=== philipballew is now known as philip [07:35] filed bug #1254986 regarding unity-mir with new Mir [07:35] bug 1254986 in unity-mir "unity-mir FTBFS against libmirserver11" [Critical,New] https://launchpad.net/bugs/1254986 [08:04] but before that the priority #1 is anyhow bug #1253685 [08:04] bug 1253685 in unity8 (Ubuntu) "unity8 crashed with SIGSEGV in QMetaObject::activate()" [Critical,Confirmed] https://launchpad.net/bugs/1253685 [08:18] Mirv: why is that critical now? it's just at shutdown, no? [08:18] Saviq: ↑↑↑ [08:18] didrocks: ↑↑↑ [08:19] tsdgeos: first reboot will take an additional 25s [08:19] tsdgeos: because whoopsie collect the crash [08:21] so we're marking something critical for 25s [08:21] not sure which qualification we'll use when we delete user data unexpectedly :D === iahmad is now known as iahmad|afk [08:22] tsdgeos: 25s on mako, seems to be 2 minutes on maguro [08:22] tsdgeos: so no, I was happy to ignore the crash, until we realized that [08:22] tsdgeos: seems it's an unity-mir issue [08:36] didrocks, FWIW I'm not even seeing such long reboot times - apport seems just interrupted here on reboot [08:37] Saviq: it's only the first boot [08:37] didrocks, re-boot [08:37] after you removed the crash file [08:37] didrocks, and yes, not seeing that [08:37] waow, can get multiple people confirming it [08:38] didrocks, I'm not saying you can't, just mentioning - tried to repro that behavior and couldn't [08:38] Saviq: the maguro tests were reported to me, I can only confirm the mako one thogh [08:38] though* [08:39] let's see once greyback is around what effort it is to fix it [08:39] oh, a greyback! :) [08:40] didrocks: which bug? The stop unity8 crash? [08:40] greyback: yeah [08:41] I think I got the other regressions fixed (just need one testing again) [08:41] didrocks: am on it now. Hopefully isn't much work [08:41] so you're the last one potentially :) [08:41] and then, we can repromote an image [08:41] ok [08:41] thanks greyback! [08:41] greyback: we'll probably just cherry-pick it FYI [08:41] didrocks: gotcha [08:41] greyback: on latest Mir version (libmirserver10 then) released in the distro [08:42] * didrocks doesn't want to risk more [09:04] greyback: no pressure, but confirmed you are the last issue :) [09:04] didrocks: yep, that was clear the first time :P [09:04] (phew, almost there ;)) [09:05] greyback: well, I didn't get my GSM dummy fix confirmed, done now :p [09:05] I just raced with you, but mine was easier TBH [09:05] ;) [09:05] :) [09:11] didrocks: where's the best place to get mir 0.1.2? the mir staging ppa? [09:11] greyback: hum, please, use 0.1.1 [09:11] greyback: I don't plan to add another variable of change on the phone [09:12] and just use 0.1.1+14.04.20131120-0ubuntu1 that we ship [09:12] with unity-mir + your patch [09:12] didrocks: ok, for some reason I thought 0.1.2 has exposed this bug. [09:12] greyback: no, it's really latest in trusty [09:12] (and latest phone promoted image) [09:13] ok === iahmad|afk is now known as iahmad [09:23] Saviq: didrocks: can either of you have a look? https://code.launchpad.net/~gerboland/unity-mir/fix-shutdown-crash/+merge/196677 [09:23] greyback, am [09:23] ta [09:23] greyback, oh, nice catch [09:24] Saviq: yeah, I've learned lambda functions as slots are dangerous [09:24] greyback, which totally makes sense [09:24] indeed [09:24] oh interested [09:24] * didrocks didn't know about that one [09:24] but they're so pretty [09:34] greyback: why dangerous? [09:35] ah [09:35] * tsdgeos reads the MR [09:36] greyback: good catch! [09:36] tsdgeos: yeah. I guess we should ensure if you use lambda slot, it does not capture anything, i.e. is of form "[]" [09:37] so there's no side effects [09:38] yeah :/ [09:38] https://code.launchpad.net/~aacid/unity8/queuedModelCountChanged/+merge/196680 anyone? [09:41] didrocks, mir not yet published, you said you want to cherry-pick unity-mir [09:41] didrocks, but unity-mir trunk already build-depends on new mir [09:41] didrocks, shall we revert that b-d? especially since it FTBFS... [09:44] greyback, ↑↑ [09:44] Saviq: I say revert [09:44] greyback, ok with me uncommitting? [09:44] Saviq: yep [09:44] greyback, you'll have to rebase your branch [09:44] Saviq: I know, it's grand [09:45] no, don't worry [09:45] I don't want to take the other commit as well [09:46] greyback: Saviq ^ [09:46] I just tested current unity-mir [09:46] + the pathc [09:46] and yeah, no crash [09:46] I'm cherry-picking directly in distro [09:46] \o/ [09:46] and get the changelog fixed [09:46] let's do it quickly [09:46] didrocks, ok then [09:46] nice work greyback! [09:46] * greyback takes rest of the day off [09:46] Saviq: but clearly saw we need trunk linked to delivery in ubuntu :) [09:46] greyback: can I as well? :) [09:46] we have -proposed stuck for another fix :( [09:47] didrocks: ask Saviq, he's always letting me take days off ;) [09:47] didrocks, sure, go for it [09:47] greyback: :D [09:47] cool to have the fix so quickly [09:48] didrocks, greyback so anyway we're still blocked for merging into lp:unity-mir due to FTBFS against new mir (and it not being published, for that matter) [09:48] you could merge the patch plus changelog manually and use bzr commit --author, while waiting for Mir [09:48] Saviq: new Mir being 0.1.2? I'd better fix that FTBFS then [09:49] greyback, yes [09:49] Saviq: feel free to merge manually [09:49] greyback, https://bugs.launchpad.net/unity-mir/+bug/1254986 [09:49] Ubuntu bug 1254986 in unity-mir "unity-mir FTBFS against libmirserver11" [Critical,New] [09:49] Saviq: it's a build-dep issue? [09:49] ah, something else [09:49] Saviq: is it available in a PPA somewhere? I'm not sure I trust mir-staging [09:49] ok [09:49] I'll let you guys figure it out :) [09:50] Saviq: in bug, nvr mind [09:50] greyback: ppa:ubuntu-unity/daily-build [09:50] Mirv: so I saw, thanks for that. I always forget that ppa [09:51] greyback: I added the changelog diff to your MP [09:51] greyback: Saviq: uploaded to unblock the image, thanks guys! [09:51] didrocks: thank you [09:51] didrocks, uh, already kicked generic-land [09:52] didrocks, care to push the changelog yourself / do an MP? [09:52] Saviq: no worry, doing it [09:53] didrocks: done [09:53] greyback: Saviq: https://code.launchpad.net/~didrocks/unity-mir/resync-changelog/+merge/196682 [09:54] uhh, to late [09:54] didrocks, greyback landing [09:54] land whatever you want, just need the changelog to be in sync :) [09:55] tsdgeos, so we were trigger-happy with countChanged were we [09:55] a bit too soon i'd say [09:56] tsdgeos, would approve, if not for bug #1254898 :/ [09:56] bug 1254898 in Unity 8 "tst_Shell::test_DashShown is unreliable" [Critical,Triaged] https://launchpad.net/bugs/1254898 [09:59] Saviq: hmm, how common is that? and what does this have to with the other thing? [09:59] or shall we not merge anything until we fix that one? [10:00] tsdgeos, it's not that we *shall* [10:00] tsdgeos, https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-trusty/ [10:00] ok [10:01] tsdgeos, it's just blocking us - think I found the issue, though [10:01] lots of yellow [10:01] haven't been paying attention lately it seems [10:06] greyback: hey, I guess the ApplicationManager has the same issue: https://code.launchpad.net/~aacid/unity8/queuedModelCountChanged/+merge/196680 [10:13] mzanetti: whoa, that's interesting. [10:13] greyback: yeah... tsdgeos discovered it [10:13] greyback: it's not always bad [10:13] but it was happening to me [10:14] that someone got the countChanged Signal before a third party (a repeater) was getting the insertedRows [10:14] yep. I think we're not running into this with applicationmanager yet as we don't do fancy stuff onCountChanged [10:14] and so when the countChanged handler wanted to do stuff on the repeater it was weirded out by the fact that the row was still not there [10:15] and i really think it makes sense all rowInserted are processed before any count changed is invoked [10:15] tsdgeos: but... shouldnt that stuff happen on Repeater.countChanged instead of mode.countChanged? [10:15] model [10:15] mzanetti: well it's the decoulping problem [10:16] a item is responsible for something and a different item fro something else [10:16] ah... right... we have that thing in the indicators [10:16] mzanetti: yes in an ideal world, yeah, and it was my first idea of fixing it [10:16] mzanetti: but then i did think about it and if we were not chaining the signals, we'd do "endInsertSignals(); emit countChanged();" [10:17] and that means [10:17] first all rowsInserted and them all countChanged [10:17] so i thought we ought to mimic that behaviour [10:18] Cimi: ping [10:21] that time of the year again :D http://ubuntuone.com/2Cf5YYEfHQ96YgIdGFyVAq [10:22] you mean that time of the decade ^_^ [10:22] tsdgeos, no, that's not radioactive fallout ;) [10:26] tsdgeos, pong [10:26] tsdgeos, ouch, your queued connections break some tests [10:26] tsdgeos, https://jenkins.qa.ubuntu.com/job/unity8-trusty-i386-ci/262/console [10:27] darg [10:27] Cimi: i am not sure i see the problem you were saying yesterday [10:27] tsdgeos, what? [10:27] Cimi: you said "open preview in carousel, switch left/right to other items" [10:27] no? [10:27] tsdgeos, y [10:27] and what should i see? [10:27] tsdgeos, look at the bottom [10:28] i am [10:28] tsdgeos, the small carousel [10:28] i even put my finger in the screen :D [10:28] seems like stuff is in the same place to me [10:28] mmmm [10:28] tsdgeos, it moves up at the first swipe [10:28] ok, i'll put my finger again [10:29] tsdgeos, it's clear, especially in tablet mode [10:30] Cimi: ok, yes, it does [10:30] big finger :D [10:34] Saviq: yeah that's rather unfortunate :-/ [10:34] tsdgeos, indeed [10:34] i can change the qcompare [10:34] to qtrycompare [10:34] and then it kind of works [10:35] but it highlights the need of a eventloop [10:35] sil2100, how are we looking on unity-scopes-api? [10:36] Saviq: so maybe i should just discard this change and as mzanetti says make the other thing able to cope with countChanged possibly happening before all the rowInserted have been processed [10:36] tsdgeos: I didn't say that :) [10:37] mzanetti: you didn't say to discard it, you did say to make the other stuff work, no? [10:37] tsdgeos: I said the onCountChanged in the view should happen onCountChanged in the view instead of the onCountChanged of the model [10:38] ok ;-) [10:40] tsdgeos, yeah, it did feel slightly funky indeed [10:41] ok, let's discard it for the moment [10:41] and let's try to make it work [10:41] everyone keep in mind that countchanged in the model may trigger before the view is updated [10:41] and that's it :D [10:42] let's see if i can do that in the code i'm trying to fix [10:43] mzanetti, https://code.launchpad.net/~saviq/unity8/fix-dashshown-test/+merge/196690 [10:43] mzanetti, it's http://s-jenkins:8080/job/unity-phablet-qmluitests-trusty/422/ minus debug logging [10:44] so the first green run since yesterday http://s-jenkins:8080/job/unity-phablet-qmluitests-trusty/ [10:46] /food, thus [10:53] tsdgeos: Saviq: FYI Qt 5.2 status update, with sensors rebuilds I got rid of the previous linker error. now it has a linker error about UI Toolkit, which FTBFS:s (test failure) and I've pinged sdk team about that bug [10:53] tx! [10:54] Mirv, cheers! [11:04] Mirv: Saviq: https://code.launchpad.net/~gerboland/unity-mir/0.1.2-supprt/+merge/196693 fixes Mir 0.1.2 [11:06] didrocks, ↑ what do we do? are we publishing new Mir first and letting that through or are we testing locally and forcing it in? [11:07] Saviq: as long as it's tested in a good variety of hardware, I have no issue with you doing a manual MP [11:07] Saviq: you clearly need feature ticket ;) [11:08] didrocks, airline CI there yet? :P I don't want to do manual, so asking what's the proper course here [11:08] Saviq: if only ;) [11:08] didrocks, I'd rather wait for mir to be released [11:08] Saviq: yeah, sounds fine [11:09] k [11:10] but one can't build the trunk in PPA without the fix? previously I've merged manually, build mir + platform-api + unity-mir + unity-system-compositor in the PPA, tested them and released at the same time [11:15] mhr3: I sent it out for preNEW review, but for now nothing [11:16] Argh, I meant: [11:17] mhr3: I sent it out for preNEW reveiw yesterday, but for now there is no info regarding how it went [11:27] mzanetti, in that branch, we should probably check if clicking on items it hides indicators? https://code.launchpad.net/~nick-dedekind/unity8/lp1238182/+merge/192965 [11:28] Cimi: yeah... that test seems a bit minimalistic [11:33] Cimi: so yeah basically is the "up" animation that messes things up [11:33] tsdgeos, up animation? [11:33] Cimi: for first show animation is still not there so Y is 158.237 and for the second is already there so it's 170.038 [11:34] Cimi: you know that thing that makes the item "grow" or whatever in the carouse [11:34] l [11:34] which tbh i don't know why it's happening [11:34] since it's the same item we had selected [11:34] am i making any sense? [11:34] tsdgeos, it's just a scale [11:34] sure [11:34] tsdgeos, it should not affect [11:35] it does [11:35] tsdgeos, scales don't affect coordinates [11:35] afaics [11:35] they don't affect coordinates [11:35] they do affect mapToScene though [11:36] hmmm [11:36] or maybe not [11:36] let me make sure [11:39] tsdgeos, you can disable the scale and see [11:39] tsdgeos, I think I did though [11:43] ok [11:52] "FAIL! : qmltestrunner::Shell::test_DashShown(in focus) Uncaught exception: Cannot read property 'width' of undefined" [11:52] does anybody knows what's that about ↑ [11:52] * dandrader is now using the arrow char as well :) [11:56] Cimi: http://paste.ubuntu.com/6478600/ does defenitely make it go away [11:56] dandrader: yes, Saviq is merging a fix for it [11:56] tsdgeos, cool. thanks [11:57] Cimi: can you confirm? [12:03] Saviq, you around? === _salem is now known as salem_ [12:11] mhr3: https://code.launchpad.net/~sil2100/unity-scopes-api/package_name/+merge/196700 <- with this in, we're ready to go with enabling in cu2d! [12:15] sil2100, thx, approved [12:16] sil2100, mhr3: +1 [12:24] nic-doffay, here [12:25] Saviq, re this: https://code.launchpad.net/~nicolas-doffay/unity8/category-transition-speed-fix/+merge/195203 [12:25] Not quite sure how to get around the speed issue without other issues occurring with this approach. === MacSlow_ is now known as MacSlow|lunch === MacSlow is now known as MacSlow|lunch [12:31] dandrader: saviq stuff merged [12:31] tsdgeos, dandrader I'll take care of restarting merges [12:32] nic-doffay, why would the speed different, when you have the same amount of travel and same duration? [12:32] Saviq, thanks. and don't forget about finishing the review of https://code.launchpad.net/~dandrader/unity8/runningAppsEndClose/+merge/196257 ;) [12:33] Saviq, that's the thing the travel isn't always the same. I had to add some clauses due to other bugs that occur. [12:33] dandrader, I never started ;P [12:33] Which you mentioned in a previous comment. [12:33] nic-doffay, why isn't it the same? [12:33] nic-doffay, it should be [12:33] nic-doffay, will have another look soon [12:34] Saviq, if you set the speed to the scope height then reset it to the collapsed/uncollapsed height as you suggested you can see a noticeable change. [12:34] Basically like a bounce. [12:39] tsdgeos, yes [12:39] tsdgeos, so what we do here is unknown :) [12:40] well now that we know what happens we can try to find out how to fix/workaround it [12:40] i'll have a look after lunch [12:49] tsdgeos, yeah I think that when it changes currentIndex position etc is screwed up maybe because it's still animating [12:49] may be, now lunch! [12:57] tsdgeos, don't answer if you're eating, that's how I do :P [13:08] Saviq, do we need something in the json card templates that specifies how many lines of text should be reserved to the title? [13:09] Saviq, otherwise apps cards would look like music ones (minus the subtitle) [13:09] Saviq, plus how are we going to specify the different card sizes? will the json include size in grid units or something? [13:09] mhr3, this might have changed, but last I checked it was "up to two lines for title+subtitle" [13:10] mhr3, so if there's no subtitle - title can span two lines, but if there's subtitle - title just one line [13:10] Saviq, so current music renderer can't be done using the templates? [13:10] music-grid that is [13:10] mhr3, meaning 2 lines for title + 1 line subtitle? [13:10] as it has two lines + subtitle [13:11] right [13:11] mhr3, yeah that was the plan last I read it, but I vaguely remember that changing [13:11] mhr3, either way, no, I don't think we should do so much for specifying it [13:12] Saviq, what about the card sizes? [13:12] mhr3, it should adapt dynamically [13:12] mhr3, card sizes [13:12] mhr3, we definitely need small/medium/large to be in the json [13:12] mhr3, which translate to GUs [13:13] mhr3, we shouldn't need to allow arbitrary sizes, but that's also something that wasn't clear in the spec [13:13] i'm just wondering whether that'll be enough [13:13] at least looking at the visual spec and what have in the jsons... might be missing stuff === alan_g is now known as alan_g|lunch [13:14] Saviq, what shall I do with this? https://code.launchpad.net/~cimi/unity8/fix-1214423/+merge/192868 [13:14] mhr3, oh yeah jsons are not complete for sure [13:14] Saviq, we have a WIP doc here https://docs.google.com/a/canonical.com/document/d/1Ded64oMdX10F5vYPoarUVle-J7qLs9-ilyPtoIZoUwU/edit [13:15] Cimi, reply to the comments at least ;) [13:15] Saviq, basically I put myself on hold on that [13:15] sdk or not? [13:15] Saviq, is someone already working on creating a renderer out of the json? [13:16] Cimi, if it's going to the SDK, probably doesn't make sense to get it into unity8 first [13:16] Saviq, cause i'm not if you're expecting me to :) [13:16] Saviq, exactly [13:16] mhr3, I'm not (expecting you to) [13:16] mhr3, and I am, when I can [13:17] Saviq, think we should sync up on what we expect to have from each other for next week ;) === MacSlow|lunch is now known as MacSlow [13:18] mhr3, I expect to have a first iteration of the card renderer that will read the json and adapt accordingly [13:18] mhr3, stand alone, tested [13:19] mhr3, so no dependency on you [13:19] good, i should have things prepared to hook it up to real scopes [13:19] then we just need real scopes :) [13:21] https://bugs.launchpad.net/unity8/+bug/1253067 [13:21] Ubuntu bug 1253067 in Unity 8 "Slow and unsable with atom n270" [Undecided,New] [13:21] really that bad??? [13:22] Cimi, icanhascommentrightsonthatdoc? kthxbai [13:22] Saviq, tak [13:22] Cimi, I'd say that bug is just something not working right, not actual issues [13:23] yeah [13:23] Saviq, I thought it would have been faster than compiz [13:23] Saviq, however, all the ubuntu shapes are consuming shaders, right? [13:24] Cimi, as does most of QML itself anyway [13:24] ok [13:24] Cimi, it's not a magic word, or a magic resource hog ;) [13:24] Cimi, and compiz/unity7 is made out of shaders just as well [13:24] Saviq, well, intel 915 has limited shader units [13:25] not sure which one is more graphically intense on this regard [13:25] Cimi, lower than Galaxy Nexus? doubt it ;) [13:25] Saviq, unity7 on the intel 915 has fallback mode [13:25] Saviq, no blur iirc [13:26] Saviq, to use less shaders [13:26] like maybe not gaussian but mipmap or not blur at all [13:27] galaxy nexus might have more updated OpenGL specs [13:27] specs/support [13:29] just let i915 die [13:29] it had a good run [13:29] but we have real gpus now [13:35] mhr3, so we should let the galaxy nexus and nexus 4 die as well? [13:35] we have so much performance work to do... [13:35] I'd stop working on unity new features just to improve performance [13:39] Cimi, i didn't say anything about galaxy nexus, i'm talking about i915 which is super old gpu that doesn't even do opengl es === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [14:03] tsdgeos, re the countchanged branch, afaict you'd have to compress those events to be safe [14:03] aaaah I got a typing machine in my phone! [14:04] mhr3: hmmm?¿ [14:05] tsdgeos, oh wait it doesn't include the actual count === alan_g|lunch is now known as alan_g [14:05] still, the qt guys had a good reason not to expose a count property on the models :) [14:05] mhr3: actually they did [14:06] just in the qml models though [14:06] because they never thought people would want to share the same interface with something that can be either a qml model or a c++ one [14:06] ... [14:06] ... [14:06] would be nice if we could get rid of them too then [14:06] now that's what I like to see http://s-jenkins:8080/job/unity-phablet-qmluitests-trusty/ [14:07] mhr3: ? what issue do you have with count? [14:07] mhr3, there's rowCount and columnCount on the models all the same? [14:08] tsdgeos, it's just weird to have it a notifiable property [14:08] why? [14:08] as a* [14:08] cause count changes when rows dis-/appear [14:08] a separate signal is just asking for all the trouble you're expriencing [14:09] not really [14:09] my problem is just because our implementation sucks [14:09] not because anything else [14:11] but it sucks cause there's no correct way to do it [14:11] right, i just need to change the Qt code to do it right unfortunately [14:12] MacSlow: is the fullscreen notifications ready for review? [14:14] mzanetti, yup [14:14] mzanetti, oh... let me push it... :) [14:14] MacSlow: ok... there's a ton of conflicts. can you merge those? [14:14] mzanetti, yeah... that's solved... [14:15] nice [14:17] mzanetti, r493 pushed [14:17] mzanetti, if you want to test it with the fullscreen simunlock, plain lp:unity-notifications will do [14:18] mzanetti, regarding the stand-alone example I mean [14:23] MacSlow: I'll run it on the device with the real sim pin [14:23] mzanetti, you're lucky you have a locked sim :) [14:24] MacSlow: ? [14:24] mzanetti, I only have an unlocked sim, which fits into my GN [14:24] MacSlow, you can lock it ;) [14:24] … [14:25] MacSlow: and: you could even make the other fit in there :D [14:25] Saviq, nope... two different form-factors on the simcards/slots and I don't have an adapter [14:25] * mzanetti has a bunch of self made adapter for all his self cut sims [14:25] MacSlow, excuses, excuses [14:25] mzanetti, hm... too fiddly for my taste [14:26] MacSlow: and: you can use qdbus to lock the sim with Ubuntu Touch [14:26] Saviq, I just don't want to mess things up :) [14:26] MacSlow, that's the only way to have things to fix, though! [14:26] anyways. you can also fiddle with all sorts of python stuff to mock it if you find that less fiddlier :) [14:27] mzanetti, well I have done that already so :) [14:27] * Saviq is pretty interested in how will www.knowroaming.com pan out [14:27] mzanetti, what's the qdbus way to lock it [14:28] MacSlow: something like "qdbus --system org.ofono /ril_0 or.ofono.SimManager.LockSim" [14:30] mterry, notes [14:30] ;) [14:33] mterry, you got off the hook ;) [14:33] Saviq, :) [14:53] Saviq, ah... just realoaded the page again (https://code.launchpad.net/~macslow/unity8/notification-fullscreen-support) and the conflicts are gone now... phew [14:53] MacSlow, cheers [14:53] Saviq, LP was a bit slow then today [14:54] Cimi: maybe we can just cheat... [14:54] Cimi, I was talking about this bug: https://bugs.launchpad.net/unity8/+bug/1213956 [14:54] Ubuntu bug 1213956 in unity8 (Ubuntu) "drag handle for an open indicator is not reliable" [Undecided,Triaged] [14:55] tsdgeos, that's why I wanted the rendererName :P [14:55]