=== _salem is now known as salem_ === salem_ is now known as _salem === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun [04:34] thomi: is there anything else you want me to change on https://code.launchpad.net/~pitti/autopilot/print-tree/+merge/192134 ? [04:35] pitti: thanks for reminding me :) [04:37] thomi: no rush, was just wondering [04:38] pitti: approved. want me to top-approve as well? [04:39] thomi: sure, if you are happy with it; thanks [04:39] thomi: BTW: wait_select_single() FTW! [04:40] thomi: all these lambda: Eventually() loops I got rid of! [04:48] pitti: yeah, I know right? [04:48] pitti: now we need a strategy to phase out the pass-a-callable-to-assertThat-with-Eventually feature [06:38] Good morning [06:56] jibel: do you find that pulling from ftpmaster.internal/ is usually that slow? the software-properties test has spent 15 minutes on dist-upgrade and installing deps now, pulling with some 30 kB/s [06:56] pitti, it is usually really fast [06:56] ah, bit faster now, ~ 300 kB/s, but still painful [06:56] jibel: ok, so perhaps it's just under particularly heavy load ATM [06:57] jibel: trying to clean up the mess ATM :) [06:57] pitti, doing that too :) [06:57] pitti, I'll monitor the connection speed and if it doesn't improve during the morning I'll notify IS [06:57] jibel: new systemd in -proposed fixes the conflicts: consolekit, so yesterday's bunch of failures related to pykde should succeed again [06:58] if s-properties succeeds, I'l retry the others [06:58] the keystone/cinder/heat bunch is something else, leaving that for the server team [06:58] pitti, I restarted ubiquity update-manager and ubuntu-release-upgrader [06:58] ack [06:59] and apport [06:59] ok, you beat me to it :) [07:00] firefox 25 fails with 13 test failures and 2 crash files :/ [07:04] new error in pyzmq:i386 ValueError: bad marshal data (unknown type code) [07:05] and same error in python-novaclient:i386 [07:07] pitti, 'bad marshal data' is usually a corrupted pyc, right? [07:09] * jibel restarts novaclient [07:09] keystone and heat failed with a crash file [07:10] that's lot of different type of failures [07:15] jibel: not sure about the reason for the bad marshall data, but it sure does sound like a corrupted pyc, yes [07:15] but they are generated at package install, so that's weird [07:16] perhaps the new python in -proposed? [07:16] jibel: I already restarted pyzmq, doesn't help [07:19] pitti, all 'bad marshal' error happened on wazn, it could be related to the disk errors retoated mentioned last week. But I don't see anything unusual in the kernel logs [07:19] I'll bring the node down, and retry [07:20] jibel: oh, i386 specific failures on wazn, that rings a bell.. [07:20] I had hoped that this would have been gone with the upgrade to saucy [07:20] it could also be a kernel or kvm bug [07:21] jibel: i. e. you'll stop wazn and retry these tests? [07:21] pitti, yes [07:21] wazn-adt actually [07:21] I need to run to the supermarket, bbl [07:22] enjoy the supermarket :) [07:22] "Jenkins Fixed - trusty-adt-pyzmq 4" ☺ [07:22] jibel: need some breakfast :) [07:23] jibel: odd, that ran on wazn-adt, too [07:48] good morning all [07:49] jibel: wow, they *all* succeeded now? [07:49] Noskcaj, Hey I just missed you when i got back yesterday. So what's the plan for testdrive? [07:50] pitti, yes, that's weird. [07:50] morning DanChapman [07:50] pitti, the only message I found is "kvm: zapping shadow pages for mmio generation wraparound" but I read it's harmless [07:50] hey elfy :-) [07:50] morning DanChapman and elfy [07:51] hi jibel [07:51] Morning jibel o/ [07:51] jibel: I wonder if we should retry mysql-5.5 on !wazn for the fun of it :) [07:51] DanChapman, so, regarding the english/non-english testcase. IMO we should always test english as reference then another testcase for non-english languages [07:53] pitti, I think it's different becaue it fails on !wazn and amd64 [07:53] +s [07:53] hey DanChapman. Since porting to python3 (something i'd like to do soon) or Qt (an eventual option if ubuntu goes heavily that was) would break quickly anyway, it's probably best just to use the existing work. If you could finish my branch (quickly stuff and gsettings) it would be great. [07:54] After that i can try and fix the issues in the coding itself, e.g. giant bin/ files and giant, complex functions inside them [08:02] jibel, ok thats easy enough to do. The custom install test currently only uses english as its a PITA trying to create the partitions not knowing the language. I'll add the other english testcases today [08:04] Noskcaj. Sure, I will try and find some time over the next week or so and see what I can get done on it :-) [08:05] Great. If using you branch is easier, feel free. But i don't think it would be [10:00] jibel, should the ubiquity runner now work for all other flavors or just certain ones? I pulled the latest from your branch and tried it with xubuntu but no luck. I see theres a lubuntu config now so i was going to try that [10:01] DanChapman, it *should* work. I'll try xubuntu. [10:02] jibel i get bsdtar: casper/vmlinuz.efi: Not found in archive for both lubuntu/xubuntu [10:02] DanChapman, ah, I see the problem. [10:03] once I'm done with the automated deployment of the jobs [10:10] jibel, cool :-) just so you know i've just pushed the additional testcases to run in english [10:13] DanChapman, I've setup jenkins for edubuntu, lubuntu, ubuntu, ubuntu-gnome and xubuntu with the same tests than the tracker. [10:14] DanChapman, edubuntu fails to download because it's called dvd instead of daily-live and desktop [10:14] that'll need a fix in the downloader [10:14] DanChapman, and I'll file a ticket to have the views created on the public jenkins [10:15] DanChapman, I splitted all the jobs, with 1 job per testcase. I find easier to read and it's easier to rerun a single test [10:19] jibel, Awesome :-) it's going to be cool to see the output of running the different flavors. Fingers crossed the tests run ok across them all :-) [10:23] DanChapman, ubuntu-gnome works. I'll fix xubuntu and lubuntu to fallback to vmlinuz if there is no efi and the downloader to make it recognize dvd variants. === chihchun is now known as chihchun_afk [11:00] Morning all [11:32] jibel, is this a bug with the runner or autopilot itself. Since i uploaded the english testcases if for instance I run the ubiquity_autopilot_tests.tests.test_default testcase it also tries to run the test_default_eng test but they are seperate .py files test_default.py and test_default_eng.py. It seems strange that one would need entirely unique names for each test file === _salem is now known as salem_ [12:14] jibel: https://code.launchpad.net/~jibel/ubuntu/trusty/paramiko/lp1248706_enable_autopkgtest/+merge/194228 > does that actually test the installed package? [12:14] jibel: i. e. do you have to call something else to run them against the build tree? [12:33] pitti, the package uses tox with its own virtualenv to test against the build tree [12:34] pitti, without tox, test.py only appends tests/ to sys.path and uses paramiko fron the system [12:34] pitti, then the tests themselves import paramiko they find in path [12:41] DanChapman, I cannot think of anything in the runner that would do the expansion of test_default to test_default_eng, but I'll check === iahmad is now known as iahmad|afk [12:43] jibel ok cheers. It actually ended up running the test_default_eng first then the test failed trying to run test_default (which is what i originally executed and not _eng) after the install. I'll try and chat with thomi later if its an autopilot issue [12:46] DanChapman, I confirm there is nothing in the runner [12:46] the list of tests is built with echo "$TESTNAME" > $workdir/custom-installation/iso-override/var/local/autopilot/testsuites [12:47] then unpacked in the live session with [12:47] (cd $SPOOLDIR; touch $(cat $AP_TESTSUITES)) [12:48] jibel, ok cool thanks for looking :-) I will catch thomi later then [12:51] jibel: splendid, thanks [12:52] DanChapman, its an autopilot bug [12:52] $ autopilot list ubiquity_autopilot_tests.tests.test_default [12:52] returns [12:53] ubiquity_autopilot_tests.tests.test_default.DefaultInstallTestCase.test_default_install [12:53] ubiquity_autopilot_tests.tests.test_default_eng.EnglishDefaultInstallTestCase.test_default_install [12:55] DanChapman, if you call it test_eng_default that should work around the problem :) but not nice [12:56] jibel, thanks. i'll file a bug for that then, and change the names about and see how it works out :-) [13:01] DanChapman, FTR autopilot does a test_id.startswith(name) that's why it returns default and default_eng [13:01] that's annoying [13:03] jibel :-( your right, that is rather annoying. [13:05] DanChapman, it'd be better to do some kind of regex matching something like if re.match(name, test_id) [13:06] that would allow of more precise selection of tests [13:07] anyway, back to fixing lubuntu and xubuntu [13:07] DanChapman, FYI gnome-ubuntu test runs fine but never finishes. I'll have a look after L and X [13:12] jibel, ive just run a u-gnome install and it looked like the installation/vm continues after the test has failed/finished. I had to Ctrl+c to terminate qemu as it looked stalled [13:13] yes same here [13:19] jibel: can you please forward the python-imaging debdiff to Debian? [13:20] pitti, yes, I've 2 others to forward too [13:20] jibel: I uploaded the paramiko one (and committed to Debian) [13:21] the others need forwarding [13:21] pitti, excellent, thank you! [13:22] jibel: I suppose you already ran them through run-adt-test, so I don't have to again? [13:22] pitti, yes, that's how I developed them. and after pushing the branch I usually do a run-adt-test -S lp:... [13:25] jibel: ah, the other three introduce the first ubuntu change, so they will stop autosyncs; so you'll be on the hook for merging them :) [13:25] jibel: or do you just want to send them to Debian and let them trickle back through syncs? [13:26] pitti, we can just let them go through debian, there is no emergency here [13:33] DanChapman, I committed a small improvement for better detection of efi. That will fix lubuntu and xubuntu [13:37] DanChapman, thinking about testcase naming: you could name the 2 default cases, test_default_english and test_default_nonenglish, that'll closely match with the labels on the tracker. [13:39] jibel, yes that sounds a good idea I will do that then :-) awesome on the lubuntu, xubuntu runs gonna give them a try now [13:39] DanChapman, I'm checking what's wrong with gnome-ubuntu [13:40] the patch for autopilot for test selection is not that simple because it would change the behavior of autopilot [13:40] and alternative would be to stop if there is an exact match and proceed with a startwith() otherwise [14:08] jibel: mind if I drop the empty "Depends:" on adduser to make it start depending on adduser? (it seems odd to not depend on the package you test) [14:10] pitti, no problem, I realized that adduser is not seeded on a minimal Debian install, so the test wouldn't pass there [14:10] jibel: I also add your stderr comment to debian/tests/control [14:11] jibel: aaand .. uploaded, merci! [14:11] pitti, \o/ more automated tests [14:12] * jibel donne une accolade à pitti [14:13] * pitti te donne une accolade en retour [14:13] jibel: we now have 0x100 tests, plus that new one [14:14] * pitti goes to finally fix gvfs, too [15:32] good morning [15:43] elopio, morning! [15:43] hey SergioMeneses. Just yesterday I remembered about your testdrive branch. I'm sorry, a lot of things came this week. [15:46] elopio, rightI I dont had time enough to work on it last days =/ maybe this weekend I might do something [15:46] jibel, hey so I have just changed the testcase names to test_english_* & test_nonenglish_*. Also Lu/Xubuntu are both running great now. Nice job :-) [15:47] DanChapman, thanks, I'll update the jobs. I'm also fixing lot of little details related to the various environment, autostart are differents, log files are in different places, ... [15:49] DanChapman, but finally we'll have something running. It is nice to have this early in the cycle. Thank you for your awesome work! [16:00] pitti I'm having a crash-course in d-bus permissions--to your knowledge, is there a way to modify permissions programatically, i.e. from Python? [16:00] alesage: no, none at all; just through /etc/dbus/ [16:00] pitti, right, I'm discovering :) [16:01] alesage: but the usual way these days is to manage permissions of individual methods through polkit [16:01] polkit, interesting [16:01] alesage: and only give some coarse-grained policy in the config [16:01] pitti, have you witnessed that permissions differ in general between phone and desktop? [16:01] pitti, I'm able to signal on desktop but not on phab [16:02] alesage: yes, the phone installs some extra policy AFAIK, as does the desktop (policykit-desktop-privileges package) [16:02] . . . but need to write a few more tests to confirm [16:02] jibel awesome, its great how it's evolved from where it started :-) I have some tidying up of catching the non-fatals and printing them nicely, then on to finishing the dual-boot and upgrade tests. We should end up with a real nice testsuite :-) [16:03] pitti hmm ok [16:03] grrr, it's painful all the DEs store their session logs in different places [16:07] * pitti waves good night === zoopster_ is now known as zoopster [16:43] DanChapman, Hello [16:44] good afternoon senan :-) [16:44] DanChapman, Night 10.15 here ;) [16:45] DanChapman, Good Afternoon :) [16:46] senan, :-) your working late tonight then [16:46] DanChapman, I used to come online at this time.. after my day job :) [16:47] DanChapman, I'm facing some difficulty in scan remote folder dialog [16:48] DanChapman, its launching NautilusConnectToServer dialog [16:48] senan, so i guess you can't see that in vis? [16:48] DanChapman, when I close and relaunch vis, sometimes its showing [16:50] DanChapman,but its showing some error while launching scan remote using autopilot [16:50] senan whats the error? [16:52] I've a screenshot [16:52] DanChapman, How do I share the screenshot [16:53] senan use something like http://imgur.com/ [16:53] or you could just paste the console output to paste.ubuntu.com [16:57] DanChapman, "Autopilot GTK interfaceloaded." is not a valid folder, could not analyze disk usage [17:02] senan hmmmmm have you pushed your latest work to launchpad so I can take a look? [17:03] DanChapman, No I haven't done anything after that [17:17] DanChapman, Can I change the name of the file and branch now ? [17:18] senan_, what do you mean change the name of the file? [17:18] robotfuel: is it called qttest, or qtest? [17:18] DanChapman, I user test_diskUsageAnalyser.py [17:19] senan_, bzr move will let you rename it, is that what you are asking? [17:19] DanChapman, like to change it to analyzer just for consistency [17:19] elopio: is what call? [17:20] elopio: if you are doing qmltestrunner import QtTest 1.0 [17:20] robotfuel: yes, that, the qt test tool. [17:20] DanChapman, I used a mix of analyzer and analyser..mix of UK and US english [17:21] senan you can either do what balloons said 'bzr move' or you can rename it with nautilus or something but you will have to run 'bzr add' to update your branch then commit & push [17:22] * DanChapman waves to balloons [17:22] * balloons waves back [17:22] btw DanChapman glad to see you and jibel connected.. +1 on everything [17:22] we have a session for vUDS on the project, so we'll get to review everything then too.. looks like we'll have everything in place by then [17:24] DanChapman,balloons, how do I use bzr move [17:24] bzr move --help [17:24] same as mv command. [17:25] bzr mv test_diskUsageAnalyser.py test_diskusageanalyzer2.py [17:25] balloons, will i be able to change the name of the branch [17:26] bzr push to a new branch name [17:26] but then you'll loose your mp, etc [17:26] balloons, what is mp ? [17:26] your merge proposal ;-) [17:27] balloons, but that I can do it agin right [17:27] you can.. but why change the branch name? [17:27] balloons, same reason spelling [17:28] balloons, thanks we seem to be making good progress now. And yes I subscribed to the blueprint for vUDS yesterday so will be good to review it all :-) [17:29] balloons, I'll change the file name and keep the branch name same [17:31] balloons, In india, we follow british english in school and college, but after joining college all the books were american edition [17:32] senan_, hehe.. yes it can be a little confusing.. [17:33] balloons, sometimes I'll use Colour and sometime color, same as analyzer and analyser [17:35] you'll see american english from me, which steams elfy sometimes.. but I think he's just given me up as incapable of spelling flavour properly [17:37] balloons: I saw what you did there :p [17:37] hehe.. [17:38] sooo - if you can do it right once :p [17:40] once.. indeed, only once [17:41] * DanChapman can never understand what the need was to take an already legible english word and change it [17:42] indeed [18:30] balloons: phablet-test-run ubuntu_weather_app works for you? [18:30] I'm using image 15, and it can't find the application's dbus [18:30] elopio, you want me to pull trunk or the image version? [18:30] balloons: the image version [18:31] 1.0.154? [18:31] balloons: yes. [18:31] k, trying [18:36] elfy, mm.. blows u for me [18:37] I'm trying 1.0.155 though [18:38] looks like some of the tests partially work tho [18:40] initctl: Unknown instance: com.ubuntu.weather_weather_1.0.154 [18:40] could that be the problem? [18:41] elopio, I've always gotten that [18:41] elopio, does it not run through the tests at all? [18:41] Ran 15 tests in 238.367s [18:41] FAILED (failures=11) [18:42] I didn't let all of them run. Let me see. [18:42] balloons: there is a no ui-toolkit bug currently that simply mean some views are not visible. if you open setting and go to time date I don't know if that is part of the issues that the tests might have. [18:42] s/no/known [18:44] davmor2, thank you [18:47] FAILED (failures=15) [18:49] balloons: what does that mean? I've obviously missed a joke somewhere :) [18:50] elfy, ? [18:50] elfy, mm.. blows u for me [18:51] elfy, lol, sorry.. intended that for elopio :-) and it should read, "blows up", as in doesn't work [18:54] balloons: ok, I don't get why the weather app is failing, but the code is too complex. I'll start cleaning it up. [18:54] but first, I'll take a long lunch to pick my motorcycle... finally. [18:54] elopio, I'm really trying to circle in on fm [18:55] elopio, still not gotten it back? [18:55] wow! [18:55] I'll be back in two hours. [18:55] morning [18:55] balloons: no, it's been one month riding bicycle, I'm really tired :) [18:56] thomi: hey, good morning. The wait_for_destroy branch failed to land. It would be great if you can kick it. [18:56] elopio, this was my bike,, pure classic, I love it ;-) http://3.bp.blogspot.com/-_l8UdJKAxiU/T_B1gE1oeYI/AAAAAAAAAB8/pn5EQWR0iZA/s1600/1975+Honda+CB+500+T.JPG [18:56] elopio: sure [18:56] elopio, gl [18:56] balloons: oh, that's pretty. [18:57] balloons: ok - was a bit worried, looked rather perverse ;) [18:58] balloons: now I hate mine and want that one :( [18:58] well, see you later o/ [18:58] elopio, lol, really? people used to turn them into cafe racers.. i love the look of those vintage japanese mfg bikes [18:59] elopio, send along a pic when you get back hehe [18:59] elfy, yea, I get the one off statement now.. out of context, hah [19:00] :) [19:40] balloons.. nice bike.. what year? [19:41] balloons.. I had this when I was 18. it was a 1982 : http://www.ridermagazine.com/wp-content/uploads/2012/11/web-IMG_0872.jpg [19:41] cgoldberg, 1975.. they only made them for like 2 years.. it's a 500 twin.. the 450 twins were really popular [19:42] interesting.. not sure I've ever seen one of those [19:43] you can see the 80's influences creeping in there :-) [19:44] I'm not a fan of all the fairing they got into during the 80's [19:44] that still looks slick [19:46] cgoldberg, this was my first bike.. I wasn't even 16 and I just rode it around the neighborhood without a license, hah [19:46] http://p1.bikepics.com/2010%5C08%5C14%5Cbikepics-2039147-full.jpg [19:47] that little black thing attached was a toolkit.. really a handy bike [19:47] it got 88 mpg :-p Gas cost me nothing.. 2 galloon tank, i joyrode all summer with 1 fillup [19:47] balloons, sweet. I wouldn't mind a small 125cc now for cruising the city [19:48] such a fun bike to ride, really thinking back now [20:03] balloons, Hello [20:04] Letozaf_, hello [20:05] balloons, what about that bug on the toolbar, have you guys decided what to do, I mean I am unble to test on my desktop as the tests fail immediatly due to this bug [20:05] balloons, should I wait for a fix or what? [20:05] Letozaf_, which bug again? [20:06] balloons, bug 1248759 [20:06] bug 1248759 in Ubuntu UI Toolkit "Once the toolbar in the Add feeds page is hidden, I can't make it appear anymore" [Critical,Confirmed] https://launchpad.net/bugs/1248759 [20:06] balloons, even if on my desktop the problem is a bit different [20:09] balloons, the toolbar disappears before you can click a button on it [20:17] balloons, also bug 1231137 is happening to me now [20:17] bug 1231137 in Ubuntu RSS Feed Reader App "Missing "Next" button while adding feed (desktop only)" [High,Fix committed] https://launchpad.net/bugs/1231137 [20:19] Letozaf_, wild.. I too am noticing toolbar issues with file manager [20:20] elopio, davmor2 you notice any issues with disappearing toolbars? [20:33] balloons: who works on the calendar app? [20:34] thomi, ap testwise, doug5 has been committing things lately ;-) [20:43] balloons, yep, why? [20:46] doug5: I'll have a MP for you soon that fixes a few of the failures [20:47] sorry, got dc'd [20:47] thomi, https://code.launchpad.net/~acerisara/ubuntu-calendar-app/failing-tests/+merge/193706 [20:49] balloons: doug5: there are a few more fixes needed. shall I base my branch off that one? [20:49] thomi, ok, cool [20:49] or off trunk? [20:49] and let you guys merge it in? [20:49] thomi, yes I would say so.. I was meaning to get back to reviewing it this afternoon [20:49] whichever works.. we can merge back your changes easy enough.. [20:50] just an fyi [20:50] balloons: OK, I'll base it on trunk, and you guys can merge it in [20:50] thomi, np [20:50] it fixes the TypeError("A float is required") errors [20:50] as seen here: http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/15:20131107.1:20131031.1/4936/calendar-app-autopilot/ [20:57] doug5: there seems to be some hard-coded assumtion that the current year is 2012? [20:57] maybe that's fixed in the other branch... [20:59] thomi, where? [20:59] doug5: actually, i think it's more that the year/month/day is incirmenting on a swipe, instead of decrementing [21:00] which test? [21:01] several of them - calendar_app.tests.test_monthview.TestMonthView.test_monthview_go_to_today_prev_year for example [21:01] and similarly for month,week,day views [21:01] bzr st [21:01] oops [21:02] I don't know about the month view, i'm checking the others [21:02] doug5: you can get my fixes here: lp:~thomir/autopilot/trunk-fix-ap1.4-errors/tests/autopilot [21:02] the tests still fail, but they no longer error with autopilot 1.4 [21:02] ok [21:02] oops, branch location is actually: lp:~thomir/autopilot/trunk-fix-ap1.4-errors [21:03] balloons: this is the issue I found with the toolbar: https://bugs.launchpad.net/ubuntu-rssreader-app/+bug/1248759 [21:03] Ubuntu bug 1248759 in Ubuntu UI Toolkit "Once the toolbar in the Add feeds page is hidden, I can't make it appear anymore" [Critical,Confirmed] [21:04] elopio, I'm hitting a similar issue with file manager [21:05] is there something underlying causing it? [21:05] balloons: when the toolbar has the lock property = True [21:06] it will start visible, then it will be hidden after 5 seconds [21:06] and then as it has locked=True, you can't open it anymore. [21:06] timp assigned the bug to himself, critical, so I suppose he will have a fix tomorrow. [21:06] elopio, perfect.. So I won't file for fm, just link it [21:07] balloons: and there's another problem, that if it takes a lot of time for you between opening the toolbar and clicking the button, it can be hidden. [21:07] he's fixing that here: https://code.launchpad.net/~tpeeters/ubuntu-ui-toolkit/ap-toolbar-open/+merge/194122 [21:07] ok doug5 time to look at your branch [21:07] since mine is bungled by the above :-) [21:08] cool [21:09] so the swiping thing.. what conclusion did you arrive at doug5 ? [21:09] which one? [21:10] balloons: http://ubuntuone.com/7RKd1HnvVGRVmjjdxis17S [21:10] next / previous year, month, day? [21:10] if now.month > 6: self.drag_page_up() [21:11] ah [21:11] now.month could probably be 10 [21:11] conclusion is that the behaviour of the component is weird [21:12] but I don't know how it works [21:12] elopio, nice once I had a Kawasaki too, sold it now :D [21:12] I don't understand why we have 12 elements, but only ten in the tree [21:13] and if you swipe, you get the remaining but one is missing [21:13] and the order has changed === salem_ is now known as _salem [21:15] Letozaf_: :) I've just got mine back, and I won't let it go anymore. [21:15] elopio, :D [21:18] sleep time [21:18] balloons, just drop me a reply on the mp [21:18] doug5, ok, I'm getting it a bit more [21:19] so, what if we swiped down and up everytime.. would that populate everything? [21:19] ohh, lol [21:19] elopio, nice.. I like the black plated exhaust [21:19] slick [21:20] looks like a nighthawk kind of [22:16] balloons: so with the calendar tests [22:17] balloons: can you manually merge my branch in? It won't land automatically while the tests fail [22:17] but I need to eradicate any autopilot 1.4-related errors from the image builds [22:17] so it's important that my fixes land [22:19] balloons: so, it happened: DanChapman is now a member :-) [22:20] balloons: nice testimonials, BTW. They did help a lot [22:30] robotfuel: I was thinking about your button emulator that's pending review. [22:30] shouldn't all the emulators have a click method? [22:30] it would be nice to be able to do textfield.click, for example. [22:30] elopio: yes [22:33] elopio: I pushed back against that the last time it was proposed [22:33] congrats DanChapman [22:33] thanks robotfuel :-) [22:33] with good reason too - we want to present a unified API. Adding 'click' methods splits the autopilot API with no good reason [22:34] DanChapman: that's good news. [22:34] If we want a 'click' method for all UI components, then we should find a way to add it *everywhere*, rather than leaving Gtk/unity7 as second-class citizens [22:34] * DanChapman learnt his lesson with using click methods in emulators ;-) [22:35] see, DanChapman knows what I'm talking about ;) [22:36] thomi: can you elaborate on "splits API" ? [22:36] thomi: that makes sense [22:36] elopio: it means that test authors writing tests against a Qt app have a very different set of tools avalable to them as they do against a Gtk app [22:37] it changes the API open to test authors depending on what type of UI toolkit they're using, which is something we've managed to avoid till now [22:37] I'm fine with abstracting away the complexities of how some UI component works [22:37] but the low-level interaction methods (typing, clicking, dragging etc) should remain the same [22:37] ...unless, of course, you can add them to autopilot, and have them available everywhere [22:37] thomi: so, we shouldn't have a Button emulator? [22:38] like we did with Mouse.click_object [22:38] elopio: you can have a Button emulator that does other things, just not one with a 'click()' method :) [22:38] elopio: it seems like we should not [22:38] makes sense. [22:39] "self.mouse.click_object(my_button)" is pretty simple already, isn't it? [22:40] thomi: yes, not as simple as my_button.click(), but you have a pretty good point. [22:40] that's a lot more letters than click() :P [22:40] it is, but I think we can assume that test authors know how to type :P