=== chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === benonsoftware is now known as clockwork === clockwork is now known as benonsoftware [14:35] elopio, talk talk talk :) [14:35] brendand: fgimenez: https://code.launchpad.net/~veebers/autopilot/revert-ext-class-fix-for-now/+merge/251025 [14:36] https://bugs.launchpad.net/autopilot/+bug/1376996 [14:36] Ubuntu bug 1376996 in Autopilot "test_customised_proxy_classes_have_extension_classes started failing all of a sudden" [High,In progress] [14:36] elopio, so merge that with AP trunk [14:36] brendand: remove the skip, reintroduce the inheritance dark magic [14:36] there are some app tests that fail then. I don't know which ones, chris is the one who knows. [14:38] elopio, brendand this should be removed from trunk's head, right? [14:41] fgimenez: I think that the idea would be to make a branch for it, make sure that all autopilot selftests pass, and then run the gatekeeper job with that branch. [14:41] then check for the new gatekeeper errors related to inheritance. Make an autopilot selftests for that error, and fix it. [14:43] fgimenez: do you have the vpn set up? [14:44] elopio, i think so, let me try to launch it [14:44] elopio, how do we reintroduce the inheritance? just removing the skip doesn't make any test fail [14:45] fgimenez: hum, so that test is not good. [14:45] fgimenez: I just know part of this story, veebers knows more. [14:46] fgimenez: about building autopilot, I think you should join #ubuntu-ci-eng [14:47] I added a row in the spreadsheet, and then sil2100 assigned a silo to me. [14:47] on this page http://people.canonical.com/~platform/citrain_dashboard/#?q=autopilot you will see the build button. [14:48] this explains this weird process: https://wiki.ubuntu.com/citrain/LandingProcess [14:49] elopio, ok thanks a lot [15:02] elopio, wait, i was executing the unit tests [15:11] elopio, but yes, it makes no difference to execute python3 -m autopilot.run run autopilot.tests.functional.test_introspection_features with or without the skip... [15:12] elopio, http://paste.ubuntu.com/10773819/ [15:12] fgimenez: I remember it was failing, but it started failing suddenly. [15:12] I don't know what caused the failure. [15:17] elopio, the previous paste was with the skip, this is without http://paste.ubuntu.com/10773866/ and test_customised_proxy_classes_have_extension_classes passing [15:24] good afternoon quality people :D [15:25] fgimenez: can you please leave a comment on the card for veebers? [15:26] jfunk: can we work on the inheritance card before the prioritization meeting? [15:27] elopio, ok will do [15:29] pitti: I' [15:29] m getting an error on adt-run but executing the test script manually works [15:29] nuclearbob: hey [15:30] I'm grabbing the error now, it's about a ( [15:30] http://pastebin.ubuntu.com/10773980/ [15:30] and I'll grab the test script [15:32] pitti: test script is here: http://bazaar.launchpad.net/~canonical-platform-qa/ubuntu-power-tests/trunk/view/head:/debian/tests/tc-powermeter-05 and if I copy it over and execute it, it works [15:37] nuclearbob: do you actually execute it, or run it through bash? [15:37] nuclearbob: because, "function foo()" is a bashism [15:38] pitti: oh right, thanks, I'll change the shebang line, I had just been running it through bash [15:38] I'll remember to check that in the future [15:38] $ sh -n /tmp/tc-powermeter-05 [15:38] /tmp/tc-powermeter-05: 28: /tmp/tc-powermeter-05: Syntax error: "(" unexpected [15:38] yep [15:38] nuclearbob: or just drop "function" :) [15:38] nuclearbob: (it seems fine otherwise) [15:39] pitti: yeah, checkbashisms reports nothing when I remove that [15:41] elopio: yes, actually can you join a call? [15:41] nuclearbob: are you ok being the vanguard from 20:00 UTC to 23:00 UTC ? [15:41] elopio: we're postponing the prio meeting to tmw [15:41] jfunk: ok. [15:41] jfunk: the fixing sanity call? [15:42] elopio: I'd prefer to do earlier than that if that's feasible, that's 7 PM for me [15:42] nuclearbob: 17:00 UTC - 20:00 UTC ? [15:43] elopio: that works fine except for today, when I have an appointment :) I can do 20:00-23:00 tonight and then 17-20 after that, if that's okay [15:43] nuclearbob: don't worry about today. [15:43] elopio: okay. I'll pick up 17-20 UTC starting tomorrow [15:47] jfunk_: brendand: tomorrow the calendar is full of stakeholder meetings. [15:48] you moved it to the same time as the unity one. === chihchun is now known as chihchun_afk [16:08] elopio, do you have a moment to go over the webbrowser test failure? [16:08] rhuddie: yes I do. [16:09] elopio, excellent. This is the issue: http://pastebin.ubuntu.com/10771280/ [16:09] elopio, there seems to be some sort of race condition when clearing the address bar text in the browser when entering a new url [16:09] it doesn't happen every time [16:10] rhuddie: is it like you clear it, and then the page is loaded, and then it's filled again? [16:11] no, the browser loads with the start page, and then we try to enter in a new url. the helpers then clear the address bar before the new url is entered. [16:12] elopio, here is the test code: http://bazaar.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/trunk/view/head:/ubuntu_sanity_tests/tests/test_with_webbrowser.py#L130 [16:13] rhuddie: yes, but I'm wondering why the field is filled after we have cleared it. [16:13] is it because we are clicking in the wrong place? [16:14] elopio, well, it is filled by default with the start page. [16:14] rhuddie: it starts filled, then we cleared. [16:14] but the clear is failing. [16:15] yes the clear is failing [16:15] elopio, but currently the clear is not called directly in the test code, but by the uitk helper which enters the text [16:16] rhuddie: yes. What I don't get is why you say there's a race condition. [16:16] elopio, well, because sometimes it works, and sometimes it doesnt :) [16:17] rhuddie: but this is sequential. For some reason we are failing to clear it. [16:17] I'm wondering if it's because we are not hiting the button. [16:17] elopio, here is the method pressing the button: http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/trunk/view/head:/tests/autopilot/ubuntuuitoolkit/_custom_proxy_objects/_textfield.py#L74 [16:18] elopio, it seems like there has been some workaround added previously for checking the visible property [16:19] elopio, I've tried adding some waits to make sure the visible property is set true, but it did still fail in the same way [16:19] rhuddie: you lost me. [16:20] when the field has a clear button but it's not visible, the field is clicked to show it. [16:20] what workaround are you talking about? [16:22] elopio, that seemed a bit strange to me, that the text field had to be pressed again to show to show the clear button? but may be its not [16:22] rhuddie: it's not clicked again. It's if you try to clear it when it's not focused, so it's the first time you click it. [16:23] there could be something wrong in there, like we make something visible that hides the clear buttonl [16:23] rhuddie: do you have a screenshot of the error? [16:24] elopio, what happens is that the existing url is highlighted, and not deleted as expected. [16:24] rhuddie: that probably means that the clear button is not clicked. [16:26] elopio, there is another way to fix the issue, which is to call browser.main_window.address_bar.clear() before we try to enter the url in the test case. [16:26] it could be because we try to click it too fast, before it's visible. But that's not likely. We could add a clear_button.visible.wait_for(True), just to be sure about it. [16:27] or it could be because we are clicking the wrong coordinates, which is also unlikely. [16:27] rhuddie: that would do the same. [16:27] just with less ifs. [16:27] elopio, I tried adding the visible.wait_for(True), but it still failed in the same way [16:28] elopio, interestingly it didn't fail when I added the address_bar.clear() to the test. - I ran it for ages but it didn't fail once. [16:29] rhuddie, what about waiting for enabled? [16:29] rhuddie: what would be nice is to understand why that works. [16:29] rhuddie, it could potentially be disabled for a period of time [16:30] what would be different is that on write we call autopilot focused_write. [16:30] if we call clear directly, the focused doesn't happen. [16:30] elopio, yes I tried enabled too, but same result [16:31] rhuddie: so, that could be a valid workaround, but I would prefer to dig deeper and understand what's wrong calling clear from write. [16:31] elopio, yes, I was trying to understand why the helper was not working correctly... [16:32] rhuddie: can you debug the issue? like putting a breakpoint, call clear, and the field doesn't get cleared? [16:33] elopio, yes, I'll do some more digging. [16:42] wxl: you know your image is not booting on vb? [17:28] elfy: say what? [17:30] couple of days ago there was a pwconv /etc/passwd issue [17:30] today I got something else - booting it now for screenshot [17:30] but - image not for today currently [17:30] http://i.imgur.com/eBDaXxd.png [17:31] just thought I'd let you know [17:32] elfy: which image? [17:33] 64bit === bladernr` is now known as bladernr_ [18:23] elfy on my desktop the amd 64 image booted and installed with manaul partitoning fine [18:23] except for the eject part which all are broken [18:24] maybe a vb thing then - I don't look particularly deeply for other flavours unless asked or needed [18:24] and yea - that's still fubar [18:24] elfy: what vbox version? [18:24] ianorlin: did you use kvm? [18:25] I installed on bare metal [18:25] oh! [18:25] wxl: 4.3.26 [18:25] elfy: from the repos or upstream? [18:26] repos thiscycle it seems [18:27] elfy: no i mean vbox [18:27] or maybe that's what you meant :) [18:27] ok I will try vbox on this new bare metal install don't want the two hypervisors conflicting [18:27] i don't use repo versions so i'm not sure i can back you up at all [18:27] danke ianorlin make sure to grab from the repos [18:27] and confirm you have 4.3.26 [18:28] repos for vb wxl [18:34] yes I have that version and virtualbox fails [18:34] awww fooey [18:34] elfy: ↑ might want to check the other flavors more completely [18:34] I am SOO glad this vbox issue isn't seen in Xubuntu [18:35] last cycle was one after the other for us :p [18:35] not in xubuntu? how weird! [18:35] wxl: I would expect other flavours have QA leads [18:36] I just happened to see this for you when ianorlin asked me something the other day so followed along a while [18:36] well we did see it a while back, so did Mate [18:37] elfy: so can you email the list? [18:37] list? [18:38] the ubuntu quality list [18:38] ok also nocompcache isn't a magic fix [18:38] wxl: ack - but it'll be later on [18:38] k [18:39] wxl: what for though? [18:39] * ianorlin will try an alternate with vritualbox [18:43] wxl: did it - I'll likely not remember now I've made you aware of the issue and won't look again [18:54] elfy: i suggested it since other leads may need to know. mean they'll probably figure it out, but … [18:55] well I would have expected them to know I guess - it's been about on lubuntu for 2 or 3 days [18:56] anyway - such is life [18:56] yep [18:58] and a bunch of questions make me wish I'd not done it lol [18:58] * wxl ducks :) [18:58] I'd rather wxl mailed the list :p [18:59] forgive me for i have sinned elfy :) [18:59] especially as I've got a real pita issue with my isp [19:00] my mails to mailing lists - get marked as spam here - so I have to go webmail to unmark them :| [19:00] i'll reply [19:13] * elfy ponders lubuntu being upstart-less, wonder if that's part of this [19:18] made you aware - my job is done :p === FunnyLookinHat_ is now known as FunnyLookinHat