[00:00] well guys this has been crazy that's for sure [00:01] tests just started running in jenkins and look to be going well [00:01] balloons, it's the end of a long road that should have been finished a month ago [00:01] this has but hopefully worth it :D [00:01] vthompson, this isn't the first time you've "finished" [00:01] * ahayzen is now rewarded with massive merge conflicts in his other branches [00:02] #520 = perfect on VM [00:02] no, errors :-() [00:02] now ...jenkins your call [00:02] MismatchError: After 10.0 seconds test on Player.isPlaying failed: True != dbus.Boolean(False, variant_level=1) [00:02] oh god [00:02] which test? [00:02] man [00:02] File "/usr/lib/python2.7/dist-packages/music_app/tests/test_music.py", line 140, in test_play_pause_library [00:03] right.... vthompson we've had something like tht before no? [00:03] ohh, btw, I run the tests now with 'debug' set [00:03] looks like that change is in the branch [00:03] vthompson, on tht exact test... [00:03] yea, that test was painful in Malta [00:03] you can watch here: http://91.189.93.70:8080/job/generic-mediumtests-utopic/573/console [00:04] balloons, vthompson it passes on my VM [00:04] as in test_play_pause_library [00:05] vthompson, u the same? [00:05] vthompson, I too have 17 tests passing on device.. it takes forever to run doesn't it? [00:05] Ran 17 tests in 931.860s [00:05] balloons, is the video of the failure up yet or do they posted at the end [00:06] ahayzen, at the end [00:06] damn [00:06] ahayzen, vthompson I see the dbus timeout error happens to jenkins too as expected [00:06] its probably clicking off the screen or something knowing us lol [00:06] sigh [00:07] test_shuffle failed as well [00:07] thts not good more like media-hub isn't working or something lol [00:07] balloons, ahayzen, given how badly the timing is for all these tests... this might put all these tests into the flaky category [00:07] ahhhhh [00:08] 'GStreamer; Unable to pause - "file:///home/phablet/Music/1.ogg" [00:08] GStreamer; Unable to play - "file:///home/phablet/Music/1.ogg" [00:08] Debug: Queue: Now has: 3 tracks [00:08] Error: "Resource not found." ' [00:08] vthompson, didn't we have this before where gst wasn't using the faked home....oh no we're directly patching wtf... [00:09] ahayzen, vthompson I'd land this regardless at this point.. it does work locally [00:09] the dashboard might just have to be red :-) [00:09] mwahahah [00:09] ahayzen, the resource issue seems to happen if the file just simply isn't there even though tthe db says it is [00:09] vthompson, do the files have the wrong permissions ;) like doing an adb push [00:09] honestly likely running on the device will yield the same results [00:09] vthompson, but surely they are in /home/phablet/Music ? [00:10] ahayzen, well then gstreamer would see them [00:10] hah exactly....except when people were doing adb push and the files were owned by root [00:10] yea, then they can't be opened [00:10] vthompson, ahayzen having jenkins hate it means pain down the line, but we can land it [00:11] balloons, vthompson the test_shuffle, which failed on jenkins, is the one i've been using to test on device so i can say it works on device [00:11] File "/usr/lib/python2.7/dist-packages/music_app/tests/test_music.py", line 399, in test_add_songs_to_queue_from_songs_tab_and_play [00:11] MismatchError: After 10.0 seconds test on Player.isPlaying failed: True != dbus.Boolean(False, variant_level=1) [00:12] balloons, and everyone was so happy to get a green image soon.... here comes us to kill that idea [00:12] balloons, basically any test that tries to play audio ... [00:12] balloons, and as we're a music-app thts a few [00:12] vthompson, I think it will be green on the dash honestly [00:12] balloons, yea, since it is the device [00:12] yeah hopefully [00:12] so anyways, what say yea? [00:12] ahayzen, let's just not play music so all the tests will pass [00:12] merge to trunk or play with jenkins? [00:13] ye, not yea.. my old english is terrible [00:13] how many tests appear to have failed? [00:13] * ahayzen is waiting for jenkins results [00:13] the 2 i posted [00:14] * balloons gives another [00:14] File "/usr/lib/python2.7/dist-packages/music_app/tests/test_music.py", line 280, in test_shuffle [00:14] MismatchError: After 10.0 seconds test on Player.isPlaying failed: True != dbus.Boolean(False, variant_level=1) [00:14] ok [00:14] balloons, i would like to see the videos before making a call [00:14] and more [00:14] haha [00:14] might be worthwhile seeing the videos to see what's up [00:14] balloons, just to check wht it is doing [00:14] everything failed like that [00:14] it's almost done [00:19] so jenkins doesn't seem to be able to start mediascanner-2.0? [00:20] 'start: Unknown job: mediascanner-2.0 [00:20] sh: 1: /usr/lib/*/mediasscanner-2.0/mediascanner-dbus-2.0: not found' [00:20] oh dear [00:20] that as well [00:20] that's what I see on the VM and also the kill doesn't work [00:20] perhaps we need to harden how we do this === chriadam|away is now known as chriadam [00:21] ....works for me.... lol [00:21] tests are cleaning up now [00:21] psssh [00:21] http://91.189.93.70:8080/job/generic-mediumtests-utopic/573/ [00:22] yeah the track simply isn't playing [00:22] everything from a UI/DB side is fine [00:23] there has to be something we can do to fix this up a bit [00:23] sure enough it never starts playing [00:23] balloons, wht is causing tht? gst is claiming tht 'Error: "Resource not found." [00:23] GStreamer; Unable to play - "file:///home/phablet/Music/1.ogg" ' [00:24] balloons, surely it is there? [00:25] so it has to be the setup right? [00:25] The setup for the failed tests look the same though [00:26] that or some permissions thing [00:26] ahayzen, yea, but that'd be universal [00:26] vthompson, yeah u would see tht on device [00:26] ahayzen, where are you seeing this? [00:26] balloons, http://91.189.93.70:8080/job/generic-mediumtests-utopic/573/testReport/junit/music_app.tests.test_music/TestMainWindow/test_play_pause_library_with_mouse_/ [00:26] balloons, just before the test log starts [00:27] ..from the syslog? [00:27] ahayzen, ohh you mean it's not playing because it hits play, but the file isn't found? [00:27] balloons, yep [00:28] balloons, thts the gstreamer error saying that it cannot find the file.... u can see we've given it the 'right' path but it thinks the file doesn't exist [00:28] oh, is /home/phablet not proper for jenkins? [00:28] ahayzen, this is easier to read: http://91.189.93.70:8080/job/generic-mediumtests-utopic/573/consoleText [00:28] oh [00:28] vthompson, lol.. yep that's it [00:29] hahah lol [00:29] do we set /home/phablet? [00:29] I thought i had a find/replace that patched up the sql file [00:29] balloons, dblocation = "home/phablet" [00:30] vthompson, and the find replace uses ^^ var [00:30] balloons, can we not just put $USER in there and it'll be fixed? [00:30] it's commented out [00:30] ahayzen, hm, we had it working with jenkins before [00:30] let's add it back [00:30] one sec [00:30] oh yeah [00:31] r524 [00:31] err523 [00:31] balloons, didn't we need the cleanups tht u removed? [00:32] ahayzen, no.. they were duplicated.. look at the top of setup [00:32] unless I'm crazy [00:32] balloons, oh yeah i see now [00:32] balloons, no i am :P [00:33] balloons, as u've added back patch home don't u now need to make dblocation = "home/phablet" use home/$USER ? [00:33] I'm confused how that worked on the VM? [00:33] oh ms2 was building the db prob? [00:33] vthompson, yeah [00:33] whereas in jenkins we can't start/stop it [00:33] vthompson, so why do we need it on jenkins [00:33] mm [00:33] should we just not start/stop it? [00:33] oh [00:33] +1 [00:34] well I thought we found it didn't work unless we did [00:34] rather than us patching it let mediascanner rebuild itself [00:34] again, the real fix was the axe to dbus :) [00:34] vthompson, if you have alot of files rebuilding isn't going to work [00:34] well, if we know what we need to patch we should do so since we dont know ms2 will do it [00:34] to be honest, our patching is a bit flaky as I said.. it should index all your media and make it availible [00:35] patching of the db is fine though it drops the table and readds a good one [00:35] balloons, can't we just let mediascanner2 patch itself...until we get the isolation working then we can patch proper [00:35] BUT that also makes us dependent on the db version as we found out [00:35] so having ms2 rebuild would allow us not to depend on that [00:35] vthompson, right.. in theory once we have isolation we won't need to patch the db [00:36] until then patching it is the only way.. and even it is little flawed imho [00:36] balloons, so can we not just leave it as it is... and blame the fact we can't restart mediascanner2 on jenkins [00:36] so I'm OK with either approach, but it sounds like we could get rid of the patching now? [00:36] but jenkins doesn't want to rescan [00:36] we could kill -9 the daemon and it would scan upon restart? [00:36] vthompson, if you let mediascanner rebuild it.. it will find the 3 files. and all the other files you have on your device [00:37] thus destroying our tests [00:37] wht is the reason for us not being able to restart mediascanner [00:37] ok, we are currently starting ms2 before our test and waiting 10 seconds... we should remove that then [00:37] you think just dbus axing is enough? [00:38] I'm not sure how music + the services work.. so I guess I should stay more silent on it honestly [00:38] mediascanner was supposed to index only the main directories in home and sure as hell should not index hidden directories [00:38] hmmm maybe ....but if we stop dbus how will the plugin work? [00:38] vthompson, I can say mediascanner scans outside of ~/Music.. I launched music and had music again [00:38] and yeah mediascanner is only supposed to scan certain dirs ... not hidden [00:39] turns out there was a phone backup I did that had a few music files in it at some point.. thus music showed them [00:39] I didn't even know it was there, but ms found it :-) [00:39] sneaky hobbit [00:39] so back to the jenkins failures [00:40] we think we have that with patching. we don't know why jenkins cant restart the service, but we can probably change the tests to rely on us hacking the db without restarting ms2 [00:41] ...or we can just ship #522 and ignore the jenkins failures... until we have proper isolation and sort it out then [00:41] the user home seemed to be /home/ubuntu. maybe mediascanner2 isn't setup as an upstart on jenkins for it? [00:42] I want clean jenkins if we can. we've come so far [00:42] yeah it would be nice.... the best solution would be to figure out why jenkins can't stop/start ms [00:43] I think 523 will fix it... we just have shit in there that isn't doing anything. killing and start which returns that nothing happened [00:43] ahayzen, I think it might be what I said [00:43] yeah but then is tht needed for the vm/device? [00:43] no tracebacks yet [00:44] device has phablet with an upstart. and most people I'd assume do for their systems/vms [00:44] I think you are right, it will pass jenkins [00:44] I guess the garbage kills and starts that don't do what we expect can stay in... I just fear they'll remain indefinitely [00:44] yeah jenkins is actually looking good on the latest [00:45] i think it is safer to leave the starts/kills for now and we can review [00:45] otherwise we need to go and retest on all out platforms again [00:46] ahayzen, balloons, I just think the way we should do this is stop ms2 if it is running and we can stop it, fix the db, and then start it if we can [00:46] vthompson, yeah tht would be good [00:46] right now we start it assuming it will build the db for us, but we also go and hack the db... one or the other [00:48] vthompson, do we want to do this now or in a future branch/ [00:48] vthompson, because if this passes jenkins we could just land this [00:49] yea, I'd be OK with that... [00:49] not starting ms2 and waiting 10s each test would be nice though [00:49] yeah [00:56] http://91.189.93.70:8080/job/generic-mediumtests-utopic/574/ [00:57] balloons, vthompson i'm confused did it pass || fail? [00:57] I need a translator [00:57] +1 [00:57] oh it failed [00:57] 'Finished: FAILURE' [00:58] that's harsh [00:58] test_mp3 didn't work? [00:58] I can't find which test(s) failed. it's the silly xml thing agin [00:58] *again [00:58] 'KeyboardInterrupt' [00:58] so it was manually stopped? [00:59] vthompson, go here http://91.189.93.70:8080/job/generic-mediumtests-utopic/574/console [00:59] vthompson, then search for test_mp3 and read from there [00:59] http://91.189.93.70:8080/job/generic-mediumtests-utopic/574/consoleText [01:00] ahayzen, yea and that's at the bottom and I see the interupt [01:00] yea, results failed to come across well [01:00] balloons, it says a keyboard interrupt why would tht happen? [01:00] btw, I'm getting the same error as jenkins on my desktop [01:00] sh: 1: /usr/lib/*/mediasscanner-2.0/mediascanner-dbus-2.0: not found [01:00] balloons, i think tht is the error if the process isn't running before the tests and we're trying to kill it [01:00] balloons, so do I... but it works on the device. one reason I think this needs to be hardened a bit [01:01] ahayzen, that's starting it though [01:01] is it working on your desktop anymore? It's not for me now [01:01] wha [01:01] hang on [01:02] I ran test_mp3 and it worked on my VM [01:02] oh wait, let me pull and rerun [01:03] #520 was when i last tested [01:03] I just pushed 525 in which I tried to simplify everything to one block.. you guys were tossing around so many ideas about needing to stop or start the services, I made it simplier to change [01:03] just want to make sure I didn't break anything [01:04] the latest works for me on my vm, but I'm also not running start ms2 to get it to rebuild [01:04] #523 works for me [01:05] yea, see nothing is working for me now.. so weird [01:06] #525 works for me as well on the VM [01:06] good it's much simpler to read.. it's in one place [01:07] i would need to run on device to check if the cleanups are working though [01:08] balloons, ok so if 525 works are we going to land tht or are we gonna continue modding it ? [01:08] lol.. nothing works for me.. device nor desktop [01:08] ur holding it wrong ;) [01:09] I get sh: 1: /usr/lib/*/mediasscanner-2.0/mediascanner-dbus-2.0: not found [01:09] on the device too [01:09] I don't think that works, lol [01:09] yeah i get tht everywhere [01:10] bwahaha, we need a new term for these tests. They aren't flaky.. they're fickle [01:10] terrible? [01:10] same [01:10] lol [01:10] I rebooted and it's running on the device at least [01:10] hah the reboot worked for u as well [01:10] but our setup is still off clearly [01:11] ok #525 is good from my point of view [01:12] balloons, vthompson, i'm gonna go to sleep...i'll check if u've merged it in the morning and propose the remove-grilo one then [01:13] ahayzen, thanks for the help. we'll try to make it the best morning ever for ya [01:13] ahayzen, indeed.. sleep well my friend [01:13] balloons, vthompson u 2 guys ... [01:13] * ahayzen sends high five across the pond [01:13] vthompson, I'm happy with 525 as well.. expect for the fact we likely don't need the dbus start [01:13] clearly we don't as it's not working lol [01:14] yea, but that was part of what I tried to remove to allow the device to have the users music after the app was started following a test run [01:15] it's weird tho.. it's like it won't work without it [01:16] but it's not doing anything? [01:16] hmm.. actually [01:16] we should kill all these os.system calls anyway [01:17] and use subprocess? [01:17] yea.. that should shed more light if nothing else [01:17] well [01:17] let me push what I have if it works [01:18] ohh, changes? [01:18] go for it [01:18] well, not sure they work on VM yet [01:18] sure.. I'll just tweak the calls anyway [01:18] pushed r526 [01:20] vthompson, so you just leave ms off for the test? [01:20] yea, since we hacked the db anyway [01:20] a single test ran on both the device and vm for me [01:20] ohh well, if we can do that ;-) [01:23] yea, the isolation version doesn't seem to agree with that but whatevs [01:24] can't the isolation depend on us modifying the db? [01:24] just thought I'd try.. I'm confused how ms and the db + music interact [01:24] maybe if we replace the db and start the dbus service? [01:24] it's confusing [01:25] the dub portion of it certain throws in a glob of who knows what [01:25] *dbus [01:26] ohh.. indeed, just launching dbus works [01:26] I think it'd really be best if we tried to make sure that ms2 wasn't required to hack things together dynamically as we run [01:26] very interesting [01:29] green? [01:30] indeed.. on 55 [01:30] 525 [01:30] which also works on my device [01:30] and andrews [01:30] I think it worked on mine as well [01:30] I just wanted to clean up the calls and timing [01:31] yea.. I'm still lost on the whole mediascanner thing.. we try things, but then when we put it to the test so to speak it breaks [01:31] it's weird [01:32] yea, we can just ship 525 if that's your preference [01:33] no if it works, I'm happier to keep the logical train going [01:33] finishing the changes to subprocess.. I would like to ship that way [01:35] not even on the desktop? [01:35] oops sorry [01:36] wrong window [01:37] vthompson, just going to check that device looks good without starting ms and we'll ship this [01:38] balloons, pyflakes [01:38] ahh yes heh, I'll check [01:38] pushed a fix [01:41] So as an Ubuntu pioneer do I get a tshirt? [01:42] asdfasdf, you do! [01:42] vthompson, how do I do that? [01:42] asdfasdf, mine has yet to arive [01:42] vthompson, hah, my subprocess calls don't seem well liked [01:42] I'm eager to rock some Ubuntu gear [01:42] asdfasdf, someone will contact you to get your shipping information [01:42] vthompson, ok thanks [01:43] asdfasdf, np [01:43] balloons, by the fickle device and/or VM? [01:43] vthompson, my music folder is gone.. desktop all the way.. no need for vm :-) [01:44] I'm just changing them to shell=true [01:44] haha.. [01:44] wow [01:45] maybe now you'll consider a VM. :\ [01:45] is it at least in the backups? [01:46] vthompson, no, I don't back up my videos or music for space reasons [01:47] which is no longer a concern, because it's all gone [01:47] I just listen to pandora now [01:47] kind of lame [01:47] balloons, you're a man who truly suffers for his work [01:48] vthompson, ok I pushed the call changes [01:48] I'm top approving [01:48] and here I am trying to build a kick ass app you can use and you cant because you've hosed your library by helping me test :P [01:48] ok cool [01:48] vthompson, lol.. [01:48] if it supports media keys I would use it [01:48] I dislike rythmnbox [01:49] ohh right, push it [01:49] ha, I pulled and was like "phantom push" [01:50] k, going straight to the top.. it's approved [01:50] 30 mins from now, fingers crossed, we'll have a merged [01:51] hell after all this, what's 30 minutes?! [01:51] see my last comment [01:52] vthompson, and yea losing a decade of music was pretty amazing.. one big swoop [01:52] balloons, that is epic! [01:52] luckily popey pointed out the internet keeps good backups [01:52] the comment that is [01:53] that she does [01:54] vthompson, you'll never guess [01:54] I deleted all my music in college once.. I vowed never to let it happen again. Friends are also a good source of backups [01:54] james pinged me back just now [01:54] #1 timing [01:54] ha, well at least he might get good news soon [01:59] I shared our implementation [01:59] lol [01:59] I'm sure he'll love it [02:00] ha, well our backs were against the wall on this one [02:18] tests take forever; http://91.189.93.70:8080/job/generic-mediumtests-utopic/578/console [02:18] * balloons stands ready to build the final click [02:20] yea, I've been casually watching them... haven't seen a failure yet... that I've noticed [02:38] balloons, woot [02:42] * balloons is building [02:42] mwahaha [02:42] I;ll leave it for popey to land [02:42] ok, maybe it'll still hit the first image built tomorrow [02:43] mm.. potentially I guess [02:43] always seems rather early [02:48] vthompson, it's pushed for review [02:48] balloons, the click? cool [02:49] vthompson, yep.. left popey a message [02:49] and now, it's time for some R and R [02:49] you too.. get off the pc man! [02:50] balloons, will do. :) good work, glad we got this working [03:04] * bluezone finished packaging and resubmitted application [03:04] * bluezone does the chicken dance [03:07] davmor2, thanks for your advice a few hours ago, i also just realized you were the person who reviewed my app. I think i fixed everything up now, just wanted to say thanks for all the help :) === chihchun_afk is now known as chihchun === Walt is now known as Guest36177 [05:31] Hi there, I'm writing a scope and would like to save some files, but I have no idea where should I store my files, any suggestions? [05:34] http://developer.ubuntu.com/api/scopes/sdk-14.04/unity.scopes.Result/#a3f2e512b10dbf2ed867d260ec33a89a1 # my issue is that I have to draw my own image, and this set_art only accept file as it image source [05:34] It would be better if there is an overload one that accept QImage or something like that [06:03] AceLan, not sure what you mean but chck the DOs and Don't here http://developer.ubuntu.com/publish/apps/other-forms-of-submitting-apps/packaging-commercial-apps-part-2-packaging-software-additional-notes/ [06:04] it's for commercial but data storage is probably the same for open license too [06:04] bluezone: reading, thanks [06:53] bluezone: the page is very useful, but it says "Please use /home/$USER/.cache// for application data storage (eg game scores, level/game saves)", how can I know the user login name? [06:53] AceLan, are you using python? [06:53] bluezone: no, scope only supports Qt C++ [06:54] Hmm can you run terminal commands with C++? [06:54] Try running uname -a and parsing the output maybe? [06:54] or maybe there's some construct in C++ that is more cross platform [06:54] i don't know any C++ hehe [06:54] bluezone: aha, right, thanks, let me try [06:56] AceLan, $USER is the login name [06:57] brendand: yes, i know [06:57] AceLan, doesn't Qt have a path expansion function? [06:57] hmm [06:58] AceLan, like python has expandvars [06:58] AceLan, but of course C++ is not python :) [06:59] http://qt-project.org/doc/qt-4.8/qdir.html#homePath # Yes, it has QDir::homePath() [06:59] brendand: thanks [07:00] AceLan, welcome :) [07:00] bluezone, btw uname has nothing to do with usernames :) [07:02] brendand: yes, I know, I was planing to use "whoami" [07:02] brendand, username != login name? [07:02] when i do my uname -a i get bluezone which is my user and login [07:03] at least 'bluezone' is in there [07:03] bluezone: I think its your machine name as well :p [07:03] o [07:03] that's possible haha [07:03] i know it sounds like it does [07:07] bluezone, yes that is your machine name [07:07] bluezone, sometimes it matches, which could confuse [07:07] 'unix name' is what it is [07:10] good morning [07:18] popey, hey [07:19] AceLan take a look at QStandardPaths http://qt-project.org/doc/qt-5/qstandardpaths.html there's no need to be digging for the $USER :-D [07:20] DanChapman: thanks, I use QDir::homePath() now [07:37] if i have an installed click package, where does 'stuff' go? [07:38] like the qml files etc [07:38]