[04:31] So I guess I'm already to start contributing, I have test drive installed and synced the latest version of raring. [04:32] Do I just work on Week 3 tests ? [04:32] Orca perhaps? [04:53] Amacidia, hello [04:53] Hi balloons [04:53] How are you? [04:54] not too bad.. About to shut down for the evening, but let me answer your questions quickly if I can [04:54] one of the was you can contribute is by writing/running tests.. [04:55] Thanks balloons. I'm thinking of starting off by work on week 3 tasks as indicated by the schedule. Am I correct ? [04:55] Amacidia, yes, indeed go for it [04:55] I do some programming at work, however I don't work directly with python. [04:55] this is the close of the cadence week for week 3 [04:56] Amacidia, ahh.. well, if your willing to give it a try, we can help.. the python isnt too bad [04:57] I can definitely look into this week for sure. I'm not sure if the classes being planned for this month cover writing tests do they? [04:57] anyways, hopefully the iso tracker makes sense to you.. I'm generally around 1400-2200 UTC so feel free to ping, or email [04:57] Amacidia, no, but we have hackfests [04:58] there are some planned for this month [04:58] it does cover this stuff :-) [04:58] Oh well that would be awesome [04:58] I don't have the dates handy, I think 3 are scheduled right now [04:58] it was sent to the list in an email [04:58] if you don't find the dates, just ask.. ;-_) [04:58] I didn't join the lists too long ago... [04:59] I'll be on my proper pc again tomorrow, so we can chat more then [04:59] its night for me here at least.. so good night! [04:59] Take care, nice chatting with you. [04:59] nice meeting you! happy testing! [05:01] Good morning [05:06] Hey pitti [05:08] I'm new to testing, so this might be a dumb question, but I'm looking at the daily iso tracker for raring and it doesn't seem like a lot of people have submitted test results, is that normal? [05:11] Amacidia: yes, it is; we have some intense testing around milestones and regularly over the cycle, but not every day [05:12] Thanks pitti, so would you say I'm wasting my time working on daily iso testing? [05:12] no, feedback on those is always welcome [05:12] we have some automatic tests to ensure that they install at all, but there are a lot of problems only humans can see [05:13] sounds good [08:08] Good morning === Noskcaj is now known as Noskcaj-afk === Noskcaj-afk is now known as Noskcaj-sleeping === _salem is now known as salem_ [12:10] jibel: FYI, restarting ubiquity autopkgtest; that's supposed to work again, right? [12:38] pitti, thanks, it's back to normal [13:35] hello :) [13:36] has anybody got time to look at my file-roller autopilot branch ? it is not listed and executed by autopilot, and I can't find my mistake [13:37] my branch is here: lp:~d-kessel/ubuntu-autopilot-tests/file-roller === fader_` is now known as fader_ === plars_ is now known as plars [15:16] roadmr: how are lp:ubuntu/$series/checkbox repositories related to the ubuntu source package for checkbox? [15:17] zyga: awesome question [15:18] zyga: I don't think there's anything that auto-packages the bzr branch and generates the package; from what I've read, uploaders are still required to build and dput the package in addition to uploading the branch [15:18] roadmr: so they are related in theory but in practice that's just smoke and mirrors? [15:19] roadmr: can we upload to lp:ubuntu/$whatever/checkbox ourselves? [15:19] zyga: yes to the smoke and mirrors (if I understand the material correctly) [15:19] zyga: and we should be able to upload there ourselves, yes [15:21] roadmr: ok [15:21] roadmr: that's pretty good actually [15:22] zyga: yes, it'll be quite useful [15:22] zyga: of course if we want to upload to non-development releases (i.e. SRU) we need to get approvals first anyway [15:23] roadmr: wait, how does that work? [15:24] roadmr: if there was a can_upload(package, repo) function, how would you specialize it for package=='checkbox' [15:24] zyga: I *think* packages are held in some sort of queue and someone has to approve them moving to the actual archive [15:25] hmm [15:25] at all times? [15:25] or only during freezes? [15:25] zyga: I'm not 100% sure really, I need to read up on it a bit [15:26] roadmr: ok, it would be good to get that understood [15:26] zyga: for the development release, there's no queue/approval needed if there's no freeze [15:26] roadmr: and I'd like to get plainbox into ubuntu as well [15:26] zyga: after ff, of course, all uploads need some sort of approval [15:26] roadmr: having plainbox in the archive would simplify some stuff for me [15:31] zyga: yay, we can go the debian route but I'm sure there's a way to upload to ubuntu first/only, man, so much stuff to read and so little time.. [15:32] roadmr: I'd rather just get it to ubuntu now [15:32] roadmr: then if needed push to debian [15:32] zyga: makes sense [15:36] roadmr: I got access to the QA lab btw [15:36] zyga: yay! I also got the vpn keys, haven't configured them though [15:37] roadmr: we should sync this week to understand how to use that stuff [15:43] spineau: did you have the time to look at https://code.launchpad.net/~zkrynicki/checkbox/zyga/+merge/141865 [15:44] zyga: I've started. Need a closer look to the tests now [15:44] spineau: thanks [15:45] spineau: I'd like to get it landed to unblock the next branch which does a few big changes [15:46] zyga: I understand [15:46] spineau: also if there is anything that I can do to make such changes easier to review, tell me please [15:52] dkessel, I'll have a look [16:04] balloons, thanks [16:06] balloons, i will be back later [16:06] kk [17:12] https://code.launchpad.net/~zkrynicki/checkbox/zyga/+merge/142164 [17:16] * zyga takes a look at all the merge requests [17:17] dkessel, hmm.. not sure what's going on with your tests [17:17] that makes two then :) [17:17] ohh [17:17] lol , I see it [17:18] btw it would seem logical to name tests by the executables or packages. which is not the case for terminal, for example [17:19] hmm.. yes, something to think about [17:19] what did we do for gnome-terminal? [17:19] ahh.. just terminal [17:20] ok, so kill the '-' in the name, then it works.. also you have some non-ascii quotes in there [17:20] ' [17:21] dkessel, the reason is you can't have a dash in the python module name [17:21] oh :) lol [17:22] yes, hence my lol when I saw it [17:22] I looked at your filename and it clicked [17:22] meh :) but the folder naming would be ok? [17:23] it should align [17:23] so you should drop the dash there also.. [17:24] you could leave it and it would work, but it would be confusing [17:24] and there goes your naming convention dkessel :-( [17:24] yup.... [17:24] underscores are allowed though [17:24] potentially you could use that instead [17:24] * zyga is unappy about https://code.launchpad.net/~sylvain-pineau/checkbox/bug1091633/+merge/141054 [17:25] i've never seen an underscore in a package name, so it might work to replace dashes with underscores. [17:25] I believe it should.. I'm not sure I'm a fan, heh [17:25] but yea, it should work [17:26] underscore in package name will not work [17:27] h01ger, why do you say that? I did a quick change and it seems to recognize the module still [17:28] because underscores in package names are forbidden by debian policy as they seperate the package name from the version [17:30] h01ger, ohh yes yes.. I was speaking about python module names [17:30] I think we crossed some terminology dkessel [17:30] ah, ok [17:30] :-) [17:33] so using underscores would enable automation of stuff - a script could replace underscores by dashes and then know which package is tested [17:34] * zyga looks at roadmr's patches [17:36] :) [17:37] thanx [17:38] dkessel, hmm.. we could look to have a 1 for 1 replacement [17:38] undersocre probably not a good idea as a character, as it's used validly for other things [17:39] hmm ok - i'll wait before i commit my renames :) [17:48] hmm python identifiers seem quite limited [17:48] and *_* is a valid unrestricted identifier [17:49] http://docs.python.org/2/reference/lexical_analysis.html#identifiers [18:12] * zyga is back from dinner [18:15] is there an autopilot helper method to select a specific value in a combobox? if not, how would i do that given the item i want to select may not be at the same position in different program versions? [18:18] dkessel, I've been working with alesage and thomi on getting introspection and gtk apps going. [18:18] that I believe would be your solution [18:18] for the moment, you have to assume the positionj [18:21] roadmr: https://code.launchpad.net/~roadmr/checkbox/1084986-report-missing-dependencies/+merge/137702 commented but not marked as needs fixing as nothing here is really a deal-breaker [18:21] zyga: let's see [18:22] zyga: thanks! I'll fix all that stuff :) [18:29] roadmr: don't feel required to test everything I comment on [18:29] roadmr: I want to keep fluid coding possible [18:29] roadmr: only when you also agree that something is better otherwise [18:30] zyga: the defaultdict suggestion looks good to me, what I'm doing is indeed kludgy [18:30] roadmr: looking at it just now, typically I add a space after # but PEP8 won't complain if you don't do that [18:30] zyga: I didn't know += was that much slower, definitely good to change to join [18:30] roadmr: it's great to see the comments BTW, they make understanding the 'why' part much easier [18:31] zyga: as for the formatting thing, I think it was for it to fit in 80-columns (otherwise pep8 tears me a new one) but I'll try to prettify it, and thanks for catching the lack of i18n [18:31] roadmr: yeah, += in a for loop is quadractic while the same "" join is not, basically python does not overallocate string buffers since strings are immutable [18:32] roadmr: yeah, I assumed as much, maybe you can put it ine a separate variable so that it's got more space [18:32] zyga: oh that sounds good, that way I'm not constrained by the surrounding indentation === Ursinha_ is now known as Ursinha [19:54] Hi guys! [20:04] hey leozaf [20:05] I'm new around here, submitted my first test result last night actually [20:06] good evening [20:09] hello all [20:10] Amacidia, welcome, well I'm not a "guru" in testing but I can try to help you :) [20:11] dkessel, did you try autopilot? [20:11] yup i did. i started with the file-roller test conversion. i have converted the first steps, but then i got stuck. [20:12] now i know i used the wrong syntax :) don't use a '-' in a test file name :) [20:12] :) [20:13] so what do you think about it ? [20:14] i think it is nice. but it need some polishing and extensions to be more useful. [20:16] but they are working on it i heard [20:21] that's great news === jackson_ is now known as Noskcaj [20:25] morning everyone [20:26] good morning (well in Italy it's night :D ) [20:26] hello [20:27] hello balloons [20:37] hello, so let's see if thomi is about and see if he made any progress on the autopilot stuff we asked about :-) [20:38] Hi [20:38] I'm about [20:38] hehe :) [20:38] Happy Tuesday! [20:38] Thanks :) [20:38] :D [20:38] we just decided a few minutes ago to add a new feature to ap, which should help you guys a lot. We'll add a new command, probabvly something like this: [20:39] 'autopilot launch gedit' -> will launch gedit with introspection enabled, so you can use 'autopilot vis' to look at it's internals. [20:39] =) [20:39] balloons knows its Tuesday, yay [20:39] regarding the gedit introspection, there was an issue with the autopilot-gtk package in raring, which I'm told has now been fixed [20:39] excellent.. so have you been able to get a successful example? [20:40] so I should be able to push that branch now... [20:40] :-) [20:40] alesage: do you have your gedit example handy? [20:40] Noskcaj, lol.. yes, Tuesday and summer [20:40] I can be taught [20:40] :) [20:40] hi thomi, yes you can find it at lp:~allanlesage/+junk/UDS_AP_session === salem_ is now known as _salem [20:42] roadmr: as for UEFI, I'll make a small efi image for netboot testing tomorrow, do you think we could try spending a few hours on pair programming? [20:43] zyga: sure! [20:43] zyga: I didn't get anything uefi-related done today :/ so tomorrow we can certainly work for that [20:44] roadmr: have a look at that email I've sent on tuesday last week please [20:44] roadmr: the one with a few links to the wiki on how to get this up [20:44] roadmr: I'd like to try that on a real machine [20:44] uefi + netboot notes [20:44] roadmr: ok, till tomorrow then :) [20:45] roadmr: if you can look at the rather big MP, feel free: https://code.launchpad.net/~zkrynicki/checkbox/zyga/+merge/142164 [20:45] roadmr: no need to review it entirely, just post your notes [20:45] roadmr: most of the diff is code motion [20:45] balloons: this is what you want: http://bazaar.launchpad.net/~allanlesage/+junk/UDS_AP_session/view/head:/tests/test_gedit.py [20:45] zyga: excellent, I'll have a look [20:46] balloons: the two key things are: 1 - derive from GtkIntrospectionTestMixin and 2 - use self.launch_test_application(...) [20:51] zyga: for pxe netboot, we usually extract the iso and then patch the initrd on the fly. As I understand it, for UEFI we'd have to then reassemble the ISO with the initrd hacks, right? [20:51] hmm [20:51] doesn't seem to blow up [20:52] zyga: just some preliminary asking around, I'd have to review the initrd patches we apply, they may be unnecessary if a full iso (rather than nfs) is used [20:53] roadmr: I honestly don't know [20:53] roadmr: the way we boot today, we netboot the old-school installer, right [20:54] zyga: yep [20:54] roadmr: but I really don't even know how that works, I understand how we pxe boot the kernel but what is after that is a mystery to me really [20:54] zyga: we also need a way to load our custom preseed [20:54] roadmr: I could try to mimick that locally but raring still has no driver for my new laptop's eth [20:55] zyga: actually feel free to ignore me, I may be talking out of my ass since I haven't read the uefi page fully [20:55] roadmr: preseeds are a part of the installer but how do we get there is something I don't yet understand [20:55] roadmr: I think you are right in describing what works now [20:55] roadmr: how efi affects that is unsure [20:55] roadmr: from what I read efi netbooting is very primitive with free tools so far [20:55] zyga: I imagine something like building a custom iso with our preseed, but an alternative would be passing a kernel parameter, which may be easier than doing the meccano thing with isos [20:55] cool stuff [20:55] roadmr: and grub 2 has none of it [20:56] roadmr: so grub somehow gets mixed with a mini iso [20:56] ok, so things are working ;-) now just have to talk about using autopilot vis and looking at your dbus session [20:56] dkessel, you still about? [20:56] roadmr: so that a few stacks of tools get to a point where you basically do what you did in bios (after booting) [20:56] yes balloons [20:56] zyga: well last time I tried any form of uefi netbooting, I didn't get past the mini iso thing (felt too kludgy back then) [20:57] roadmr: native-aware boot would be better but is not implemented by grub, efilinux (IIRC) has some code for that but it's still far cry from working [20:57] zyga: since it now seems to be the recommended way, I guess we will have to do it that way :) [20:57] roadmr: we should ping cjwatson to know if there's some movement there [20:57] so, want to try doing some introspection on fileroller? [20:57] you needed it for something yes> [20:57] roadmr: yeah, unless there's a more "native" solution that's what was recommended last time I've checked [20:57] zyga: yep. It's not even that late so I may have time to do a bit of experimentation today [20:57] roadmr: ok, I'll take a break now [20:57] i wanted to manipulate a combobox, yes [20:58] roadmr: I'll talk to you tomorrow :) [20:58] zyga: enjoy, we can continue this tomorrow [20:58] roadmr: have a good evening :) [20:58] is a new autopilot version up? [20:58] zyga: enjoy your rosca :) [21:00] alesage, the fix is also in the ppa right? [21:00] If I remember your not on raring dkessel [21:00] balloons, correct--please ping if you find different :) [21:00] balloons, i have setup a quantal vm [21:03] alesage, autopilot-gtk_0.1-0ubuntu0 is the old version, right? [21:04] apt-cache show libautopilot-gtk0 [21:04] dkessel, yes-- [21:04] I have Version: 0.4-0ubuntu1+bzr14pkg0raring1 [21:04] dkessel, install libautopilot-gtk-dev [21:04] got that [21:04] and actually dkessel you may want to remove autopilot-gtk_0.1bla [21:05] as it's from before a renaming [21:05] oh ok [21:07] bug: autopilot is missing --version ;) [21:07] ok got the same version as balloons now [21:09] :-) [21:11] now how do i run file-roller with introspection? [21:12] ok, so autopilot has this tool called autopilot vis [21:12] yeah, i tried that [21:12] dkessel: very soon you'll be able to do 'autopilot launch file-roller' [21:12] you can try running it now.. by default, you'll see unity as a connection [21:13] dkessel: right now you need this: [21:13] alesage: I forgot what the env variable is... remind us please? [21:14] thomi types faster than me ;-) [21:14] thomi GTK_MODULES=autopilot-gtk:$GTK_MODULES file-roller, I believe [21:14] dkessel ^^ [21:14] * dkessel tries [21:15] alesage, aaah :) and there is a new "Root" connection [21:16] dkessel, and a pile of GTK+ widgets underneath :) [21:16] dkessel: yeah, that's the one [21:16] all sorts of stuff [21:16] * dkessel spends the next day poking around in gtk stuff :) [21:16] dkessel: please let us know if you find problems [21:17] * thomi writes 'autopilot launch' now [21:17] shouldn't take too long [21:18] hmm - hard to identify a program. both name and title of the gtkwindows are empty ? [21:19] dkessel: Gtk seems to create a lot of garbage in it's widget tree [21:19] dkessel: the bits you're looking for will be in there somewhere [21:19] dkessel, it takes some digging [21:19] ok [21:19] but it's a lot less clean that a Qt app for sure [21:20] the standard practise is to write a class called an 'emulator' that retrieves the useful bits from the introspection tree for your tests [21:21] you have some unity emulators in this way [21:21] but it would be good to get some gtk examples.. so we can work on that know [21:22] thomi, actually does it make sense to have emuators for us? [21:22] would we be able to avoid writing one for each app? [21:23] balloons: it makes sense to have functions that get commonly used components.... so if the component chsanges it's position in the introspection tree you don't need to update all your tests [21:23] 'emulator' is a terrible term for non-unity tests I realise [21:24] so yeah, I think it makes sense [21:33] ok, it is a start. i have the dialog and i have identified some of the components in it. however, for the test, i guess the "id" attribute values are generated at runtime? so i would have to guess and write some code like "get the first combobox in this dialog and select entry xyz in there" [21:37] dkessel: alesage: the way to do this is to find something unique to that widget. If the app is well written, you can say "select the ComboBox with the name 'someUniqueName'" [21:38] dkessel, Im also having a look around [21:38] if that's not possible, you may need to get a bit more creative - select the dialog box, then iterate over it's children or something similar [21:38] pro tip: autopilot's select_single, select_many and get_children_by_type are all very powerful - you can specify several filters... [21:38] thomi, balloons : i guess the real thing would be submitting patches upstream to get some real names in there... [21:38] for example: [21:39] combo_box_widget = app.select_single('ComboBox', name='foobar', someOtherProperty=123) [21:39] dkessel is this bug relevant? https://bugs.launchpad.net/autopilot-gtk/+bug/1082391 [21:39] the first parameter is the type name. The other parameters are filters to apply to the selection [21:39] Launchpad bug 1082391 in autopilot-gtk "Needs proper widget identifiers" [Undecided,New] [21:40] I guess having the widget ids visible would certainly help [21:41] alesage: did ted come back to you with anything useful WRT that? [21:41] thomi yes I got a brief intro, we'd have to decide how to expose 'em [21:41] yes, alesage - i guess that "affects me" ;) [21:42] alesage: easy to do? [21:42] thomi unknown, I'll have to review :/ [21:43] thomi will study for AWG and discuss then? [21:44] a search would be helpful :-) [21:45] balloons: in the vis tool? [21:46] yea, that would work around the annoying widget spam [21:46] in the interim [21:47] hm autopilot -h says: "Opens visualisation in xdot when not passed an output path to write file to. is it still possible to dump the entire tree to a file? [21:47] oops... missing " after "to." [21:48] dkessel: oops, that's a bug [21:48] dkessel: we haven't used xdot for a while, I'll clean up the help string [21:49] ok :) [21:49] i don't like dot files anyway [21:49] graphvis can't handle trees of that size anyway [21:49] ...which is why we stopped using it [21:50] yup - it crashed when i handled it a graph of our producation software once too.... [21:51] thanks for your work so far guys! [22:16] ...and i'm off [22:58] bah.. I cannot find the popup in the tree ;-( [22:59] any thoughts on finding a pop-up window.. in this case, the gedit prefs window? [23:01] I'm guessing it's a child somewhere.. hmm [23:13] sadface.. I simply can't find it.. [23:16] balloons, yeah this can be quite frustrating :( [23:16] I thought I had it.. maybe spawning under the gtkmenu tree.. but nope [23:19] balloons beware having to 'refresh' the autopilot vis tree--thomi sez it's possible to do so by clicking on a root node, then the child nodes are refreshed [23:20] balloons, if you're looking at a modal dialog then it's probably all the way up the tree [23:20] alesage, ok.. not sure what you mean [23:20] that's another feature we need. balloons if you file a bug I'll try and get around to doing it... [23:20] I have to click the top level root node perhaps [23:20] thomi, a search you mean? It would make this much easier [23:20] and I could ignore the nonsense spawning [23:21] balloons: both a search function, and a way to explicitly refresh the tree [23:21] alesage, lol.. I'm not going to look through it again, but I'm noting it [23:21] thomi, ok, let me file request [23:22] I'll go through it again tomorrow [23:22] I want to see if what I've got works without that [23:26] AttributeError: 'BamfWindow' object has no attribute 'select_single' [23:27] alright.. good night all.. good stuff, glad this is working now.. [23:27] have to parse through it a bit, and the enhancements your making will clean this up nicely [23:27] thanks again