[00:00] mterry: I know! [00:00] josharenson, if you follow that dep error chain to its root, you should find the real error [00:00] josharenson, I WISH apt-get did that for you [00:00] josharenson, or sometimes apt-get -f install is a magic bullet [00:00] mterry: last night, I built/installed unity-settings [00:00] fixed a lot [00:00] mterry: tried -f, does nothing [00:00] mterry: I'll do it the sad way [00:00] josharenson, I haven't installed manually built software into system paths in... like a decade [00:01] Oh, but you meant build a deb and install that [00:01] That I've done for sure [00:01] I was thinking you were doing just a make install [00:01] mterry: nah [00:03] mterry: liboxide is complaining that it can't install a dep, but that dep is already installed... this is dumb [00:04] josharenson: watch the version strings, they're probably different. You may just be unlucky trying this at a time that packages are migrating [00:05] josharenson, yeah I bet a version mismatch somewhere [00:05] greyback_: was same error almost 24 hours ago, but I'll try waiting it out [00:05] josharenson, maybe look and see if there is a newer version of something in proposed? [00:05] "apt-cache showpkg package" will help you see if package has multiple versions [00:05] ack * 2 [00:07] greyback_, huh interesting, I usually use apt-cache policy package for that, but this output is interesting too [00:07] mterry: ah that's less verbose than mine, thanks :) [00:08] greyback_, "policy" is an awful name for it, though I can get where they are coming from [00:08] yeah [00:21] phew, solved [00:21] purged liboxide and downgraded, everyone is happy now [00:32] josharenson, \o/ [00:35] mterry: and mouse works now! woot === duflu_ is now known as duflu [08:50] tsdgeos, do you know if the soundcloud scope is broken because of a unity bug or their? [08:50] Cimi: i didn't know it was broken at all :D Is there a bug or you found it while using it? [08:52] tsdgeos, for me, the cards have a black background [08:52] tsdgeos, they always had [08:53] Cimi: investigate it if you think it's a bug :) [09:13] Cimi: errr [09:13] Cimi: black background? [09:13] Cimi: it's the sound wave of the song, isn't it? [09:17] tsdgeos, let me try again [09:19] tsdgeos, some of my soundcloud songs are completely black, no title [09:20] well [09:20] tsdgeos, "Stream" are all black [09:20] investigate it then :) [09:20] tsdgeos, Explore are fine [09:20] tsdgeos, looking at another bug [09:20] currently [09:20] I just asked in case you knew [09:50] pete-woods: there? [09:51] tsdgeos, Cimi: it looks to be a bug in unity8, or maybe lower level than that? [09:51] pete-woods: what? [09:51] oh, sorry, thought you were asking about what Cimi was asking about [09:52] pete-woods: no :D [09:52] pete-woods: so i was trying https://code.launchpad.net/~unity-api-team/unity8/scopes-active-depends-on-screen/+merge/250610 [09:52] and the logcat still never shuts up [09:52] if i'm on the weather channel of the dashboard [09:52] do i need some other branch? [09:52] i do [09:52] obviously [09:53] it's linked on the bug [09:53] ah no [09:53] it's the same branch twice [09:53] * tsdgeos needs to learn to read [09:53] tsdgeos: all that bug fixes is the dash part. I've verified that I'm sending the correct signals to the location service [09:53] I suspect that the location service is broken also [09:53] I think there are two bugs [09:53] one that we weren't sending the right stuff when the screen goes off [09:54] and the other that the location service isn't correctly updating the GPS status [09:54] i see [09:55] pete-woods: is anyone looking at the location service side? [09:57] tsdgeos: tvoss has provided a branch for the location service https://code.launchpad.net/~thomas-voss/location-service/be-verbose-about-starting-and-stopping-updates-for-testing/+merge/250600 [09:57] but it just enables logging, like my branch for the shell plugin [09:58] right [09:58] ok [09:59] i'll code approve the branch but i don't think we should land it until the whole puzzle is fixed [09:59] in case it is shown later that this is not the right approach [09:59] and we end up with half-fix in the dahs [10:26] tsdgeos: yeah, that makes sense. the silo should include the complete fix [10:26] or QA will probably reject it, anyway [10:32] morning pete-woods, quick q - how do you write tests for scopes? That is, which type of tests do you write (unit, integration, etc.) and which frameworks are you using to write them? [10:33] dpm: well at the minute, we write unit tests for them. however we have a harness in development to enable more complete functional testing of them [10:34] pete-woods, ok, thanks. What framework do you use for writing the unit tests? [10:36] dpm: google test http://bazaar.launchpad.net/~unity-api-team/unity-scope-youtube/14.10/view/head:/tests/unit/youtube/scope/test-youtube-scope.cpp [10:36] they look like that atm [10:36] which is pretty horrible [10:36] this is why we are making the scope test harness [10:39] pete-woods, ok. The reason I'm asking these questions is because we're updating the Quality pages on developer.ubuntu.com. We seem to have a good story for QML apps, but I'm not too familiar with scope testing. What will the scope test harness cover (e.g. can you still use autopilot for functional testing)? Is it based on an existing framework? And does it have a name :-)? [10:48] dpm: the idea was to avoid using autopilot for testing scopes. we talk directly to the QML model behind unity8's dash [10:48] you can still write the tests in python, so it's nice and easy + quick [10:48] but it doesn't use autopilot [10:49] and you have the option of either controlling the actual installed scopes, or starting up a private instance of the scope machinery and testing in isolation [10:49] was there any particular reason for not using AP? Or was that the one? ^ [10:50] dpm: yes. we considered it vital to be able to functionally test scopes without having to have a real user session, and unity8, etc [10:51] so this scope test harness, what does it look like? does it cover unit and functional testing? which languages do you write tests in? [10:51] dpm: you can either write tests in C++ or Python [10:53] dpm: an example of usage http://paste.ubuntu.com/10406012/ [10:53] C++ test [10:54] dpm: http://bazaar.launchpad.net/~stolowski/unity-scope-click/scope-harness-tests/view/head:/scope/scope-harness-tests/apps-scope-harness.py [10:54] an example of python tests [10:54] * dpm looks [10:55] those last tests are still WIP afaik [10:55] but they do work [10:55] I understand this is all still wip and there is no documentation? [10:56] well we obviously plan to release documentation [10:56] some is definitely going to be written for the python interface [10:57] but right now, I don't think we've done that, no [11:02] thanks a lot pete-woods, that was very helpful [11:02] :) === greyback__ is now known as greyback [12:08] MacSlow: answered back on https://code.launchpad.net/~macslow/unity8/unbreak-notification-positioning-fix-1422711/+merge/249998 we can have the discussion here if you prefer :D [12:10] tsdgeos, already put the anser there... to not dump qml-code here on irc [12:10] right, I've got a vivid VM now [12:11] but I still don't seem to be able to run unity8-dash [12:11] pete-woods: what porlb [12:11] do you have? [12:11] porlb = problem [12:11] just going to paste the log :) [12:11] MacSlow: if it's already happening why we need to new code? [12:12] tsdgeos: http://pastebin.com/C8W1xJhk [12:13] maybe I'm supposed to set some env var or something? [12:13] pete-woods: and what do you get? [12:14] tsdgeos, the reset of the topmostIsFullscreen is not working correctly otherwise. [12:14] which makes no sense [12:14] and means your other code is not doing it correctly [12:14] or i'm not understanding how notifications work [12:14] tsdgeos: just a white window (in the correct shape, i.e. a portrait rectange) [12:14] pete-woods: have you started the registry? [12:14] tsdgeos, wait... I'll pull away from the ap-test for a moment... [12:15] tsdgeos: yes. unity-scope-tool works normally [12:15] but I'm trying to do something with temporary scopes, so need to run the real unity8-dash [12:17] pete-woods: now that you mention [12:17] i get nothing either [12:17] but i guess i'm missing stuff [12:18] since i get http://paste.ubuntu.com/10407027/ [12:18] ah [12:18] tsdgeos, start smart-scopes-proxy [12:18] starting the smart-scopes-proxy fixed it [12:18] yeah [12:20] this is basically just a fresh vivid install, updated, and unity8 + scopes installed [12:20] I haven't built anything myself [12:20] pete-woods: have you started the smart-scopes-proxy too? [12:21] yep [12:21] then there's probably broken dependencies somewhere [12:21] you can get away with just starting smart-scopes-proxy, it also starts the registry [12:21] i don't have anything other than vvid installed [12:22] I don't think it's anything specific to scopes, as unity-scope-tool works [12:22] it makes me think there's something more related to the QML-y UI side [12:22] tsdgeos, *sigh* my branch is broken again... setting it back to wip :/ [12:23] MacSlow: ok [12:29] Saviq: do you have any other suggestions? packages I could try installing? I'm kinda stuck here [12:30] pete-woods, so you can't get unity8-dash working? [12:30] yeah [12:31] pete-woods, anything in the console? [12:31] Saviq: http://paste.ubuntu.com/10407027/ [12:31] sorry, wrong pate [12:31] http://pastebin.com/C8W1xJhk [12:31] Saviq: ^ [12:31] ok so nothing interesting [12:32] I've tried commenting out that line, just in-case it was the problem [12:32] and it didn't fix it [12:32] * Saviq kicks off a vivid chroot [12:32] gimme 5 [12:33] thanks! [12:35] Saviq, tsdgeos, any of you guys get this failure with trunk, running "make testShell"? http://paste.ubuntu.com/10396487/ [12:36] dandrader: test passes fine here [12:37] dang it [12:37] dandrader: is it clean unity8? [12:38] tsdgeos, yes [12:38] tsdgeos, here, it passes with "make xvfbtest" but fails with "make test" [12:38] weird [12:39] yeah... :( [12:39] dandrader: try making it a tryCompare maybe? maybe my pc is faster/slower than yours [12:40] food [12:41] Saviq: interestingly, running it all through "strace -etrace-file" shows that it's loading all the scope header pngs, etc [12:41] so it's certainly getting a fair way through loading up [12:41] it kinda feels like there's just some top level property that's just set visible: false or something [12:48] pete-woods, is it blank as in white, or can you see the paper background? [12:49] Saviq: you can see the grey folder paper [12:49] *folded === MacSlow is now known as MacSlow|lunch [12:52] pete-woods, are you seeing «Unable to add overview scope, can't find with ID: "scopes"» ? [12:52] pete-woods, it seems to be some timeout [12:52] pete-woods, as every 4 or 5 runs I actually get stuff on screen [12:52] in which case I don't see the above [12:54] right, but I'm getting a timeout then, that you don't [12:54] Saviq: that error is usually caused by not having the "scopes scope" (unity-scope-scopes) not installed I thought [12:55] Unable to add overview scope, can't find with ID: "scopes" [12:55] and yes, I don't get any timeout errors [12:58] pete-woods, yeah I have it installed, but registry times out: [12:58] ERROR! Caught unity::scopes::TimeoutException: Request timed out after 5000 milliseconds (endpoint = ipc:///run/user/1000/zmq/Registry-R, op = list) [12:58] Saviq: to be honest, I've never tried running all this in a chroot [12:58] I've no idea if the scopes support working like that [12:59] pete-woods, they (sometimes) work here after a bit of twiddling [12:59] in any case, that's not your issue [13:01] pete-woods, can you try installing ubuntu-desktop-next and see what packages are being installed, maybe that will spring an idea [13:02] pete-woods, or just build-dep unity8 [13:03] Saviq: http://paste.ubuntu.com/10407643/ [13:04] installing the build-deps hasn't helped [13:06] pete-woods, add "packagekit" to the ubuntu-desktop-next install line [13:07] tsdgeos, it's actually because my computer is too fast. adding a wait() in the test fixes it [13:07] Saviq: http://paste.ubuntu.com/10407685/ [13:07] that's a lot of packages! [13:08] tsdgeos, so I have to refactor those tests to make them more stable. but it's still annoying that I seem to be the only guy to get them failing :/ [13:08] pete-woods, well, it's the whole set of packages we have on the desktop-next install [13:08] yeah, fair enough [13:09] installing now to see if it helps [13:09] pete-woods, but meh, nothing there jumps out, really not sure what's going on / [13:10] I guess I'd have expected actual errors (symbols fail to resolve, QML imports fail) really if a package was missing [13:11] yeah, well... [13:12] unfortunately the install hasn't helped :( [13:23] Saviq: is there anything else I can do to help debug this? [13:24] pete-woods, what if you "start unity8-dash" instead of just running it? [13:24] pete-woods, oh, and also - how's your favorite list? [13:25] gsettings get com.canonical.Unity.Dash favorite-scopes [13:25] Saviq: I've also tried that. I get the same result [13:25] I thought maybe there was something special in the upstart config [13:26] going to try killing the fav list [13:26] pete-woods, tried on a guest user? [13:26] I have a feeling you've just hit the nail on the head [13:27] we should have UI saying you have no favourited scopes ;) [13:27] pete-woods, also, to be able to use the bottom edge, pass -mousetouch if you don't have a touchscreen [13:27] haha! [13:27] Saviq: it was the favourites! [13:27] great idea [13:28] we should have a UI to show you've no favourites... owait! [13:28] manage dash anyone? ;) [13:28] dandrader: so wait -> tryCompare and it'll work [13:29] tsdgeos, no, that's not the proper fix. I'll post it shortly [13:29] dandrader: ok [13:30] Saviq: thanks v. much for the assistance! have been stuck on this for nearly a day! [13:30] pete-woods, sounds like some wasted time for a stupid thing :( [13:31] well I made the mistake of assuming it was a code failure [13:31] Saviq: no worries. my own fault for forgetting I'd played with the favourites manually a few months ago [13:32] pete-woods, top tip: try in the guest account to see if stuff's maybe b0rked in your user account [13:32] yeah [13:32] definitely a good idea [13:32] for any issue, realy [13:32] +l [13:34] tsdgeos, there you go: https://code.launchpad.net/~dandrader/unity8/stabilizeShellTest/+merge/250920 [13:58] dandrader: interesting, which machine do you run? [13:58] tsdgeos, a desktop pc with an i7 [13:59] i have an i7 too, i guess either it being a laptop or probably the awful disk make an impact [13:59] dandrader, brilliant stabilize branch, will test / review [13:59] dandrader, oh did you hand to tsdgeos? nm then. will just appreciate them [14:00] dandrader, that might want to be a generic helper function: "LightDM.Greeter.showGreeter(); wait-for-shown; removeTimeConstraints()" [14:01] mterry, I thought about that, but in test_suspend the greeter is summoned in a different way [14:02] dandrader, ah right, more indirectly [14:02] dandrader, but still, in other tests I think we do similar showGreeter stuff [14:02] mterry, ok, will look into making a helper out of it [14:02] dandrader, but only these tests failed, so I doubt it's urgent [14:03] just some convenience cleanup [14:04] mterry, because only those two tests reload the greeter and them swipe it away [14:04] mterry, others just reload and do some checks, etc [14:05] Sure === MacSlow|lunch is now known as MacSlow [14:31] ChrisTownsend: branches approved, sorry it took so long [14:32] tsdgeos: Thanks! And no worries. I know you guys have been under the gun. [14:40] mterry, it's all there now: https://code.launchpad.net/~dandrader/unity8/stabilizeShellTest/+merge/250920 [14:41] dandrader, cool [14:41] dandrader, is the verify(spy.valid) just paranoia? [14:42] mterry, makes it easier to spot test failures due to code changes that leave the test outdated [14:43] because the sinal name has changed, for instance [14:43] well, you could call it paranoia :) === dandrader is now known as dandrader|afk [14:50] tsdgeos: Hey, are those MP's of mine candidates to land in Vivid? [14:50] ChrisTownsend: i understand yes [14:51] they're bugxies to me [14:51] but i'm not the landing guy [14:51] Saviq: ↑↑↑ ? [14:51] tsdgeos: Ok [14:56] tsdgeos, ChrisTownsend, yeah, totally [14:56] Saviq: Cool, thanks === dandrader|afk is now known as dandrader [15:32] tsdgeos, I added more stuff to https://code.launchpad.net/~dandrader/unity8/stabilizeShellTest/+merge/250920 and jekins results have arrived [15:33] tsdgeos, so need a review of the new stuff and can now be top approved if there no review concerns? :) [15:33] dandrader: ok, will have a look after meeting [15:34] tsdgeos, sure, thanks [15:54] josharenson, I split the greeter card into 3 checklist items, which may help for at least hitting some targets this cycle [15:54] mterry: cool ill look them over [15:54] josharenson, and maybe that means we should simplify the branch we have going to only do the first checklist item instead of trying to do all at once like I was [15:55] mterry: /me likes things that are nice and discrete [15:56] josharenson, first one means just the --mode argument and some Shell.qml finagling to only load some components [15:56] josharenson, second one is linking against real liblightdm and using a SessionManager plugin etc [15:56] josharenson, third one is just shipping some .desktop files and testing everything together [15:56] mterry: how much #1 _isn't_ already done? Are there still some loose ends in Shell? [15:58] josharenson, #1 is still a large chunk of work. Like, we could stuff the Greeter object under a loader, but there are still greeter.XXX references in Shell.qml that need to do something smart in that case. Same for stage.XXX if the stage isn't running. Etc [15:58] mterry: gotcha [15:59] josharenson, but it is mostly contained to Shell.qml at this point. And the stage object is already basically a Loader [16:00] josharenson, landing #1 will be a little odd though, because we'd be supporting modes that don't make sense -- like running just a greeter as your desktop session isn't very functional. But maybe we just don't actually expose the functionality yet [16:01] mterry: sounds good to me, and if its not well received, wouldn't be so bad to have to implement all the checklist items at once [16:02] josharenson, yeah, I just figure the smaller the items are, the more likely we will have landed something by end of three weeks :) [16:03] mterry: a sure === dandrader is now known as dandrader|lunch === alex_abreu is now known as alex-abreu === alan_g is now known as alan_g|EOD === dandrader|lunch is now known as dandrader [20:28] mterry, could you please review this one when you have the time? https://code.launchpad.net/~dandrader/unity8/unifyShellTests/+merge/250986 [20:29] mterry, it's not high priority [20:29] dandrader, ok [20:30] thanks [20:33] Hello :) I googled this and I can't find an answer, so gonna ask you experts. Is there a currently a way to ask unity to move open applications above unopened applications in the sidebar, so they're easier to access?