=== JanC_ is now known as JanC [08:46] greyback_: https://code.launchpad.net/~aacid/qtmir/no_double_search/+merge/272707 ? [08:48] mzanetti: https://code.launchpad.net/~aacid/unity8/stabilize-launcher-test-more/+merge/272708 [08:48] tsdgeos, thanks [08:48] would've never found this without being able to repro :/ [08:51] yeah, repro is nice in these cases [08:52] well mostly always :D [08:53] tsdgeos: nice find, thanks [08:56] tsdgeos, did you intentionally set my branch as target? [08:58] mzanetti: i did intentionally set your branch as prerequisite, and failed :D [08:58] fixing [08:58] ok :) [09:01] mzanetti: now https://code.launchpad.net/~aacid/unity8/stabilize-launcher-test-more/+merge/272709 === marcusto_ is now known as marcustomlinson [09:39] tsdgeos, looks like your clazy run MP reduced our memory usage significantly enough that Victor could not get webbrowser to get OOM-killed on arale ;) [09:40] lol [09:40] really? [09:40] wat [09:40] he struggled to get it killed indeed [09:40] and nothing else in the list looks to me like could have any impact https://requests.ci-train.ubuntu.com/#/ticket/410 [09:41] maybe all the QStringLiterals? [09:41] well, arale is also not the best example to try to repro that [09:41] that'd be dumb seeing what QSL is [09:42] well [09:42] mzanetti, yeah, but he went: silo → can't get it killed; no silo → killed; silo → can't get it killed [09:42] the nice thing about it [09:42] is not that it saves memory [09:42] it's that it doesn't fragment it [09:42] so it saves much more memory than what it really saves [09:42] since doesn't need to allocate/deallocate it [09:42] all the time [09:43] right [09:43] yeah, I would have expected it to have a bigger impact on perf than on actual usage [09:43] but glad to hear it also gets us better there === marcusto_ is now known as marcustomlinson [09:54] Saviq, hey, I got a svg icon for the dash from design (to be used in unity8-dash.desktop), do you have any recommendations where to put it? qml/graphics/applicationIcons? [10:14] pstolowski, data/ [10:19] hmm can't access launchpad and other Ubuntu servers... [10:20] anyone else have problems? [10:21] Saviq, work here [10:21] works here too [10:21] tx [10:21] i have problems accessing the yahoo servers though :D [10:22] so maybe the internet is breaking little by little [10:22] might be [10:22] yeah it's back again for me [10:22] and yahoo is back here too :D [10:22] weird [10:22] I am struggling too [10:23] me too [10:23] launchpad, IRC... [10:23] google stuff seems ok [10:38] ltinkl, hey, greyback pointed me at https://launchpad.net/libqtdbustest that could be useful for the dbus mocking (although in the case of logind it's probably easier to just use the template provided by dbusmock) [10:43] Saviq, hi, thanks for the ref, I can have a look if it provides something better [10:51] kgunn: dpm tells me you had fun with xorg recently - a bad update? (I'm getting xorg exploding randomly on my intel laptop) wondered if there was a bug to follow or some other detail? [10:52] * tsdgeos got a xorg update in vivid overlay [10:52] now i'm scared to reboot :D [11:12] tsdgeos: I'm on wily :) [11:12] ok [11:25] unity8 requires webbrowser nowadays? [11:44] Mirv, we need the web view [11:45] Mirv, to display the HERE terms === alan_g is now known as alan_g|lunch [12:11] Saviq: ok. I just noticed it since a) Oxide just (re)gained dependency on private Qt ABI, needing a rebuild, b) Oxide fails to rebuild against Qt 5.5, c) webbrowser needs oxide d) unity8 needs webbrowser [12:11] sorries :) [12:11] this after I was finally able to hack UITK far enough to get packages out and I was rejoicing at the possibility of rebuilding webbrowser + unity8 to be able to run Qt 5.5 on the phone for the first time since July :) [12:12] or was it June [12:12] Saviq: no problem, life is just complicated [12:12] Mirv, did you see mterry's questions yesterday about an upstream LP project for Qt? [12:12] to be able to assign upstream bugs to it [12:13] Saviq: yes, I replied to his e-mail about this another patch and said to do whatever he thinks would help. I wasn't aware of anything that could be done aside from coding Qt bug report support into LP [12:13] ack [12:14] @unity please merge trunk in your branches, most of the MPs had conflicts with new trunk [12:14] ack [12:15] Saviq, I didn't see that, but it's basically bug #157488? [12:15] bug 157488 in Launchpad itself "Add bugwatch support for the JIRA bugtracker" [High,Triaged] https://launchpad.net/bugs/157488 [12:15] now that launchpad has some more hackers maybe they can do that? ;-) === alan_g|lunch is now known as alan_g [12:32] Saviq, which is indicator-bluetooth not working on wily? (just saw that in the unity8 changes) [12:35] seb128, you can't enable it [12:35] seb128, I thought it was a known issue that we've not ported to BlueZ 5 yet... [12:35] Saviq, it should work, it got ported to bluez5 and works on unity7, do we have a bug? [12:36] Saviq, you though we would release ubuntu desktop without working bluetooth? ;-) [12:36] seb128, not desktop, but wily phone is quite ignored [12:36] oh, that commit doesn't state phone [12:36] what about unity8 desktop session? [12:37] though in theory it should work on the phone, I'm unsure it got tested and about the kernel side though... [12:39] seb128, indeed we didn't pay enough attention to desktop, sorry about that, I just thought this is an expected state at this point in time, let me verify and file a bug [12:39] Saviq, thanks [12:39] I'm going to update my test machine and try there as well [12:39] it should work, but if it doesn't it's good to know and we should fix it ;-) [12:50] Saviq: so this landing means the tests passed in that new test-y thing we added right? [12:50] autopkg tests i mean [12:51] tsdgeos, at the moment they're "always failed" because of the missing logind mock [12:51] oh right [12:51] and armhf looks broken in terms of dependencies, so need to look into that too [12:53] seb128, can't get a unity8 desktop session, just get a black screen and broken vtswitching :/ [12:55] Saviq: arhmf on CI or when building? [12:55] tsdgeos, autopkgtest [12:55] ah [12:55] which is != than CI, because it installs packages from -proposed, doesn't build the whole thing [12:56] right [12:57] seb128, bug #1500855 [12:57] bug 1500855 in indicator-bluetooth (Ubuntu) "Bluetooth can't be enabled on devel-proposed phone" [Undecided,New] https://launchpad.net/bugs/1500855 [12:57] tsdgeos, but we should be getting green CI now [12:58] let's see [12:58] i proposed another clazy run with 3 small fixes [12:58] should be green [12:58] Saviq, thanks [12:59] Saviq, is the desktop session not working an unity8 regression? [13:00] seb128, doubt it (the cursor disappears, too, which suggests unity-system-compositor), will need investigation [13:00] OTOH I can see an ibus and unity8 crash [13:00] let me see if it does that here [13:01] unity8 segfault doesn't sound too good [13:01] * Saviq tries again with a clean /var/crash [13:08] seb128, no unity8 crash this time, but ibus yes, stacktracetop mentions mirclient [13:11] seb128, here's the ibus oops https://errors.ubuntu.com/oops/79161cec-66ab-11e5-8bc8-fa163e22e467 [13:12] Saviq, thanks [13:24] seb128, quick glance suggests it's similar to bug #1439202 [13:24] bug 1439202 in ibus (Ubuntu) "/usr/lib/ibus/ibus-ui-gtk3:11:XKeysymToKeycode:keybinding_manager_bind:panel_keybinding_manager_bind:panel_bind_switch_shortcut:panel_construct" [High,Confirmed] https://launchpad.net/bugs/1439202 [13:25] Saviq, yeah, not likely the unity8 session issue [13:25] seb128, indeed === pat_ is now known as Guest41182 [13:37] dandrader: greyback_: since the clazy fixes for unity8 seem to have been so good i've done one for qtubuntu too https://code.launchpad.net/~aacid/qtubuntu/clazy_fixes/+merge/272747 [13:38] ok [13:59] ltinkl, dandrader's mousePointer branch already bumps application api to 9, you might wanna rebase on that [14:00] Saviq, yikes, noted; I feared that :) === davidcalle_ is now known as davidcalle [14:03] * ltinkl on it [14:04] Saviq, on my wily test laptop unity8 desktop session works and indicator-bluetooth loads and shows devices and has working items [14:08] seb128, glad to hear [14:13] tsdgeos: thank you [14:19] hey, seb128 sounds familiar [14:20] seb128: I think you reclassified some bug I looked at, gvfs breaking magnet links [14:21] dandrader, did you merge qtmir as well? [14:21] Saviq, qtmir/mousePointer? [14:21] dandrader, yeah [14:21] debian/control has a conflict there too [14:21] Saviq, not today [14:21] please do [14:23] Saviq, done [14:23] tx [14:38] * guest42315 ubuntuonair in 20 min http://ubuntuonair.com/ [14:38] biot, yes, I did === dandrader is now known as dandrader|afk [15:15] ouch [15:15] we had a function in qmenumodel not returning anything [15:15] when it should be returning a bool [15:20] dednick: https://code.launchpad.net/~aacid/qmenumodel/clazy_run/+merge/272788 [15:25] greyback, heyo -- we talked about X-Ubuntu-Touch the other day -- assuming that's the road we go down, is this MP roughly acceptable? https://code.launchpad.net/~mterry/qtmir/no-touch-no-lifecycle/+merge/272791 [15:26] tsdgeos, how is that not a compile error? :) stupid lenient compile modes [15:26] mterry: kinda. First thing I see if you're implementing policy in qtmir - I'd rather you export the "isTouchApp" as a bool to the shell, so it can decide what to do [15:26] mterry: yaeh :/ [15:27] mterry: there is no second thing :) [15:27] greyback, it looked like that policy decision was already in qtmir? [15:27] tsdgeos: looks fine. suppose i should test it... :) [15:27] greyback, but I can do the round trip way -- that's something that mzanetti needs eventually anyway to show a different close button [15:28] tsdgeos: what's with the reserve? [15:28] greyback, if I'm doing that though, I feel like we ought to pull the whitelist out of qtmir [15:28] mterry: well it is the shell which decides if an app should be suspended or not. [15:28] tsdgeos: just dynamic alloc? [15:28] mterry: yep, that I agree with [15:29] dednick: so makes sure the list is the szie we need from start instead of growing as it "fills" and needing to realloc and stuff [15:29] dednick: it's minor optimization [15:29] greyback, right -- that's why I stuck it in qtmir. Because it was doing the policy via whitelist already. But sure, I'll move all out [15:29] since we know the size it doesn't harm to reserve [15:29] mterry: cool, thanks [15:29] tsdgeos: mkay [15:30] figured [15:31] dednick: "replaceme" in a debian .symbols file does what I think it does? [15:32] greyback: i think the ci bot comes along and changes it when released [15:32] cool [15:32] greyback: looked at history and seems to be what happens :| [15:32] I didn't know that [15:33] greyback: i've requested a review from ricmm, who should know the details [15:34] ack. he's a busy bunny tho [15:34] dednick: this Holder thingy is weird [15:34] greyback: yeah. doesn't do anything that i can see [15:35] was it supposed to be a QScopedPointer/unique_ptr [15:35] but it's got no deleter afaics [15:36] greyback: the other sensors are all handled by android registration keeping shared pointers. so they aren't destroyed till the sensor controller is. the haptic is a special case. [15:36] larsu: yeah it's ugly, but saves lots of allocs [15:36] that's why it's cool having a tool that does the replacement for you [15:36] we can code and then just run the tool :D [15:36] dednick: I see [15:36] greyback: i'm thinking it probably should be shared, but it isn't here. [15:36] tsdgeos: ya I just read about it, pretty neat [15:37] dont see a point of 2 things asking for it creating 2 dbus connections. [15:37] dednick: to be absolutely correct, maybe yeah. Let's see what a papi maintainer says [15:38] I'm hoping every button doesnt do that :) [15:38] i think the Qt sensor may be shared [15:38] yeah, but qtsensors does that itself [15:39] IMO papi should be the one dedupicating such resources [15:39] +l === dandrader|afk is now known as dandrader [15:56] dednick, about your polit-close branch: ApplicationManager test is still failing [15:56] dandrader: hm weird. give me a minute [15:57] dandrader: passes for me :/ [15:58] dednick, and item 2 here still applies https://code.launchpad.net/~nick-dedekind/qtmir/polite-close/+merge/262188/comments/683011 [16:01] dandrader: ok. fixing in a minute [16:01] greyback, could you check if you get the test failure I describe here? https://code.launchpad.net/~nick-dedekind/qtmir/polite-close/+merge/262188/comments/683011 [16:03] dandrader: you on mir 0.16 [16:03] ? [16:04] dednick, yes [16:05] dednick, but it's a test, shoulnd't make a difference [16:05] well i just upgraded things have just gone shitty [16:06] Application test wont run either. [16:06] dandrader: ^ [16:07] dednick, great. so you're getting the same test failure now [16:07] dandrader: no. Application, not ApplicationManager [16:08] dandrader: ApplicationTests.checkResumeAcquiresWakeLock [16:08] dednick, that one passes here :) [16:09] those wake lock tests have always been flaky for me [16:12] dandrader: tests passed on jenkins wily a couple of days ago. [16:12] dandrader: i'll kick off another run [16:15] dandrader: deleted all my build files and rebuilt. all tests pass again [16:35] Saviq, could you check if you get the same test failure I describe here? https://code.launchpad.net/~nick-dedekind/qtmir/polite-close/+merge/262188/comments/683011 [16:36] Saviq, I'm getting it but dednick is not === alan_g is now known as alan_g|EOD [16:58] dandrader, wouldn't ci show a test failure? [17:02] Saviq, right, and it's passing there. damn it [17:02] dandrader, in any case, all passed here as well [17:03] I'm on wily, mind you [17:07] me too [17:07] I do out of source builds. But I would be really surpised if that's the cause === dandrader is now known as dandrader|afk [18:29] mzanetti, I assume you're not around, but if you are, I updated my no-touch-no-lifecycle branch to have an isTouchApp bool you can consume in unity8 [18:29] https://code.launchpad.net/~mterry/unity-api/no-touch-no-lifecycle/+merge/272829 [18:29] https://code.launchpad.net/~mterry/qtmir/no-touch-no-lifecycle/+merge/272791 [18:29] mzanetti, I'm working on the unity8 side, but it may already be useful to you [18:30] mterry, cool! thanks. not working on it any more today, but will check it out tomorrow morning [18:30] mzanetti, good answer :) === dandrader|afk is now known as dandrader [20:01] greyback, still around? [20:02] dandrader: for brief questions, yes [20:02] greyback, why not having polite-close as a pre-requisite of https://code.launchpad.net/~gerboland/qtmir/dont-delete-qml-cache-on-good-stop/+merge/272761 ? [20:02] greyback, as you said you took a bunch of code from it [20:02] dandrader: I will tomorrow. I didn't realise polite-close was going to be approved so fast [20:03] rebase should be easy === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [20:50] josharenson, https://plus.google.com/105839534016416729197/posts/BSnrzs65r7i [20:50] there's some feedback in the comments [20:51] mzanetti: ah good to know... I'll merge trunk into the ppa branch and try it on wily (just finished upgrading to wily about 10 min ago) === Cimi_ is now known as cimi