it all happen which may not be pleasant to the eye.
[14:32] <karni> if in contrast one could push page, and pop the one below it, you wouldn't have to see it "go away", but you'd see the n-2 once you pop the top one.
[14:33] <dpm> popey, I was disconnected for a bit, so I'm not sure if you already sent me the FM/Nautilus screenshot. No rush, just checking that I didn't miss it
[14:34] <mcphail> hmm - apparently the hash function in rsync is based on the Adler-32 checksum which is implemented in zlib... ? worth exploring as something faster?
[14:34] <mihir> popey: here is the link of the bug , https://bugs.launchpad.net/ubuntu-calendar-app/+bug/1464457
[14:34] <ubot5> Launchpad bug 1464457 in Ubuntu Calendar App "Delete an event from the Edit Event view" [Medium,In progress]
[14:34] <popey> mihir: on it
[14:35] <popey> dpm: http://imgur.com/6uudmQj
[14:35] <dpm> nice, thanks!
[14:37] <sverzegnassi> kenvandine, understood. iirc the lastModified date-time does not change in the imported file (while created date-time does), so we have also a third parameter that would make the comparison strong enough
[14:37] <kenvandine> cool
=== chihchun is now known as chihchun_afk
=== javiercrowsoft1 is now known as javiercrowsoft
=== javiercrowsoft1 is now known as javiercrowsoft
[17:35] <SturmFlut> Hrmpf, I can no longer create an schroot on my 15.04 development desktop. See http://paste.ubuntu.com/11794281/ . Does it have problems with my ecryptfs homedir?
[17:35] <SturmFlut> Regardless what I try, it won't close the session because it can't unmount /home/sturmflut
[18:09] <nikwen> SturmFlut: I guess this doesn't help you but yesterday it still worked perfectly for me. Gonna try it again now.
[18:09] <SturmFlut> nikwen: zbenjamin pointed me to https://bugs.launchpad.net/ubuntu/+source/click/+bug/1427264, it is actually broken
[18:09] <ubot5> Launchpad bug 1427264 in schroot (Ubuntu) "using ecryptfs, creating frameworks fail to bind mount issues" [High,In progress]
[18:10] <nikwen> Sturmflut: Ah, ok.
[19:06] <ahayzen> kenvandine, we had this report against music the other day which is related to content-hub but I've just realised that the files get deleted on a restart anyway so is this expected behaviour? Or would you expect the contentItem.move() to delete the source if a copy was used? bug 1469452
[19:06] <ubot5> bug 1469452 in Ubuntu Music App "After importing files via ContentHub files won't be deleted" [Undecided,New] https://launchpad.net/bugs/1469452
[19:08] <kenvandine> ahayzen, when the content-hub-service gets killed, it should clean up all those
[19:08] <ahayzen> ah when does that happen?
[19:08] <kenvandine> device reboot usually
[19:08] <ahayzen> ok so shall i close the bug as it is 'expected' behaviour really?
[19:09] <kenvandine> no...
[19:09] <kenvandine> something's wrong
[19:09] <ahayzen> remember the contentItem.move() fails on confined apps so it uses a copy instead
[19:09] <kenvandine> yeah
[19:09] <dubstar_04> davidcalle: This page has two definitions for text:string. Its a bit confusing.
[19:09] <kenvandine> doesn't matter
[19:10] <kenvandine> when the device gets rebooted, those files should get removedc
[19:10] <dubstar_04> davidcalle: https://developer.ubuntu.com/api/apps/qml/sdk-15.04/Ubuntu.Components.TextField/
[19:10] <ahayzen> kenvandine, yeah they do after a reboot i confirmed that :-)
[19:10] <kenvandine> ah
[19:10] <kenvandine> ok...
[19:10] <kenvandine> then that is the expected behavior
[19:10] <kenvandine> ahayzen, there is something you can do though
[19:10] <kenvandine> after you move them
[19:10] <kenvandine> set the transfer state to finalized
[19:11] <ahayzen> ooo ok :-)
[19:11] <kenvandine> that'll tell the hub you are done with the cached files
[19:11] <kenvandine> and clean them up right then
[19:11] <ahayzen> sweet thanks, i'll update the bug and make that change, thanks :-)
[19:11] <kenvandine> anytime!
[19:16] <dubstar_04> davidcalle: if you come across this later or in the IRC logs I've filed a bug . https://bugs.launchpad.net/developer-ubuntu-com/+bug/1469852
[19:16] <ubot5> Launchpad bug 1469852 in Ubuntu Developer Portal "Textfield has two entries for text:string " [Undecided,New]
[19:30] <davidcalle> dubstar_04, oooh
[19:31] <davidcalle> dubstar_04, thanks for the report! I'll see with the sdk team if they have a small issue in their code that could trigger that doc error (doc is auto-imported)
[19:44] <dubstar_04> davidcalle: no worries.
[19:44] <dubstar_04> Is it right to set the text of a textfield: textfield.text = "Text"
[20:34] <ahayzen> kenvandine, the finalize() works perfectly thanks :-)
[21:28] <alecu> ahayzen: Hi! We have a new silo with the thumbnailer rewrite (silo 10), and we've issued a "call for testing" to the phone mailing list.
[21:28] <alecu> ahayzen: I understand the Music app is using the thumbnailer, right?
[21:31] <alecu> ahayzen: we plan to land that silo later this week, so it would be great if you guys can try the app with the silo installed, and let me know of any issues.
[22:13] <ahayzen> alecu, yeah music-app uses thumbnailer :-) i was literally talking with victor about that silo a few hours ago, thanks for the heads up :-)
[22:39] <alecu> ahayzen: great! Thanks a lot, and let me know of any issues with it
[22:39] <ahayzen> alecu, first issue...no thumbnails ;-)
[22:40] <ahayzen> just doing an apt update; apt install thumbnailer-service
[22:40] <alecu> ahayzen: did you reboot after installing it?
[22:40] <ahayzen> yup twice
[22:41] <ahayzen> but i had thumbnailer-service : Depends: libleveldb1 but it is not installable
[22:41] <ahayzen> The following packages have been kept back:
[22:41] <ahayzen>   thumbnailer-service
[22:41] <alecu> Ah, good point
[22:42] <ahayzen> then i've done the update and install and its doing this http://pastebin.ubuntu.com/11796040/
[22:42] <ahayzen> seems to pull down part of snappy :-) .. so i've installed snappy via apt on a click-based system \o/
[22:43] <ahayzen> alecu, thumbnails are loading now :-)
[22:46] <alecu> ahayzen: I've no idea what's with that libsnappy1 dependency :-/
[22:46] <ahayzen> magic \o/
[22:47] <ahayzen> alecu, what pulled it down lol ? http://pastebin.ubuntu.com/11796064/
[22:49] <ahayzen> so i don't think i've lost any thumbnails :-) and it seems pretty quick once they have been cached
[22:49] <alecu> ah, it's the snappy compression library: https://launchpad.net/ubuntu/trusty/+package/libsnappy1
[22:50] <ahayzen> ahh
[22:50] <ahayzen> alecu, do you know if we'll ever see this supported in the new thumbnailer service at some point? bug 1372000
[22:50] <ubot5> bug 1372000 in Thumbnailer "Album art from images in the same folder not exposed" [Undecided,In progress] https://launchpad.net/bugs/1372000
[22:50] <ahayzen> i feel it is important for convergence especially
[22:51] <ahayzen> as we will need to tackle much larger and varied music library structures
[22:55] <alecu> ahayzen: I'm reading the bug, and from what I understand this looks like something that should be provided by mediascanner rather than thumbnailer
[22:56] <alecu> ahayzen: mediascanner should provide the url to the local image when no album art is on the media file but there is some art on the folder.... and then thumbnailer can work with that url
[22:56] <ahayzen> alecu, ah yes maybe i'm getting confused :-) i've just been waiting ages on it lol
[22:57] <alecu> ahayzen: sorry to hear about that. I'll ping around to try to understand who can fix it.
[22:57] <ahayzen> yeah exactly that, which the thumbnailer probably already supports ;-)
[22:57] <ahayzen> thanks :-)
[22:59] <alecu> ahayzen: now, what about the "online art"? who's providing that?
[22:59] <alecu> ahayzen: is it mediascanner too?
[22:59] <ahayzen> alecu, yeah it gives us a path that has the album and/or artist in it IIRC
[23:00] <alecu> ok, sounds great.
[23:01] <SturmFlut> When I start an app from within QtCreator to run on the device (arale), it seems to spawn a python3 process that doesn't want to die. After starting a couple of apps on the phone, I had five running python3 processes left which consumed so much CPU that the device heated up considerably.
[23:01] <ahayzen> either image://albumart/artist={artist}&album={album} ... or image://albumart/file:///path/to/audio/file
[23:02] <ahayzen> image://thumbnailer/file:///path/to/audio/file .. not albumart and there is an artistart one or something as well IIRC
[23:05] <alecu> SturmFlut: can you look at the cmdline parameters of that python3 process?
[23:05] <alecu> SturmFlut: if it's apport it means that something crashed: try looking in /var/crashes
[23:06] <SturmFlut> alecu: A moment please, it's very late here in europe and I'm a bit tired
[23:07]  * SturmFlut has been hacking on his phone for 18 hours
[23:08] <alecu> :-)
[23:08] <SturmFlut> Okay, QtCreator crashed. Maybe it's time to go to sleep.
[23:11] <SturmFlut> alecu: "python3 /tmp/qtc_device_applaunch.py /tmp/com.ubuntu.developer.sturmflut.floodit_0.9.19_all.click --hook floodit --force-install"
[23:11] <SturmFlut> alecu: Consumes a whole core on the device
[23:12] <ahayzen> "floodit"
[23:12] <SturmFlut> alecu: That's my app
[23:12] <ahayzen> ah
[23:13] <SturmFlut> "Flood It", it's in the store and on github, https://github.com/Sturmflut/ubuntu-touch-floodit
[23:13] <ahayzen> i was thinking it was a command to flood something lol should have read the click path first ;-)
[23:13] <SturmFlut> Haha
[23:15] <SturmFlut> The app itself doesn't do much, it only reacts to touch input, so there's no reason to consume 100% CPU
[23:19] <SturmFlut> ahayzen: Arent't we in about the same timezone, and isn't it about 12 PM where you live
[23:19] <ahayzen> its 0019
[23:19] <SturmFlut> ahayzen: Okay, it's 0119 here, so I'm in no position to tell you to go to sleep ;)
[23:20] <ahayzen> hehe
[23:22] <SturmFlut> I have to get up in four hours. Stupid work ;)
[23:24] <ahayzen> yeah same in 6 :-/
[23:24] <ahayzen> s/work/internship/
=== salem_ is now known as _salem