[00:02] digital, im not actually that sure, as you are no longer using unity :) [07:34] didrocks, ping [07:35] davidcalle: pong [07:35] didrocks, ça va bien ? [07:35] davidcalle: ça va, et toi, les vacances bientôt? :) [07:35] didrocks, ça va :) Oui, dans quelques semaines... Et toi ? [07:36] fin août, donc il y a le temps! [07:37] didrocks, hehe, en effet ;) Au fait, j'ai oublié ça hier https://code.launchpad.net/~davidc3/cupstream2distro-config/online-raring-to-saucy/+merge/172476 [07:38] davidcalle: il n'y a pas d'importance en fait de ça, puisque cette stack n'est pas sous daily releases [07:38] (tout ce qui est dans online) [07:38] davidcalle: tu peux même le supprimer en fait :) [07:39] didrocks, hum... J'ai l'impression que Jenkins ne peut pas trouver les build deps, ça ne dépend pas de ça, donc... [07:39] intéressant [07:40] dac [07:40] done! [07:41] didrocks, par hasard, est-ce que Jenkins utilise la distro du dernier changelog ? [07:41] davidcalle: possible, aucune idée pour la partie non daily release, ça a pas mal changé depuis mon premier prototype :) [07:42] mais à l'époque, ça utilisait la distro du changelog [07:42] didrocks, ok, je vais regarder de ce côté là aussi. Merci ! ;) [07:44] de rien! [07:55] pstolowski: thostr_: hey! [07:56] didrocks: hi! [07:56] FYI https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1189088/comments/20 [07:56] Launchpad bug 1189088 in unity (Ubuntu) "Unity Dash, left mouse button starts preview instead of directly the application" [High,Triaged] [07:56] so I think we need to revert the behavior for the application lens [07:56] as well as adapating surely some integration tests, right? [07:58] didrocks: yes. and most probably extending the scope interface with a kind of activationBehaviour flag [07:59] thostr_: right, would make sense [07:59] do you know how long we can expect this to take before landing? [07:59] would be good to have the definitive action soon enough [07:59] didrocks: that's good news... the bad thing is it has been implemented in the dash globally [08:00] argh [08:01] didrocks: so we need to make an 'exception' for apps in the dash... or implement something along the thostr_'s idea [08:01] just give me a rough approximation for landing those :) [08:07] the fix shouldn't take that much time (hours) [08:07] the challenge is predicting the integration and fixing the tests [08:13] thostr_: right, we need a plan for that. Can you preparing one and get back to me once you get the estimate? [08:13] pstolowski, thostr_ couldn't this work roughly how action activation works in previews? i.e. on activation you either come back with a preview (or a hint that a preview should be requested?), notice of scope-side activation or no-op (as it is now, right?) [08:14] didrocks: I'll implement the change across the stack today (libunity, unitycore, dash). not sure about AP test atm, I've a feeling it's simple change [08:14] pstolowski: ok, let's cross fingers! Thanks :) [08:15] Saviq: it could... but then you need to change every single scope. so we can think of it as a 'hint' about how to handle activation click [08:15] pstolowski, you guys managed to make alt-f2 skip previews, isn't the "framework" already there? [08:16] or how is that different? [08:16] pstolowski, right... [08:17] seb128: for alt+f2 we added a special pseudo-schema "x-unity-no-preview:". no use for this case. [08:18] ok, I was mentioning it in case... ;-) [08:18] you probably know what you are doing, /me goes back to stuff he's doing :p [08:18] seb128: that was a good question, it got me puzzled for a second and I had actually to lookup the sources ;) [08:18] ;-) [08:18] seb128: see, they don't know what they are doing OMG /o\ ^ [08:19] haha [08:29] Saviq, tsdgeos: are you ok with https://code.launchpad.net/~stolowski/unity8/activation-and-previews/+merge/172102 for the time being (I've just merged trunk as there was a text conflict)? [08:30] pstolowski, I didn't look at the code much, tsdgeos your call - I'm ok with leaving LocalResult for now [08:30] pstolowski, btw, can you add a TODO/FIXME around there? [08:31] pstolowski: i'll have another look onow [08:31] Saviq: sure. I plan to do the change today/tomorrow [08:31] pstolowski, the index approach looks sane, either with a "master" index role passed down from DeeListModel [08:32] pstolowski, or maybe even saner with mapToSource() [08:32] Saviq: do you think we are safe wrt to potential races? should be ok as long as we're single-threaded? [08:33] pstolowski, think so, but just to be safe maybe the index role approach makes more sense [08:33] we did that somewhere already, yoo [08:33] too [08:34] pstolowski, I think in the Categories object [08:34] Saviq: ok [08:34] pstolowski, yeah, case RoleId: return QVariant::fromValue(index.row()); [08:35] Saviq: ok, nice, I didn't see that [08:52] is testDragHandle failing for you too? [08:52] Saviq: greyback: mzanetti: ↑↑↑ [08:53] * mzanetti tries [08:53] Actual (actualFunc()): 100 [08:53] Expected (expectedValue): 0 [08:53] Loc: [/home/tsdgeos_work/phablet/lolo/tests/qmltests/Components/tst_DragHandle.cpp(139)] [08:54] tsdgeos: yep. failing here too [08:54] * mzanetti wonders how that passed CI [08:54] yeah :D [08:54] we were unlucky in that it passed CI [08:54] and now it's blocking the rest of the stuff :D [08:54] hm. that passed when i tested before approve [08:55] didn't dandrader file an MP to fix? [08:55] https://code.launchpad.net/~dandrader/unity8/fixDraHandleHint/+merge/172426 [08:55] it was probably my fault [08:55] 'cause I merged lp:unity8 and lp:unity/8.0 [08:56] happroved [08:56] ahhh [08:56] makes sense [08:56] tsdgeos, dednick, mzanetti ↑ [08:56] ah ok [08:56] Saviq: ah ok. yeah, i remember that line there [08:56] bzr f***ed up [08:57] Saviq: tsdgeos: still, the commit that introduces the failure was approved by PS Jenkins Bot [08:57] mzanetti, nope [08:57] and also dednick and tsdgeos... [08:57] mzanetti, there were two commits in parallel to lp:unity8 [08:57] mzanetti, and lp:unity/8.0 [08:57] mzanetti, to sync them, I merged one into the other [08:57] hmm... so much for the history in bzr [08:57] mzanetti, and pushed [08:57] mzanetti, it's fine [08:58] mzanetti, only I didn't test [08:58] /mea culpa [08:58] but bzr didn't complain... :/ [08:58] question [08:58] why https://code.launchpad.net/~stolowski/unity8/activation-and-previews/+merge/172102 CI passed ? [08:58] Saviq: would you be apposed to me creating cmake procedure for qmltypes files for the plugins? so we can get type info in qtcreator? [08:59] dednick, no, I wouldn't be apposed :) [08:59] dednick, assuming it doesn't take your whole day [08:59] Saviq: nope. should be quick [08:59] dednick, yup [08:59] dednick: yay. nice [09:00] tsdgeos: oooops [09:00] tsdgeos: you found a bad bug [09:01] where? when? how? [09:01] tsdgeos: the jenkins job still merges to lp:unity/8.0 before testing [09:01] wops :_D [09:02] fixed. tests may start failing now :D [09:04] yay! [09:04] not! [09:06] huh? [09:08] Saviq: any news on the unity-api releasing? [09:08] didrocks, ↑? [09:08] mhr3: hey, around? [09:09] mzanetti, that's no bug, lp:unity8 == lp:unity/8.0 [09:09] mzanetti, well, bug, maybe - but nothing tragic yet [09:09] mzanetti: what about it? it's released and in the next ppa [09:09] Saviq: well, tests fail with unity8 but pass with unity/8.0 [09:09] as the whole unity8 stack [09:10] didrocks: ah ok... I thought we shouldn't use any ppas any more [09:10] mzanetti: well, apparently, we can't push unity8 to distro right now [09:10] to we have to use a ppa for releasing [09:10] mzanetti, unity8 will remain in a PPA until mir integration is complete [09:10] didrocks: ok. thanks [09:12] didrocks, in hangout [09:12] mhr3: there is just the start of a hang [09:12] mhr3: on the ati machine [09:12] mhr3: so run if you want to look at it :) [09:15] didrocks, are we merging all the qtubuntu and unity8 packaging changes today [09:15] ? [09:16] Saviq: waiting for Mirv to deal with the unity stack to get libunity in distro and then, yeah :) [09:16] didrocks, cool [09:16] mhr3: letting the -check running for you [09:24] mzanetti, you have a moment? [09:26] nic-doffay: meeting starting in 4 minutes [09:26] nic-doffay: what up? [09:26] Does anyone here have an always-displayed global menu bar working? I've installed various PPAs and they never seem to work.. I used to have this with ASD's global menu hack and gnome panel widget... [09:26] oh, interesting. [09:27] mzanetti, unknown issue I'm facing. [09:27] QML related. [09:28] nic-doffay: that's not exactly the most precise description of a problem... [09:28] nic-doffay: have some code and a error message? [09:29] mzanetti, it's in the gallery of the SDK. [09:29] ./gallery.sh [09:30] The templates are displayed using a delegate. [09:30] Everything on the component is repeated, even though it follows the same trend as all other gallery components. So I must be missing something. [09:31] It's fine in a normal qmlscene. [09:32] nic-doffay: not following [09:32] nic-doffay: I've opened the gallery [09:32] now what? [09:32] mzanetti, you'll need my branch. [09:32] mzanetti, lp:~nicolas-doffay/+junk/option-selector [09:33] mzanetti, to see the issue go to list items, then scroll down to option selector. [09:33] That's handled is ListItems.qml [09:34] *in [09:34] didrocks, eh, lost me scrollback, which one was it? [09:34] my* [09:34] mhr3: ati machine [09:34] thx [09:34] yw :) [09:36] didrocks, can't attach to it again, so no good :/ [09:36] jibel, ^ ati dead [09:37] mhr3, ack, oom [09:37] can I have that machine for 15min, I'll collect information to report a bug [09:37] jibel, sure [09:38] ehm, i mean, not my call :) [09:38] but surely 15minutes won't kill anyone [09:38] even more if needed :) [09:38] nothing else to run [09:39] mhr3: when you are free can you have a look at #1196878 #1196800 and #1196822 [09:39] you know both the components well :) [09:42] [76766.931095] lxc-list[26171]: segfault at 0 ip b762c1a6 sp bfd80674 error 4 in libc-2.17.so[b75a7000+1ad000] [09:42] not good [09:43] m4n1sh, what's the benefit of moving those to libzg2? [09:44] jibel, so, did i make it crash with the list call? [09:44] nic-doffay: so. meeting has ended. checking out your branch now [09:45] mhr3, no compiz and recordmydesktop crashed before that [09:45] I'm disabling the node in jenkins, so it doesn't reboot under my feet [09:50] pstolowski: ping [09:50] tsdgeos: pong [09:51] pstolowski: the data of a unityPreview never changes? [09:52] tsdgeos: yes [09:52] damn yes to a negative question :D [09:52] :) [09:52] yes as in "yes, it does change" or "yes, it does not change"? [09:52] tsdgeos: it doesn't change [09:52] oki === shiznix_ is now known as shiznix [09:53] pstolowski: i was thinking that you could reuse a single signal then [09:53] instead of one per property [09:53] since you're only going to emit it at setPreview(unityPreview) anyway [09:53] no? [09:55] mhr3: moving everything to libzg2 [09:56] to avoid long term maintenance of libzg1 [09:56] some fine day they will have to be moved [09:58] Saviq: can't install libunity-api from the next ppa. requires libboost 1.49 while we require libboost-1.53 [09:58] tsdgeos: in therory yes, but doesn't NOTIFY protocol mandate that notification of every property follows Changed naming? [09:59] not that i know [09:59] but i may be wrong [09:59] afaik the qml side of things is all defined by the property name [09:59] mzanetti, we shouldn't require boost at all, do we still? [09:59] and the rest is just the c++ names of the "bindings" [10:00] Saviq: apparently... installing libunty-api-dev and then doing ./build -c removes libunity-api-dev again. because of conflicting boost [10:00] Saviq: oh... comes with libunity-core and libnux I guess [10:00] mzanetti, yeah, probably [10:01] mzanetti, still, needs fixing in unity-api [10:01] mzanetti, MP incoming [10:04] * Saviq hooked sbuild up to `bzr bd`, feels smug ;) [10:05] pstolowski: i mean http://paste.kde.org/~tsdgeos/787022/ in case i wasn't clear [10:05] Saviq: mzanetti: ↑↑↑ this should work, right? [10:05] tsdgeos, yeah, should [10:06] tsdgeos: yes, that's what I understood [10:06] what you guys think about that? [10:06] am i being silly? [10:07] tsdgeos, pstolowski question is - do they actually ever change? [10:07] should they just be constant instead [10:07] because you create a new Preview every time anyway? [10:08] they never change [10:09] or so said pstolowski a few lines above [10:09] well, they change from empty to something [10:09] and then they never change [10:09] Saviq: in theory you could create a preview object with default ctor, and then do setUnityPreview [10:09] in which case they would change [10:09] well, it's what you do, no? [10:09] pstolowski, right [10:09] yes, but only in factory method [10:09] so it's not exposed to qml before that [10:10] right [10:10] pstolowski, tsdgeos so it's never possible that only one of those change [10:10] pstolowski, tsdgeos in which case I'm fine with tsdgeos's thing [10:10] Saviq: exactly [10:13] pstolowski, tsdgeos only thing to make sure is to only emit the signal once [10:13] pstolowski, so in setUnityPreview [10:13] after having updated all the members [10:14] mzanetti, https://code.launchpad.net/~saviq/unity-api/any-boost/+merge/172525 [10:15] Saviq: cheers [10:15] Saviq, tsdgeos: ok [10:15] dednick: you have a lot of "\No newline at end of file" in your MP [10:16] pstolowski: will give us a somewhat smaller binary ;-) [10:16] mzanetti: grr. ok, fixing now [10:17] lol [10:17] * Saviq sees latestsnapshots [10:17] Saviq: yep, let's wait for it to migrate to the release pocket to do the merges? [10:18] didrocks, yup [10:18] Saviq: on the qtubuntu transition, they wanted to wait for the flipped image to be switched on though [10:18] Saviq: and I think it's not there yet? [10:18] didrocks, yeah, might be [10:19] Saviq: so we'll have to postpone the packaging merge until it's in, I'll ask rsalveti and sergiusens once they are around [10:20] didrocks, yup [10:34] Saviq, unity on the phone is constantly using 25% cpu with yesterday's image, known? [10:36] I get as high as 46% continuous hog on Nexus 4 for unity8 [10:36] mhr3, om26er is that with the flipped image? [10:37] Saviq, yeah, I have flipped image installed [10:37] Saviq, i don't [10:37] hmm I'm not seeing that... [10:37] on 20130701.2 [10:37] there's also 10% in surfaceflinger, so clearly something's repainting [10:38] dunno how to give you more info [10:38] mhr3, we'll try and reproduce [10:39] this is even with screen off btw [10:40] on another note the NetworkManager was consuming 99% yesterday, had to kill the service. But that's clearly for someone else to figure. [10:42] woa, i'm getting 40% of unity8 in my desktop [10:42] that's bad :D [10:46] uh oh, i just dist-upgraded and unity just crashes [10:48] mhr3: apt-cache policy unity? [10:49] Installed: 7.81.2 [10:50] ah, unity8 [10:50] as we just published a version to saucy, you frightned me :) [10:51] didrocks, you're frightening me, you don't care about working system on the phone?! :P [10:52] mhr3: not that, but as long as we don't run integration tests and if that passed tests, it would mean we have big issues in our infra :) [10:56] tsdgeos, I can't reproduce, can you try and find out which commit introduced that? [10:57] amy not be our fault [10:57] i did a qml-debug thing on the phone [10:57] tsdgeos, right [10:57] and it has no repaint [10:57] so may be some other thing [10:57] i was going to run callgrind on it now [10:57] ok, I'm not seeing it anyway [10:57] modprobe nouveau [10:58] oh figures [11:01] mzanetti, seems to work but all tests are screwed :D will have to dig but thanks [11:01] Cimi: ok, cool [11:04] tsdgeos: updated my MP [11:07] oki [11:07] Saviq: gone back to r37 and i still get around 40% cpu usage when i'm at the videos scope [11:08] tsdgeos, flipped or not? [11:09] tsdgeos, and is that on your desktop still? [11:11] Saviq: desktop yes [11:11] tsdgeos, interesting [11:11] callgrind didn't gave me much [11:11] gdb seems to imply there's something drawing all the time [11:11] since everything i ctrl+c there's a renderer thread going on [11:11] but i may be wrong and the thread is just waiting [11:12] anyways, if you guys don't get it let's forget it for the moemnt ;D [11:13] tsdgeos, well, mhr3 and om26er seem to get it on devices [11:13] tsdgeos, do you? [11:14] let me see === MacSlow is now known as MacSlow|lunch [11:22] Saviq: no, can't repro on the phone [11:23] tsdgeos, k, I'll have a shout in -touch, too === alan_g is now known as alan_g|lunch [11:55] didrocks, mhr3 ati is back, but I couldn't collect data I needed. I killed the machine by removing the wrong modules. It's all yours, ping me when it crashes again === MacSlow|lunch is now known as MacSlow [12:38] didrocks, ping again. Who should I talk to for the Jenkins issue ? I now have both a saucy changelog and cupstream2distro switched to saucy, but the CI pass is still trying to build with raring (and obviously fails). [12:48] davidcalle: fginther should be your contact [12:48] jibel: ok, thankx [12:48] didrocks, thanks === alan_g|lunch is now known as alan_g === _salem is now known as salem_ [13:06] pstolowski: why setUnityPreviewBase + setUnityPreview instead a virtual setUnityPreview where children call the parent implementation too? [13:13] tsdgeos: because I didn't like to require subclass to call base class method in the overriden method to work correctly, I preferred to ensure this from the outside (the factory method) [13:15] ok [13:17] paulliu: shall i top-approve https://code.launchpad.net/~stolowski/unity8/activation-and-previews/+merge/172102 or you want to have a look too? [13:18] tsdgeos: I've looked already. No special comments. [13:18] ok [13:20] thanks for reviewing! [13:25] mterry: and is that "regression" accepted? [13:26] tsdgeos, it's only a regression if we consider the touch image to be just a demo. Otherwise it's getting closer to actually working [13:27] sure [13:27] but still it'll look worse :D [13:27] tsdgeos, it'll look the same as when a user gets the device for the first time [13:27] ok [13:27] tsdgeos, which is what's happening when you install [13:43] dandrader: you need a review for panel drag handle branch? [13:43] dednick, oh yeah [13:44] dednick, a big diff, unfortunately [13:44] dandrader: indeed. i'll get on it [13:45] dednick, thanks a lot! [13:46] dednick, biggest changes are in the tests though [14:01] greyback, I'm getting this error when running unity8+mir on the flipped saucy http://paste.ubuntu.com/5836289/ [14:01] greyback, how far did you get? [14:02] greyback, btw, I've updated that wiki page [14:02] dandrader: not even that far :) I get crash [14:02] dandrader: there's no working Ubuntu.Application plugin yet. So you need to use hte fake one [14:03] mzanetti, can I ask you to please look at https://code.launchpad.net/~nick-dedekind/unity/8.indicators-client/+merge/168022 from autopilot PoV? [14:03] greyback, ah... it's weird the library is actually installed, so it should at least find and load it [14:03] greyback, have you gone the ppa:phablet-team/mir way? [14:04] dandrader: no, I'm using the mir staging ppa. To my detriment I believe [14:04] but I need to work on cutting edge [14:04] dednick, are the autopilot tests actually run in the indicators-client branch? [14:04] dednick, they're not part of the unity8 autopilot module, are they? [14:04] mzanetti, should they be ↑? [14:05] dednick, is there a custom target to run them maybe? [14:05] dednick, I'm unable to get them to work, I'm afraid [14:07] Saviq: hm. apparently they're not. [14:08] Saviq: should i make them part of the unity8-autopilot package? [14:08] Saviq: or make a indicators-client-autopilot one? [14:09] Saviq: is there an easy way to only load the dashhome? [14:11] tsdgeos, it's a dconf setting [14:11] greyback, ok it's working now! [14:11] dednick, I think part of unity8 [14:11] dandrader: excellent! [14:11] dednick, as that's the only autopilot package we're running in CI [14:12] mzanetti, any comment on ↑? [14:12] greyback, but doesn't mean much as it's using fake plugins, just like when you "./run" on the desktop [14:12] dandrader: yes I know. I'm working on a real application manager plugin so we get real behaviour back [14:13] fginther, hey, is it correct that http://10.97.2.10:8080/job/unity-phablet-qmluitests-saucy/ is bound to ps-saucy-server-amd64-1 ? [14:13] fginther, we're kinda blocked by mir on that one :/ [14:14] Saviq, I think I can help there, just give me a few minute [14:14] s [14:14] fginther, thanks [14:26] fginther, btw, would it be possible to have unity-8.0-autolanding and unity8-autolanding to block each other? i.e. so they don't run in parallel? [14:29] Saviq, cant' we forget about lp:unity/8.0 ? [14:29] dandrader, we will, soon [14:29] Saviq, maybe. I don't know of a simply way to do this off the top of my head, but it may be possible. Can we discuss a little later (an hour or so)? [14:30] fginther, probably not worth it :) [14:30] Saviq, ok [14:30] dandrader, OTOH, we could just resubmit against lp:unity8 and be done with it... [14:31] dandrader, I think I was just hoping we'll reduce the queue against lp:unity/8.0 quicker [14:31] Saviq, exactly [14:31] Saviq: i've added indicator-client to the autopilot target. should now run with make autopilot. [14:31] and added to install. [14:31] dednick, cheers [14:33] dandrader, ok, doing [14:33] dandrader, there's only 4 branches now [14:34] dandrader, right, but... "lp:~unity-team/unity/unity8-packaging-cleanup is not mergeable into lp:unity8" [14:34] * Saviq needs to pull/push [14:36] Saviq: sorry... have been eating [14:36] mzanetti, don't be === alan_g is now known as alan_g|tea [14:39] dandrader, ok, did it, if anything happens, I'll blame you :P [14:40] didrocks, resubmitted against lp:unity8 https://code.launchpad.net/~unity-team/unity8/packaging-cleanup/+merge/172578 [14:40] dednick, resubmitted against lp:unity8 https://code.launchpad.net/~unity-team/unity8/indicators-client/+merge/172582 [14:40] Saviq: ah great! let's wait from the phone fundation team for the greenlight :) [14:41] Saviq: ta [14:41] didrocks, yup, with that we'll be able to drop lp:unity/8.0 altogether [14:41] Saviq: debian/unity8.install merge conflict btw :) [14:41] didrocks, will merge [14:41] thx! [14:42] Saviq, hahaha [14:46] lol, someone using the LVWPH exposed a bug in QLimitProxyModelQML :D [14:46] took me a while to trace === alan_g|tea is now known as alan_g [14:49] tsdgeos, ;) [14:50] tsdgeos, test, fix, ship it :) [14:50] on it === dandrader is now known as dandrader|afk [14:57] Saviq: dednick: commented regarding autopilot tests [14:57] mzanetti, yup, thanks [14:59] mterry, uh, the meeting got pushed back to tomorrow, can you please ping sforshee to have a chat about the blanking? [15:03] mzanetti: ta. i'd prefer to tackle the further dev of autopilot in a future revision. This one is taking a bit long, and I'd like to get it in so we can link up with the new backends sooner rather than later. [15:03] dednick: sure... It wasn't meant to add it here [15:04] dednick: but as you guys asked me on the general opinion for the autopilot suite, I wrote it there as a comment [15:04] mzanetti: i'll fix up the targets. I'm sure i tried and the unity8 ap tests were run as well as the indicators_client. [15:04] dednick: oh... I didn't do that [15:04] dednick: I killed the unity8 one after a while [15:04]