=== jono is now known as Guest7396 === wgrant_ is now known as wgrant === Mirv_ is now known as Mirv [06:06] Hey [06:06] I need some help, how can I add something to the mail icon in the notification area? [06:07] Like say I wanted a program that interfaced with a website like twitter to show the notify about PMs or something [06:07] How could I add it to the notification icon? [07:26] Hey [07:27] Anybody there? [07:30] Hello? [07:30] Can someone tell me how can I add an indicator inside the default messaging menu? [07:32] sil2100, ping [07:33] odraencode, best to wait for tedg or larsu_ [07:33] Q___Q [07:33] Oh [07:33] It's a menu not an indicator [07:33] Why the eff is it a messaging menu, that is confusing D: [07:36] odraencode, you found the docs? [07:36] odraencode, http://developer.ubuntu.com/resources/platform/api/12-10/ under "messaging menu" [07:36] I thought http://developer.ubuntu.com/resources/technologies/messaging-menu/ was an indicator [07:38] tvoss_: pong! [07:40] didrocks, sil2100 for info autopilot with my patch to pass arguments to recordmydesktop has been uploaded to the archive, I'll update autopilot testsuite to run with a lower fps (6 should be good enough to view failures) and with --no-wm-check to avoid running in fullshot mode [07:40] mhr3: just got the e-mail about the dbus issues, and wow [07:40] jibel: \o/ let's see how it goes :) [07:41] didrocks, we'll see ... [07:42] jibel: we have as well good hints as sil2100 told from mhr3 about the dbus hang [07:42] jibel: yay! hm, we'll see about those 6-fps, I wonder how the vids will look like then [07:42] didrocks, also, I'm wondering if recordmydesktop really likes to be savagely killed, instead of being kindly aborted [07:42] sil2100, did you talk to Wellark_ about it? maybe hud itself is not the root cause [07:42] it's the hud spawing 120 dbus signals a second [07:43] mhr3: not yet, but will be poking him and Ted just in case [07:43] jibel: hum, yep, maybe SIGABRT would be better, they send a SIGKILL, right? [07:43] and btw good morning :) [07:44] sil2100, Wellark_, here's the full stats btw http://paste.ubuntu.com/5863196/ [07:44] good morning mhr3, nice work! [07:45] didrocks, i've been broken :( [07:45] mhr3: urgh? :/ [07:45] didrocks, had to read haskell code when i was trying to figure out the bustle file format [07:45] Good morning and indeed good work ;) [07:45] didrocks, yes, on success kill() is executed, which as the side effect to leak directories in /tmp/ [07:45] s/as/has/ [07:46] mhr3: I feel your pain then! but but monads are good for you ;) [07:46] didrocks, i don't think so :P [07:46] heh [07:47] jibel: if we send SIGTERM, it's going to save the video? [07:47] didrocks, SIGABRT terminates the program and removes the specified output file. [07:51] * didrocks wonders why they didn't look at the gtk-recordmydesktop code, it sends SIGINT [08:12] mzanetti, ping [08:18] Saviq: pong [08:21] launcher backend merged... can't believe it :D [08:21] Wellark_: ^ [08:21] mzanetti, yeah [08:21] mzanetti, so... unity8 autopilot suite runs in different configurations every time now [08:22] mzanetti, when ran on device, that kinda beats the purpose [08:22] mzanetti, ideas? [08:22] not following yet... [08:22] ah... you mean the size and input method? [08:22] yeah [08:22] geometry mostly [08:22] Saviq: there is also a szenario when running on device [08:22] szenario? [08:22] :D [08:22] Saviq: in that case it'll run fullscreen [08:23] f*** you [08:23] :P [08:23] mzanetti, it's just that "sz" is a valid sound in PL [08:23] in german too... [08:23] mzanetti, yeah, but much different [08:24] mzanetti, sz is ~ german sch [08:24] heh... [08:24] actually... there is a difference [08:24] so it just sounded fine in my head :D [08:24] sz like it was before is ß [08:24] n no [08:24] ah maybe before somewhen [08:24] yeah [08:24] anyways.. [08:24] yeah :D [08:25] so check out the header of the testcases.py [08:25] mzanetti, so how can we make it use that szenario? [08:25] Saviq: its hud.py by now [08:25] in there you find the scenarios :P [08:26] 3 of them [08:26] 2 for desktop, 1 for device [08:26] mzanetti, ah, so will it Just Work™? [08:26] Saviq: in theory yes [08:26] Saviq: haven't tried it in a while tho [08:26] mzanetti, nice, will know soon [08:28] 1 [08:28] 2 [08:28] 3 [08:29] hah [08:29] byobu's been breaking my workflow lately [08:33] mzanetti, I wonder if we should run without the fake Ubuntu.Application [08:33] mzanetti, and it doesn't kill the system unity8, so we're launching both the fake and the real camera app [08:36] Saviq: yeah... we'd need to kill the running shell [08:36] Saviq: are you using phablet-test-run ? [08:36] Saviq: it has an argument -n [08:36] that'll stop the running shell teporary [08:37] mzanetti, ah! [08:40] jeez, I could swear this was ubuntu: http://www.engadget.com/2013/07/10/playstation-3-youtube-send-to-tv/ [08:40] dandrader, it's UbuntuTV :D [08:42] mzanetti, re: Ubuntu.Application then - I'm inclined to split the fake out into a separate package with Provides: qtubuntu-shell, so that it won't get installed on the devices [08:43] mzanetti, we'd then test the real apps, only not sure how autopilot would cope with apps forked from our process [08:43] Saviq: autopilot should be fine with that [08:43] and yeah, the tests are mightly unreliable :/ [08:43] Saviq: hmm... we might need to change tests to close the launched apps again [08:43] on device, that is [08:44] mzanetti, yeah, ok, second step, then [08:44] Saviq: I kinda expected that [08:44] mzanetti, it fails to unlock like every other time [08:44] Saviq: oh, that bad [08:44] didn't expect that tbh [08:48] mzanetti, other times the shell doesn't switch to Home by itself [08:49] mzanetti, so it can't launch the app [08:49] huh? thats weird [08:49] I mean... it works when we normally run the shell [08:49] mzanetti, I think it doesn't if you unlock it quickly enough [08:50] Saviq: iirc we have a check that waits for that to happen before trying to unlock :/ [08:52] Saviq: this line used to do that at least: [08:52] self.assertThat(dash.showScopeOnLoaded, Eventually(Equals(""), timeout=30)) [08:52] mzanetti, no way to run just a single test? [08:52] Saviq: yes [08:52] phablet-test-run unity8.tests.testhud.test_show_hud [08:52] Saviq: run "autopilot list unity8" [08:52] just comes up empty [08:52] Saviq: you'll get a list of all tests [08:53] ah TestHud [08:53] Saviq: then run "autopilot run unity8.test.foo.bar." with the full path of the test you want to run [08:53] huh? [08:53] mzanetti, yeah, I was missing the class name [08:53] ah [08:58] mzanetti, something's weird [08:58] $ phablet-test-run unity8.tests.testhud.TestHud.test_show_hud [08:59] mzanetti, it completes the test case AFAICS [08:59] mzanetti, relaunches the shell and fails after some 30s [08:59] but only reports one test being run?? [09:00] ah no [09:00] Ran 2 tests in 232.057s [09:00] but just 1 failure [09:00] but why would it run two? [09:00] hmm... Saviq, what does autopilot list tell you? [09:00] Saviq: does the test have an * ? [09:01] *1 unity8.tests.testhud.TestHud.test_show_hud [09:01] Saviq: in that case it found 2 (or more) scenarios for the test [09:01] mzanetti, yeah, all of them do [09:01] running them all... [09:01] Saviq: can you add some print's in setUp to see which scenario it runs? [09:01] mzanetti, yeah, will do [09:02] Saviq: maybe the check for model() == Desktop doesn't work [09:02] mzanetti, otoh the failed test is (Fullscreen) [09:02] mzanetti, and we only have Fullscreen if != Desktop [09:02] hmm...yeah. so that one seems to do [09:03] Saviq: I'd need to run it myself to be of more use... give me minute and I'll try [09:04] dednick, fix conflict in indicators-client please? it's a small one in debian/control [09:04] Saviq: yup, doing now [09:06] mzanetti, ah, it actually ran "test_show_hud_button_dont_open(Fullscreen)", too [09:07] Saviq: yeah.. its a filter [09:07] mzanetti, right === larsu_ is now known as larsu [09:07] (which imho sucks sometimes) [09:08] mzanetti, so there's no way to limit it to just one in that case... [09:09] Saviq: pick another one where the filter only matches one :/ [09:09] :P [09:10] mzanetti, so... basically if it runs two tests, it fails to unlock the second time [09:10] mzanetti, ok, will investigate [09:10] Saviq: yeah... can reproduce it here too [09:11] the first test run actually doesn't look that bad [09:11] so if we're luck its a micro fix and it'll start working [09:11] yeah [09:11] mzanetti, any idea if we have a way to unblank the screen? [09:12] * mzanetti searches for the bug report [09:13] Saviq: https://bugs.launchpad.net/autopilot/+bug/1185065 [09:13] Ubuntu bug 1185065 in Unity "autopilot should prevent screen from locking during test runs" [Undecided,New] [09:14] Saviq: unfortuantely thomi decided for Wontfix [09:14] mzanetti, I kinda agree with him there... [09:14] Saviq: uh... I left it running while searching for the bug and now, after a couple of minutes it actually succeeded to unlock the screen [09:14] mzanetti, next test maybe? [09:15] mzanetti, that's what I meant by "every other time" [09:15] ok... might be [09:15] Saviq: you really agree every test suite should care by itself about the screen locking prevention? [09:16] * mzanetti doesn't [09:16] mzanetti, dunno, would have to think more [09:22] mzanetti, also, it should actively _unblank_, not only prevent timing out [09:22] Saviq: yep [09:22] Saviq: but again. thats an issue all tests will suffer, especially in jenkins [09:34] Saviq: package 'unity-shell-launcher=2' not found [09:34] dednick, ppa:ubuntu-unity/next [09:35] dednick, /me needs to fix the build scripts [09:35] Saviq: ah. is that a replacement, or a new one? [09:35] dednick, a replacement [09:35] desktop-deps? [09:35] dednick, not used anymore [09:35] dednick, we'll only do with ppa:ubuntu-unity/next [09:35] dednick, qt5 and sdk go away, too [09:35] Saviq: awesome. ta [09:36] mzanetti, D09add_ppa~ubuntu-unity~daily-build-next hmmm that's not the one we should be using [09:37] mzanetti, it should be ubuntu-unity-next [09:37] Saviq: right... lemme fix it [09:37] mzanetti, and we can get rid of all the others [09:37] mzanetti, just leave ppa:ubuntu-unity/next in there [09:37] mzanetti, actually, leave desktop-deps there until indicators-client merges [09:38] as that will fail autopilot [09:38] mzanetti, so, it looks like unlock actually passes, as the actual failure is: [09:38] MismatchError: After 10.0 seconds test on HudButton.opacity failed: 1.0 != dbus.Double(0.0, variant_level=1) [09:39] * Saviq traces unlock_greeter [09:39] this failure doesn't really sound like a unlock_greeter thing tho... [09:39] mzanetti, yeah, because the greeter didn't unlock [09:40] mzanetti, and the test continued assuming it's unlocked [09:41] Saviq: where did you see those ppa hooks? [09:41] mzanetti, http://s-jenkins:8080/job/unity8-ci/232/parameters/? [09:45] mzanetti, right, unlock_greeter doesn't actually fail if it fails to unlock [09:46] Saviq: yeah... just seen that too. well. its a helper. not a test... that's probably the reason [09:46] mzanetti, still, it needs to error out [09:46] mzanetti, and in the test, if you don't care - you catch and ignore [09:46] I remember the first version I wrote did fail [09:47] Saviq: does this look ok? http://s-jenkins:8080/job/unity8-ci/234/parameters/? [09:48] Saviq: the desktop deps is not in there... I'll just add that temporary in the webiface until we can drop it [09:48] mzanetti, ok [09:48] Saviq: this would be what the c2d config would look like [09:48] mzanetti, yeah, looking good [09:52] mzanetti, btw catch-all "except:" are evil [09:52] :) [09:53] Saviq, 100 revisions in lp:unity8, champagne time! ;) [09:53] mhr3, ;) [09:54] Saviq: https://code.launchpad.net/~mzanetti/cupstream2distro-config/drop-old-hooks/+merge/174146 [09:55] mzanetti, I top-approved, good? [09:55] Saviq: I'd prefer fginther to test-depoly it on his local jenkins instace [09:56] k [10:01] mzanetti, feels like an autopilot bug :| [10:01] mzanetti, the exact same drag doesn't work the second time [10:01] mzanetti, maybe it doesn't release or something [10:02] Saviq: really... thats weird [10:02] Saviq: you can check that with getevent on android side [10:04] mzanetti, not sure how to interpret it ;) [10:04] Saviq: log in to the android side on the phone (no clue how that works with flipped images tho) [10:05] Saviq: then run gevent -someparamicantremember [10:05] mzanetti, I'm there, dunno how to interpret the output [10:05] ah [10:05] Saviq: there should be at least the pressed and released quite obvious [10:06] mzanetti, right, so I'm missing the -someparamyoucantremember [10:06] Saviq: and there is another command line arg that prints it more readable [10:06] doesn't it have a -h? [10:06] I think it was -e [10:06] Saviq: ^ [10:06] mzanetti, there's no -e ;) [10:07] dammit :D [10:07] and -h isn't really -helpful [10:07] mzanetti, btw, adb shell; android-chroot [10:08] mzanetti: commented on the MP FYI [10:09] didrocks: hmm... yeah. I wasn't entirely sure about this. for some reason this seems to cause the usage of the daily-build-next too which is not what we want, right? [10:09] mzanetti: not sure on the upstream-merger, I think TBH the upstream-merger should never use any ppa, just a local repository [10:10] maybe it's to avoid the missing local repos? [10:10] but "ppa" is used as where to dput packages waiting validations for daily release [10:10] once built and the tests (using that ppa) pass [10:10] they are copied to "dest" [10:10] didrocks: ah, I see [10:10] (if no "dest" -> distro) [10:11] didrocks: but then there is no way NOT to use daily-build-next in the job right now? [10:11] mzanetti: you mean, upstream merger job? [10:11] didrocks: no, I mean our ci jobs [10:11] didrocks: ah.. .yeah... I guess you refer to our autolanding job. [10:11] didrocks: so yes [10:11] yeah, so upstream merger I guess ;) (but some people use "ci" for daily release as well, this is all confusing ;)) [10:11] it is :) [10:12] mzanetti: autolanding is even worse, more people confused, because landing == landing to distro :p [10:12] if you have better terms than upstream merger and daily release, I'm interested btw :) [10:12] so yeah, I think the ci/upstream-merger should have a local repo only [10:12] so that you can grab latest unity-api package for example [10:12] when building unity8 [10:12] didrocks: it has a local repo by now... we added that a few days ago [10:12] (without waiting next daily) [10:13] ok, so I think the hook (if it's a separate hook) to add the ppa from header can be removed [10:13] didrocks: but the problem is that if I keep "ppa: ubuntu-unity/daily-build-next" it will also get included in the used ppas while building [10:13] didrocks: thats the thing... it is not a separate hook. it gets generated from this [10:13] yeah, in practice, that shouldn't matter that much as the local repo should have higher version, but better to remove it [10:13] argh ;) [10:13] so yeah, should be a hook [10:14] requires a change in c2d then I guess [10:14] I'll talk to fginther ones he shows up [10:14] sounds the right plan :) [10:14] mzanetti, it's "-l" [10:14] didrocks: thanks for the info... [10:14] mzanetti: yw! ;) [10:23] mzanetti, I dunno, it's like input is somehow broken every other time :/ [10:23] mzanetti, even if I "help" it unlock [10:24] mzanetti, it will then fail launching the app [10:24] om26er: hey, what's your experinece with autopilot on the phone? [10:24] om26er: does it work reliably for you? [10:25] mzanetti, yeah these days its working fine, except for a new bug which veebers is working on [10:25] om26er: is that input related? [10:25] om26er: like input not working every other time? [10:25] mzanetti, we do need to have certain tricks to make sure the same test passes on the devices as they on desktops [10:25] mzanetti, no, that issue is with autopilot not being able to connect to apps [10:26] mzanetti, bug 1198277 [10:26] om26er: but you don't have problems with input not working at all? [10:26] bug 1198277 in Autopilot "Failed to create proxy object for gallery app" [Critical,In progress] https://launchpad.net/bugs/1198277 [10:26] mzanetti, by input you mean keyboard input or touch ? [10:26] om26er: touch [10:27] mzanetti, well I have not seen that in tests though in the morning i was not able to unlock my screen as touch was not working, had to turn off and turn on the screen again [10:27] Saviq: could this sound like the same? ^ [10:28] om26er: you mean with autopilot or even with your real finger? [10:28] real finger :D [10:28] mzanetti, yeah without autopilot, i.e. real world testing [10:28] hmm.. no. its something different then [10:29] Saviq: maybe we miss some dtor or the like that causes /dev/uinput to be closed? [10:29] Saviq: seems to work fine for apps [10:29] mzanetti, no, events come through [10:29] mzanetti, at least getevent says so [10:29] * Saviq instruments shell.qml [10:34] I've been having issues flashing my phone... can anyone help? http://paste.ubuntu.com/5864475/ [10:35] katie: is it a new phone? [10:35] katie: i.e. is still android on it? [10:35] mzanetti, no [10:35] mzanetti, its been flashed many times before [10:35] dednick, debian/control doesn't look good there [10:36] katie: is that everything you get as output? [10:36] katie: the output doesn't really contain an error [10:36] dednick, we need the | alternatives there [10:37] dednick, [10:37] 1233 - unity8-fake-env | qtubuntu-shell, [10:37] 1234 - unity8-private | unity-launcher-impl, [10:37] should've remained like thi [10:37] s [10:37] mzanetti, no, this is everything http://paste.ubuntu.com/5864482/ [10:37] Saviq: ah. crap. [10:37] katie: and then it just sits there and waits forever? [10:38] mzanetti, yes [10:38] ! [10:38] katie: you can try with phablet-flash -b [10:38] katie: that will erase settings tho [10:39] katie: i answered you elsewhere [10:39] you have outdated phablet-tools [10:39] mzanetti, that's what i tried :s [10:39] you need to update them _first_ [10:39] ah... thanks popey [10:39] popey, ok, i'm just going to try and update them first [10:39] popey, mzanetti thanks for your help [10:39] np [10:39] ☻ === dandrader is now known as dandrader|afk [10:39] Saviq: unity8-private is a dep higher up [10:39] dednick, that's fine [10:40] dednick, it's temporary to say "we need the default implementation of launcher, which is unity8-private now" [10:40] dednick, later it'll be something else [10:40] katie: feel free to ping us in #ubuntu-touch if you need further help [10:41] Saviq: i c [10:44] mzanetti, !!!! [10:45] mzanetti, enabling output made it pass [10:45] Saviq: ??? [10:45] and slow as hell [10:45] Saviq: ok, i think i've fixed it. can you take another look [10:45] what output? [10:46] mzanetti, -o [10:46] mzanetti, /me tries again [10:46] Saviq: that doesn't make any sense [10:47] hmm hmm hmm [10:47] Saviq: unless its really a bug in ap that bails out somewhere if that arg is missing [10:47] mzanetti, I think what made it pass is I had a left-over shell [10:48] that's definitely what made it slow [10:49] Saviq: hmm... it writes the xml file instead of dumping stuff to sdout/stderr [10:49] mzanetti, leave the real unity8 running → no more missing input [10:50] Saviq: weird thing is, apps stop the shell too :/ [10:51] Saviq: or did we release a change in that area lately and om26er might just not be up-to-date enough to hit the problem? [10:51] mzanetti, "apps stop the shell"? [10:52] mzanetti, you mean it gets SIGSTOPped? === dandrader|afk is now known as dandrader [10:55] Saviq: no... they run tests with phablet-test-run -n [10:55] Saviq: which translates to removing it from device-services and restarting the session [10:55] mzanetti, yeah, but they're not system sessions [10:56] mzanetti, well, I'm not sure yet what cause that [10:56] as I can't freakin' get output from a passed test [10:56] Saviq: yeah... ap doesn't think you would ever need that :P [11:01] greyback, ping [11:02] vesar: ping [11:03] mzanetti, really? there's no way to get that output :ADsdsd [11:03] Saviq: not that I would know of [11:03] [11:03] mzanetti, yep. [11:04] vesar: so, the Popover does not support manual positioning [11:04] vesar: SDK guys said I would need to open a request [11:04] vesar: are we sure we want to use the Popover for quicklists? if yes, I'd go ahead and create the request [11:05] greyback, ping [11:05] mzanetti, why do you mean it doesn't support manual positioning? [11:05] dandrader: pong [11:05] Saviq: does it? [11:05] mzanetti, check PageHeader out, it's doing *some* things manually [11:05] Saviq: thats not enough [11:05] mzanetti, k [11:06] greyback, I got a platform-api and mir that makes the surface created by maliit have mir_surface_type_inputmethod [11:06] mzanetti, wait still. I'm having a visual review with Marcus today afternoon. I'll cover using popover as a quicklist in that meeting. Will let you know right after the meeting. [11:06] greyback, do you want to try it out with your qml-demo-shell [11:06] vesar: ok, cool [11:06] dandrader: yeah definitely === alan_g is now known as alan_g|errands [11:06] greyback, awesome! I will put the debian packages somewhere for you to download [11:07] dandrader: I've one piece of work to complete before I can properly test it (need to Input shape the shell surface, so events will pass to the surfaces underneath) - should be only a few hours more work [11:07] dandrader: excellent :) [11:08] mzanetti, so anyway, there's completely no input coming in to the shell after the first run [11:09] mzanetti, and I've no idea why that would be :/ [11:09] Saviq: thats really weird... [11:09] mzanetti, but it's limited to autopilot touch === MacSlow is now known as MacSlow|lunch [11:10] mzanetti, as real finger, as you put it, does work [11:10] * Saviq creates a failure [11:10] Saviq: did you try with an app? [11:10] mzanetti, not yet, will do in a sec [11:15] * greyback going to buy usb wifi, tired of continually dropping out [11:18] greyback, ooh that bad? [11:20] mzanetti, ok, I take that back... [11:20] we do get input [11:20] * Saviq is lost [11:20] greyback: which chip? [11:22] * Saviq doesn't get it [11:22] Saviq: this is really strange indeed [11:22] mzanetti, when I put a raise Exception("bla") at the end of launch app helper [11:23]