=== pitti_ is now known as pitti [06:51] good morning === mmrazik is now known as mmrazik|afk [08:54] xnox, could I pick your brain about the ubiquity autopilot test for a second? [08:55] sure. [08:55] what's up? [08:56] :-) in the launch_application method the get_autopilot_proxy_object_for_process method is asking for two arguments but only one given. What else is it looking for as I can't find it in the docs [08:57] DanChapman: do you have a full error message? is that from my wrapper scripts or something else? it could be that newer autopilot does things differently. [08:58] ok two secs ill get the message [09:01] xnox, http://pastebin.ubuntu.com/5783022/ [09:02] DanChapman: let me push a patch, it looks like second argument is emulator_base which can be set to None [09:03] let me first check if that works and/or something else is needed there. [09:03] xnox, great thanks :-) === mmrazik|afk is now known as mmrazik [09:12] DanChapman: pushed a fix. So just pull the branch and try again. [09:12] should work now. [09:19] xnox, awesome thanks mate [10:09] jibel: do you have an alioth account? [10:10] jibel: I asked about NMU'ing autopkgtest in Debian, and they said it would be fine to do our changes directly in git, so that we can get everything back in sync (git, debian, ubuntu) [10:11] jibel: or do you prefer routing changes through me for review? [10:12] jibel: ah, seems you don't [10:24] pitti, I don't hava an alioth account [10:24] have [10:53] thomi: do you happen to have an idea what provides the "window-mocker" program? It's being used in autopilot's test suite (but the functional parts don't run during package build) [10:57] thomi: oh, python-wndowmocker, sorry (doesn't seem to be in command-not-found yet) [11:25] we haven't updated command-not-found cache in ages..... there was somebody doing it for raring release, but as far as I know that was not uploaded. [11:26] pitti: i don't mind uploading changes to autopkgtest ;-) xnox on alioth. === mmrazik is now known as mmrazik|afk [13:04] stgraber, do you think you would have some time next week to go through the ISO tracker code with me to prepare for stuff proposed at https://lists.ubuntu.com/archives/ubuntu-quality/2013-June/003941.html ? [13:07] balloons, re: the bug icons, we probably should use the LP bug icons === _salem is now known as salem_ === mmrazik|afk is now known as mmrazik [13:52] balloons, ping [14:04] knome: so I posted some stuff to balloons last night but I don't think he read it yet. As a summary, I'm opposed to moving things to the sidebar as implementing this would be horribly hacky in Drupal and would potentially double the number of SQL queries required for the page [14:06] blocks are completely separate from the page and can't share data easily, there's also no reliable way for a Drupal module to force a specific block to be displayed and on-screen placement isn't guaranteed (varies between themes). Other communities have showed interest in using the qatracker for their testing, so I'd rather not make any assumption on the theme or exact Drupal configuration in the module. [14:18] stgraber, right. what about this: i rework the submit-form to suit the main area as well [14:18] stgraber, we've been talking about some js-enhancements, and one of those would be to move the form to the sidebar [14:19] DanChapman, pong [14:19] stgraber, (pick up the element with jquery, remove it, then readd to sidebar) [14:20] stgraber, if we had the css ready, js-enabled browsers could show the style currently proposed without the form being in the sidebar programmatically [14:20] knome, stgraber yes I read your notes stgraber. I assume some of the stuff we imagined would not be doable, which is ok. It's still safe to imagine :-) [14:20] balloons, hey have you still got examples of the _find and _walk functions I cant seem to find them [14:20] DanChapman, I believe so.. just a sec [14:20] cheers pal [14:20] stgraber, that's also one of the reasons why i don't use drupal ;) [14:21] knome: I guess that'd work. You'll have an hard time making the JS code work with the various layout configurations that are possible in Drupal, but I guess you can have it check that there's enough space and that it's a supported them and only do it in that case [14:21] vis doesn't give much help with ubiquity :-( [14:21] stgraber, the other option is to work on a generic drupal theme... [14:22] stgraber, it's a loooong time since i did that, but i can't imagine it being way too complex [14:22] knome: hmm, no, please no ;) we had that in the previous version of the tracker (custom layout, sidebars, ...) and it was a nightmare to maintain and port to newer version of Drupal [14:22] heh, ok [14:22] stgraber, in that case, is drupal a hard requirement? ;) [14:23] DanChapman: sup? I might be able to help you. as well, I know ubiquity inside out & can change ui element names as best suited for autopkgtests. [14:23] most of them are either unnamed or named rather mechanically. [14:23] stgraber, and re: various layouts, yes, the point would be to make it gracefully degrading, even if somebody didn't have JS at all [14:24] xnox, DanChapman has been the one continuing the hacking on the autopilot ubiquity tests.. If you two haven't met.. ;--) [14:24] stgraber, i don't know if you read all of my and balloons' discussions, but the main enhancement i'd like to do with js is being able to "detach" part of the page, eg. have the testcase in a popup rather than sticky in the page [14:25] balloons we met earlier xnox awesomely did a fix for me :-) [14:25] DanChapman, kk :-) good [14:25] knome: Drupal isn't an hard requirement, however the only allowed alternatives are Zope and Django and any plan to move away from Drupal needs to also be able to migrate all the existing data and user accounts (all 30 thousands of them) [14:25] brb, testing box says it's reset time [14:26] stgraber, ugh. in that case, we probably should stick with drupal ;) === salem_ is now known as _salem [14:26] balloons, DanChapman hey:) === _salem is now known as salem_ [14:27] knome: that was my thought last time I rewrote it ;) I would have been tempted to go with flask+storm in python but flask hasn't been approved by IS so I'd likely have ended up having to use Django which for some reason I really don't like... [14:27] stgraber, do we have the same problem (sidebar) with the testcase links? [14:27] xnox I have a few issues, one is to get the focus back to ubiquity once I start the test in term. Also I can't seem to get the GtkLabel which on the first screen is Welcome, I want to assert it through the test that it changes but using select_single('GtkLabel', label='Welcome') doesn't get me anywhere and can't find it in vis. [14:27] stgraber, well IS has been reluctact to allow stuff, but have done it before... [14:28] smartboyhw hey :-) [14:28] stgraber, anyway, if you have some time next week, we could go through some of the changes right away, and see what we need to change [14:29] knome: hmm, actually looking at those two blocks, it's possible we don't need SQL queries after all, let me take a look in the code [14:29] sure [14:30] xnox, I just thought would the label string be inside a element or something, I remember coming across that in a fileroller dialog. [14:31] DanChapman: I can select it with: self.app.select_single('GtkLabel', name='page_title') [14:32] DanChapman: and then test: .label to be Equals to whatever. [14:32] DanChapman: I open files with glade under /usr/share/ubiquity/gtk/*.ui [14:33] knome: right, so we already force to have one relatively large side block for the filters, making it a bit more generic and changing to your proposal when on */results should be doable. However I'd prefer we stick to a single block so you'd have to merge Testcase Links and Submit result somehow [14:33] DanChapman, I think this is more than you asked for, heh, but here: http://pastebin.com/s21ca2HJ [14:33] DanChapman: ubiquity.ui is the main one. And then you can click on different objects and see their class (e.g. GtkWindow, GtkLabel, etc) as well as their name (well, id='page_title') for example. [14:34] xnox, great thanks for that I didnt think of using glade to view it :-) [14:34] DanChapman: you have the source! use it. Or since you have the branch checkout out you can use glade to view gui/gtk/*.ui files instead =) [14:35] stgraber, that sounds doable [14:35] balloons, cheers mate perfect [14:35] knome: so to make sure I get this right, the sidebar is always adding a new result right? if you want to change an existing one you need to click the edit icon? [14:36] stgraber, yes, i think that's sane [14:36] right, so in theory the sidebar doesn't need any DB access then, good [14:36] stgraber, we also dropped the "in progress" -state, so... [14:37] stgraber, there was some discussion if that could be automatically probed in some way, but that's for later [14:37] knome: hmm, yeah, that's going to annoy the release team quite a bit I'd expect [14:37] stgraber, eg. check with JS if somebody is on the page and has inputted anything to any textform [14:37] stgraber, the dropping of "in progress" ? [14:37] yeah [14:37] we can always add it back [14:38] but that state should be sent to the tracker without a page load/submitted result anyway [14:38] during milestone testing I usually fire a dozen install at the same time, each will take around 1h to complete, so I mark them all in-progress so other tests can avoid doing those [14:38] (at least if you have JS, again...) [14:38] if we don't have (and use) in-progress, everyone will rush on Ubuntu Desktop i386/amd64 full disk install [14:39] do you think it would be sensible to do that with a JS call? [14:39] (or as a fallback, by saving a result) [14:41] well, we definitely can't assume the user will keep the page open while testing (I don't) and I'm not sure we can assume that starting to type something means that the test is in-progress either [14:42] DanChapman: no idea about focus thought. [14:42] hm, what if you check the in-progress checkbox and the tracker will remember that state until you submit a result? [14:42] as if you had submitted a result - just do that automatically [14:43] how would you indicate you're running that test multiple times (on VM + on hardware)? [14:43] stgraber, i suppose no way to indicate that. [14:43] stgraber, knome using comments? [14:43] so that'd be a regression from what we have currently [14:44] We can report multiple testing comments, and indicate where we have it tested (for example Virtualbox, or real machine) [14:44] comments could work, but i can see that's not optimal === mmrazik is now known as mmrazik|afk [14:45] knome, we do have comments in our original one. Besides, for package testing we really WANT comments... [14:45] knome: I think we should keep in-progress as a state, pretty much as it's currently. However to make it less painful, a user should be able to mark a bunch of tests as in-progress directly from the /testcases page without going to /results and we should have a separate view, similar to /subscription to list all the tests the user is doing, so they can easily mark those as past/fail or enter a complete result if needed [14:45] smartboyhw, we're not discussing that [14:45] knome, I know. [14:46] But still, comments should be kept. [14:46] smartboyhw, we're not discussing dropping them. [14:46] And I do use in-progress sometimes. [14:46] stgraber, can you shed a bit more light on how you use in progres... [14:46] stgraber, would it be good enough if there was two checkboxes, "i'm running this on hardware" and "i'm running this virtualized" [14:47] stgraber, (i know that's a non-working idea as is, but to generalize) [14:47] knome: grab a dozen ISO, figure out what I'll run on what, start around 20 installs at the same time (mix of virtual and hardware), update the tracker to indicate as much so I can remember what I'm actually testing [14:47] stgraber, i mean, we have hardware profiles... [14:48] stgraber, if you select hardware profile 1 and check the in-progress checkbox, you have one test running [14:48] stgraber, select hardware profile 2 and check the in-progress checkbox, you have another test running [14:48] right, except that for Edubuntu final testing, we run each test on 2-3 different hardware configuration and I often do those at the same time [14:49] VM 512MB of RAM, VM 1GB of RAM, VM with UEFI, HW with UEFI+SB being my usual ones [14:49] what if the hardware profile wasn't a dropdown but a series of checkboxes? [14:49] and checking any indicated you're running the test on that hardware [14:50] would that both make us able to drop the in-progress state and make it easy enough to report in progress tests on multiple setups at a time? [14:50] how would you add a bug report for one of those configuration but not the others then? [14:50] heh, you're right. [14:50] anyway, i need to go for a while now [14:50] will come back with the definite solution. [14:50] ...or not [14:50] see you later :) [14:51] later [14:52] balloons, knome: btw, as far as implementation, I might be able to find some time for that between the 22-26th of July as I've got a release sprint in London and will be doing some work on the tracker during that time anyway [14:53] stgraber, that's great news [15:09] pitti, apport, update-manager and software-properties adt tests failed with unmet deps on python3-pykde4, any idea why ? [15:10] jibel, what day did it fail? [15:10] there was a library issue with kde4 a couple days ago [15:10] balloons, Jun 20, 2013 2:51:25 PM :) [15:11] balloons, it is unrelated to the sip4 api mismatch [15:13] ahh no easy explanation then [15:20] balloons, ah, it is kind of related [15:20] The following packages have unmet dependencies: [15:20] python3-pykde4 : Depends: sip-py3api-9.2 [15:20] hehe [15:20] nailed it [15:31] jibel: oh, so it's pykde, not gtk [15:32] jibel: I guess there was some upload to -proposed causing uninstallability [15:32] jibel: yeah, http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt is full of that; new sip4 ABI at first sight [15:34] pitti, yes, xnox confirmed there is transition happening. My understanding was that britney would not submit tests for packages that are uninstallable. [15:35] pitti: jibel: yeah see http://people.canonical.com/~ubuntu-archive/proposed-migration/update_output.txt for sip4,python-qt4 at the bottom. I'm expecting britney trying to migrate those once we upload any of those reverse-depends, until eventually it succeeds. [15:38] xnox, no problem. We just deployed the interface between britney and autopkgtest and I'm making sure it is not a false positive caused by some nasty bug there, just me being paranoid :) === mmrazik|afk is now known as mmrazik [16:36] so is anyone else on saucy having unity issues? === mmrazik is now known as mmrazik|afk [16:36] my unity is completely dead.. rendering me quite crippled ;-( [16:37] when I say dead, I mean nothing will land.. no panel, dash, nothing === om26er is now known as om2 === om2 is now known as om26er [16:57] balloons: have you tried turning it off and on again? [16:58] balloons: i. e. even after a session restart? nothing like that here [16:58] pitti, IT Crowd <3 [16:58] pitti, yes, session restart, even purged it.. I'll dig deeper.. but I guess I'm the only one with those issues apparently [16:58] happened after yesterdays' updates [17:46] balloons, everything crashed after yesterday's update but a restart fixed it, and no crash today [18:02] total nuking time === francisco is now known as Guest74254 [19:23] whew.. ok, unity is back :) [19:29] balloons, hi [19:31] Letozaf_, hello.. my box hates me today [19:32] how are you? [19:32] balloons, fine, need you help again :( [19:32] hey Letozaf_ [19:32] balloons, for the calendar app, have you got a couple of minutes ' [19:32] ? [19:32] hey DanChapman [19:34] not sure I found a bug or "I am the bug": my assertion in test_calendar.py", line 82 gives me this errror: [19:34] balloons, self.assertThat(title_label.text, Eventually(Equals("Test event"))) [19:34] balloons, AttributeError: 'NoneType' object has no attribute 'text' [19:34] If I am not getting it wrong in autopilot vis I see the text attribute [19:34] balloons, i pushed the app on lp [19:35] Letozaf_, ohh.. so you can see it in vis yes? [19:35] Letozaf_,sure.. let's talk some calendar app [19:37] balloons, hope its not me getting it wrong again :p [19:39] Letozaf_, ok I'm pulling the branch again [19:39] balloons, ok thanks [19:40] lp:~carla-sella/ubuntu-calendar-app/calendar? [19:43] balloons, yes [19:45] ohh look.. I *think* my desktop will stay up for more than 10 seconds now [19:45] fingers crossed [19:47] balloons, http://ubuntuone.com/0rCv1987RU47H5JaMjdTxJ [19:47] balloons, oh! hope so :) [19:47] balloons, I mean fro your desktop [19:51] Letozaf_, ok so you can see it :-) hehe [19:52] balloons, do you think it's a bug ? [19:52] i think I don't know yet.. [19:52] balloons, :p [19:53] ok, so in your branch, it's missing stuff [19:53] this can't be the proper branch [19:53] oh hwait [19:53] balloons, argh! [19:53] no no.. I' updated, lol.. [19:54] balloons, fiew! [19:57] Letozaf_, ohh sorry. I meant, yes there is an issue. [19:57] bzr pull isn't pulling anything else and it can't find the emulator file [19:58] bzr log shows last commit of 6/17 [19:58] balloons, oh! did I push it wrong ? [19:59] Letozaf_, see where you pushed it [19:59] let's look [19:59] https://code.launchpad.net/~carla-sella [19:59] * Letozaf_ is looking [19:59] ahh, see [19:59] lol, I was wrong [19:59] lp:~carla-sella/ubuntu-calendar-app/calendar-app [20:00] balloons, I see all the files .. [20:01] balloons, is everything ok now ? [20:01] yes that branch is ggod [20:01] so I'm looking now [20:01] balloons, ok [20:07] balloons, as for Bug #1188833 what is the list view ? I was searching for it... :p cannot find this list view... [20:07] bug 1188833 in Ubuntu Calendar App "Autopilot Testcase Needed: Test event list view" [Undecided,In progress] https://launchpad.net/bugs/1188833 [20:07] We can ask the calendar guys what they mean [20:07] your right, I see no "list view" [20:08] ok Carla so this looks simple [20:08] balloons, tell me please [20:09] but apparently it's not SO simple, haha [20:09] balloons, oh :p [20:09] the first is to understand the timing issue [20:09] balloons, oh right! the timing... always forget about the timing.... [20:09] don't grab the label until after your finished inserting it.. it's not going to exist or be the right value :-) [20:10] that said, I changed the test to grab the label, then assert it [20:11] in general it works well to grab the object right before you need it, instead of at the beginnig of the test [20:11] that will save heartache [20:12] balloons, ok I must remember this [20:13] k, the other piece is to remember to use objectName and assign a name to the asserts you want to go after [20:13] it makes it easier to search for them, so you don't get the noneType errors (that means the search failed :-) ) [20:14] balloons, do you mean this: def get_title_label(self): [20:14] return self.app.select_single("Label", objectName="titleLabel") [20:15] balloons, in the main_window.py [20:15] ok so let's put it altogether and see if it works [20:15] yes, but I didn't see where you defined the objectName="titleLabel" in the qml source file [20:17] so with those changes it runs, but doesn't seem to pass [20:17] i notice that the event doesn't appear on the right date, heh [20:18] balloons, I reported the other Bug #1193090 [20:18] bug 1193090 in Ubuntu Calendar App "[calendar app] event time is wrong when creating a new event" [Undecided,New] https://launchpad.net/bugs/1193090 [20:18] balloons, the one you say did not happen to me, but let me check again [20:18] add a new event for today, it shows up yesterday [20:20] balloons, I swear it has the right date for me, how can it be ? [20:20] balloons, I inserted an event at 23:00 [20:20] balloons, could it depend on the time you put ? [20:20] balloons, what time did you put -? [20:20] Letozaf_, it might just be a bug for me then [20:20] balloons, could be... [20:20] i'm more worried about your test at the moment [20:20] balloons, sure [20:22] ok, so one more try ;- [20:23] I'm going to go for it specifically [20:23] self.app.select_single("Label", text="Test Event") [20:23] does that make sense? [20:24] balloons, yes [20:24] balloons, the "old" way [20:26] bah, it worked but I'm getting many events [20:26] we need something more defninitive [20:26] sadly we don't have it I don't think.. date, time, attendees, location [20:27] could still be many duplicate events [20:27] balloons, true [20:31] argh this is so annoying [20:33] well Letozaf_ I don't know.. this app has too many views, lol [20:33] balloons, :( don't you think that just date, time and title are enough? we could have to identical events, but how many could there be ? [20:33] sorry two not to [20:34] balloons, right ! [20:34] I have like ten test events at the same time [20:34] but it's showing at the wrong day [20:34] there's a lot wrong [20:34] balloons, I took a look at vis, but found nothing that could help [20:34] ohh it's certainly there [20:34] we can get it easy enough [20:35] grab label with text = 'event title' [20:36] I wanted to try the visible property, but that doesn't work [20:36] it may be enough to say it exists to be honest [20:37] balloons, yeah sound good [20:38] so i've got a version that should do the trick [20:38] let me push it [20:39] what I did was to just check to make sure we can get the label object.. nothing more [20:39] I made the title unqiue by adding a timestamp to it, hag [20:39] *hah [20:39] balloons, great [20:40] lp:~nskaggs/ubuntu-calendar-app/autopilot-title-label [20:41] balloons, I'm getting it [20:41] that was kind of silly to work out [20:44] balloons, I think you had a great idea, but I get errors, I'm on Sacy [20:45] balloons, no wait, had another calendar app opened [20:45] balloons, let me try again [20:45] :-) [20:46] http://paste.ubuntu.com/5784933/ [20:47] balloons, sorry but I get the error I paste, naughty Saucy [20:47] wild [20:47] that should happen because the events are timestamped [20:48] we can put in a fix though [20:48] so you can change to select_many [20:48] balloons, let me try [20:49] balloons, great! it worked [20:50] balloons, magic balloons :-) [20:52] lol.. not really [20:52] so, whew [20:52] balloons, I got no errors [20:56] I'm really confused now.. Can I use a VM through TestDrive to run ISO testcases, or not? (I'm not sure if this is the right place to ask either..) [20:58] sethj, yes TestDrive syncs the ISO's and then launches them in VM using Virtualbox, KVM or Parallels [20:59] sethj, http://www.youtube.com/watch?v=noDAsZjvmSA [21:00] Letozaf_, Thanks. I noticed most of the testcases start with "Boot up the iso using a CD/DVD or USB Key to a Live Session" though.. Is that the same as booting a VM? [21:01] sethj, yes [21:02] sethj, the VM boots from the ISO so it's the same [21:02] Letozaf_ Okay, great. I think that answers all my questions :) [21:02] sethj, :) [21:02] sethj, hmm good point. We could probably change the text to read or a VM :-) [21:04] That might be a good idea. I know I was confused there for a bit. [21:05] balloons, http://paste.ubuntu.com/5784976/ [21:06] balloons, for me the test looks fine and working [21:06] sethj, so the project has a place to report bugs and submit changes or new tests [21:06] https://launchpad.net/ubuntu-manual-tests/ [21:06] i'll file something on changing the text for you now [21:08] balloons, OK thanks. I did see the button "File a big against this testcase", should I do that next time? [21:08] *bug [21:08] https://bugs.launchpad.net/ubuntu-manual-tests/+bug/1193128 [21:09] Launchpad bug 1193128 in Ubuntu Manual Tests "Include VM in image test wording" [Undecided,New] [21:09] sethj, your more than welcome to file if you see something out of place.. or discuss it first :-) [21:09] Okay :) [21:10] I'm pretty new to all this... [21:13] balloons, I think the test is fine now, do you think I can propose merge ? I will check it again before doing it [21:13] sethj, hehe, no worries [21:14] Letozaf_, indeed, propose away [21:14] balloons, oh, no! first I have to find out what that list view is [21:14] balloons, that's missing [21:14] Letozaf_, ohh right.. we don't have to close every bug with this one merge [21:14] balloons, ok fine [21:14] balloons, thank you very much for your help [21:15] balloons, I going to bed now [21:15] balloons, 'night [21:19] g'night carla === hggdh_ is now known as hggdh === salem_ is now known as _salem [23:31] balloons: you about? [23:36] ah well, balloons have a look at https://wiki.ubuntu.com/Testing/Activities/Classroom/Saucy/Zsync In future, we should be careful what we wish for! I've got some serious homework to do!