=== shuduo-afk is now known as shuduo [09:57] dednick, good morning [09:58] dednick, I think the shell_chrome branches have quite a lot of failing tests still [10:00] mzanetti: i'll take a look. [10:02] mzanetti: have you got a link for the test run? [10:07] dednick, check your email [10:07] (have just been in a meeting with like every manager, asking me when silo 41 lands and then the color palette fixes) [10:14] mzanetti: oobe doesnt seem to even be reviewed yet... [10:15] dednick, it is [10:16] dednick, just needs the final stamp but mike and me reviewed it already and we tested it like a 100 times [10:18] pstolowski, could you put the social card branches on top of the filters one? I guess the filters will land first so it kinda make sense [10:18] cimi, ok [10:32] mzanetti: i'm a bit scared you mentioned silo 41 is all or nothing, it'd seem to me filters could land even if the rest wasn't ready, no? [10:33] tsdgeos, we have to get all the things ready... it works fine now. it's "just" about fixing the tests [10:33] we really want the side stage stuff, and the oobe too... [10:34] if anything, I guess the shell chrome could be dropped for now... [10:34] sidestage broke a few tests. recursive touchPress function [10:34] dednick, yeah... are they just to be fixed or is there a bigger issue with it? [10:35] mzanetti: no, it's easy. i just called the wrong function within the UnityTestCase impl. [10:35] ok, cool [10:35] when i added multiTouchPress [10:35] fwiw, things passed in my manual testing on phone and tablet now... [10:36] desktop looked mostly ok too, need to check out some details still [10:37] I'm fixing some more color palette issues, please let me know when something is to be rebuilt and tests ran on it [10:39] hmmm... I have the greenish background now in the panel too [10:41] cool [10:41] i've never seen that [10:42] * ltinkl neither [10:53] mzanetti: do you know if make tryNotifications is how the notifications really look on the phone? if so probably need some fixing since it's almost black on black [10:54] tsdgeos, got a branch on that [10:54] ah cool [10:54] tsdgeos, https://code.launchpad.net/~lukas-kde/unity8/fixNotificationsColorWithNewUitk/+merge/288869 [10:55] tsdgeos, I even added an "Inverse mode" checkbox to the controls so that you can really see how they look like [10:55] cool, daniel is taking care of that branch, right? [10:56] mzanetti: are we including ↑↑↑ on the landing for ota10? [10:56] tsdgeos, yeah, mterry mostly [10:58] tsdgeos, the fixes should be done in a silo following silo 41, still for OTA10 [11:01] tsdgeos, the color fixes? yes [11:01] tsdgeos, yeah, prepping a silo with that atm [11:01] it's in 27 [11:09] ah so we'll still be landing another silo [11:09] just 41 is FFE [11:09] got it [11:21] cimi, ok, I think we've fixed all unity related things from this bug: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1554616 [11:21] Launchpad bug 1554616 in unity8 (Ubuntu) "Shell and dash visual issues with new UITK and palette" [Critical,In progress] [11:21] cimi, any findings on the green panel? [11:21] mzanetti, haven't started yet - is critical enough to stop doing code reviews and move to that? [11:22] depends on the code reviews. if they are for somthing in silo 41, then no, otherwise yes [11:22] also do we have a proper bug report? [11:22] cimi, so i merged filters changes in unity-api & shell plugin in silo 71. you will need to do the same for unity8 if you want this silo in a working state [11:22] pstolowski, gotcha [11:23] mzanetti, ok I will stop code review, was for something else [11:23] ok [11:23] mzanetti, do we have a bug report? [11:25] cimi, do't think so [11:25] cimi, but I count that into the palette issues [11:26] as it seems to have started breaking with that [11:31] mzanetti, so I'll test it with silo 41 [11:31] cimi, the color issues? [11:31] yes indeed [11:32] cimi, use silo 27 rather [11:34] ltinkl, nick had some comment about the touched files not having 2016 in the copyright. please update that. I'm about to approve the branch [11:34] it looks ok and tests well [11:34] mzanetti, ok, will update the (c) [11:39] mzanetti, done [11:40] ta [11:45] afternoon Saviq. Quick question: would it be possible to get an up-to-date .pot file for the unity8 and indicator-session in silo 41? The reason I'm asking is because bq are preparing a translated manual, and they are waiting for translations for new strings in that silo. Without them being in LP, it makes it a bit difficult to do them, so I've thought I could manually upload the new .pot files to Launchpad and thus we don't need to block on the landing [11:45] to do the translations [11:46] really.... [11:46] a criss-cross in side stage [11:46] dednick ^ [11:47] mzanetti: eh. i'm updating sidestage now. [11:47] mzanetti: oh. hm. sidestage. [11:47] dpm, Saviq is out [11:47] i'm fixing chrome now... [11:47] dpm, I'll take care of it [11:47] one sec [11:48] mzanetti, awesome, thanks!. All I need is the resulting .pot files for the unity8 and indicator-session builds in the branches of silo41 [11:48] dednick, let me know when you go back to the side stage branch so I can explain how to do it, and what you do wrong so you get criss-cross merges all the time [11:49] ltinkl, can you update the .pot file in your indicator-session branch [11:49] mzanetti: i've just pushed [11:49] mzanetti, ah right, ok [11:50] mzanetti: i didnt do anything... ltinkl updated oobe [11:50] dednick, so... the problem is that you have merged it with trunk [11:50] dednick, the history says "merge with trunk" [11:50] yesterday [11:50] yeah, I think you only pushed that today, could that be? [11:50] dont think so. [11:50] maybe... [11:51] dednick, in any case, once your branch has a prerequisite set, you must *never* again merge it with trunk [11:51] dednick, always merge it with the prerequisite it has [11:51] mzanetti: ok [11:52] if you need to get trunk in somehow, merge the prerequisite with trunk (or it's prereq) and get changes through the chain like that [11:52] ok, now lets see how to get out of this again [11:53] mzanetti, complaining about missing unity-shell-scopes when trying to compile more-palette-fixes, which package version you have installed? [11:53] dpm, indicator-session doesn't contain any POT file, how does that work? :) [11:53] ltinkl, it gets created on build [11:54] cimi, 10 [11:54] mzanetti, of the deb package [11:55] dpm, right, with "make pot"; do you want that file? [11:55] 7.107+16.04.20160223-0ubuntu1 [11:55] ltinkl, exactly [11:59] mzanetti, you have any silo installed? [11:59] I also need 54? [12:00] sorry 41 [12:02] wow [12:03] loading the search icon on my desktop takes 1696 stat calls until it finds the correct one :D [12:03] http://paste.ubuntu.com/15401036/ [12:03] performance \o/ [12:04] tsdgeos: worth logging a UITK bug about that [12:04] their icon loading logic could definitely improve [12:05] not sure if it's glib related though [12:05] let me try it on the phone and see how many hits the phone gets [12:05] greyback, tsdgeos: remember a lengthy ML thread about KIconLoader... we used to have the same issue... ages ago [12:06] ltinkl: heh, why am I not surprised [12:06] tsdgeos, there is no glib code involved afaik, just QFile,QDir & co., all in nice loops :) [12:07] http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/trunk/view/head:/src/Ubuntu/Components/plugin/unitythemeiconprovider.cpp [12:07] see? :) [12:07] lots of loops! O(n^2) at least [12:07] QSettings in a foreach loop, nice [12:08] cimi, don't think so, no... [12:09] on the phone it's only 1016 [12:10] ltinkl, ? [12:11] not sure why everyone keeps on picking on QSettings [12:11] it caches all the stuff and is very lightweight to create [12:11] even for write access it caches things and write them to disk when idle [12:12] being clever enough to sync through memory if in between another instance is created and ready that value [12:12] reads [12:12] mzanetti, ok, it's just that those nested loops feel heavy [12:13] the only place where QSettings becomes slow is with a huge amount of ungrouped keys or endless bytearrays in values... [12:13] that can really take it down [12:13] but creating/destroying it is not the issue [12:20] mzanetti, ltinkl, do you happen to know which package the "Displays" string in here http://i.imgur.com/VzhS2bE.png comes from? It doesn't seem to be in indicator-session [12:20] dpm, you shouldn't see that indicator [12:20] dpm, where do you see that string? [12:21] mzanetti, it's on a screen capture bq sent us while they are preparing the m10 manual [12:22] right... that was the MWC image [12:22] and pointed out to us they need translations for those strings (while they haven't landed, they'll photoshop the translated strings on the screen caps) [12:22] yeah [12:22] this has changed. that string should not be around any more [12:22] oh, what does it look like now? [12:22] it's in the session indicator now [12:23] dpm: a dead parrot [12:23] one sec, will get you a screenie [12:23] mzanetti, awesome, thanks [12:23] dpm: man mzanetti spoils all my fun :) [12:23] davmor2, well, not quite dead yet, the rest of strings are in indicator-session. A half-dead one :) [12:23] well, I can still send him a picture of a parrot I guess [12:24] :D [12:24] hahaha [12:24] mzanetti: good morning! I see your email, is the situation basically the same? [12:25] mterry, it should be a bit better by now, but yes [12:25] mterry, good morning [12:25] mterry, I might just have found another issue for you tho... when booting the tablet, it starts up with "Try again" in the login field [12:25] mterry, requires 2 taps to be able to enter the password [12:26] mzanetti: OK... sounds like a post silo 41 bug though [12:28] dpm, not sure if you saw it... the screenshot is in your telegram inbox [12:29] got it , thanks! [12:30] mterry, oh... and... you don't want to hear it [12:30] but... it just doesn't pick up that username for me [12:30] mzanetti: you said that! I don't get it [12:30] mzanetti: another post silo issue I guess [12:31] yeah... at least that's not a regression... [12:31] I'm not sure about the double tap tho [12:31] a bit worried QA will block us on that [12:31] mzanetti: ok, i think i've got all the side-stage low-chrome test fixes. pushed. [12:31] cool!!! thanks a lot dednick. kicking abuild [12:31] mzanetti: oh really? OK, I can look into that then [12:35] mzanetti: shall I also remove camera-app, gallery-app, and mediaplayer-app from the ppa? I see you dropped the MPs [12:36] mterry, yes, please [12:37] done [12:47] mzanetti: is there any way for me to fake tablet mode these days on my phone, without editing qml? [12:47] mzanetti: (to reproduce the "try again" bug) [12:48] these days == with silo 41 [12:50] mterry, /etc/ubuntu/devices.conf [12:50] mterry, add a section: [12:51] wait, gotta create a pastebin [12:52] mzanetti: /etc/ubuntu/ doesn't exist for me, I assume that's normal [12:52] mterry, http://paste.ubuntu.com/15401213/ [12:52] mterry, yeah, create it [12:52] mterry, replace krillin with mako or whatever you have [12:52] mzanetti, what's the ETA for the current silo? [12:52] dandrader, when the tests are passing [12:53] mzanetti, something like mid next week? [12:54] mzanetti, or you expect this week already? [12:54] definitely this week [12:54] mzanetti, ok [12:54] mzanetti: that didn't fix it for me, will just edit qml [12:54] QA is just waiting for this [12:54] lets see, got a build running. I expect there might be some more tests failing, but shouldn't be many any more [12:56] tsdgeos, what would be the best approach to update teh .pot? check out the merged silo 41 branch, run the update script and then copy the po/ folder over to a checkout of trunk? === alan_g is now known as alan_g|lunch [13:09] mzanetti: but the pot gets updated on landing, no? [13:09] mzanetti: or you mean locally update it? [13:09] what do you need it for? [13:10] dos it? [13:10] yes it does [13:10] there's a hoook for it [13:10] ok. problem solved... dpm asked me to do so [13:10] thanks [13:10] mzanetti: how reliable is the "try again" issue for you? I got it once testing on my mako, but can't reproduce now [13:10] mzanetti: "override_dh_auto_clean:" in debian/rules [13:10] mterry, I see it all the time basically [13:11] humph [13:11] mterry, steps to repro: [13:11] * unlock the device [13:11] * lock it (power button presS) [13:12] * wake it up, all is fine [13:12] * turn screen off again [13:12] * wake it up => issue triggered [13:12] mzanetti: ooohhhh, it's not on boot? i thought I had read that [13:12] mterry, I guess there too... not sure [13:13] mterry, thing is, for me the screen wake up on his on very often because of telegram notifications [13:13] mterry, so I might just bootet it and by the time I looked at it the screen had been on/off twice already [13:14] mzanetti: ok, reproduced with your steps, thanks [13:30] morphis: ping, I'm just checking in again to see if aethercast is any closer to working on the Nexus 4 this week [13:31] mzanetti: I see why this 'retry' bug is happening, trying to figure out best fix [13:31] mzanetti, is the new screenshotting scheme part of that silo? [13:31] mterry, do you confirm it is a regression of the silo? [13:31] mzanetti: you sure it's a regression? [13:31] dandrader, no [13:31] mzanetti: my guess is that's it's been there. Let me check on another phone [13:31] mterry, heh, I'm not 100% sure, but I noticed it last night for the first time [13:31] mzanetti, too late to the party? [13:31] dandrader, yes [13:32] dandrader, really trying to tie down the very last bits in this silo [13:32] dandrader, in theory we wouldn't be allowed to land this atm as there is the Feature Freeze for OTA-10 going on. However the things in this silo have an exception granted already [13:32] dandrader, so even if I wanted I can't add other features now... and have like 10 people staring at me and asking when this silo will land [13:33] mterry, flashing my flo with trunk now to see if it's there too [13:33] mzanetti, sure, just asking :) [13:34] dandrader, yeah, no prob, just wanted to explain the situation instead of just saying "no - go away" [13:34] mzanetti, right, thanks [13:40] mterry, confirming the issue on trunk [13:40] mterry, you're off the hook :) [13:40] mzanetti, looks like the devices.conf trick doesn't work anymore since your better-windowed-logic branch, fyi [13:40] mzanetti: awesome, you flash faster than me [13:40] whaaaat? [13:41] mzanetti: I'll still work on it for post-silo though [13:41] can't be :D [13:41] mzanetti: you dropped "return deviceConfiguration.category;" as a backup for usageScenario [13:41] uh oh [13:41] well, but wait [13:41] how come the tablets are still tablets then? [13:42] the only reason for flo/frieza to show a side stage is because of the deviceconfig [13:42] mzanetti: ah, you check deviceConfiguration.category == "phone" [13:43] mzanetti: so the trick may still work... not sure why it didn't worked for me [13:43] mterry, maybe I got you wrong syntax... I didn't try it tbh... [13:43] mzanetti: whatever, I just hacked qml [13:43] or maybe I screwed something else up (enabled windowed mode) [13:45] mzanetti: can I help with an autopilot test? Do you know the latest set of failing ones? [13:45] mterry, running the adt tests atm [14:00] mzanetti, tsdgeos, not yet solved. I got the indicator-session .pot from ltinkl, but I don't have the unity8 one. If you guys can point me to a branch where it's up-to-date, I can also grab it from there. [14:01] dpm: why do you need it? [14:02] tsdgeos, bq are preparing a translated manual, and they are waiting for translations for new strings that are only in silo 41. Without them being in LP, it makes it a bit difficult to do them, so I've thought I could manually upload the new .pot files to Launchpad and thus we don't need to block on the landing [14:07] dpm: you mean upload it to the unity8 codebase? or? [14:07] tsdgeos, no, upload it to translations.launchpad.net [14:08] or more specifically, to https://translations.launchpad.net/ubuntu-rtm/15.04/+source/unity8 [14:08] I can upload a .pot file there in advance of any packages landing, so that the strings are translatable without having to wait on the landing [14:08] dpm: seems like a lot of hassle for somethign that will hopefully land soon, but ok, i guess [14:08] mzanetti: will you revive the branch you created? [14:09] kk [14:09] dpm, lp:~mzanetti/unity8/silo-41-pot [14:11] mterry, wheeee! [14:11] mterry, https://unity8-jenkins.ubuntu.com/job/test-ppa-autopkgtest/label=amd64,release=xenial,testname=qmluitests.sh/lastCompletedBuild/testReport/ [14:11] just some tutorial ones left [14:11] davmor2, just 4 test failures (probably just one thing to change to make all 4 pass) and we're good with 41 [14:13] mzanetti, thanks! Where can I find the .pot file in the branch, or do I need to build it? [14:13] dpm, po/unity8.pot [14:13] mzanetti, argh, sorry, misread the output of ls, all good, thanks [14:13] kk [14:14] mzanetti, perfect, that's all I need, thanks a lot! [14:21] mzanetti, this branch seems like is kinda touching that code https://code.launchpad.net/~unity-team/unity8/indicator-color-fixes/+merge/288175 [14:21] mzanetti, I dont have a micro sim to test the call :/ [14:22] I'll ask my gf to borrow me hers for a bit [14:22] cimi, I've tested it [14:22] cimi, but wait... what's with that branch? [14:23] ah, you mean the green panel [14:23] mzanetti, replaces fixed colors with colors from the palette [14:25] cimi, still strange how it never reverts, no? [14:25] I can't see an assignment somewhere [14:26] and of course https://code.launchpad.net/~unity-team/unity8/palette-fixes/+merge/288560 [14:26] mzanetti, unless something doesnt work in the applicationmanager [14:26] mzanetti, as you can see in active call hint [14:30] cimi, I'm struggling a bit to follow your thinking tbh [14:31] mterry, you about? [14:31] mzanetti: ye [14:31] mzanetti: sorry didn't see irc ping [14:31] mzanetti: I'm close to done with these tutorial failires [14:31] ah ok, great [14:31] sorry for being a pain today [14:34] mzanetti: no I get it :) [14:37] mzanetti, that if something broke recently, it might well be due to the only two branches that touched that code [14:37] mzanetti, so we should try to reproduce without those branches no? [14:37] fair enough [14:39] mzanetti, if we can reproduce without those two branches, is worth investigating more [14:49] I'm seeing https://launchpadlibrarian.net/235779719/ThreadStacktrace.txt with unity8 on a desktop (bug 1539614) and I've been seeing it for a while. Does anyone have an idea what it might be? [14:49] bug 1539614 in unity8 (Ubuntu) "unity8-dash crashed with SIGABRT in QMessageLogger::fatal()" [Medium,Confirmed] https://launchpad.net/bugs/1539614 [14:52] mzanetti: ok, I think I fixed em all, final testing [14:52] mterry, should I build yet? [14:52] mzanetti: not yet, will ping [14:52] dholbach, seems you're running a debug build and hit a qFatal() [14:55] dholbach, if this is when starting unity8, it's very weird. it seems to try to start as a client, not as a server [14:57] mzanetti, I installed unity8-desktop-session-mir and ran it from there [14:58] mzanetti, the debug output it simply because the dump was retraced in launchpad [14:58] "ran it from there" = in lightdm, pick unity8 session, try to login [14:59] right... [15:00] dholbach, do you have multiple libmirserver.so.XX around? [15:00] no, just so.38 [15:01] dholbach, ok.. is this a new user? or did you log in with unity7 on it already? [15:02] mzanetti, it's my regular user, the one I use in unity7 too [15:02] ok [15:02] guhh [15:03] mterry, no.... [15:03] mzanetti: nothing terrible, I just... there's some timing issue. Like some waitForRendering call missing or something. Having trouble finding the right spot [15:05] oh ok... [15:05] mterry, not in the mood for terrible things any more :D [15:05] mzanetti: I could disable this one test... :-/ and fix it later, if I'm the last test waiting [15:05] test works if I add a wait() call... [15:06] mterry, I can help searching [15:06] tell me which one [15:09] mzanetti, is there anyone else I could ping about this? [15:09] mzanetti: tutorial-redesign, test_tutorialBottomOnlyCoversMainStageOnTablet [15:09] mzanetti: pull first to get my other test fixes [15:09] dholbach, duflu usually knows quite a lot what's going on in that area [15:10] mzanetti: and if you disable all other test functions first, make sure to at least re-enable test_tutorialBottomFinish, this test only fails if it's not the first one... [15:10] but I don't think it's a case of failure to reset state... [15:11] thanks mzanetti [15:17] hmpf... fails to build here locally... /home/micha/Develop/reviews/tutorial-redesign/plugins/Wizard/timezonemodel.cpp:215:69: error: ‘geonames_city_get_country_code’ was not declared in this scope [15:17] ah, I guess I need teh patched geonames [15:20] something is focusing u8-dash on top of my dialer-app. it appears to be a race on setup, but I check on setup that dash is focused.... [15:25] hrm.. the mock AppManager is delaying dash focus... [15:27] mterry, so... I've built your branch [15:27] mterry, ran make xvfbtestTurotial [15:27] and have 0 failures [15:28] mzanetti: you might need the rest of silo 41 then... [15:28] mzanetti: I've been testing on top of full silo [15:29] right... [16:05] mzanetti, dednick: how are app stages determined in silo 41? [16:06] mterry: all start in main stage. state gets saved when they close [16:06] mzanetti, dednick: (especially in test mode with our mocks). Most seem to open main stage, and setStage seems to have odd sideffects [16:06] s/state/stage [16:06] dednick: can I just call setStage on a mock object? No other test does so [16:06] (other than tutorial) [16:06] mterry: mmm. [16:07] mterry: maybe... [16:07] dednick: how is this tested in the side stage tests? [16:07] not sure it'll work. will probably get overwritten with loaded stage. [16:08] test_applicationLoadsInCorrectStage_data [16:08] dednick: ah... WindowStateStorage.saveStage [16:17] ok... may have fix === dandrader is now known as dandrader|lunch === shuduo is now known as shuduo-afk [16:22] mterry, please tell me :) [16:23] mzanetti: ok, pushed. [16:24] mzanetti: it's a little janky (relies on a requestFocus to workaround a focus race that I still don't understand). But seems to work [16:24] And uses new more proper way to set stages [16:25] kk [16:25] mterry, you've seen my keymapSwitching tests? had to use requestFocus too... [16:25] ltinkl: no, didn't notice [16:25] mterry, https://code.launchpad.net/~unity-team/unity8/keymapSwitching/+merge/288842 at the very end [16:26] mterry, hey, do you know why ubuntu-settings-components are in the silo? [16:26] ltinkl: well there at least you use it for switching to apps like I'd expect [16:27] ltinkl: in my case, I started an app and immediately have to request focus as well [16:27] mterry, I mean, there's a package in the ppa, but no MP for it [16:27] mzanetti: ? [16:27] mzanetti: oh ubuntu-settings-component [16:27] mzanetti: ... no? [16:28] mzanetti: the diff in the silo points to some vpn changes [16:28] mzanetti: but the vpn stuff isn't landing in this silo [16:28] also seems to be a reverse diff [16:28] yeah, that doesn't look right [16:29] mzanetti: I'll drop from ppa [16:29] ack [16:29] mzanetti: also the silo probably should drop those manual click download urls [16:29] mterry, also, the silo says camera-app etc in source package name [16:29] right [16:29] mterry, not sure if they would land [16:29] mterry, I thought they are just for testing [16:29] mzanetti: I don't know either [16:30] mzanetti: ok settings-component out [16:30] cool [16:31] mzanetti: manually removed them from packages list too (and apps) [16:31] mterry, ok... build running... when it's done, I'll do another quick check on my devices, run adt on it and if it passes, off to QA [16:31] * mterry crosses fingers [16:33] mzanetti: I'm heading out to lunch [16:33] kk [16:33] enjoy [17:07] nooo [17:07] the silo build failed [17:08] but seems for a LP timeout === dandrader|lunch is now known as dandrader [17:32] mterry, failed to build :'( [17:32] not sure yet why [17:32] LP keeps timing out [17:32] gaaaaaa [17:32] whitespace test [17:33] mterry, /«BUILDDIR»/unity8-8.12+16.04.20160316.3/tests/qmltests/Tutorial/tst_Tutorial.qml: bad whitespace in line 266 [17:36] mterry, I've fixed it [17:42] mzanetti: guh thanks, sorry [17:43] should have run the test first [17:56] ltinkl: oh I see why setting name isn't working -- wizard doesn't make right call. Oddly enough, AS doesn't let you directly set those properties. It makes you call a special method like SetRealName, so that it can use policykit to more tightly control access [17:57] mterry, aha... did that come with the as-refactor? [17:57] ltinkl: it shouldn't have? That's just a property of how the AS server works [17:58] ltinkl: I can whip up a fix [17:58] for post silo 41, I'm guessing, but pre-ota10 [17:58] mterry, yeah I meant in the code where we talk to AS over DBUS [17:58] mterry, thanks! [17:58] * ltinkl bbl [17:58]