=== ubot5` is now known as ubot5 [03:51] Good morning === chihchun_afk is now known as chihchun [04:04] thomi: hey Thomi, how are you? [04:05] hey pitti [04:05] pitti: spent all day bashing my head against this: https://bugs.launchpad.net/mir/+bug/1233944 [04:05] Ubuntu bug 1233944 in Mir "Unity8/Mir is unable to open autopilot uinput devices" [Critical,New] [04:05] about to hand it over to the european contingent... [04:06] thomi: ah, I'll have a look today [04:06] thomi: but /dev/input/eventX != uinput [04:07] the latter has its own udev rule for permissions [04:07] pitti: sure, the bug title may be a bit misleading [04:07] thomi: I wanted to ask you for a quick opinion about DanChapman's GTK emulator branch [04:07] thomi: (already pinged yesterday) [04:07] ahh, I saw that go by, I'm a bit conflicted about it TBH [04:08] thomi: my gut feeling is that we should put them into lp:autopilot into some autopilot.emulators.gtk submodule, instead of replicating an entire second python build system and packaging in libautopilot-gtk (that would need another binary package either way, though) [04:09] pitti: my concern isn't where to put them. I worry that we're basically creating two separate APIs, which will be confusing [04:09] thomi: how do you mean? [04:10] I think I'd rather see us improve the existing autopilot API for *all* UI toolkit backends, rather than improve it for Gtk only [04:10] "enter_text" method for text entry widgets... well, why not extend the keyboard instead (we actually already did that, but the example proves my point I think) [04:11] thomi: ah, I didn't even get that far with code review yet [04:12] so, replicating the generic click() methods and the like certainly sounds redundant === chihchun is now known as chihchun_afk [04:12] i. e. I don't really understand why we need so many wrappers [04:13] so I think a good thing would be to see what ideas are applicable to all of gtk/qt/unity, and apply those in autopilot 1.4 [04:13] I do see why it's useful to have emulators for GtkFileDialog and some standard bits [04:13] pitti: agreed [04:13] indeed [04:13] where it's nontrivial to pick out particular buttons etc. [04:13] and obviously those things are going to be specific to each ui toolkit [04:14] GtkTreeView._item(self, labelText) also seems highly useful [05:00] Good morning [05:01] bonjour jibel, ça va ? [05:02] bonjour pitti, ça va et toi? More manual phone testing for me today :/ [05:03] jibel: un peu fatigue, mais j'attends avec impatience le long week-end à Lyon [05:03] (pour le mariage de Didier) [05:05] pitti, ah great! I didn't know you were going. Have you been there already? [05:08] jibel: no, that will be my first tiem [05:08] time [05:08] jibel: last time I stayed at Didier's he still lived in Paris === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [05:52] good morning all [05:55] hey DanChapman, good morning [05:56] good morning pitti, thanks for finishing the review I'm just reading through it now :-) [05:56] DanChapman: not quite finished yet [05:57] DanChapman: I had a quick chat with thomi this morning about the structure, I gave a summary in the MP [05:57] once this gets simplified, I think we can put the remaining emulators into ap itself, and generalize the enter_text() bits etc. [06:58] pitti, ok cool thanks. I will start trimming it back. Ive left a comment on the MP if you could take a look for me. [07:05] DanChapman: so you are saying the two lines of code that you have for getting the second GtkTreeView would work with your branch, but not with current AP? [07:05] DanChapman: it certainly should work with AP, if not that sounds like a general bug [07:06] DanChapman: I mean, the solution certainly cannot be to duplicate the entire Gtk/Qt API with stub classes for emulators? there must be something better [07:10] pitti yeah it will only work if there is a stub class. [07:11] DanChapman: how does it fail without? The class names should be exported by ap-gtk? [07:15] pitti, let me get output for with and without so you can see :-) It says something like Class Gtk* has not attribute * without the stub. === chihchun_afk is now known as chihchun [07:26] pitti http://paste.ubuntu.com/6182845/ [07:28] DanChapman: I'm afraid that's a question for Thomi, i. e. how to make it possible to create an emulator for one particular class only [07:29] DanChapman: so you mostly created those stub classes so that everythign gets an emulator, to work around that bug? [07:29] at least now I understand the motivation :) [07:30] DanChapman: can you please explain this in the MP, so that it doesn't get lost on IRC for thomi? [07:48] elopio: hey, how are you? [08:55] pitti, sorry had to take my son to school. Yeah i had to create the stubs to workaround the issue, i didn't realise it was a bug at the time. I'll explain it all in the MP now [09:05] DanChapman: thanks; as I said, I'm not entirely sure how emulators work, whether they should indeed be this "I magically replace this class" thing or whether they need to be wrapped around the AP objects manually [09:05] pitti, hey [09:05] hey knome [09:06] pitti, i see you're mentioned in bug 1231978, is it possible it's a regression that's related with something you changed lately? [09:06] bug 1231978 in gvfs (Ubuntu) "Thunar 1.6.3 locks when browsing Trash with Xubuntu 13.10 Beta 2 and following dailies" [Critical,Confirmed] https://launchpad.net/bugs/1231978 [09:07] apparently that's not thunar only, but other file managers as well [09:07] it's a relatively nasty bug, so if you'd have any idea why that might be happening.. [09:07] I saw the bug, yes; there was a new upstream microrelease around that time [09:07] * pitti looks at https://git.gnome.org/browse/gvfs/log/ [09:07] cheers [09:11] knome: replied [09:12] ok, cheers, i'll get that test done ASAP === vrruiz_ is now known as rvr [09:22] morning all [09:23] hey slickymaster [09:23] hey slickymaster [09:24] DanChapman, Any progress on testdrive? Or autopilot for ubiquity [09:24] Hey DanChapman, Noskcaj [09:28] pitti, i don't seem to be one affected by that bug, so can't confirm if the older version fixes :/ [09:35] could anyone confirm bug 1234051? [09:35] bug 1234051 in ubuntu-system-settings (Ubuntu) "timezone settings are reset on reboot" [Medium,New] https://launchpad.net/bugs/1234051 [09:35] bug 1234051 [09:35] oops [09:35] jibel, ^ [09:38] om26er, sorry, my device is not in a suitable test to verify this bug at the moment. But I'll confirm later if nobody beat me to it [09:38] jibel, ack [09:38] s/test/state [09:43] Noskcaj: Hey the ubiquity tests are coming along now. Should have new ones finished later today/tomorrow and should work for any flavor (minus kubuntu) and any locale. [09:45] DanChapman, Awesome, great work. [09:46] pitti, so change copyright to "Copyright (C) Canonical Ltd 2013"? [09:47] Noskcaj: thanks :-) [09:55] DanChapman: right (but that's more like a finishing touch, don't worry about these small bits too much :) ) [10:03] Morning all === mzanetti is now known as mzanetti|otp === mzanetti|otp is now known as mzanetti === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === iahmad_ is now known as iahmad|afk [11:28] pitti: hello. [11:28] I'm about to leave to the gym. [11:29] can I help you quickly, or can you wait 1.5 hours? [11:32] oh, sorry. I misread "are you here" instead of "how are you" :D [11:32] I'm great, a little asleep still :) === iahmad|afk is now known as iahmad|afk|reall [11:38] elopio: no hurry; enjoy the gym! === iahmad_ is now known as iahmad| === iahmad| is now known as iahmad === _salem is now known as salem_ [13:24] I'm back, but I need to leave again because I'm stupid when I get up this early. I should stop doing it :) [13:25] I forgot I had to take the motorcycle to the mechanic. === Ursinha_ is now known as Ursinha [14:30] elopio, ping [14:30] cgoldberg, ping [14:37] jfunk, pong [14:57] cgoldberg, how Is testing click GOING? [14:58] pardon the caps lock [14:59] jfunk, good. i evaluated what exists for unit tests... and I found some issues in the Autopilot tests (they don't work) [14:59] jfunk, how should I deal with the broken AP tests. try to implement/fix them myself... or just file a bug? [15:00] cgoldberg, are they broken with Mir on, or were you testing with Surface Fliger SF? [15:00] jfunk, broken period. I don't see how they ever ran with the current code [15:00] im referring to click-update-manager [15:04] cgoldberg, hmm, do you know whether or not the tests are in the CI [15:05] they must have been removed [15:05] jfunk, i'm not sure.. if they are, they are failing [15:05] right got that [15:06] cgoldberg, so, at the moment, all eyes are on mir, click packages, system settings and multi media (mir should be on for the other 3) [15:06] jfunk, i should prob enter a defect against click-update-manager.. so at least it's being tracked and we can refer to it [15:07] cgoldberg, right, and you and elopio need to run the broken AP tests manually with mir on in the meantime :/ [15:07] i also MP'ed a branch to click-update-manager cleaning up the test code a bit [15:07] jfunk, ok.. will do [15:08] cgoldberg, it is vital that we are detecting the defects for the click apps and related cases often coming toward this release [15:08] ok [15:09] I know elopio is off getting his bike fixed again :p - perhaps when he returns you can coordinate with him: testing on proposed images with mir on, hit the click apps story [15:10] jfunk, sure thing.. i will coordinate with Leo [15:10] cgoldberg, elopio - I have a meeting with Ubuntu Engineering leads tomorrow morning, if you could create a report similar to Jean-Baptiste's Mir report - http://goo.gl/JT0ALH [15:11] including the automated tests that were broken which you ran manually instead [15:11] (see the bottom) [15:11] kk [15:14] balloons, so, how did your brainstorming go?:P [15:16] doanac: I tried all that with sudo -i etc. [15:16] doanac: I added the DBUS export as it was complaining about not having a session bus [15:17] ah, maybe I screwed up the syntax (I didn't use the double quotes at first), trying again [15:18] pitti: you might look at phablet-test-run. it does those gymnastics :) [15:19] smartboyhw, heh, I'm in firefight mode [15:19] balloons, LOL [15:20] doanac: ooh, got it now, nice! [15:20] doanac: the tests still fail and the dialer doesn't come up, but at least they start running [15:21] fun, I see like a miniature version of the dialer on the bottom edge of the screen [15:37] i'm back, now for real. [15:39] jfunk: not fixing again, fixing finally. Just today the parts arrived, so finally it's going to look like new. [15:39] in the mean time, I'm stuck traveling by bus. I hate life outside the internets. [15:39] cgoldberg: ping. [15:39] there's one thing with our idea of the mock server. [15:39] elopio, hey [15:40] now all our client code is c++. So if we make the mock on python, it will be only called in an easy and direct way from the autopilot tests. [15:40] cgoldberg: but maybe that's not a problem. Maybe the unit tests should mock the server responses on a lower level. [15:40] maybe it's just that I don't like c++ and I'm trying to justify using python :) [15:41] elopio, good idea :) for a start we could use it for AP tests... but it should be technology agnostic right? it's just an http server right? [15:42] you just need to hit it with HTTP requests to setup a response, then call the API [15:43] cgoldberg: yes. Maybe if we need it from c++, we write the files for the response, and call a binary. [15:44] if we need it from python, we can set up the responses on the fly, or even write an object that encapsulates all the possible responses and prepares them. [15:45] elopio, also.. read scrollback.. jfunk wants a report on the click testing. and we need to manually run tests where autopilot is not working [15:46] cgoldberg, elopio + mir on [15:46] jfunk, cgoldberg: yes, ack to all you said. [15:47] yesterday I ran all the tests again to check the documentation I made on moztrap. I'll do it today with mir enabled. [15:47] jfunk, cgoldberg: the document I sent to you with the notes to set up testing on trunk and the tests to run and the problems found is almost what jibel did on that other doc. [15:47] we just need to clean it a little. [15:48] cool [15:53] nuclearbob, hey got the video ready ? [15:54] om26er: it should be uploaded, I just need to share it [15:54] working on that now [15:55] I can't find it in the interface, unfortunately, but my machine says file sync has completed [15:55] I'll do the manual upload [15:56] yeah could be a bug in UbuntuOne [15:58] om26er: how about this: http://ubuntuone.com/0NOmNJbkg2noc3RAFrojQp [15:58] nuclearbob, yes, its downloading [16:22] jfunk: yesterday I focused on testing the U1 accounts plugin for system settings. [16:22] you are including that as part of "click" right? [16:24] elopio, I may be mistaken but I believe it should be more focused on the install / upgrade / uninstall [16:24] jfunk: you can't update click packages without an account. [16:24] and I think it's also used for the search. [16:25] I think it's important to get it right, because if people start adding a u1 account on a wrong way, next cycle we will see lots of funny errors. [16:25] elopio, ah, there's much I don't know about click I defer to your good judgement [16:25] * jfunk nods [18:56] morning all [19:05] morning thomi [19:10] hey cgoldberg [19:18] elopio, is there a standard or shared 'UbuntuTouchAppTestCase' that apps are using? or does each app use it's own? there's one here inside click-update-manager [19:35] * cgoldberg bbiab [20:25] afternoon letozaf [20:25] balloons, hello [20:25] well buonasera actually [20:25] balloons, buonasera :) [20:26] so I saw the rssreader got fixed -- did you see my note? [20:26] fixed=merged :-) [20:26] balloons, yes [20:26] balloons, great, I made the change you asked me [20:26] balloons, I am now working on the remove topic part of rssreader app test [20:26] balloons, nearly finished it [20:26] letozaf, I think you'll have to do another mp.. ahh, right, so you can fix it with that one ;-) [20:26] great thank you! [20:27] balloons, yw [20:28] balloons, I also changed the initial part where you check network activity [20:28] I saw, I take it that works a bit better now [20:28] balloons, it looks like [20:29] balloons, oh, but wait I changed it again now, you cannot see it until I propose for merge [20:29] balloons, it's even better [20:29] balloons, I think :p [20:30] cgoldberg: I've been writing one for ages now. [20:30] thomi, if you get a chance today, can you review the questions on this merge? https://code.launchpad.net/~dpniel/autopilot-gtk/autopilotgtkemulators/+merge/187673 [20:31] balloons: sure, I'll add it to my list [20:31] ty :- [20:31] cgoldberg: https://code.launchpad.net/~elopio/qtcreator-plugin-ubuntu/uitk_base_class-2/+merge/188924 [20:31] it just occures to me that I can simplify it on just two cases, so it needs a littme more work. [20:31] letozaf, :-) I'm sure it is. Glad rssreader is getting some love again [20:31] balloons, :-) [20:36] elopio, should we have a common 'ClickAppTestCase' inside ubuntuuitoolkit? [20:38] cgoldberg: I wanted that, but after some reviews it seems it would be too much magic to be launching the application on the base class. [20:38] so, we went jsut with ubuntuuitoolkit.base.UbuntuUIToolkitAppTestCase, that's pretty simple. [20:39] and we will duplicate what I have on my branch up there ^^ on every project. [20:39] it's not as bad as it seems. We will have just two cases, running on the desktop or running on the phone, so the code will be like half of what I currently wrote. [20:40] ok [20:43] cgoldberg: you can see that this new style removes the MainWindow :) [20:49] ohh cgoldberg going to join in hacking on the emulator? We'd love it1 [20:49] balloons, I will push the test so you can see what I did if you want, but I will not propose merge as I have a timing issue just at the end of the test, I put a sleep now, just for debugging [20:49] you can propose a work, and mark it work in progress [20:49] *propose a merge [20:49] balloons, good so I can finish it tomorrow [20:51] letozaf, right, and at the same time you can easily get a diff, et [20:53] balloons, should I push this in the same branch as the previous work or create a new one ? [20:53] letozaf, your choice [20:53] i usually go for new I guess [20:53] balloons, ok I will also go for new [20:57] balloons, where is it that you mark it WIP ? [20:57] * letozaf is looking for WIP [20:57] when you propose uncheck the box "ready for review" [20:57] or, click the exclamation point at the top under ready for review and set it [20:59] balloons, argh! I had already proposed for merge, the status now is pending review [20:59] letozaf, yep, just change it at the top [20:59] no worries [20:59] balloons, there is a need review choice should I select that ? [21:00] balloons, wait no... [21:00] letozaf, yep, the first one, work in progress [21:00] balloons, needs fixing sounds goog [21:00] sorry good [21:00] it's right at top, under status [21:01] balloons, LOL didn-t see it as it was written in black [21:01] balloons, lol [21:01] balloons, I was looking at it but could not see it :p [21:02] balloons, ok so I'm going to bed now, buona notte :D [21:02] letozaf, ciao! [21:13] elopio, is this going to land? https://code.launchpad.net/~elopio/ubuntu-ui-toolkit/fix_tab_switch/+merge/188774 [21:14] balloons: it's waiting for the prerequisite branch. [21:14] elopio, so everything is still broke eh? Well I got one MP in [21:15] balloons: the error that this fixes happens sometimes. So, it's more or less broken. [21:15] what's the URL of your MP? [21:19] elopio, I believe this one is most affected: https://code.launchpad.net/~nskaggs/sudoku-app/convert-ap-tests-to-sdk/+merge/187945 [21:19] I realize I can modify it to click tab names instead, and that should work eh [21:20] elopio, just a friendly reminder from julien... make sure you flash with the devel-propsed image for any click testing [21:23] balloons: yes, that's likely to be fixed with that branch. [21:24] I'm not sure clicking by tab names will fix it. [21:24] cgoldberg: ok. === salem_ is now known as _salem [21:55] thomi: once a click package is installed, if we do from the desktop phablet-test-run name_of_the_package, will it work just like it did when we had debs? [21:57] elopio: only if the AP test suite uses launch_click_package rather than launch_test_application [22:34] thomi: can you take a look? https://code.launchpad.net/~elopio/qtcreator-plugin-ubuntu/uitk_base_class-2/+merge/188924 [22:35] sure [22:57] elopio: LGTM