[08:33] mornin' [08:33] what did I miss? ;) [08:36] [GitHub] Subscribed to credativ/puppet-anysyncd notifications [08:36] ups [08:45] Mirv, btw do you have mediascanner2 in the qt5.2 transition? we added a qml plugin to it very recently, so it probably needs rebuild [08:45] Saviq: not much i'd say [08:46] mhr3, does it use private headers? if not - it should not need rebuilding [08:46] tsdgeos, cool beanz [08:47] Saviq: i don't think what Thomi says makes any sense, or at least any sense for my problem [08:47] Saviq: so i guess all i have left is debug this stuff myself? [08:47] Saviq, no, but doesn't it need to relink against the core5a binary? [08:47] mhr3, ah, right, damn ABI break [08:48] tsdgeos, so you think your issue is not what I described? [08:49] Saviq: you say objectName evaluation is "not taken into account" or something [08:49] i say "the object is simply not there" [08:49] or not updated at all [08:49] i can scroll a list and autopilot vis will still show the old children [08:49] even after restarting autopilot vis [08:50] tsdgeos, I might not have read into your email properly [08:50] well i did not add that extra info (the objects not updating) [08:50] but still it can't be waht you say [08:50] since i have [08:51] objectName: foo + index [08:51] and all children names are [08:51] fooNumber [08:51] and the index is not really changed there [08:51] so all items have the evaluation [08:51] just one at least is missing [08:52] of course it still could be something similar to what you say but i don't think it may be the easiest explanation [08:53] tsdgeos, what I was saying (the problem we encountered) was that ap managed to select a object0 before it became object1 after the binding got updated [08:54] Saviq: right, but how can that cause object0 not existing at all? [08:55] tsdgeos, sure, it can't [08:55] and that's what i'm seeing [08:56] tsdgeos, so yeah, as I said I might not have read your email properly - was Sunday evening after all ;D [08:56] sure :-) [08:57] tsdgeos, so yeah, please reply to my email saying I'm full of shit [08:57] oh yay, no WiFi after flashing... [08:58] mhr3: incidentally, I found out about mediascanner2 on Friday and I've just built it [08:59] Mirv, very well [08:59] Saviq: so back to my question, do you have any clue how/where to start debugging that? [09:01] tsdgeos, /me looks [09:09] Saviq: ah, and we discovered some compiz regression that made some tests fail [09:09] Saviq: https://code.launchpad.net/~aacid/unity8/vjog_compiz_workaround [09:09] basically it's not obeying window sizing sometimes [09:11] tsdgeos, hmm after applying the patch on top of new-scopes-cleanup it doesn't seem to unlock the screen even... does it work for you? [09:12] it was defenitely working on friday [09:13] yeah works today too [09:13] http://paste.ubuntu.com/7066602/ [09:14] Saviq: i'm doing this with 5.0 though [09:15] tsdgeos, I'm on 5.0 currently, too [09:15] ok [09:15] and you get stuck on greeter when rnning autopilot? [09:17] tsdgeos, yes [09:17] try again? :D [09:17] tsdgeos, looks like it's looking for the home scope and fails [09:17] Saviq: are you up to date? [09:17] or maybe i didn't push [09:17] that can very well be [09:17] tsdgeos, did you have any luck with autopilot? [09:18] tsdgeos, ah, actually it didn't switch to apps scope on startup [09:18] tsdgeos, s/didn't/doesn't/ [09:18] Cimi: well, i thinking i have found an autopilot bug, proving it is hard though [09:18] Cimi: did you manage to get past the "everything fails" point? [09:19] Saviq: nothing left for pushing here [09:19] tsdgeos, I stop when you told me you were dealing wih it :) [09:19] Cimi: ok [09:19] I can try again now [09:19] tsdgeos, yeah, the "switch to apps on startup" seems somewhat flaky :/ [09:20] tsdgeos, just read your mail [09:21] Saviq: weird, never had that problem on friday [09:21] tsdgeos, when did this dashcategory error start? [09:21] Cimi: i have no clue [09:21] tsdgeos, it's a new scopes error or also in trunk? [09:26] Cimi: the test works in trunk but the whole QML tree is different, so that proves nothing [09:48] biab [09:56] tsdgeos, are you taking new-scopes introduced regressions as bugs or should I wait for it to first land in trunk ? [09:57] om26er: if it is indeed a regression/bug, sure [09:58] sure, the scrolling in the Apps scope have apparently become much jerky on my mako -- will report in lp. === pete-woods is now known as pete-woods-afk [11:17] tsdgeos, ok they fail for trunk here as well [11:17] tsdgeos, on a different machine [11:17] tsdgeos, I have this QQuickLoader.isCurrent failed error [11:18] no idea : [11:18] you'll have to try to dig more yourself [11:18] otoh it should not be that hard for us to run the autopilot tests [11:18] yeah :( [11:36] tsdgeos, are you trying to [11:36] understand as well? [11:36] Cimi: no, it works for me, i'm fixing the tests in new-scopes that don't work because of the code changes [11:37] @unity who else has issues with autopilot run unity8 ? [11:37] * mzanetti has *always* issues with AP :D [11:37] mzanetti, can you run [11:37] Cimi: I fiddled a while with it last week and got it to run eventually [11:37] autopilot run unity8 > log [11:38] * mzanetti tries [11:38] then a nice grep FAIL log | pastebinit :) [11:44] Cimi: anything you're interested in particular? [11:44] Cimi: I don't want to run all of them now because they either fail or It'll mess up my desktop if I kill my notification system [11:50] tsdgeos, Cimi, the QQuickLoader.isCurrent thing, AFAICT, is dash not switching to Apps on startup [11:51] I'm replying to a mail, will be back soon :) [11:51] and reply to you :) [12:05] Saviq, so we have a fix for that [12:05] ? === MacSlow is now known as MacSlow|lunch [12:21] Cimi, no, that's something I found just today with tsdgeos's cleanup branch [12:21] Saviq, weird [12:21] Saviq, I have that bug with trunk [12:21] ;) [12:22] Cimi, well, if you're going "autopilot run unity8", you're not actually running trunk most probably, but the system-wide installed version [12:22] Cimi, you need `make -C install`, `PYTHONPATH=tests/autopilot autopilot run unity8` to run the local build [12:23] always check out what autopilot says ni "loading tests from..." [12:23] in [12:24] Saviq, oh!! [12:24] Saviq, ok, nevermind :D === _salem is now known as salem_ === alan_g is now known as alan_g|lunch [12:54] Saviq: actually now i'm having a problem with the correct scope not being loaded, it's sad, isCurrent is true, but it's not really the current one, having a look [13:10] didrocks: when you have some time, could you please review the packaging changes in here and related branches? https://code.launchpad.net/~mzanetti/unity-api/new-screenshot-and-focusing-api/+merge/199810 [13:11] mzanetti: sure, will do in ~20, that's ok? [13:11] didrocks: not urgent at all [13:11] ok ;) === MacSlow|lunch is now known as MacSlow [13:43] mzanetti: hum, where are your packaging change? :) [13:43] or the diff isn't correct? [13:46] didrocks: well, it's only bumping unity-application-iml version [13:47] mzanetti: ah ok, that part that we marshmall [13:47] didrocks: and the others bump their "Provides" and "Requires" accordingly [13:47] so yeah, that's the question. is that enough? [13:47] mzanetti: yeah, that's perfect :) [13:47] nice, thanks [13:48] mzanetti: yw, commented (to +1) [13:48] purrfect [13:48] greyback: jfi ^ [13:48] mzanetti: ack from me so === alan_g|lunch is now known as alan_g [14:00] mhall119: mornin', hey i put in for 2 sessions at uds...and they're still in "proposed" can you help get them on the calednar? [14:00] http://summit.ubuntu.com/uds-1403/kgunn72/meetings [14:13] kgunn: what track? [14:13] kgunn: you need to get a track lead to approve it and schedule it [14:13] track leads are listed here: http://summit.ubuntu.com/uds-1403/tracks [14:15] mhall119: thanks [14:20] mterry, did you have better luck with the newer version of the egl-spinner or didn't you try yet? [14:20] MacSlow, I didn't try yet... Let me do that [14:20] tsdgeos: is there something I can do to help with the autopilot updates? [14:21] elopio: i'm on it, just need some more time to debug some stuff [14:21] tsdgeos: ok, let me know if you need a hand. [14:21] mterry, vanvugt tried it earlier today on his mako too and it worked there too. [14:21] tx [14:22] MacSlow, hrm [14:22] mterry, I hope you'll have better luck this time around [14:22] mterry, when it crashes again a backtrace would be helpful if you can get one [14:23] MacSlow, I did find where it crashed... But will see if it happens again [14:23] MacSlow, works like a charm :) [14:24] mterry, rev7 with the text-rendering?! [14:24] MacSlow, yup. might as well merge that into the shared USC branch then [14:24] mterry, awesome! [14:24] MacSlow, has Christina seen this yet / had a chance to comment? [14:25] MacSlow, I know she was OK with the general idea, but curious about details [14:25] mterry, I cc'ed here in the last email (~30 minutes ago) which has a link to the latest video... so I hope she'll reply with some feedback [14:25] cool [14:26] mterry, from my side it basically just cleaning up the code and making it nicer to maintain... so I would (after that) hand off to you and trun back to notification-duties. [14:27] MacSlow, sure [14:27] mterry, although I _love_ GL! :) [14:27] :) [14:27] MacSlow, well if Christina wants changes you can dive back into it :) [14:28] mterry, I take any excuse to do that :) [14:28] kgunn, you didn't read that *waves.hands* ;) [14:33] kgunn: in "alsamixer", if you hit F5, there's a "mic boost" option, maybe you can turn it down? [14:34] mhr3, dednick btw, tomorrow we can meet 1pm in front of brixton tube station [14:35] Cimi: ok. cool [14:39] Cimi, ok [14:42] kgunn, if you want to stay in mumble, maybe we can try different settings, or later when you have time [14:44] tsdgeos, Saviq [14:44] cimi@vostro:~/Development/unity8/new-scopes-cleanup/builddir$ grep -B 3 FAIL LOG | pastebinit [14:44] http://paste.ubuntu.com/7067998/ [14:45] thats' bad [14:45] stuff like test_passphrase_screen_wrong_password should be passing [14:45] yeah, something weird with shift, too - "FFFFFf" instead of "FFFFFF" [14:46] i'm down to http://paste.ubuntu.com/7068005/ failing [14:46] tsdgeos, some context? [14:46] grep -B3 ? [14:46] Cimi: what do you mean some context? [14:47] tsdgeos, apart the error [14:47] sorry [14:47] the name of the failing test [14:47] tsdgeos, I run tests with > log [14:47] Cimi: they fail because they are checking toatlly wrong stuff [14:47] then I grep -B3 FAIL log [14:49] * greyback needs to run to bank, bbiab [14:49] dandrader: so, this Model of apps, that's still the ApplicationManager, right? [14:51] Cimi: i got about 10 min...wanna try real quick [14:51] mumble [14:52] and actually the 4 are basically the same [14:52] the preview opening doesn't work [14:52] because we changed the way preview opening works [14:52] mzanetti, no, it's a simple ListModel of qml items [14:52] mzanetti, it's a list model of MirSurfaceItems (qquick items holding mir surface gl textures) [14:52] mzanetti, then ListView has a Window.qml delegate that takes a suface and reparent it to itself [14:53] dandrader: ah, but shouldnt the MirSurfaceItems be the Delegates? [14:53] and the model still hold ApplicationInfo objects? [14:53] Cimi, l4d or l4d2? :) [14:54] dandrader: at least that's what I though of when doing the right edge [14:54] mhr3, l4d2 [14:54] dandrader: so all we'd need to do in that case would be to replace "Image" with "MirSurfaceItem" in the SpreadDelegate.qml [14:54] mzanetti, no, MirSurfaceItems are created from the qpa/cpp/appMan side. it essentially wraps a gl texture that's bound to a mir (EGL) surface [14:55] mzanetti, so that the qml side can manipulate that texture as a QQuickItem fairly easily, [14:55] ok... so the ApplicationInfo objects will become MirSurfaceItems? [14:55] mzanetti, no, they're are still there [14:55] but then we have 2 models [14:56] mzanetti, yes [14:56] ok... I think I understand [14:56] but is that really way to go? [14:56] mzanetti, the concept of an application is one thing, more abstract, a client surface is another one [14:57] yeah. [14:57] that's why I said, the MirSurfaceItems should be the delegates [14:57] mzanetti, and stage wants to arrange an shuffle around surfaces/client windows [14:57] but the stage still needs to represent all running apps [14:58] and not some other model of surfaces [14:58] dandrader: so we can't create such MirSurfaceItems in QML as we like? [14:58] for example, one in the app-spread, one in the runningappsgrid etc [14:59] but still showing the same surface [15:03] mzanetti, at the moment no. mir tells the QPA that there's new surface and then QPA creates a MirSurfaceItem that makes use of it, so that the mir surface can be used by the QML scene [15:03] mzanetti, then QPA informs the QML side that there's a new MirSurfaceItem available [15:04] mzanetti, QML side (unity8) then reparents it to the view delegate to have it on the scene [15:04] dandrader: what I mean is this: I'm ok if we need to create such MirSurfaceItems somewhere in C++, but I don't think we should only have that model available in qml... Can't we wrap that somehow? So that I can just say this in QML: [15:04] AppSurface { appId: "foobar" } [15:05] and that then just finds its MirSurfaceItem somehow and paints its contents in the AppSurface {} item [15:05] so we can still do things like: [15:05] ListView { model: ApplicationManager; delegate: AppSurface { appId: model.appId } } [15:06] otherwise this sounds like a hugely complex thing to me to match surfaces with apps in QML all over the place [15:07] mzanetti, I think it would be good if you checked out the current stage code to get an idea on how that all fits together [15:08] dandrader: what is the current stage code? [15:08] mzanetti, lp:~unity-team/unity8/mirCompositor [15:08] ok. I'll check it out [15:10] mzanetti, also an application can have many surfaces (on the desktop at least) [15:10] dandrader: true... but that doesn't really invalidate my point [15:11] just something to keep in mind [15:11] yeah, sure === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [15:51] someone do me a favour please: run "xev | grep -A 3 KeyRelease" and hit SysReq [15:53] greyback, http://paste.debian.net/86897/ [15:53] Zhenech: perfect, thank you [15:54] it's a thinkpad, so I actually pressed Fn+PrtSc if that matters [15:54] * greyback has almost managed to remap his macbook's eject key to sysreq [15:55] Zhenech: I've a macbook, no such key so playing with remapping [15:55] I've mapped eject to printscreen, which is close :) [15:55] post it on the web, newer thinkpads miss that one too [15:55] will do [15:57] eean: is it ok for you if i kill the get_details autopilot tests? they don't do much, do they? [15:57] what? [15:57] eean: sorry [15:57] elopio: read ↑↑↑↑ [16:03] tsdgeos: I need that to check the application details shown on the click scope [16:03] if you remove the test, I might break the helper without noticing [16:03] ?¿ [16:06] elopio: ↑↑ [16:06] not sure i understand what you men [16:06] mean [16:07] mterry, pushed rev8 of the egl-spinner, which should be easier to read and alter. Until we get feedback from Design, I'll leave it as-is. If you run into any issues just ping me. [16:08] MacSlow, thank you! [16:08] mterry, yw [16:16] tsdgeos: I mean that if something changes on unity and get_details doesn't return what I expect it to return [16:16] then the click scope tests will start failing without a clear reason [16:16] why do you want to remove them? [16:17] elopio: because i don't see the point [16:18] i mean it's not testing anything else than "we're showing stuff on screen" [16:19] tsdgeos: it's testing that get_helpers returns the title, the subtitle and the description. [16:19] the real user tests is in the click scope, where we check that the title, subtitle and description match the ones of the application we are opening. [16:19] but lets say we remove this test in unity. [16:20] and for some reason, unity no longer shows the description anymore. [16:20] without the test, we are with no clue about how it will affect other tests that might be using the helper. [16:22] ah [16:23] so you're using that helper from somewhere else?¿? [16:23] elopio: ↑ ? [16:24] tsdgeos: yes. On the unity-scope-click suite. [16:24] tsdgeos: all the things you will see with tests in test_emulators.py are because I'm using them somewhere else. [16:24] elopio: well [16:24] that sucks [16:24] i mean i suck probably :D [16:24] they are not user tests, just to make sure that the helpers remain working. [16:24] is that documented? [16:24] because i had no idea [16:25] elopio: maybe it's better you fix those tests then [16:25] those = unity8 [16:25] because i'm not going to go over unity-scope-click suite and fix them to work with new-scopes [16:25] tsdgeos: yes, mzanetti told me to add a big comment on test_emulators.py explaining it. [16:25] it's after the copyright. [16:26] poor elopio... first needed to convince mzanetti and now tsdgeos :D [16:26] tsdgeos: you won't need to go over unity-scope-click suite as long as you keep the signatures of the helpers. [16:26] elopio: i can't keep the signature of the helpers [16:26] because everything has changed in previews [16:26] it's a different world [16:26] and if you have to break a signature, then you talk to the devs of the scope tests, and coordinate with them the updates. [16:26] right [16:26] in this case, I'm the dev of the scope tests :) [16:26] i don't want to do that [16:27] tsdgeos, don't want to do what? [16:27] be a good citizen [16:27] ;) [16:27] feeling moody today [16:27] sorry [16:28] tsdgeos, those helpers are ours, we need to commit to them, and yeah try and maintain the signatures, but obviously in this case we'll need to break them [16:28] click scope tests will need to be redone just as well [16:28] tsdgeos: I'm here to help with whatever you need. [16:28] I'm used to moody devs, don't worry ;) [16:29] elopio: well basically [16:29] the get_details thing doesn't apply anymore [16:29] since you can have a random number of stuff in a preview [16:30] there is not just one title [16:30] tsdgeos: what I suppose we should do is to make it specific for the application preview. [16:30] you can have 10 of them [16:30] tsdgeos: the application preview will still return title, subtitle and description, right? [16:30] elopio, I don't think so, that knowledge needs to move towards the click apps scope [16:30] the application preview doesn't exist in our autopilot world [16:30] elopio, it's the scope that decides what will be displayed, we have no control over it, effectively [16:30] is just a preview the same it was [16:31] that sentence made no sense [16:31] is just a preview as the others [16:31] so that's why i want to remove the whole get_details thing [16:31] tsdgeos, Saviq: yes, so we need to move this get_details to the autopilot helpers that live in unity-scope-click [16:31] i don't think it makes sense [16:31] tsdgeos, and we need to facilitate that ↑↑ [16:31] tsdgeos: then, I agree with you that it needs to be removed. [16:32] just, we need to be in sync so the tests in unity-scope-click are broken just for a little time. [16:32] elopio, or not at all, we can land them together with new-scopes [16:32] tsdgeos: remove the helper, remove the test, and when we have all the other helpers working, I will prepare a branch for the scope. [16:32] elopio: cool thanks [16:33] tsdgeos: thanks to you. [17:04] Saviq: elopio: i just pushed something that makes the last tests i had written down as not passing pass again, hopefully it should not regress the rest of the tests [17:04] will check first thing tomorrow in the morning === dandrader is now known as dandrader|lunch === alan_g is now known as alan_g|EOD === dandrader|lunch is now known as dandrader [19:34] mzanetti, hey, tried hi-dpi Unity7? worth checking out probably :) [19:34] mzanetti, it's in the archive already, iiuc [19:43] Saviq, yup it is, dont open the dash or hud though :) [19:55] is there an easy way on current build of unity 8 on nexus 7 to get the sidestage, or do I need the bleeding edge stuff shown at MWC? === dandrader is now known as dandrader|bbl [20:42] popey, no, no easy way, and it won't be enabled soon either, it was using a hack to rotate the shell, which we need to replace with a proper implementation [20:55] Saviq: thought so, thanks === jhodapp_ is now known as jhodapp|brb === jhodapp|brb is now known as jhodapp [22:10] mhall119: ok...so i got my sessions approved but not scheduled [22:10] http://summit.ubuntu.com/uds-1403/all/ [22:10] see i'm listed here...but no date/time [22:10] any ideas (day before :) [22:13] kgunn: still up to the track lead to schedule it [22:13] ug...they've all gone home [22:26] kgunn: any one i've clicked on has a time [22:28] ahh, yours is the only one without a time [22:35] mhall119, do you use your @canonical mail? === dandrader|bbl is now known as dandrader === salem_ is now known as _salem === jhodapp is now known as jhodapp|afk [23:28] mhr3: sometimes, why?