[07:20] hello [07:20] I'm teaching unity about logind, and wrote some new test cases [07:20] right now, they only say "fail" on "make check", which isn't particularly helpful [07:20] is there a way to make them more verbose? like showing the particular assertion that failed, plus unity's log output? [07:21] and can I run an individual test case? [07:31] pitti, just running ctest should help [07:33] pitti, does that help? [07:36] tvoss: it says "No test configuration file found!", in both the source root and in obj-x86_64-linux-gn [07:36] ...u [07:36] (back in 10) [07:37] pitti, ack, branching unity :) [07:57] ./obj-x86_64-linux-gnu/CMakeCache.txt:CMAKE_CTEST_COMMAND:INTERNAL=/usr/bin/ctest [07:57] that's the only hit [07:59] but no hit for CMAKE_CTEST_COMMAND, so apparently it doesn't use ctest? [08:02] pitti, might well be for unity current [08:03] Saviq, ping [08:03] ah, obj-x86_64-linux-gnu/tests/CMakeFiles/check.dir/build.make has some clues [08:03] * pitti can't believe that he's the first guy wanting to debug a failed test case [08:04] pitti, the unity cmake setup redefines the make test iirc [08:04] right, see above build.make comment, I think that's it [08:04] it calls gtest and dbus-test-runner [08:05] hm, calling gtester manually like this fails all over the place [08:07] and calling dbus-test-runner as build.make specifies is invalid: option parsing failed: Unknown option --gtest_output=xml:./ [08:07] so perhaps that's some old stuff [08:07] pitti, yup, think so [08:07] Trevinho, ping [08:09] tvoss, here [08:09] ah, I did find the verbose output, it just scrolled way off the screen [08:09] pitti, \o/ [08:09] so I guess my question reduces to "how can I run a single test case"? [08:53] mzanetti: do we want the -v1 here? https://code.launchpad.net/~mzanetti/unity/phablet-make-check-hook/+merge/153868 or is it a "typo" [08:53] s/typo/leftover [09:03] Hi guys [09:04] does anyone here uses gnome3/unity with a double screen? [09:05] men, this really sucks sometimes, I'm sorry but its true :) [09:07] Saviq: should we kill ./Components/QuitResetDialog.qml ? [09:07] I'm still getting this build-error on quantal... http://pastebin.ubuntu.com/5627627/ (used ./build -s ; ./build) any ideas what's missing? [09:07] tsdgeos, yes [09:07] * tsdgeos kills [09:08] MacSlow, that means UnityCore didn't build/install properly [09:08] pg_pt: I do sometimes, are you having issues with it? [09:08] MacSlow, go into ../unity_build/unity/UnityCore [09:08] Saviq, what checks are recommended in that case? [09:08] MacSlow, go make; make install; [09:08] MacSlow: or you have an old CMakeCache, try removing it [09:09] yeah, that too [09:09] MacSlow, go ./build_unity 2>&1 | tee unity.log [09:09] MacSlow, and paste unity.log somewhere [09:10] veebers: some issues yes, but the most hateful is that when you remove the large screen and everything just "Insert random position" [09:10] pg_pt: ah, yeah. I hate that too :-) [09:11] veebers: gnome shell didn't had that issue, but gnome shell is for fedora users it seems... [09:12] veebers: other issues are folders going to the black, unacessible part of the screen (when you have two monitors with diferent sizes) [09:13] veebers: and I cant see all my spaces in the bigger (on my case on the rigth) screen, because it puts a balck space at the rigth of the monitor. WHY??? [09:14] Saviq: https://code.launchpad.net/~aacid/unity/killQuitResetDialog/+merge/154012 [09:14] pg_pt: that is odd, I haven't seen them before [09:14] tsdgeos, shall we kill Notifications already [09:14] tsdgeos, now that we're cleaning things up [09:15] Saviq: my grep showed that it was used somewhere, but maybe not really used :D [09:15] ah, in the CMakeListst.txt ^_^ [09:15] tsdgeos, ;) [09:16] ok, will put it in the same MR [09:18] Saviq, tsdgeos: there you go http://people.canonical.com/~mmueller/unity.log [09:19] MacSlow: looks fine, did you try clearing the CMakeCache? [09:19] MacSlow, rm CMakeCache*; ./build [09:20] Saviq, tsdgeos: doing that now [09:20] cache wasn't wiped yet [09:23] mzanetti, can you comment on https://code.launchpad.net/~mzanetti/unity/phablet-verbose-ctest/+merge/152213 please? [09:23] veebers: http://imgur.com/kuHIqJo [09:23] mzanetti, also, can you try generating coverage? [09:25] my build suddenly doesn't find hud-client.h anymore :-S [09:25] * tsdgeos cleans everything [09:28] and it works again :D [09:31] tsdgeos, Saviq: you want to wipe the Notifications directory from unity/phablet?! I intended to use that for the new stuff. [09:31] MacSlow, well, you'll add it then, won't you ;) [09:32] MacSlow, or will there be common history between the "old" and the "new"? [09:32] tsdgeos, Saviq: sure... just wanted to hint you this :) [09:32] MacSlow, I actually wanted to clean it to prevent any conflicts between your work and the old stuff [09:32] Saviq, not really... just wanting to keep the Notifications directory... everything else will change [09:32] MacSlow, then yeah, it's best to just drop and re-add [09:33] Saviq, in a custom branch I wiped it already. [09:33] Saviq, ok [09:38] * tsdgeos doesn't know how to do a for loop :-/ [09:38] Saviq: MacSlow: mzanetti: https://code.launchpad.net/~aacid/unity/fixTestMouseFlick/+merge/154017 [09:39] tsdgeos, taking a look... [09:39] Saviq, tsdgeos: btw... build runs now on my desktop... thx [09:39] MacSlow: great :-) [09:40] MacSlow, cool, the occasional `bzr clean-tree; bzr clean-tree --ignored` helps [09:40] Saviq, oh... ok [09:41] MacSlow, especially when you get build failures [09:42] Saviq, tsdgeos: I'll try to get it also working on my raring-based laptop.. and then see how I can get my humble beginnings of nosd in [09:42] MacSlow: i'm using raring so it should also "just work" [09:42] btw... "Number of leaked pixmaps: 82" anything to fix this? Wrong setting? [09:46] that happens [09:46] tbh not sure who's fault is it [09:46] whose [09:48] tsdgeos, I even get that with a very simple stand-alone test-program exercising preliminary notification-drawing... so it might be a QML-issue/bug?! [09:48] may be [09:54] Cimi: ping [09:55] Saviq: I think the MP can be deleted. Cimi will add this properly in as he requires the same functionality as we have in qmluitests... [09:56] mzanetti, k [09:56] Saviq: if not, I'll fix the MP to not overwrite the file [09:56] mzanetti, which one, then? ;) [09:57] Saviq: https://code.launchpad.net/~mzanetti/unity/phablet-verbose-ctest/+merge/152213 [09:57] mzanetti, yeah, but delete or not? ;) [09:57] lets wait for Cimi to reply my ping. when I know if he already started I'll take care of this one [09:57] Saviq: ^ [09:57] k [10:05] davidcalle: hey [10:05] davidcalle: unity-scope-tomboy shouldn't dep on tomboy, I'm removing it [10:05] didrocks, suggests, I presume? [10:06] didrocks, all: When I change something in lp:unity and send a MP, I don't add debian/changelog, right? that'll be autogenerated from the commit logs [10:07] davidcalle: yeah, fixing that [10:07] pitti: its not autogenerated for lp:unity/phablet I don't know for lp:unity [10:07] pitti: depends, if you linked to a bug, right [10:07] pitti: if not and you think you should mention it, please add to debian/changelog [10:08] didrocks: I did link to a bug, yes; I can add debian/changelog if that's better, I just don't want to collide with the automatic changelog generation [10:08] davidcalle: https://code.launchpad.net/~didrocks/ubuntu-scopes/suggest-tomboy/+merge/154027 [10:08] didrocks, if you prefer, you can send me a list of these "invasive" deps, I'll make the mps [10:08] pitti: well, if you add to debian/changelog and mention the same bug than the one you linked with, it will detect you already mentionned it (with your name), and so won't repeat it :) [10:09] didrocks: ah, great; manual changelog it is then [10:09] pitti: if you think that the bug title is not enough for the description, you can go manual :) [10:09] oh, it takes the bug title, nice [10:09] pitti: otherwise, let it automatic :) [10:09] yeah [10:09] * pitti adjusts that then [10:10] thanks, sorry for my n00b questions [10:10] no worry ;) [10:10] pitti: basically, it should just do the right thing, if you mention it, it won't duplicate the mention. If you didn't, it will take the associated bug title [10:10] davidcalle: it's the only one I noticed ^ [10:11] davidcalle: ensure when you send the 3 remaining ones (pypi, sshsearch and evolution) and that they are ready for you, that they don't drag evolution for instance :) [10:11] didrocks, all the music players :) [10:11] didrocks, not sure where you put the limit, the calculator scope depends on gnome-calculator [10:11] didrocks, hey, but that would be fun :) [10:12] didrocks, just wanting to avoid pulling stuff in the default install? or do you try to slim down what unity brings in? [10:19] Saviq: you know why the home lens is broken? [10:19] mzanetti, broken how? [10:19] Saviq: empty [10:19] mzanetti, isn't, here [10:19] * Saviq checks [10:19] Saviq: its here since yesterday evening [10:19] mzanetti, indeed [10:19] Saviq: mterry was the first to mention it [10:19] yesterday morning or so [10:19] mzanetti, raring? [10:20] mzanetti, sorry had irc closed [10:20] mzanetti, indeed [10:20] mzanetti, I was just doing that [10:20] Cimi: ok. thanks. I'll delete the old one then. [10:20] mzanetti, unity-lens-mock doesn't work in raring [10:20] seb128: I think we should have the same policy for all scopes [10:20] seb128: the scope is just a hook [10:20] so a feature available [10:20] then, if you install the apps you have a new functionality in the dash [10:20] I don't think that's what should pull gnome-calculator in the distro [10:20] davidcalle: I don't see the banshee scope anywhere [10:20] davidcalle: not a rhythmbox one (called rhythmbox) [10:20] davidcalle: any news on my trivial MP? :p [10:20] Saviq: yes... but I'm on raring since 2 weeks now and it broke only yesterday [10:21] mzanetti, might've been the smart scopes update [10:21] * pitti sends his first unity MP, wohoo! [10:21] heh pitti ;) [10:22] pitti, congrats ;-) [10:22] Saviq: hmpf... and it crashes when searching the music lens :/ [10:22] seb128: belay that until it gets merged :) [10:22] mzanetti, yeah, we need to run the mock lenses against the locally installed libunity [10:22] crap [10:23] Saviq, smart scopes aren't in R yet [10:23] hmm [10:23] * Saviq drops the PPA [10:24] mzanetti: Saviq: latest home scope update is in the ppa [10:24] mhr3 fixed the home scope not working [10:24] 6.8.0daily13.03.19.1ubuntu.unity.experimental.certified-0ubuntu1 [10:24] Saviq: mzanetti ^ [10:24] * mzanetti checks [10:24] didrocks, yeah, unrelated [10:25] (I think) [10:25] 11:19:47 mzanetti | Saviq: you know why the home lens is broken? [10:25] didrocks, in lp:unity/phablet [10:25] Saviq: the result is that no result in home scope ^ [10:25] why do you speak about the smart scopes ppa then? [10:25] didrocks, we're not using the system-wide lenses [10:25] didrocks, complicated ;) [10:25] ok ;) [10:25] believing you then :p [10:26] seb128: merci! [10:27] didrocks, de rien (free karma \o/ :p) [10:27] ;) [10:28] mzanetti, everything works here with current raring [10:29] mzanetti, but smart scopes from the PPA do break it [10:29] ah ok... [10:31] Saviq: next problem: [10:31] mzanetti, how can I replace set_tests_properties(tst_QmlTests PROPERTIES ENVIRONMENT "QT_QPA_PLATFORM=minimal") in a similar cmakelist as the one in qmluitests? [10:31] I'm helping nic to test the PageHeader... Turns out we access shell all over the place in there [10:32] Saviq: ah... nvm... just came to the solution myself [10:32] mzanetti, ;) [10:33] mzanetti, ultimately we need a global object that we proxy stuff through [10:33] Saviq: I would define a SearchListModel that offers the addQuery() function [10:33] Saviq: when unset, every pageheader will have its own searchhistory [10:33] mzanetti, ah this [10:34] Saviq: or, it can be set to be the one in Shell.qml to share the same searchHistory [10:34] mzanetti, that's temporary code anyway (we need persistent storage there) [10:34] preferably u1db [10:34] Saviq: yeah. that can be added in the SearchListModel at some point I'd say [10:34] mzanetti, yup, works for me [10:35] ok [10:35] Cimi: so... your turn now :) [10:35] haha [10:35] Cimi: whats the problem? [10:35] mzanetti, a doc for cmake might be good :) [10:35] Cimi: well... its pretty easy actually [10:36] Cimi: every cmake's functions first argument is a free-form string: target [10:36] mzanetti, https://pastebin.canonical.com/87100/ (for the records) [10:37] doesn't work [10:37] Cimi: so you do "add_executable(myTarget binary)" or "add_library(myTarget libfile)" [10:37] Cimi: that defines a new myTarget [10:37] Cimi: after that you have a bunch of functions that can modify that target [10:38] like set_target_properties(myTarget ENV foo=bar) to export a env variable when running this target for example [10:39] ah I see [10:39] Cimi: the paste looks good [10:39] Cimi: haven't ran it tho [10:39] mzanetti, with target instead tests work [10:40] actually not [10:40] no tests were found [10:41] http://paste.ubuntu.com/5627815/ [10:41] Cimi: you can add "message(blabla)" to cmake [10:41] Cimi: to have some debug prints === mmrazik is now known as mmrazik|afk [10:43] all: autopilot tests are enabled in jenkins. every MP must have test passing now to be approved by Jenkins [10:48] mzanetti, ah ok it does work, but requires calling make testCarousel etc etc [10:49] Cimi: then somethings wrong with "add_dependencies(unittests test${COMPONENT_NAME})" [10:49] I guess I need to not use custom targets? [10:49] * tsdgeos realizes eh can't compare anchors in the tests [10:49] :./ [10:49] tsdgeos: there is a way... [10:49] tsdgeos: some special command I can't recall right now [10:50] mzanetti: any clue where to find it besides google? === mmrazik|afk is now known as mmrazik [10:50] tsdgeos: hmm... no... [10:53] tsdgeos: hmm... maybe I'm wrong... this is the anchor test from qtdeclarative: http://www.qt.gitorious.org/qt/qt/trees/a7493235a70f9e60d5d25d84b0782ee0a2e5c5fd/tests/auto/declarative/anchors [10:54] yeah, seen that one [10:54] compares x, y [10:54] yeah :/ [10:54] can do the same i guess [10:54] tsdgeos: probably... just make sure its flexible enough for different GU's and form factors [10:55] my current idea is doing something like [10:55] compare(hud.__searchBarContainer.x + hud.__searchBarContainer.height, hud.height, "AppStack is anchored at the top") [10:55] so test the anchors by testing the x [10:55] err [10:55] obviously i need to change that to [10:55] y [10:55] but you get the idea :D [10:58] mzanetti, so if I run make unittests cycles through all tests [10:58] mzanetti, (I understood much more of cmake now) [10:59] Cimi: all == tests/unittests/* or all == tests/* ? === mmrazik is now known as mmrazik|lunch [10:59] mzanetti, I was running from the dir [10:59] oh... that works? [10:59] mzanetti, so from tests/unittests [10:59] it always worked [10:59] but before make test was enough [10:59] now make unittests [10:59] Cimi: make check [10:59] thats for the unittests [11:00] mzanetti, it doesn't [11:00] mzanetti, shall I add a dependency to main cmakefile? [11:00] or do you have a fix for the previous cmakelist? [11:01] Cimi: yes please, add that [11:01] no. I don't have a fix here [11:02] ok works [11:03] Saviq: what is smartscopes (the package that breaks stuff)? [11:04] mzanetti, unity* [11:04] mzanetti, http://www.omgubuntu.co.uk/2013/01/unity-to-ship-with-smarter-scope-searching-in-13-04 [11:06] Saviq: I need to remove unity* ? that doesn't sound right [11:09] mzanetti, what PPAs do you have installed? [11:09] Saviq: yeah... I have the phablet-team/ppa [11:10] mzanetti, well, that wouldn't break the phablet shell - it would break your desktop one [11:10] Saviq: don't care about that [11:11] mzanetti, then nothing in raring does break the phablet shell (yet) [11:11] still it doesn't work here :D [11:12] mzanetti, is your home completely empty or just the favorites / recently contacted are empty? [11:13] Saviq: completely [11:13] mzanetti, not even Apps?? [11:13] Saviq: nothing [11:13] Saviq: it doesn't even switch to the home lens any more [11:13] Saviq: stays on the music lens at startup [11:13] mzanetti, console output please [11:14] Saviq: I removed all sorts of stuff and am reinstalling right now [11:16] Saviq: haha... now all my lens are empty... seem we're not pulling needed deps [11:17] mzanetti, yes we are (at least since yesterday) ;) [11:17] mzanetti, demo-assets; unity-lens-mock [11:17] Saviq: ah ok... but not the package in the ppa [11:17] yeah, already installing [11:17] mzanetti, no, they're not deps [11:18] mzanetti, so :D what's the difference between unittests dir and qmluitests? [11:18] Cimi: try running them and find out :D [11:18] Cimi, unittests can run without GL [11:18] :D ok boys [11:19] Saviq, DraggingArea requires gl?? [11:19] Cimi, QML requires GL, remember? ;) [11:20] Saviq, so the Carousel is in unittests [11:20] Cimi, 'cause it's only doing JS tests [11:20] and uses QML [11:20] Cimi: yes [11:20] Cimi, if you need to draw anything [11:20] indeed [11:20] ok guys [11:20] I blame myself [11:20] Cimi: rule of thumb: do you use the line "when: windowShown" somewhere? [11:21] Saviq, crossfadeimage? :D [11:21] ratingstars as well [11:22] Cimi: whats with those? [11:23] mzanetti, they are in unittests [11:23] Saviq: https://pastebin.canonical.com/87104/ <-- I removed everything unity* installed qml-phone-shell from the ppa and unity-lens-mock and demo-assets [11:23] mzanetti, but they are qml, not js [11:23] Cimi: so? [11:24] mzanetti, so I don't understand why draggingarea is in qmluitests [11:24] [12:20] Cimi: rule of thumb: do you use the line "when: windowShown" somewhere? [11:24] only that? ok [11:25] Cimi: that causes a window to be openend wich requires openGL [11:25] Cimi: you can also compile QML without opening a window [11:25] dragging area doesn't [11:25] I really don't understand why draggingarea should be in qmluitests [11:27] unless mousearea requires opening a window [11:27] for events [11:28] mzanetti, your system is b0rked :P [11:29] Saviq: seem sso [11:30] Cimi: draggingarea does open a window [11:30] Cimi: try running the test [11:31] mzanetti: what you think of https://code.launchpad.net/~aacid/unity/test_hud_element_positions/+merge/154048 ? [11:32] Saviq: actually our packages are too... installing qml-phone-shell on a system with absolutely nothing unity related should give you a working binary :/ [11:32] no? [11:32] mzanetti, working binary, yes, but not stuff like demo-assets or unity-lens-mock [11:32] tsdgeos: don't export all those __ properties [11:33] Saviq: I think it should pull in unity-lens-mock. but anyways, I've installed those two myself which causes the shell to crash === mmrazik|lunch is now known as mmrazik [11:33] mzanetti: so how do i get to them? want me to do child[1].child[2].child[]? [11:34] mzanetti, no, I don't think it should pull in unity-lens*, maybe recommend, but not depend for sure - you should have a working shell without any lenses installed [11:34] tsdgeos: review this and you'll have findChild(obj, objectName) https://code.launchpad.net/~mzanetti/unity/phablet-greeter-tests/+merge/153963 [11:34] mzanetti, did you install the *unity* packages from phablet-team/ppa? [11:34] that was in needs work this morning :D [11:34] i.e. did you apt-get upgrade from phablet-team/ppa [11:34] Saviq: I have... just no lens... if I install the mock-lens and demo-assets it crashes [11:35] Saviq: yes [11:35] mzanetti, well, it doesn't crash on the devices, does it ;) [11:35] mzanetti: ah the evil crash of demo-assets `mock-lens [11:35] it happened here when i was in quantal [11:35] the update to raring fixed me [11:35] probably some crap lying around?¿ [11:35] yeah, the mock lens story is problematic [11:36] it will use the system-wide GI for lenses/scopes [11:36] we need to build them locally asap :/ [11:37] * mzanetti tries ./build -s for the first time :D [11:38] mzanetti: that keyClick looks baaaaaaaaaad [11:38] tsdgeos: you mean the typeString() function? [11:38] mzanetti: in c++ you can do QTest::keyClick(myWidget, 'a'); [11:38] that doesn't work in qml? [11:39] tsdgeos: tbh I didn't try it, but all the "docs" I could find sad, no, its an integer [11:40] mzanetti: right, they only export the Qt::key version to qml :-/ [11:40] mzanetti: i think we should add the patch to support the char version [11:40] it's a 5 line patch [11:40] because your typeString will break once i try to write "Barça" [11:40] tsdgeos: yeah... but until that's in upstream can we haz this? [11:41] tsdgeos: yes, I know... it doesn't support special chars [11:42] biab [11:43] mzanetti: i guess, if there's nothing better :D [11:43] tsdgeos: I'll add a todo [11:43] plz [11:47] tsdgeos: pushed [11:48] mzanetti: tryCompare(pathview.currentIndex, data.uid) [11:48] that's wrong syntax, no? [11:48] tsdgeos: do'h [11:48] * mzanetti wonders why it passes [11:49] because it does [11:49] compare(obj[prop], value) [11:49] obj is pathview.currentIndex [11:49] prop is data.uid [11:49] and value is undefined [11:49] so pathview.currentIndex[data.uid] is probably undefined too [11:50] and undefined and undefined match [11:50] :D [11:53] damn java script [11:53] anyways... pushed the fix [11:53] tsdgeos: ^ [11:54] tsdgeos: ^ [11:54] yep [11:54] ooops... [11:54] was rerunning qml-phone-shell [11:54] wrong window for ↑+Enter ;-) [11:54] and lost focus to the terminal [11:55] mzanetti: so you think that the findObj thing is better than aliasing? why? [11:55] tsdgeos, I've replied to your comments: https://code.launchpad.net/~dandrader/unity/phablet_tst_Revealer/+merge/153937 [11:55] tsdgeos: because it does not pollute the public api of a copmonent [11:56] mzanetti: ok, fair enough :-) [11:56] dandrader: cool, i'll check asap [11:57] mzanetti: why is the background always grey in the tests? [11:57] because Graphics/ is not there? [11:57] tsdgeos, ah, the mouseFlick fix was already merged. I'll have to modify my merge proposal them. wait a minute [11:57] tsdgeos: yes. [11:58] tsdgeos: but there is a branch from mterry in the works that integrates with lightdm [11:58] tsdgeos: that will fix this issue [11:58] mzanetti: should we care? it makes me think "it's not working!" when seeing the test [11:58] ok [11:58] mzanetti: objectName: "wallpaper" we don't really use, but i guess mterry's branch will? [11:59] tsdgeos: nope... I added it and then decided not to use in the tests... will remove [11:59] okidoki [12:00] done === tvoss is now known as tvoss|lunch === MacSlow is now known as MacSlow|lunch [12:14] Saviq, so I still have the same issues as yesterday, I do import "../../plugins/Unity" [12:14] in my test file [12:14] and I don't have the right library imported I believe [12:15] http://paste.ubuntu.com/5628002/ [12:16] mzanetti, or you know? [12:16] some cmake magic required here? [12:18] mzanetti: https://code.launchpad.net/~aacid/unity/test_hud_element_positions/+merge/154059 fixed to use findChild [12:19] tsdgeos, can you help me? :) [12:19] Cimi: seems you have the wrong libunity-core [12:20]