=== pgraner` is now known as pgraner === chihchun_afk is now known as chihchun [07:51] good morning [07:52] morning DanChapman [07:52] hey elfy :-) [07:53] busy busy busy I see :p === spineau is now known as spineau_akf === spineau_akf is now known as spineau_afk [09:03] elfy indeed :-) [09:05] jibel, thomi has said the snapshot bug is intended behaviour :-| ?? I will comment back explaining its not, I should probably have included an autopilot.log where it doesn't occur [09:08] DanChapman, right I disagree with his explanation. furthermore the behavior changes from one run to the other [09:10] jibel indeed, I think he missed that ubiquity is launched long before and outside of the testcase [09:11] DanChapman, I think he didn't read the bug report :) [09:11] jibel :-D [09:12] hm, another crash but in libglib this time :/ === spineau_afk is now known as spineau === zequence_ is now known as zequence [10:03] jibel all greens for ubuntu :-) Shouldn't the *_devel_daily-run go green as well? [10:03] bbl [10:04] wrong channel lol [10:04] DanChapman, no because I re-ran all the jobs 1 by 1 and jenkins doesn't re-aggregate results of children jobs [10:04] I'm still trying to catch this segfault [10:13] jibel ah i see, it's still being a pain then! Nice to see them all green though === xnox_ is now known as xnox === abner_ is now known as abner === chihchun is now known as chihchun_afk [13:55] jibel, yesterdays lubuntu custom install caught bug 1261745 :-) have you had any luck yet reproducing the segfault? [13:55] bug 1261745 in ubiquity (Ubuntu) "Cannot create JFS partition" [Undecided,New] https://launchpad.net/bugs/1261745 [13:58] DanChapman, not locally. I uploaded a crash file and waiting for the retracer. [14:05] DanChapman: jibel: i must admit, the care factor for JFS is very low =))))) but it's a nice catch ;-) e.g. a good reason to discuss and pull jfs support out of the desktop installer ;-) [15:06] good morning. [15:24] well, not quite what was expected, but I guess we're moving forward [15:25] jibel, DanChapman do you think we could just mail Evan to coordinate the details on turning this on? [15:27] balloons, yes, that should be enough, and ask him to assign someone from his team to feed the dashboard with test results [15:28] jibel, done.. Do you have the autopilot bug in question? [15:29] balloons, yes sounds good to me. also this autopilot bug needs fixing as well [15:29] 1260012 [15:29] bug 1260012 [15:29] bug 1260012 in Autopilot "app snapshot during test setup is being set to an empty list" [Critical,Incomplete] https://launchpad.net/bugs/1260012 [15:29] * DanChapman got there in the end [15:29] balloons, bug 1260012 and the other is a crash not yet retraced bug 1261685 [15:29] bug 1261685 in ubiquity (Ubuntu) "ubiquity crashed with SIGSEGV in in ??() [non-native amd64 package]" [Undecided,New] https://launchpad.net/bugs/1261685 [15:31] ok, so both bugs we would like solved before turning it on right? [15:31] balloons, jibel, what will be the process of adding new tests? Is it just a case of contact the CI team to add them? [15:31] I would still want the work to go forward, so basically we just have to flip the switch when they close [15:31] balloons, thats correct [15:37] balloons, agreed, we can push to the dashboard even if tests are not green [15:38] jibel, I didn't mention the bugs directly to evan, we'll just sort it out with whomever he assigns. I don't want them to hold on doing the work :-) [15:39] DanChapman, I am not sure they have anything automated but there is a manual procedure to have them added https://wiki.ubuntu.com/CI/AddingTests [15:39] balloons, sure [15:50] jibel: the autopilot test results are a bit spammy at the moment. 41 messages over two days, which is like quadriple of the average monthly traffic for ubuntu-installer@ list =) [15:51] jibel: i've discarded those notifications for now, i'll whitelist them once they are more stable. [15:51] xnox, I'll tune it a bit to send only a notification on regressions and back to fixed. [15:58] jibel thanks [15:59] DanChapman,Balloons : Good Eve!!! [15:59] senan howdy o/ [16:00] DanChapman ; Still couldn't find the solution of the problem :( === Ursinha-afk is now known as Ursinha [16:09] senan, ok so did you manage to make any more progress with it, i need my mind refreshing, just going to take a look at your branch [16:12] DanChapman : No dan, By the way what is state not found eror [16:12] senan, that's where nothing is returned when selecting an object. i.e object doesn't exist [16:18] DanChapan : The scan folder menu comes like this GtkMenu->GtkModelMenuItem [16:20] hey senan :-) [16:20] senan lets start with improving the working tests, for instance the test_home_list test does some crazy stuff where the mouse tries to escape off the screen :-| rather perculiar we should try and find a way to wait for the UI to change before the mouse even starts moving [16:20] balloons : hey [16:21] DanChapman : Means we will put an explicit wait ? [16:22] senan so atm once you click home folder the mouse is moving before the UI has changed, you should try using Eventually on the dropdown treeview on the left side of the next UI try using its visible property [16:23] * DanChapman is going to get this gtk emulator finished right now, it will make this sooooo much easier [16:24] senan you want to be sure everything has loaded before the mouse starts moving [16:26] ping rhuddie: do you want to meet in 5 minutes? [16:27] elopio, sure. I think it is just you and me today. [16:29] DanChapman: That is happening in some other tests also [16:30] senan which ones? [16:31] Mouse going out of the screen [16:40] DanChapman : that issue is resolved using self.assertThat(lambda: self.app.select_single('GtkSpinner',BuilderName='spinner').sensitive,Eventually(Equals(True))) [16:42] senan very nice :-) [16:44] DanChapman : But I've no idea how to resolve "More than one item was returned for query" issue [16:47] senan you have a couple of options either use select_many()[index] using the index of the one you want but do some tests to make sure they are always in the same order or you can try selecting it from the tree the object is in instead of from root [16:50] DanChapman : I think I tried that but It was not working as expected [16:53] DanChapman: like thisself.app.select_many('GtkModelMenuItem',action_name= menu_item_name)[index] ? [16:58] DanChapman : even with that there is only one index.. it will work only for index 0 [17:00] senan yes either like that or select an object in the same tree as your menuitem but a bit further up and then use that object to select the menuitem i.e menu = self.app.select_single('object') then menu.select_single(menu_item) [17:00] ok. let me try that [17:11] DanChapman , when I first select menu using select single, its saying more than one item returned. [17:15] and Dan, the menus are not under baobabwindow tree node.. its actually a separate GtkWindow node [17:16] senan yes all menu windows have there own GtkWindow tree and a corresponding *accessible tree. You need to try and find something unique to grab on to when selecting menu's it difficult as there isn't many keyword args available [17:17] Yes there is no keyword available :( [17:20] DanChapman, is it possible to select using accllabel ? [17:28] senan you can select any object you can see in vis its just not easy to get to alot of them using select_single. [17:30] DanChapman, when I tried to selectt AccelLabel, its throwing state not found error [17:37] senan try self.app.select_single('GtkCheckMenuItemAccessible', accessible_name='Scan Folder...') you need to make sure the menu is open first as so the globalRect can come back to a sensible value === shadeslayer_ is now known as shadeslayer [17:44] DanChapman, I'm getting state not found error [17:44] senan can you push your latest tests so I can take a look :-) [17:45] ok [17:48] DanChapman, its done [17:50] DanChapman,balloons, going to sleep.. Good Night [17:50] night senan [17:51] DanChapman, I've pushed the latest changes [17:51] senan great thanks ill take a look [17:51] ok. thanks :) [18:03] o/ [18:04] balloons, where's the link to the wiki page with the calendar, again? Google's not responding to my request for email so I"m missing the link [18:05] https://wiki.ubuntu.com/QATeam/Calendar [18:05] balloons, thank you kindly === Noskcaj10 is now known as Noskcaj