=== WebbyIT is now known as rpadovani === rpadovani is now known as WebbyIT === chihchun_afk is now known as chihchun [05:25] Good morning [07:07] rbasak: so I ran all (but two) of our > 300 autopkgtests on ARM with --ephemeral, and do all my local tests with that, too [07:07] rbasak: I never ran into any issues with it; would you mind much if I add a --clone and make --ephemeral the default? [07:43] Good morning [08:05] utlemming, rbasak: oh, to wrap up my "ancient container contents with trusty cloud images" issue, I didn't use the "daily" argument; that probably got me alpha-1 instead? [08:19] pitti, FTR, notifications for adt are working again [08:19] jibel: yay [08:27] utlemming, rbasak: hm, I just tried that again with "daily" and that still gives me alpha-2 [08:40] jibel: FYI, I just tried "sudo HTTP_PROXY=apt lxc-create -t ubuntu -n trusty -- -r trusty" on an ARM node [08:40] jibel: that DTRT and creates a fully working container that will work with autopkgtest 2.7 and doesn't require any further setup [08:41] jibel: and it's only 330 MB instead of the cloud image (640 MB), i. e. less stuff preinstalled which is good for testing [08:41] pitti, \o/ and it is smaller than a cloud image, right? [08:41] :) [08:41] right [08:42] still takes much longer to build of course [08:42] and faster when packages are already cached [08:42] but we can do it on one node and rsync it to the other three [08:43] jibel: perhaps we can sit down at the sprint and turn this into production? [08:44] pitti, sure. I hope that after the system apps project I'll have more time to spend on autopkgtest. [08:45] jibel: at least you need to show me the britney setup; this stuff desperately needs some fixing, I'd like to help you out while you are working on apps [08:45] jibel: so, we won't get bored at the sprint :) [08:46] pitti, I think we need to rethink this interface anyway. [08:47] *nod* [08:47] pitti, it is not complicated but there are lot of case to consider [08:47] +s [08:48] who would I talk to regarding potential apparmor issues? [08:49] veebers: jdstrand and jjohansen [08:49] veebers: tyhicks, too [08:49] pitti: awesome, thanks [08:56] pitti: the php5 dep8 test reliably fails using adt-virt-lxc and --ephemeral on our internal openstack deployment [08:56] I have had a 100% success rate reproducing it. [08:56] rbasak: yes, I know, but that seems to be the only one so far? [08:57] It has the biggest build tree that I know of. And is perhaps a little memory contrained [08:57] rbasak: we can still run with --clone on openstack, but it seems for local use its an option which you always want [08:57] I suspect it's a race to do with tar. It's because it's tar that I'm so sceptical about changing the default. Feels essential. [08:57] or on our ARM nodes; cloning containers there is rather expensive [08:58] the ARM nodes survived the firefox and mysql-5.5 tests, which are rather sizable too [08:58] I favour the default being the one least likely to trip you up first, and what people generally want interactively the second. [08:59] rbasak: hmkay [08:59] So I'd prefer you to use --ephemeral if you want to do so on any automation, since there isn't really any cost to doing that. [08:59] But, I'm not precious about it. Change it if you feel appropriate - you have to deal with it probably more than I do, and you've been looking after it, so it's your decision. That's just my opinion. [08:59] no, but it's an "always have to specify this" thing if you run it manually [08:59] Yeah, I don't like the "always have to specify this" thing either. But for me, less likely to trip you up is a higher priority. [09:00] rbasak: btw, http://people.canonical.com/~pitti/tmp/autopkgtest-arm/adt-php5/ [09:01] rbasak: perhaps this trips up due to the "deeper" stack of precise -> KVM -> lxc? [09:01] pitti: I believe you. It's just on openstack it happens. It's clearly some kind of race, possibly triggered by different memory/cache/buffer contention. [09:02] rbasak: so if I make --ephemeral the default and add a suitable warning to it, would that be ok for you? I don't want to change it if you don't want [09:03] pitti: in my opinion, we shouldn't change the default. Perhaps we could make it easier, like a one-letter option to do common things that you might want to do? However, if you want to change it anyway, then I'm fine with that. It's not a strong objection. [09:03] pitti: for example, in rsync, -P is a shortcut for --partial --progress. [09:04] rbasak: ok; I'll add an -e for it at least [09:04] Sure. That sounds like a good idea. [09:46] morning all [10:15] jibel: oh, we could even use lxc-create -t download to use stgraber's daily pre-built ARM trusty images (if we can access images.linuxcontainers.org, or shovel it to the lab in another fashion) [10:20] pitti, I didn't know this template. The lab doesn't have access to this site, but if we can allow it or find a way to pull images into the lab, that should be really fast. [10:21] jibel: well, the 10 mins of building it ourselves per day won't kill us either, but I just read about it and found it very interesting and useful [10:21] right, as you suggested we could as well create a base container on one of the node and sync it to the others [10:30] jibel: it seems the nodes can't currently IP each other, but I guess that can be fixed somehow === _salem is now known as salem_ === vrruiz_ is now known as rvr [11:55] jibel: ah, I'm finally getting adt mail again [11:56] meh, wazn's VMs drive me crazy [11:56] * pitti rebuilds the amd64 VM and disables wazn-adt for the time being [12:08] ok, VMs work again, wazn-adt reenabled [12:25] mmm - long time since I had anything to do with an LTS - anyone know if it's normal to set the notify of new version to For Any New Version when upgrading from LTS to LTS [14:07] good morning [14:08] ping rhuddie. Tell me where should I start with the browser today. [14:09] hey elopio, I was taking a look at ChromeDriver [14:09] elopio, don't know if you have any experience with it already? [14:10] rhuddie: yes, plenty. That's selenium, what we used on SST and the ubuntu one web servers. [14:10] rhuddie: but is it available while using oxide? [14:11] elopio, well that is what I was trying to establish... [14:12] rhuddie: I'd just ask the devs. [14:12] rhuddie: where is osomon on IRC? [14:13] elopio, I was talking to to him yesterday on #qa [14:14] rhuddie: I think #ubuntu-touch would be better. I'll ping him. [15:29] rhuddie: sorry. That was a long standup in unity, with many followup questions. [15:30] elopio, hey no problem [15:30] Not sure if this is a little off topic, but I was wondering how to use autopilot with PyGObject? It seems as though autopilot translates all Glade objects to Gtk2 (since glade is Gtk2) and I can't seem to find a way to "translate" them into pygobject (since my python code is pygobject and not pygtk2). [15:30] so, osomon said he will check about the web driver. [15:30] ah, crap, and he has just left, I was going to continue asking. [15:31] fully_human, I'm a little confused by your question. But perhaps pitti could shed some light here [15:31] rhuddie: do you have time on friday to meet with thomas strehl? at 14:00 UTC [15:32] elopio, yes that time is fine [15:32] jfunk: and you? ^ [15:32] fully_human, if you launch your app does in introspect with autopilot properly? [15:32] and jibel ^ [15:32] elopio: in meeting -- will catch up on backscroll later [15:32] mgr meeting [15:34] elopio, yes it is fine [15:35] balloons, I've created an interface using glade and am using python as my code base (with pygobject as the library). I'm trying to test using autopilot but the problem I run into is that the objects are all PyGTK2 objects, and not PyGObjects ("Gtk.Window" vs. "GtkWindow"). Not only that, but it's using C's type name (meaning, without the "."). [15:37] The application *launches* just fine but I have to call select_many with the C Gtk type name and not the PyGObject type name. (select_many('GtkWindow') works while select_many('Gtk.Window') does not) [15:38] Is there a way to translate the C-type Gtk objects to python GObjects? [15:38] Sorry, I feel like I'm not asking this correctly and confusing others and myself. :-\ [15:40] I think I understand now fully_human.. That said, I don't have an answer for you :-) thomi or pitti or veebers should know it, but none are around atm [15:41] balloons, Okay,t hanks. :-) [15:41] fully_human, email works well for this stuff [15:43] fully_human, I'd suggest sending something along to ubuntu-quality mailing list. [15:43] Okay, thanks. [15:43] Is this a quality question, though since it doesn't deal with the Ubuntu OS? [15:43] (not yet, at least :-) ) [15:44] fully_human, there is also #ubuntu-autopilot. But sure :-) [15:44] autopilot is a qa tool [15:49] I think I figured out most of my answer with the question "What is a Proxy Object?" in the docs. [15:50] mm.. those work too :-) === mhall119_ is now known as mhall119 [16:28] fully_human: no, AP uses GI a little, but in any way autopilot-gtk *only* uses GTK3 [16:28] fully_human: we don't support GTK2 at all [16:28] balloons: ^ [16:28] Ah, thanks. [16:28] fully_human: so if your application uses PyGTK, then we don't support that with AP [16:28] ty pitti.. I missed that tidbit he mentioned :-) [16:28] right pygtk = gtk2 [16:29] fully_human: also, GTK 2 and pygtk have been abandoned years ago, you shouldn't write new stuff with it [16:29] pitti, Right...I'm using PyGObject. :) [16:29] fully_human: oh - yes, the type name is still the Gtk name (GtkWindow), not the PyGObject transliteration [16:29] I discovered that autopilot got what's called a "proxy object." [16:29] Man, documentation is really helpful if you actually read it! :P [16:29] fully_human: from the POV of autopilot it doesn't know in which language you wrote your app [16:30] And then doing more research I found that to get an object from a glade file I need to pass in the BuilderName argument to select_single. [16:30] fully_human: Gtk.Object is just between pygobject and gobject-introspection, then everything is translated to the actual libgtk functions [16:30] Ah, thanks. [16:30] fully_human: yes, that's usually the best way to select an object: select_single(BuilderName="foo") === chihchun is now known as chihchun_afk [16:40] Where is this API documented? [17:34] balloons: bu 1274096 [17:34] bug 1274096 even [17:34] bug 1274096 in Ubuntu Manual Tests "Upgrade tests ignore LTS" [Undecided,New] https://launchpad.net/bugs/1274096 === bfiller is now known as bfiller_afk [17:47] elfy, we have upgrade tests I believe that we should toggle on [17:48] oh [17:48] oO, I see.. this might be different than what you are saying [17:48] let's go look regardless [17:48] in that case we'd best toggle them on then - I started shouting for our lot too then [17:48] yea.. 14.04 = lts [17:48] yep [17:49] I did one today - left with Prompt=normal as well - so I'm not sure if something that needs turning on with upgrade manager hasn't happened yet [17:54] balloons: oic what you mean - I'm talking about the wording which doesn't check what you get left with after upgrading [17:55] I can't login :-( [17:55] I hate sso ... :p [17:55] yes.. I noticed that after reading the bug a bit [17:55] something about ff hates it [17:56] yes, so I see the testsuite I was thinking of uses the same upgrade casse [17:56] so yes you are correct [17:57] ok - I'm half finished with that - I'll assign it etc and do mp [17:57] but - the other issue then is toggling the lts tests on? [17:58] yes, we need to add one testsuite to test specifically for the upgrade from precise [17:58] we've been in alpha mode, so it's fine to have left it alone for now [17:58] so let's see [17:58] yea - I'm just quick off the mark :p [17:58] or really annoying :) [18:00] balloons, in calc app there are a lot of old bugs related to autopilot mark as triaged, do you have time to veirfy if they are still vaild? Or explain me how to verify it [18:00] WebbyIT, link? and sure [18:01] so I added the upgrade.. I suppose I should include both upgrade tests. I guess it should become it's own entry under upgrade [18:01] balloons, https://bugs.launchpad.net/ubuntu-calculator-app/+bug/1256865 [18:01] Ubuntu bug 1256865 in Ubuntu Calculator App "test_simple_page.TestSimplePage.test_three_negative_numbers_multiplication fails on #40 on mako" [Undecided,Triaged] [18:01] Thanks :) [18:02] and bug 1256867 and bug 1256869 [18:02] bug 1256867 in Ubuntu Calculator App "test_simple_page.TestSimplePage.test_sign_button fails on #40 on mako" [Undecided,Triaged] https://launchpad.net/bugs/1256867 [18:02] bug 1256869 in Ubuntu Calculator App "test_simple_page.TestSimplePage.test_one_negative_number_multiplication fails on #40 on mako" [Undecided,Triaged] https://launchpad.net/bugs/1256869 [18:02] yes, old stuff like this can be closed. We should have caught it when we fixed the bugs [18:02] aka me.. since I commented on the cause :- [18:03] Ok, thanks :) [18:03] I mark them as fix released [18:03] I'm marking them as invalid actually [18:03] calc didn't change, it was a bug in mir [18:03] ok [18:03] I don't touch anything then :-) [18:03] assuming i remember it correctly [18:04] thanks for cleaning those up :-) [18:04] I'm doing some earlier spring clean :P [18:05] balloons: https://code.launchpad.net/~elfy/ubuntu-manual-tests/1274096/+merge/203808 [18:09] elfy, so I'm adding "LTS Desktop Upgrade (Precise)" suite for trusty to upgrade tests for ubuntu [18:11] like so: http://iso.qa.ubuntu.com/qatracker/milestones/308/builds/57251/testcases [18:12] I don't want to automatically add it to all the flavors I don't think [18:12] not all will support lts -> lts (not all shipped lts) [18:17] elfy, can you explain this line a bit more in the test?

Echoes Prompt=normal for non-LTS and Prompt=lts for LTS upgrades

[18:18] can we make it less confusing for folks? [18:18] That means more wordy [18:18] hehe [18:18] Also, it looks like spacing is a little off.. *twitch* [18:24] terminal echoes Prompt=normal or Prompt=lts [18:26] elfy something like "if you have installed an LTS release, then:" [18:26] I can make it as blatantly obvious and wordy as you want :p [18:26] elfy, lol.. yes, let's do that [18:26] okey doke if you're sure ... [18:26] ohh biy [18:26] I'm not sure ,I'm not sure! [18:26] had to lose 12000 word from my dissertation ... [18:26] lol [18:26] is it too late? [18:26] no - not done anything at all yet lol [18:27] seriously though, yes I think a little more explaining would be good [18:27] I'm still prevaricating [18:27] *would be useful [18:27] look at me bastardizing English again [18:27] I'll look a bit later - off to school - little one choosing options ... [18:28] I know - it is a S you keep mizzing out :p [18:28] cheers elfy ty [18:39] balloons: the 1310 upgrade is missing the "Proceed in your native language if you wish" bit - I'll add that as well [18:39] bbl === bfiller_afk is now known as bfiller [20:08] afternoon Letozaf_ [20:08] morning veebers :-) [20:08] balloons, hello [20:08] Hi balloons o/ [20:10] balloons: I'm back and it's pushed again [20:19] balloons, do calendar app tests work ? what I mean is that on my PC I have 8 tests failing :O [20:19] elfy, ty [20:19] Letozaf_, yes, they should be working [20:20] balloons, :? maybe I've got something wrong on my PC [20:20] sorry if I'm a bit slow.. I just took a big hit to the head.. :-( [20:20] balloons, :( ouch! [20:21] balloons, for instance http://paste.ubuntu.com/6839921/ [20:21] Letozaf_, fully updated? [20:22] balloons, yes [20:22] balloons, let me check though [20:22] balloons, never know [20:25] balloons, to make it simplier I just re-branched ubuntu-calendar-app and got the same error [20:27] balloons, If I wanted to cancel all the data and start from a pristine environment, wasn't the directory ~/.local/share//com.ubuntu.calendar/Databases the one to cancel ? [20:29] balloons, I haven't got it :? [20:34] Letozaf_, the directory is correct [20:35] balloons, oO then how come I haven't got it :? [20:35] * Letozaf_ thinks there's something weired on her PC [20:37] balloons, it's not that I want the tests to run at any cost on my PC, but I was wondering if it's a locale problem, that error sounds familiar to me [20:38] Letozaf_, it will be created when you ru nthe app [20:38] if you don't have one [20:38] Letozaf_, i agree that it seems like the locale error is still occuring [20:39] balloons, I will try to run the tests on my notebook, just in case ... === salem_ is now known as _salem [20:49] balloons, same error also on my notebook [20:55] stgraber, tracker visual updates? :) [21:14] balloons, I have a bug I think https://plus.google.com/photos/110772571847074999454/albums/5974426967349405889?authkey=CJH6tpHdxICPXQ you cannot see the evernote account when you launch reminders app on a device [21:16] Letozaf_, did you add it? [21:16] or are you talking about the text going offscreen [21:17] balloons, yes in settings, accounts [21:17] because that's a bug too it seems :-) [21:17] balloons, you cannot see the account label [21:17] Letozaf_, turning the phone sideways help? [21:17] balloons, but if you tap on it it enters the account [21:17] oH [21:17] gotcha [21:18] balloons, should I report a bug or is it "too early" as it's a "new app" [21:20] Letozaf_, no report it. .it's in the store and released [21:20] are you running the newest image? [21:20] and newest store release [21:20] of the app that is [21:21] balloons, I just updated the device I've got r152 [21:21] ahh.. you are running proposed then or stable? [21:21] I believe 152 would be proposed atm [21:22] balloons, proposed [21:22] balloons, ping [21:22] hey knome [21:22] we need to look at QA logos [21:22] the check mark in the circle of friends? [21:23] https://blueprints.launchpad.net/ubuntu/+spec/community-1311-quality-bugsquad [21:23] [knome] Help update the QA team logos to fit the Ubuntu Brand Guidelines: TODO [21:23] hmm.. indeed [21:25] so first things would be how many and what logos do we have / use [21:25] I can think of a few off the top.. and they do need updating [21:25] yep [21:25] well one of them is the hideous old ubuntu logo with the green checkmark [21:25] if you remember, i even did a new logo for you earlier [21:25] let me dig it up and upload [21:26] http://temp.knome.fi/qa/qa-final.svg [21:31] yes.. That logo should be appearing in places now [21:31] good [21:31] for instance; https://launchpad.net/ubuntu-manual-tests [21:31] one place i still see the old logo is https://launchpad.net/~ubuntu-qa-website-devel [21:31] although, I just noticed it's cut off a little [21:32] ugh.. that logo :-p [21:32] i can't see that (cut off) [21:33] hmm, go google [21:34] * knome wrote: bugsuqs, google suggested "bugsquad" [21:34] ok, I'll change the branding on those [21:34] possibly want to extend to ~bugsquad too [21:34] yes, ty [21:35] so branding wants 14x14, 64x64 and 192x192 [21:35] want me to exporT? [21:35] yea, if you could.. that'd make it simple [21:35] btw, 192x192 is unused... [21:35] heh.. I've not seen it [21:35] mostly 64x64.. [21:36] not sure I've seen the 14x14 much [21:37] it's used quite a lot [21:38] https://launchpad.net/~nskaggs [21:38] under "member since" [21:38] ahh.. cool [21:39] oh bah, will take some time [21:39] will poke the icon size a bit, direct export is too fluffy [21:43] knome, sorry I'll bbl. I'm not feeling so well atm [21:43] ok, np [21:44] i'll PM you the links [21:44] ty [21:44] get well soon :)