[00:02] <Saviq> fginther, yeah, that got added in http://bazaar.launchpad.net/~unity-team/libunity/trunk/revision/289?start_revid=297
[00:03] <Saviq> fginther, released as 7.1.1+13.10.20130918.1-0ubuntu1
[00:04] <Saviq> fginther, libunity gets upgraded with unity8 tests 'cause we're requiring > 7.1.1 now, but unity-common does not get upgraded (should be a version dependency on libunity9 maybe?), so it bails out
=== shiznix_ is now known as shiznix
[00:34] <kgunn> thomi: sorry for the confusion...sounds like you've got a plan of attack at least
[00:34] <kgunn> i'll check back in a bit
[00:34] <thomi> kgunn: maybe. I have very little idea what I'm doing :-/
[01:04] <fginther> Saviq, hey, are you around?
=== _salem is now known as salem_
=== salem_ is now known as _salem
=== jhodapp is now known as jhodapp|afk
[07:06] <tsdgeos> what happened in CI
[07:06] <tsdgeos> suddenly it's all unstable again
[07:21] <om26er> tsdgeos, people are working on it
[07:21] <tsdgeos> good :-)
[07:40] <mzanetti> paulliu: hi
[07:41] <paulliu> mzanetti: hi
[07:41] <mzanetti> paulliu: I think the previews are mostly done: lp:~unity-team/unity8/fix-genericpreview/
[07:41] <mzanetti> paulliu: its all your branches merged together and adjusted
[07:42] <mzanetti> paulliu: feel free to give it a try and let me know if you find bugs
[07:42] <paulliu> mzanetti: ok
[07:42] <mzanetti> paulliu: ah, one question: What is the difference between VideoPreviews and MoviePreviews?
[07:43] <paulliu> mzanetti: VideoPreview is for mockvideo
[07:43] <mzanetti> paulliu: afaics we don't use the VideoPreview at all any more
[07:43] <paulliu> mzanetti: yes.
[07:44] <paulliu> mzanetti: So I'll fix the moviepreview.
[07:44] <mzanetti> paulliu: that's fixed already in that branch
[07:44] <mzanetti> paulliu: I'm just thinking about deleting the VideoPreview
[07:44] <paulliu> mzanetti: also for the playback?
[07:45] <mzanetti> paulliu: do you mean inline playback in the preview?
[07:45] <paulliu> mzanetti: no. brings up the browser or player.
[07:45] <paulliu> mzanetti: anyway I'll check it.
[07:45] <mzanetti> paulliu: that seems to work
[07:46] <mzanetti> paulliu: so can I delete the VideoPreview? According to PreviewDelegateMapper.qml it's not used anywhere any more
[07:46] <mzanetti> not even for mockvideos
[07:46] <paulliu> mzanetti: yes. please delete it.
[07:46] <mzanetti> ack
[07:48] <tsdgeos> mzanetti: https://code.launchpad.net/~aacid/unity8/updatepot/+merge/188524
[07:49] <mzanetti> tsdgeos: why this?
[07:49]  * mzanetti thought this happens automagically
[07:50] <tsdgeos> it doesn't afaics
[07:50] <tsdgeos> i.e. we've outdated stuff over there and our translators are translating old stuff
[07:50] <mzanetti> tsdgeos: looking the the commit log there are some autocommits doing this... maybe something broke in that machinery?
[07:51] <tsdgeos> i don't see any autocommit doing this
[07:51] <tsdgeos> which revision?
[07:51] <mzanetti> tsdgeos: 314 for example
[07:51] <mzanetti> tsdgeos: but you're right. it didn't happen in quite a while
[07:51] <mzanetti> tsdgeos: I remember like 2 of those commits a day a month back
[07:51] <tsdgeos> mzanetti: 314 is updating the catalan translation
[07:52] <tsdgeos> not the .pot
[07:52] <mzanetti> mhm... ok
[07:52] <tsdgeos> the translations get autocommited from the launchpad translations thing
[07:52] <tsdgeos> that part is right
[07:52] <tsdgeos> mzanetti: bzr qlog po/unity8.pot
[07:53] <mzanetti> tsdgeos: well, approved then
[07:54] <tsdgeos> thankyousir :-)
[07:57] <pete-woods> guys, does anyone know why a directory /var/lib (created by package maintainer scripts) could have changed from being owned by the correct user to being owned by root?
[07:58] <pete-woods> this is only on some of the images (ones I'm not familiar with, numbered 70, 71, etc, rather than a timestamp)
[08:23] <tsdgeos> Saviq: do we have somewhere written that with Mir we seem to lock the screen when we press the button to unlock and with SF it's done on screen blank?
[08:23] <Saviq> tsdgeos, that'd be powerd's bug, are you sure that's the case?
[08:24] <tsdgeos> i'm sure of the first
[08:24] <Saviq> tsdgeos, wait a few seconds after pressing power to blank - there's some lag
[08:24] <tsdgeos> Saviq: i've waited like minutes :D
[08:24] <Saviq> tsdgeos, and you still saw the greeter coming in on resume?
[08:24] <tsdgeos> yep
[08:25] <Saviq> tsdgeos, there's "powerConnection" in Shell.qml
[08:25] <tsdgeos> i can try again just to confirm
[08:25] <Saviq> tsdgeos, can you grab some debugging out of there?
[08:25] <tsdgeos> sure
[08:25] <Saviq> tsdgeos, to see when onDisplayPowerStateChange is fired
[08:26] <Saviq> MacSlow, re: API, the shell-facing API changed - the type of the hints property changed from flags to a variant map
[08:26] <Saviq> MacSlow, but that's ok, just keep that in mind for later
[08:26] <MacSlow> Saviq, right.
[08:29] <Saviq> MacSlow, btw, does the spec say that extended snap decisions need longer timeout than standard ones?
[08:30] <MacSlow> Saviq, it says nothing about it... but it's a suggested value.
[08:31] <Saviq> MacSlow, let's just flatten the logic then? for all snap decisions 60s?
[08:31] <MacSlow> Saviq, I'm ok with that
[08:31] <tsdgeos> Saviq: so the onDisplayPowerStateChange thing triggers on time
[08:31] <tsdgeos> correctly
[08:32] <tsdgeos> but i still see the greeter on pressing the button
[08:32] <Saviq> tsdgeos, maybe the proximity hint / thingie is not set correctly?
[08:32] <tsdgeos> maybe
[08:32] <tsdgeos> let me print that
[08:33] <MacSlow> Saviq, pushed change
[08:35] <Saviq> MacSlow, k
[08:36] <Saviq> MacSlow, btw v
[08:36] <Saviq> https://code.launchpad.net/~saviq/unity-notifications/fix-themed-icons/+merge/188483
[08:37] <MacSlow> Saviq, talking a look
[08:47] <dpm> morning Saviq, when you've got a minute, do you think you could look at this bug to update the translations template for Unity 8? It should only take a few minutes to fix: http://pad.lv/1232374
[08:47] <ubot5> Launchpad bug 1232374 in Unity 8 "Please update translations template" [Undecided,New]
[08:48] <Saviq> dpm, sure
[08:48] <dpm> great, thanks :)
[08:49] <larsu> does unity8 still have CI problems on jenkins? I can't see anything wrong in the logs relating to my MR
[08:49] <larsu> https://code.launchpad.net/~larsu/ubuntu-ui-toolkit/remove-gicon-provider/+merge/188306
[08:50] <Saviq> dpm, https://code.launchpad.net/~saviq/unity8/update-pot-file/+merge/188528
[08:52] <tsdgeos> Saviq: the signals are correct and we enter the showGreeter part
[08:52] <tsdgeos> but i can still see the greeter coming later
[08:52] <Saviq> tsdgeos, ok, so completely our fault
[08:52] <pstolowski> Saviq: I've just finished implementing a scope for popular music results and having a weird issue with icons - http://ubuntuone.com/4bl6mKBlFlHpm2oXcRKzMX
[08:52] <Saviq> tsdgeos, file bug for mterry please?
[08:52] <tsdgeos> i wonder if it has anything to do with us getting suspended too or something
[08:53] <tsdgeos> Saviq: sure
[08:53] <pstolowski> Saviq: happens only on the phone; works for on the desktop with same result set
[08:53] <Saviq> pstolowski, you mean they're not loaded?
[08:53] <Saviq> pstolowski, how many results are there?
[08:53] <dpm> nice one, thanks Saviq
[08:53] <pstolowski> Saviq: yes, but only "left" column
[08:54] <Saviq> pstolowski, anything interesting in ~/.cache/upstart/unity8.log ?
[08:54] <pstolowski> Saviq: 20 results
[08:54] <pstolowski> Saviq: ahhh, ssl handlshake failed
[08:55] <pstolowski> * handshake
[08:55] <Saviq> pstolowski, :)
[08:55] <pstolowski> Saviq: file:///usr/share/unity8/Dash/Music/AlbumTile.qml:35:16: QML QQuickImage: SSL handshake failed
[08:55] <Saviq> pstolowski, what's your datetime?
[08:55] <mhr3> eh, wifi dropped out and now i see ssl handshake issues
[08:56] <tsdgeos> Saviq: https://bugs.launchpad.net/unity8/+bug/1233564
[08:56] <mhr3> i think i it got connected to a parallel universe
[08:56] <ubot5> Ubuntu bug 1233564 in Unity 8 "Greeter is seen animating when pressing the side button to wake up" [Undecided,New]
[08:56] <Saviq> tsdgeos, thanks
=== ejat- is now known as ejat
[09:00] <Saviq> greyback, I'll need your input on https://code.launchpad.net/~saviq/unity8/fix-frequent-app-ids/+merge/188532
[09:00] <Saviq> greyback, problem: facebook and gmail in Dash Home doesn't launch, 'cause they've been converted to click apps
[09:00] <pstolowski> Saviq: right, my time setting was incorrect; fixing it didn't help though
[09:01] <Saviq> pstolowski, can you wget them?
[09:01] <Saviq> greyback, but we have like 3 places where we parse the application:/// part or something
[09:01] <pstolowski> Saviq: but I'm still getting ssl errors
[09:01] <Saviq> pstolowski, can you wget the images?
[09:01] <greyback> Saviq: great. Ok, let me see.
[09:02] <dednick> Cimi: ping
[09:03] <Cimi> dednick, pong
[09:03] <Saviq> greyback, I'd like to simplify all of that and have just one place where we parse the url
[09:03] <Cimi> dednick, I'm working on the calendar
[09:04] <Cimi> dednick, I took the logic from the calendar app, but I discovered there's a bug that's why tests are failing
[09:04] <Cimi> dednick, I'm trying to see where the logic is broken
[09:04] <dednick> Cimi: ok. i was thinking we may as well get the MP merged and fix the calendar in a later merge. It's not being used at the moment anyway.
[09:05] <Cimi> dednick, if Saviq is fine
[09:05] <dednick> Cimi: then we can get the landing sorted as well.
[09:05] <Saviq> Cimi, yeah, I'm feeling fine, why?
[09:05] <Cimi> hah
[09:06] <Cimi> Saviq, ok to approve a MP that contains tested all commented out because the real code contains a bug?
[09:06] <Saviq> Cimi, skip() the test, don't comment it out
[09:06] <dednick> Saviq: tests for calendar in ubuntu-settings-components fail.
[09:07] <Cimi> because there's a bug on dates
[09:07] <Cimi> I think it's on the logic that finds the start of the month...
[09:07] <dednick> Cimi, Saviq: ok, i'll put a skip in
[09:07] <Saviq> Cimi, dednick, and well, obviously fixing the logic would be best ;)
[09:07] <Cimi> Saviq, it's what I am doing
[09:07] <Cimi> Saviq, but at the moment I don't have an ETA
[09:08] <Cimi> Saviq, this stuff is so confusing :)
[09:08] <Cimi> because you have to deal with all timezones
[09:09] <dednick> Saviq: i want get ubuntu-settings-components into archive. But my MP needs to go in first.
[09:09] <Saviq> dednick, k
[09:09] <Saviq> Cimi, is that not a solved problem? i.e. isn't there some library that we could use instead of doing it ourselves?
[09:10] <Cimi> Saviq, nope… the designs of the calendar app differ so much from the indicators that we decided to go separate
[09:11] <dednick> Cimi: what does timezone have to do with calendar?
[09:11] <Cimi> dednick, it needs to display the right month
[09:12] <Cimi> dednick, if you're in japan you switch to the new month before someone here
[09:12] <dednick> Cimi: yeah, but if it's based on current time, then TZ should be included
[09:13] <paulliu> mzanetti: yeah.. almost ok. I'll add some logic in MoviePreview.
[09:13] <dednick> Cimi: nevermind. i'm sure you know what you're talking about
[09:13] <mzanetti> paulliu: ok, cool
[09:13] <mzanetti> paulliu: what exactly?
[09:14] <paulliu> mzanetti: the playback should use previewData.result.uri and base on that, open the browser of player depends on the prefix.
[09:14] <mzanetti> paulliu: ah, I see. ok, great
[09:16] <dednick> Cimi: ok, i've changed the failing test to skip.
[09:17] <paulliu> mzanetti: btw, do you know if there's any function to get an preview from imageSourceUrl? A image.
[09:17] <mzanetti> paulliu: not sure if I understand the question
[09:17] <paulliu> mzanetti: ok, the preview provides either image or imageSourceUri.
[09:18] <paulliu> mzanetti: if preview provides iamgeSourceUri, we have to generate an image by ourselves.
[09:18] <mzanetti> paulliu: but it's still a url, pointing to a image, no?
[09:19] <paulliu> mzanetti: No, to mpg or pdf.
[09:19] <paulliu> mzanetti: so I'm actually asking if we have any functions to get a snapshot of mpg or pdf.
[09:19] <paulliu> mzanetti: or other file formats.
[09:19] <mzanetti> paulliu: oh... hmm. no idea if there is something. I'm not aware of anything
[09:19] <mzanetti> Saviq: are you? ^
[09:20] <Saviq> paulliu, we're getting the thumbnailer from Satoris
[09:20] <Saviq> mzanetti, ↑
[09:20] <mzanetti> cool, thanks
[09:20] <Saviq> paulliu, but it probably won't support pdfs initially
[09:21] <Saviq> paulliu, but we don't have support to *display* them anyway
[09:21] <Saviq> paulliu, for movies / videos we'll be good
[09:21] <paulliu> ok.
[09:22] <Saviq> paulliu, https://code.launchpad.net/~jpakkane/ubuntu-ui-toolkit/thumbnailer/+merge/187730
[09:22] <paulliu> Saviq: ah.. thanks..
[09:30] <Cimi> dednick, found the bugger
[09:37] <Cimi> dednick, works for me http://paste.ubuntu.com/6178936/
[09:40] <Saviq> grr my manta discharged and doesn't let me build stuff
[09:41] <tsdgeos> Saviq: comments really appreciated on https://code.launchpad.net/~aacid/unity8/enable_height_animation_when_visible/+merge/188536
[09:42] <Saviq> tsdgeos, was just there ;)
[09:43] <Saviq> tsdgeos, 'undefined'?
[09:43] <Saviq> tsdgeos, why the quotes?
[09:44] <tsdgeos> javascript for "this thing doesn't exist"
[09:44] <tsdgeos> damn me
[09:44] <tsdgeos> i forgot the typeof
[09:44] <tsdgeos> when reordering the code
[09:44] <Saviq> tsdgeos, you can just do blah === undefined
[09:45] <tsdgeos> can i?
[09:45] <Saviq> tsdgeos, I believe so
[09:45] <mhr3> Saviq, is there a bug about updating the music grid renderer?
[09:45] <tsdgeos> afair i tried that and compleined
[09:45] <tsdgeos> let me try again
[09:45] <Saviq> mhr3, yes
[09:45] <mhr3> i remember one for videos, but not for music
[09:45] <Saviq> mhr3, or not
[09:45] <Saviq> mhr3, yeah, I think there is none
[09:45] <mhr3> the two are actually quite similar
[09:46] <mhr3> just one has background
[09:46] <mhr3> Saviq, updated https://bugs.launchpad.net/unity8/+bug/1231937
[09:47] <ubot5> Ubuntu bug 1231937 in Unity 8 "Video & music renderers are not to spec" [High,Triaged]
[09:47] <Saviq> mhr3, k
[09:47] <Saviq> tsdgeos, https://code.launchpad.net/~aacid/unity8/enable_height_animation_when_visible/+merge/188536/comments/431201
[09:47] <Cimi> dednick, what shall we do?
[09:48] <dednick> Cimi: sorry. was just busy with something. will look now.
[09:49] <tsdgeos> Saviq: not sure i understand "Should we maybe check if the item has a "culled" property and only set it then?"
[09:49] <tsdgeos> you mean the LVWPH checking the existence of the property and setting it if it's there?
[09:50] <dednick> Cimi: I can merge that into my changes if you want.
[09:50] <Cimi> dednick, changes to the dates maybe are not required, was of previous tests
[09:50] <Saviq> tsdgeos, yeah
[09:50] <Cimi> dednick, anyway, what fixes is the changes to midnight prototype
[09:50] <Saviq> tsdgeos, so that the item needs to explicitly "want" that set
[09:50] <Cimi> dednick, it was changing day..
[09:50] <dednick> Cimi: ok
[09:51] <Saviq> tsdgeos, by having that prop, instead of LVWPH adding it regardless
[09:51] <tsdgeos> Saviq: it's unfortunately that easy though
[09:51] <dednick> Cimi: i'll try it with just the midnight/addmonth changes
[09:51] <Saviq> tsdgeos, this way the component can work in both a ListView and LVWPH, but not as good
[09:51] <Cimi> dednick, u can merge
[09:51] <Saviq> tsdgeos, :/
[09:51] <tsdgeos> since RunningApplicationsGrid is not the item the LVWPH knows about
[09:51] <nic-doffay> pete-woods, ping
[09:52] <tsdgeos> LVWPH only has ListItems.Base inside
[09:52] <Saviq> tsdgeos, ah, the loader?
[09:52] <Cimi> dednick, test on selectedte has to be removed
[09:52] <tsdgeos> and ListItems.Base has loaders inside
[09:52] <Saviq> tsdgeos, how about an attached prop?
[09:52] <Saviq> tsdgeos, ListViewWithPageHeader.culled?
[09:52] <dednick> Cimi: ok
[09:52] <tsdgeos> Saviq: what i did is an attached prop
[09:52] <tsdgeos> ah wait not relaly
[09:53] <Saviq> tsdgeos, yeah, not really
[09:53] <pete-woods> nic-doffay: hi
[09:53] <tsdgeos> tbh i find attached props confusing, last time i tried them i discovered they're available for all the items even if they're not in a LVWPH
[09:53] <tsdgeos> not sure what they give us over a context property like this
[09:57] <Saviq> tsdgeos, available in what sense?
[09:58] <Saviq> tsdgeos, they're only valid on the delegate, though
[09:58] <tsdgeos> Saviq: available in the sense that you can use it and the engine won't complain and will be defined
[09:58] <Saviq> tsdgeos, true
[09:58] <Saviq> tsdgeos, but the value would be invalid (undefined, hopefully)?
[09:58] <dednick> Cimi: test fails when we dont change the tests
[09:59] <Cimi> dednick, dates?
[09:59] <dednick> Cimi: yes, didnt change dates, failed
[09:59] <Cimi> dednick, so change dates :)
[09:59] <Cimi> dednick, current date probably needs to be changed
[10:00] <Cimi> dednick, because otherwise currentDate is todat
[10:00] <Cimi> *today
[10:00] <Cimi> dednick, and tests need to be the same every time
[10:00] <tsdgeos> Saviq: i.e. see http://paste.kde.org/~tsdgeos/p573bd894/
[10:00] <tsdgeos> it gives you
[10:00] <tsdgeos> false false
[10:00] <dednick> Cimi: yeah. you should really be putting the Calendar into a well known state before running a test. Otherwise it depends on the state the previous test left it in. That's what the init function is for. I'll fix that up.
[10:00] <tsdgeos> imho that's lame
[10:01] <dednick> Cimi: if you dont, when somebody adds a test which runs before one of your tests, it will break.
[10:01] <Cimi> dednick, use init without creating new dates
[10:01] <Cimi> reuse the same vcariables
[10:02] <Cimi> dednick, I ihad issues when date() was still creating the dates while the yet was proceeding
[10:02] <Cimi> dednick, that's why they are at the beginning
[10:02] <dednick> Cimi: any reason for that?
[10:03] <Cimi> dednick, dunno
[10:03] <Cimi> dednick, multithreading?
[10:03] <dednick> eh. all the dates should be created in the _data functions. I'll look into it
[10:07] <Saviq> nic-doffay, two minor comments https://code.launchpad.net/~nicolas-doffay/unity8/search-indicator/+merge/188008/comments/431219
[10:07] <Saviq> tsdgeos, right, well, it's correct isn't it ;D
[10:08] <Saviq> tsdgeos, I think that's the idea though - that you should be able to have a component that integrates with a ListView as well as works standalone
[10:09] <tsdgeos> sure
[10:09] <tsdgeos> and it does
[10:09] <tsdgeos> with the check for undefined
[10:12] <Saviq> mhr3, pstolowski did you ever get the info you needed re: libunity?
[10:12] <pstolowski> Saviq: I didn't
[10:12] <mhr3> nope
[10:12] <Saviq> vila, can you help getting the version numbers for libunity9 and libunity-protocol-private0 from the mediumtest runners?
[10:13] <mhr3> Saviq, but i'm pretty sure it's that, there was abi break in protocol... and protocol is called -private cause libunity9 wants it always compatible
[10:13] <Saviq> mhr3, yeah, so what do we do in debian/control to make it work?
[10:14] <mhr3> not enough apparently :)
[10:14] <Saviq> mhr3, pfft
[10:15] <tsdgeos> Saviq: actually we did something similar for the header cliping
[10:15] <vila> Saviq: may be (doing apt-cache policy XXX on the relevant slaves ?), do you have slave names ?
[10:15] <tsdgeos> Saviq: see Base.qml lines 76-77
[10:16] <Saviq> tsdgeos, right, that's a poor man's attached prop :D
[10:16] <tsdgeos> well
[10:16] <tsdgeos> it's a better one imo
[10:17] <Saviq> tsdgeos, the biggest problem I have with that is that we're reaching out of scope to get the heightToClip
[10:17] <tsdgeos> tbh i remember discussing this when we introduced it
[10:17] <tsdgeos> not sure if it was you i was discussing with though :D
[10:17] <Saviq> tsdgeos, as it's set in the Loader's scope
[10:17] <Saviq> ;)
[10:17] <tsdgeos> sure
[10:17] <tsdgeos> it's set in the context of the parent
[10:18] <tsdgeos> and that i don't like
[10:18] <Saviq> tsdgeos, yeah, so we should pass it down through props
[10:19] <Saviq> tsdgeos, at which point the attached prop has the advantage of being obvious where it comes from
[10:19] <Saviq> tsdgeos, instead of "just showing up" in the context
[10:19] <Saviq> tsdgeos, because LVWPH puts it there
[10:19] <tsdgeos> you mean like an attached prop :D
[10:20] <tsdgeos> anyway
[10:20] <tsdgeos> if you want an attached prop
[10:20] <tsdgeos> i'll do one
[10:20] <vila> Saviq: hmm, from http://10.97.2.10:8080/job/generic-mediumtests-builder-saucy-amd64/137/console that would be kinnara, but libunity9 is not installed there so that would be in the pbuilder env ?
[10:20] <tsdgeos> just say it :-)
[10:20] <Saviq> vila, that's the builder
[10:20] <Saviq> vila, we need the runner
[10:21] <Saviq> vila, so any of http://s-jenkins:8080/job/autopilot-testrunner-otto-saucy/
[10:21] <Saviq> vila, ps-nvidia-gt630||ps-intel-4600||ps-radeon-hd8350
[10:23] <tsdgeos> pstolowski: ok, so waht was the code i had to look at and what was the problem?
[10:24] <Saviq> dednick, https://code.launchpad.net/~nick-dedekind/unity8/indicator-titles/+merge/186049/comments/431233
[10:27] <vila> Saviq: otto stuff ? Damn, not up to date with that stuff yet :-/
[10:27] <Saviq> vila, k, let's wait for fginther, then
[10:28] <vila> Saviq: yup, sry about that, will try to shadow him to learn
[10:29] <pstolowski> tsdgeos: there is the branch that enables reordering: lp:~stolowski/unity8/category-reordering
[10:30] <pstolowski> tsdgeos: and remember, before you went on vacation, you said you thought your branches were fixing list view, but then found out you forgot to commit all changes, and it sill wasn't working correctly
[10:30] <tsdgeos> i do remember i thought it was fixing everything
[10:30] <pstolowski> tsdgeos: what was your last day before vacation? I'll dig in the logs
[10:30] <tsdgeos> and you found out not
[10:30] <tsdgeos> don't remember if i was sure to reproduce it not working on my side
=== alan_g is now known as alan_g|afk
=== hikiko is now known as hikiko|lunch
[10:33] <tsdgeos> pstolowski: can you merge unity8 to your branch?
[10:33] <tsdgeos> it doesn't merge cleanly
[10:34] <pstolowski> tsdgeos: yup, doing
=== alan_g|afk is now known as alan_g
[10:49] <pstolowski> tsdgeos: pushed
[10:49] <tsdgeos> tx
[10:54] <tsdgeos> pstolowski: so do you remember the steps to make it fail?
[10:55] <pstolowski> tsdgeos: I don't.. let me test it again; all your changes were already merged, right?
[10:55] <tsdgeos> yep
[10:55] <tsdgeos> all's in
[10:57] <mzanetti> Saviq: any idea what the thread named "SignalSender" is?
[10:57] <Saviq> mzanetti, nope :/
[11:00] <mhr3> clearly it's a thread sending signals :)
[11:00] <mhr3> Saviq, i just had indicators hide, is that a new feature? :)
[11:00] <Saviq> mhr3, you mean fullscreen?
[11:00] <Saviq> mhr3, it's been there since January ;)
[11:01] <mhr3> Saviq, no, i was browsing in the dash and they decided to hide
[11:01] <mzanetti> yeah, there is something new indeed
[11:01] <mzanetti> I remember having reviewed a branch that can do this while saviq was on holiday
[11:02] <om26er> mzanetti, any clues what went bad here https://jenkins.qa.ubuntu.com/job/unity8-autolanding/516/console ?
[11:02] <mzanetti> om26er: yeah 404
[11:03] <om26er> mzanetti, ofcourse but what really ? all the tests passed fine. what was it not able to download ?
[11:03] <om26er> artifacts ?
[11:04] <mzanetti> om26er: yes... looks like that
[11:05] <mzanetti> Saviq: this looks suspicious too: http://paste.kde.org/pf627c745/
[11:05] <mzanetti> Saviq: are we actually running the runtimeprofiler in release mode?
=== MacSlow is now known as MacSlow|lunch
[11:09] <Saviq> mzanetti, no, in debug
[11:09] <mzanetti> Saviq: this is when attaching to the installed unity8
[11:13] <Saviq> mzanetti, not sure what you mean/
[11:14] <mzanetti> Saviq: that stack trace above..
[11:14] <mzanetti> Saviq: it shows some runtimeprofilerstuff
[11:14] <Saviq> mzanetti, we're not running the profiler unless you pass the argument
[11:15] <Saviq> mzanetti, must be something Qt-internal that enables that
[11:15] <mzanetti> Saviq: yeah... found a thread of a guys seeing the same in chrome
[11:16] <mzanetti> Saviq: starts spinning cpu with same strace output and also this same thread bt
[11:16] <mzanetti> Saviq: seems v8 waiting to free up memory. and indeed my device seems to be out of mem right now
[11:17] <Saviq> mzanetti, so garbage collector?
[11:17] <mzanetti> Saviq: seems so
[11:17] <Saviq> yay
[11:18] <mzanetti> Saviq: lol... clsoed all apps, unity8 stops spinning the cpu
[11:19] <mzanetti> Saviq: which explains why it happens mostly when using the gallery app (the heaviest biest we have)
[11:19] <Saviq> mzanetti, yeah
[11:19] <mzanetti> so... I'm a little bit lost how to handle this one right now
[11:20] <mzanetti> anyways... noting down my findings in the bug report
[11:20] <Saviq> mzanetti, yeah, "use less memory"....
[11:21] <mzanetti> hm... now cpu usage went up again
[11:21] <mzanetti> oh... only 500KB free mem again
[11:21] <Saviq> fginther, ping
[11:25] <tsdgeos> pstolowski: any luck reproducing the problem?
[11:25] <Saviq> tsdgeos, so what do we do with the "don't animate if not visible"? attached prop that we pass down through Loader and Base? or do we just say fook it?
[11:27] <tsdgeos> Saviq: interesting question, will using the attached prop in RunningApplicationsGrid work? Or that's as unclean as the current solution? And how do we pass it down propetly for the clean solution? We ought to create another "base"  for all the items the loaders load, to make sure it has the property, or otherwise it means another "if culled !== undefined" in the code, just in a different place
[11:27] <Saviq> tsdgeos, no, it won't work
[11:27] <Saviq> tsdgeos, attached prop is only available on *the* delegate
[11:27] <Saviq> tsdgeos, not down the tree
[11:28] <Saviq> tsdgeos, so we'd have to pass it through from the Loader
[11:28] <tsdgeos> well, technicaly it's available "everywhere", just the correct value is just on the delegate
[11:28] <Saviq> tsdgeos, and yeah, what you say - a LVWPHDelegate base component
[11:28] <Saviq> tsdgeos, yes, that
[11:28] <tsdgeos> passing it down from the Base to the Loader is ugly unless we create the LVWPHDelegate thing
[11:28] <Saviq> tsdgeos, indeed
[11:28] <tsdgeos> I can do that if you want, to be honest not sure it's worth the hassle at this point
[11:29] <tsdgeos> I'd create a "think the whole thing again"
[11:29] <tsdgeos> since for example
[11:29] <Saviq> tsdgeos, yeah I agree we should look at this again
[11:29] <tsdgeos> Components/ListItems/Base.qml is only used (and i'd say useful) inside the LVWPH
[11:29] <Saviq> tsdgeos, +1
[11:29] <tsdgeos> so not sure it makes sense to be inside Components
[11:29] <tsdgeos> so that could become part of the  LVWPHDelegate thing
[11:30] <tsdgeos> otoh if we say "we'll have a look later"
[11:30] <tsdgeos> that later may end up being never
[11:30] <Saviq> tsdgeos, btw, do you have in the back of your mind the fact that we should make the header common between lenses at some point? to enable tab-like navigation?
[11:30] <tsdgeos> you know how this works :D
[11:30] <Saviq> tsdgeos, yeah, I'm afraid we'll need a look as part of ↑↑ anyway
[11:30] <tsdgeos> i had forgotten to be honest
[11:30] <Saviq> tsdgeos, so let's just go with what you have now - add a FIXME though, please
[11:31] <tsdgeos> ok
=== hikiko|lunch is now known as hikiko
[11:34] <Saviq> dednick, can you try and track down https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1233245 please?
[11:34] <ubot5> Ubuntu bug 1233245 in unity8 (Ubuntu) "Volume up/down keys not working in music-app with Mir" [High,Triaged]
[11:35] <Saviq> mzanetti, btw, one thing to mention on the bug - 5.2 comes with v4, which will hopefully be much better, in that it's tailored to QML's needs, not to browsers'
[11:36] <Saviq> or "should come"
[11:36] <mzanetti> Saviq: oh is that in with 5.2 already?
[11:36] <mzanetti> that'd be grand
[11:36] <Saviq> mzanetti, I think so
[11:36] <Saviq> mzanetti, yup
[11:36] <Saviq> mzanetti, http://blog.qt.digia.com/
[11:36] <mzanetti> Saviq: altough I had a chat with Lars about this too
[11:36] <mzanetti> Saviq: there is a slight chance that it introduces really bad stuff
[11:37] <mzanetti> Saviq: for example autmatic type coercion in case of a integer overflow can behave differently :D
[11:37] <mzanetti> good luck in hunting those
[11:38] <Saviq> mzanetti, not like int overflow is a normal situation, so ;)
[11:39] <mzanetti> Saviq: any ideas how to find more about this issue?
[11:39] <Saviq> mzanetti, about the CPU hogging you mean?
[11:39] <mzanetti> Saviq: yes
[11:39] <Saviq> mzanetti, not really, not unless we dig deep into v8 (which I don't think we will...)
[11:40] <tsdgeos> Saviq: fixme pushed
[11:40] <Saviq> mzanetti, it's something that our app lifecycle should cater for - make sure the shell and focused app(s) have enough memory
[11:41] <Saviq> tsdgeos, thanks
[11:41] <tsdgeos> mzanetti: that's why someone should start testing our apps with 5.2
[11:41]  * tsdgeos hides
[11:43] <Saviq> /food
[11:44] <tsdgeos> Saviq: the qt guy didn't answer to my review, i've asked around and he seems to be at a Blink meeting or soemthing making him a bit unreliable on reviewing so we may have to just go with the patch in our Qt and adapt later :-/
[11:44] <tsdgeos> what's your take?
[11:44] <tsdgeos> ah
[11:44] <tsdgeos> enjoy food :)
[11:44] <Saviq> tsdgeos, please do
[11:44] <Saviq> tsdgeos, just distro-patch our own
[11:45] <tsdgeos> ok, i'll open a bug and poke Mirv
[11:50] <Mirv> poke poke
[11:51] <mzanetti> anyone knows there the upstart user services are stored?
[12:10] <Cimi> mzanetti, help
[12:10] <Cimi> mzanetti, I fixed a part of the test which was failing
[12:11] <Cimi> mzanetti, but now it complains randomly of this
[12:11] <Cimi> http://paste.ubuntu.com/6179310/
[12:11] <Cimi> and sometimes another test
[12:12] <Cimi> if I run the test directly (just that test), no error
[12:13] <mzanetti> Cimi: well, the app_grid = ... went wrong
[12:13] <Cimi> mzanetti, yeah
[12:14] <Cimi> mzanetti, but I didn't change that in my hud tests
[12:14] <Cimi> mzanetti, maybe we need to add a wait?
[12:15] <mzanetti> Cimi: isn't there a 40 secs timeout already before selecting the app_grid?
[12:16] <Cimi> mzanetti, not that I can see in the code
=== jhodapp|afk is now known as jhodapp
[12:18] <Cimi> mzanetti, just pushed latest fixes lp:~cimi/unity8/unity8.hud-2_hint-reveal-commit
[12:18] <Cimi> mzanetti, but I have this app grid error
[12:21] <mzanetti> Cimi: I'm not sure any more if we wouldn't have been faster by writing those 5 tests in qml
[12:21] <Saviq> vila, any idea about https://code.launchpad.net/~mzanetti/unity8/launcherbackend-no-more-pinning/+merge/186603/comments/431293 ?
[12:22] <Saviq> vila, everything is SUCCESS, but the end result is FAILURE anyway?
[12:22] <mzanetti> Saviq: I think that's the one om26er posted before
[12:22] <mzanetti> Saviq: collecting artifacts went wrong
[12:22] <Saviq> mzanetti, there's more like that I'm afraid
[12:22] <Saviq> https://code.launchpad.net/~mterry/unity8/skip-intro-chevron/+merge/188182/comments/431253
[12:23] <vila> Saviq: https://jenkins.qa.ubuntu.com/job/unity8-ci/1189/console
[12:23] <Saviq> mzanetti, so it's not a one-time failure
[12:23] <mzanetti> Saviq: yeah... I assume the otto jobs have some new configuration which prevents just using the VM's as a drop in replacement
[12:24] <kgunn> mzanetti: just sent you a mail, on the latest proposed build Mirv is able to get 100% phenomena pretty easily it seems, you 2 might want to connect for some co-debug
[12:24] <Cimi> mzanetti, with hindsight...
[12:24] <Cimi> mzanetti, maybe yes maybe not
[12:24] <Cimi> mzanetti, anyway fix might be easy now, no?
[12:25] <mzanetti> kgunn: this looks like yet another one :/
[12:26] <pstolowski> tsdgeos: it fails for me after typing a few characters; I see reordering for first 1-3 letters, but after a while the Dash becomes empty and doesn't recover
[12:27] <pstolowski> tsdgeos: I'm running unity8 on the desktop
[12:29] <mzanetti> Cimi: I don't know whats happening and I'm sort of not in the mood to waste another 2 hours getting my device into a state where I can execute autopilot tests
[12:31] <mhr3> sil2100, thumbnailer in universe? ;)
[12:32] <dednick> Cimi: I've made some changed the tests/calendat. You may want to look at the changes; I had to make some fairly significant changes in the Calendar population code to fix some bugs I found.
[12:37] <pete-woods> Saviq, nic-doffay: there's a bug in the infographics in unity8, it doesn't handle the colors correctly on the first day of the month (see Greeter/Gradient.js, it doesn't handle the case where total = 0)
[12:37] <mzanetti> Mirv: do you have a way to reproduce the CPU hogging issue?
[12:40] <pete-woods> nic-doffay: I fixed it like this myself (http://pastebin.ubuntu.com/6179422/)
[12:41] <mzanetti> dafuq... we have running 32 threads when running with mir
[12:41] <Mirv> mzanetti: I basically just boot my device. let me try again, for the last reboots it happened every time.
[12:42] <mzanetti> Mirv: yeah... I can see it too.. but it seems a different one that the one I was investigating
[12:42] <Mirv> mzanetti: you were investigating the blank cpu usage issue?
[12:42] <mzanetti> Mirv: yeah
[12:43] <Saviq> dednick, did you see me asking about tracking down https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1233245 ?
[12:43] <ubot5> Ubuntu bug 1233245 in unity8 (Ubuntu) "Volume up/down keys not working in music-app with Mir" [High,Triaged]
[12:44] <dednick> Saviq: nope
[12:44] <dednick> i do now
[12:45] <dednick> Saviq: will take a look after lunch.
[12:45] <Saviq> dednick, thanks
[12:47] <mzanetti> Saviq: could be the same as the HUD button. (if we have something like inputfilterarea for kbd too)
[12:48] <Saviq> mzanetti, it doesn't work when shell's on top either
[12:48] <mzanetti> ok
[12:48] <Saviq> mzanetti, so I'm feeling we don't get VolumeUp/Down from Mir at all
[12:48] <mzanetti> yeah, makes sense
[12:49] <mzanetti> Mirv: this looks a bit suspicious http://paste.kde.org/pf955fae0
[12:50] <Saviq> tsdgeos, https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-saucy/1962/testReport/junit/%28root%29/ListViewWithPageHeaderTest/testMaximizeVisibleAreaMoveUpAndShowHeader/ looks familiar, eh?
[12:54] <kgunn> dandrader: hey check out mzanetti 's pastebin
[12:54] <dandrader> kgunn, what about it?
[12:55] <kgunn> dandrader: just wonder if you might have some thots about it, its cpu pegging...looks android input related
[12:55] <kgunn> wondered if you might have thots on it as you've been doing some excavation in that area
[12:56] <dandrader> kgunn, you mean mir is consuming 100% of cpu time constantly?
[12:56] <fginther> Saviq, pong
[12:57] <Saviq> fginther, sooo....
[12:57] <Saviq> fginther, the switch to vm didn't really work :/
[12:57] <Saviq> fginther, http://s-jenkins:8080/job/unity8-autolanding/516/console
[12:57] <Saviq> fginther, artifact collection failed
[12:57] <dandrader> kgunn, mzanetti   that stack trace shows that the dispatcher thread is waiting to be woken up to dispatch the next pending events that came from the InputReader
[12:57] <fginther> Saviq, yeah, I noticed that too. I patched it just a moment ago
[12:58] <fginther> the current autolanding job should catch the update
[12:58] <Saviq> fginther, to fix the runs on otto, we'd just need to upgrade libunity9 there
[12:58] <pete-woods> Saviq, nic-doffay: I raised a ticket for the bug I mentioned earlier (https://bugs.launchpad.net/unity8/+bug/1233657)
[12:58] <Saviq> fginther, and well, make mhr3 and/or pstolowski fix the deps on libunity9
[12:58] <ubot5> Ubuntu bug 1233657 in Unity 8 "[greeter] Infographic colors are black on the first day of the month" [Undecided,New]
[12:58] <fginther> Saviq, would unity need to restart?
[12:58] <fginther> Saviq, unity 7 that is
[12:58] <Saviq> fginther, probably
[12:58] <Saviq> fginther, but maybe let's fix it properly instead
[12:59] <dandrader> kgunn, mzanetti so it's a pretty normal state to have a thread waiting there
[12:59] <Saviq> fginther, can you let mhr3 and pstolowski know which versions of libunity9 and libunity-protocol-private0 are installed on there?
[12:59] <fginther> Saviq, sure, one moment
[12:59] <mzanetti> dandrader: maybe... I just attached gdb while it was spinning cpu and this one looked the most like the one waking up too often
[13:00] <Saviq> tsdgeos, https://code.launchpad.net/~aacid/unity8/enable_height_animation_when_visible/+merge/188536/comments/431335
[13:00] <tsdgeos> Saviq: damn 0 != almost0
[13:00] <Saviq> tsdgeos, yeah
[13:00] <fginther> Saviq, pstolowski: http://10.97.2.10:8080/job/autopilot-testrunner-otto-saucy/534/artifact/results/sysinfo/dpkg-l.postsetup/*view*/
[13:00] <dandrader> mzanetti, it would be pretty easy to check it out. just add a printf there and see if its flooding the terminal
[13:01] <fginther> Saviq, pstolowski, package versions in use at otto run time
[13:01] <tsdgeos> Saviq: sure, you can still see the animation if you're unlucky enough that it takes too much to "create" and when created it's already inside the view
[13:02] <Saviq> tsdgeos, so I'm unlucky enough to easily trigger that
[13:02] <tsdgeos> i guess
[13:02] <tsdgeos> nexus4 or galaxynexus?
[13:02] <Saviq> tsdgeos, n4
[13:02] <Saviq> tsdgeos, we could either have the Behavior enabled/disabled, or just have an arbitrary Animation that we trigger on transitions
[13:02] <Saviq> tsdgeos, feels like a more robust solution
[13:03] <dandrader> mzanetti, so, that thread is woken up for two reasons: 1 - new input events coming  2 - some client has sent a reply
[13:03] <tsdgeos> Saviq: what do you mean on transitions?
[13:03] <Saviq> tsdgeos, onAdd, onRemove
[13:04] <pstolowski> fginther, Saviq : thanks. so the versions of libunity and proto are in sync? confused.
[13:04] <Saviq> pstolowski, on *start*
[13:04] <Saviq> pstolowski, but then http://10.97.2.10:8080/job/autopilot-testrunner-otto-saucy/534/consoleFull
[13:04] <pstolowski> Saviq: aah
[13:04] <Saviq> pstolowski, protocol-private gets upgraded to 7.1.1+13.10.20130927-0ubuntu1
[13:04] <fginther> pstolowski, sorry, that wasn't the full details, there is more in the console log
[13:04] <Saviq> pstolowski, but libunity9 remains
[13:04] <Saviq> mhr3, ↑
[13:04] <tsdgeos> Saviq: onAdd of what?
[13:05] <Saviq> tsdgeos, there are add and remove transitions in RunningApplicationsFilterGrid
[13:05] <Saviq> or well, there's a move only
[13:05] <nic-doffay> pete-woods, thanks I'll sort that out today.
[13:06] <mzanetti> paulliu: hey. what's the status on that movie preview url? can we put the merge up for review?
[13:06] <pstolowski> Saviq: ack
[13:06] <Saviq> tsdgeos, what I mean is: we only want the behavior to be active when there are items added or removed to the GridView, right?
[13:06] <pstolowski> mhr3: you've a fix for the control file, don't you?
[13:06] <Saviq> fginther, can we just dist-upgrade the machines or is that too risky?
[13:07] <Saviq> tsdgeos, not even when the whole category gets added/removed, 'cause that transition should be handled by LVWPH for all categories added/removed, right?
[13:08] <tsdgeos> LVWPH doesn't do transitions
[13:08] <Saviq> tsdgeos, I know, not yet
[13:08] <tsdgeos> that a hell lot of code i did not do
[13:08] <Saviq> tsdgeos, so what I mean is we should trigger the animation onCountChanged, effectively
[13:08] <Saviq> tsdgeos, and only then
[13:08] <tsdgeos> i'm afraid
[13:08] <tsdgeos> that since you're creating it
[13:08] <tsdgeos> it'll ahve a count changed
[13:08] <tsdgeos> i'll check
[13:09] <Saviq> tsdgeos, yeah, just have a look, the current solution isn't working good enough I'm afraid
[13:09] <tsdgeos> Saviq: btw are you sure it's animating?
[13:09] <Saviq> tsdgeos, yes, can make a video if you want
[13:09] <Saviq> tsdgeos, but I can clearly see it uncover the last items
[13:09] <tsdgeos> can you put a onHeightChanged in teh RunningApplicationsGrid.qml ?
[13:10] <Saviq> tsdgeos, easiest if you launch at least 4 apps
[13:10] <tsdgeos> because honestly i don't see how you could be that unlucky that many times easily
[13:10] <fginther> Saviq, that should be ok, otto uses an lxc container, should get a fresh one each time
[13:11] <Saviq> fginther, that would solve the issue for us without waiting for the fix in libunity
[13:11] <Saviq> fginther, and we could get back to otto mediumtests
[13:11] <tsdgeos> ok, let's see if i can repro with 4 apps
[13:12] <paulliu> mzanetti: wait, I'm checking why moviepreview doesn't have any imaegs..
[13:12] <Saviq> tsdgeos, http://pastebin.ubuntu.com/6179535/
[13:13] <tsdgeos> Saviq: ah
[13:13] <tsdgeos> obviously
[13:13] <tsdgeos> silly thing that is created small
[13:13] <Saviq> tsdgeos, yeah, gets on screen and then animates
[13:14] <tsdgeos> together with unity taking 100%
[13:14] <tsdgeos> obviously doesn't help to make stuff smooth
[13:14] <mhr3> pstolowski, yea
[13:14] <fginther> Saviq, let me give that a try, just need to make a test job
[13:14] <mhr3> pstolowski, so what's the issue?
[13:14] <Saviq> fginther, thanks
[13:14] <tsdgeos> someone should investigate why the hell unity+mir takes 100% cpu
=== _salem is now known as salem_
[13:14] <pstolowski> mhr3: can you MP it?
[13:14] <Saviq> fginther, so you want to dist-upgrade on each job? shouldn't we just dist-upgrade once in a while? like now?
[13:14] <tsdgeos> anyhow yes, let me find if using count changed we can do better
[13:15] <tsdgeos> Saviq: you on mir or on sf?
[13:15] <Saviq> tsdgeos, sf
[13:15] <Saviq> mhr3, libunity-protocol-private0 gets upgraded without libunity9
[13:15] <tsdgeos> ok, so can't blame unity taking 100% of cpu, because yours probably isn't :D
[13:15] <Saviq> lol
[13:16] <mhr3> Saviq, why?
[13:16] <fginther> Saviq, shouldn't it be both? One to update the container each day/week and the one inside the job to update it for each run
[13:16] <Saviq> fginther, we shouldn't need the per-job update
[13:17] <Saviq> fginther, everything should be handled by deps - and when it isn't (like now)
[13:17] <Saviq> fginther, it's a bug
[13:17] <Saviq> mhr3, you tell me
[13:17] <fginther> Saviq, ack, that would mask the deb bug
[13:17] <Saviq> mhr3, libunity-core pulls in protocol-private? but that doesn't pull in libunity9?
[13:17] <paulliu> mzanetti: https://code.launchpad.net/~paulliu/unity8/fix-genericpreview-2/+merge/188594
[13:18] <mhr3> hmmm... :/
[13:18] <mhr3> dunno really
[13:18] <Saviq> mhr3, libunity-core-6.0-8 depends on libunity-protocol-private0 (>= 7.1.1+13.10.20130927)
[13:18] <mhr3> Saviq, actually... yea, cause libunity is client lib, protocol is protocol
[13:19] <Saviq> mhr3, well, *something* is broken, so that we're getting "no such key in gsettings: hidden-scopes"
[13:19] <Saviq> mhr3, pstolowski suspected protocl-private0
[13:22] <pstolowski> Saviq, mhr3 : yes, because protocol now uses hidden-scopes key (which introduces a dependency on libunity which we didn't have before at all?)
[13:23] <Saviq> pstolowski, sounds like a problem
[13:23] <mhr3> eeeh :/
[13:24] <mhr3> so proper fix is to split out the pkg with the gschema and make the key optional inside protocol
[13:24] <mhr3> s/and/or/
[13:25] <pstolowski> mhr3: testing your fix
=== MacSlow|lunch is now known as MacSlow
[13:26] <pstolowski> tsdgeos: any luck with reproducing reordering issues?
[13:27] <tsdgeos> pstolowski: got busy again with the stuff i was previously working on, sorry
[13:27] <mhr3> pstolowski, i don't it fixes it really
[13:27] <mzanetti> paulliu: why the item at the bottom?
[13:28] <paulliu> mzanetti: it prints out that the item is missing?
[13:28] <paulliu> mzanetti: is that ok?
[13:28] <mzanetti> paulliu: my mistake there. I removed it by now in the main branch
[13:29] <mzanetti> paulliu: if you merge with the main branch that message should be gone
[13:29] <paulliu> mzanetti: ok
[13:29] <pstolowski> mhr3: shouldn't protocol depend on libunity?
[13:29] <mhr3> pstolowski, no
[13:29] <mzanetti> paulliu: so I'd say remove that item again and you can push this change directly to the main branch
[13:29] <mhr3> libunity is built on top of protocol, not the other way around
[13:29] <Saviq> mzanetti, going into a mtg, won't be there for standup
[13:29] <mzanetti> Saviq: ack
[13:30] <nic-doffay> Saviq, 359	+ pageHeader.triggerSearch()
[13:30] <nic-doffay> 360	+ typeString("test")
[13:30] <nic-doffay> Why is that needed? Shouldn't you triggerSearch before you start testing for visibility altogether?
[13:31] <nic-doffay> Why is what needed exactly? Could you elaborate on this a bit?
[13:31] <Saviq> nic-doffay, before .triggerSearch, the search entry is offscreen
[13:31] <mzanetti> dednick: standup?
[13:32] <Saviq> nic-doffay, so feels like you should get it onscreen first
[13:32] <Saviq> nic-doffay, but no need to type anything
[13:32] <paulliu> mzanetti: ok.. pushed.
[13:32] <mhr3> pstolowski, updated the branch, has proper fix now
[13:32] <pstolowski> mhr3: I know, but this change alone won't trigger update of protocol
[13:33] <mzanetti> paulliu: thanks.
[13:33] <mzanetti> pstolowski: do you think it's good to go now?
[13:34] <mzanetti> sorry... paulliu ^
[13:34] <paulliu> mzanetti: yes.
[13:34] <mzanetti> perfect, thanks
[13:34] <pstolowski> mhr3: the schema *is* installed by an old libunity, so it will still fail on hidden-scopes, no?
[13:35] <nic-doffay> Saviq, righto
[13:36] <mhr3> pstolowski, well yea
[13:37] <mhr3> pstolowski, then we need to make it conflict with old version of libunity i guess
[13:37] <pstolowski> mhr3: the problem as I understand it from the logs is: both libunity and protocol are initially in sync, but then an update of unitycore pulls newer protocol which needs hidden-scopes. but this update doesn't pull matching libunity
[13:38] <dandrader> mzanetti, I wonder if running unity8-mir with perf might help you identifying the hot spot
[13:38] <pstolowski> mhr3: yes, conflict rule would do
[13:39] <mzanetti> dandrader: did you manage to get perf running on the phone?
[13:39] <dandrader> mzanetti, didn't try yet. was playing with it on my test laptop
[13:39] <dandrader> mzanetti, it doesn't work on the phone?
[13:39] <mzanetti> dandrader: I couldn't get it to run there
[13:40] <dandrader> :(
[13:40] <mzanetti> dandrader: well, give it a shot. maybe its just me
[13:40] <mhr3> pstolowski, so just a conflicts?
[13:40] <Cimi> mzanetti, to have assert that to emulators (dash.py), shall I have to write it in try except?
[13:41] <Cimi> mzanetti, because Hud class doesn't have that function I think
[13:41] <mhr3> pstolowski, isn't that usually used with the conflicts+replaces+provides?
[13:41] <pstolowski> mhr3: you got me wondering.. would conflicts alone trigger an update of libunity?
[13:42] <mzanetti> Cimi: hmm... let me have a look
[13:42] <Cimi> mzanetti, I want to wait there...
[13:42] <Cimi> until its not none
[13:42] <didrocks> kgunn: greyback: if you have the fix in trunk before EOD, just ping us, so that we can get it in ASAP
[13:43] <greyback> didrocks: ok
[13:43] <Cimi> mzanetti, like __init__.py does
[13:43] <kgunn> didrocks: much appreciate that offer - thanks for the help
[13:43] <Cimi> mzanetti,     def get_dash(self):      dash = self._proxy.select_single(Dash)        self.assertThat(dash, NotEquals(None))        return dash
[13:43] <mzanetti> Cimi: I think you can use someVariable.waitFor(condition)
[13:43] <mzanetti> Cimi: oh. this won't work
[13:43] <didrocks> yw!
[13:43] <mzanetti> Cimi: you need some lambda function, let me search for an example
=== dandrader is now known as dandrader|afk
[13:46] <mzanetti> Cimi: try something like this: http://paste.ubuntu.com/6179651/
[13:48] <Saviq> dednick, when back - I've actually pushed the bug up to mir, we're not getting the key events at all
[13:50] <pstolowski>  mhr3 what if protocol recommends libunity=binaryversion ?
[13:50] <mhr3> pstolowski, eek, no reverse deps pls
[13:52] <dednick> Saviq: yay
[13:52] <Cimi> mzanetti, wait_for doesn't exist
[13:52] <Cimi> http://paste.ubuntu.com/6179669/
[13:52] <Saviq> dednick, lol :D
[13:52] <dednick> now i dont have to figure out how to enable mir
[13:53] <Cimi> AttributeError: 'function' object has no attribute 'wait_for'
[13:53] <dednick> probably should do that anyway
[13:53] <mzanetti> Cimi: this is in emulators/hud.py, right?
[13:53] <Cimi> mzanetti, dash.py
[13:53] <mzanetti> Cimi: check out hud.py. there are some wait_fors
[13:54] <Cimi> mzanetti, but those are for properties
[13:54] <Cimi> mzanetti, not lambda functions
[13:54] <mzanetti> Cimi: yeah... I hoped it would work
[13:54]  * mzanetti isn't a python expert either
[13:55] <Saviq> fginther, whoa, 1:03 for amd64 package build? http://10.97.2.10:8080/job/unity8-saucy-amd64-autolanding/102/
[13:56] <mzanetti> Cimi: ah.. I think you can use "from testtools.matchers import Equals, Eventually" and then try that with the lambda function
[13:56] <mzanetti> Cimi: just the assertThat() wouldn't work, but you probably shouldn't use that in emulators anyways
[13:59] <Cimi> mzanetti, this lambda doesn't work...
[14:00] <tsdgeos> Saviq: there?
[14:00] <Saviq> tsdgeos, here
[14:00] <Cimi> mzanetti, or you meant without wait for?
[14:00] <Cimi> mzanetti, ah use eventually not equals?
[14:00] <tsdgeos> Saviq: so i can probably use countChanged, but not for the "creation" scenario (i.e. first app launched) because i don't get a count changed there
[14:00] <tsdgeos> are we ok with that regression?
[14:00] <tsdgeos> probably not, right?
=== dandrader|afk is now known as dandrader
[14:01] <Saviq> tsdgeos, ideally no, but you should be able to rework it somehow
[14:01] <Saviq> tsdgeos, by having the behavior on on creation, and disabling it as soon as it finishes for the first time
[14:01] <Saviq> tsdgeos, WDYT?
[14:02] <mzanetti> Cimi: check out tests/test_notifications.py
[14:02] <tsdgeos> Saviq: but next time it'll also be "onCreation"
[14:02] <mzanetti> Cimi: we use that lambda stuff there
[14:02] <mhr3> pstolowski, https://code.launchpad.net/~mhr3/libunity/same-ver-dep/+merge/188596
[14:02] <Saviq> tsdgeos, right <facepalm>
[14:02] <tsdgeos> that's what makes it so hard
[14:02] <Cimi> mzanetti, but with assertThat
[14:03] <Saviq> tsdgeos, so we're back to square one...
[14:03] <mzanetti> Cimi: yeah... doesn't Eventually() work without it?, mabe something like Eventually(get_dash) != NoneType or something like that?
[14:03]  * Cimi tries until it works
[14:03] <fginther> Saviq, that hurts, the builders have just been kept fully loaded the past week. probably doesn't help build times (some help is in the works, but no eta)
[14:04] <Saviq> fginther, k
[14:04] <mzanetti> Cimi: if you can't get it to work you probably can write your own loop
[14:04] <Saviq> mzanetti, Cimi what's going on?
[14:04] <tsdgeos> Saviq: i guess we somehow need to throw DashApps.qml into the equation
[14:04] <tsdgeos> since there the model is being kept all the time
[14:04] <tsdgeos> so we know if the change to 1 is coming from 0 or not
[14:05] <tsdgeos> but it's all so decoupled that it's not trivial
[14:05] <Cimi> Saviq, I don't know :) some autopilot tests fail if you run them in sequence but not if you run them separately
[14:05] <Saviq> tsdgeos, or well, how about actually looking at the geometry?
[14:05] <Cimi> Saviq, might be a race or so, I'm trying to add protections
[14:05] <tsdgeos> Saviq: that's the same, there's no difference between creation because you had 0 and now have 1
[14:05] <tsdgeos> and you had 1 and were out of view and came back
[14:06] <mzanetti> Cimi: something like this might help: http://paste.ubuntu.com/6179709/
[14:06] <Saviq> tsdgeos, yeah, we want transitions in LVWPH don't we ;)
[14:06] <Cimi> mzanetti, I think i'll do something like that
[14:06] <tsdgeos> Saviq: we do :D
[14:07] <Saviq> tsdgeos, that would be it - just an add: transition
[14:07] <Saviq> tsdgeos, and behavior onCountChanged
[14:07] <Saviq> tsdgeos, but yeah, I know...
[14:07] <tsdgeos> transitions are ahrd
[14:07] <tsdgeos> not going to happen for this bug
[14:08] <ritz_> Hi, wrt https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1051921 . Where did this commit go ?
[14:08] <ubot5> Ubuntu bug 1051921 in Unity 5.0 "lens-bar-keynavigation periodically writes to /tmp/wut.png" [Medium,Fix committed]
[14:09] <pstolowski> mhr3: top-approving, ok?
[14:09] <mhr3> pstolowski, yep, that's why i pinged
[14:12] <ritz_> okay, nm. unity uses bzr branch
[14:14] <nic-doffay> Saviq, mind taking another look at the search indicator branch?
[14:14] <Saviq> nic-doffay, will do
[14:15] <ritz_> bregma, hi, https://launchpad.net/unity/+milestone/5.26.0 . when is this scheduled to be publish ?
[14:16] <pstolowski> Saviq, fginther : so the fix for libunity should be merged soon
[14:16] <bregma> ritz_, are you looking for when Ubuntu will release a new update for 12.04 LTS?
[14:16] <ritz_> yes
[14:16] <Saviq> pstolowski, yeah, need to release it first :
[14:16] <Saviq> pstolowski, :/
[14:16] <Saviq> pstolowski, since we're not in the same stack
[14:17] <ritz_>  wrt https://bugs.launchpad.net/ubuntu/+source/unity/+bug/1051921 .
[14:17] <ubot5> Ubuntu bug 1051921 in Unity 5.0 "lens-bar-keynavigation periodically writes to /tmp/wut.png" [Medium,Fix committed]
[14:17] <ritz_> I see this is marked for 5.26, with fix committed
[14:18] <ritz_> I am unable to locate where they are committed
[14:23] <pstolowski> mzanetti: ping
[14:23] <mzanetti> pstolowski: hi
[14:24] <pstolowski> mzanetti: forgot yo tell you; my branch with preview cancellation was merged a few days ago; it also adds 2 cancel methods for programatic cancellation of preview and actions from the shell
[14:25] <mzanetti> pstolowski: ah, cool. will add that
[14:26] <dholbach> hiya
[14:26] <dholbach> how much work would it be to offer a unity8 desktop session?
[14:28] <tsdgeos> Mirv: https://bugs.launchpad.net/ubuntu/+source/qtdeclarative-opensource-src/+bug/1233705
[14:28] <ubot5> Ubuntu bug 1233705 in qtdeclarative-opensource-src (Ubuntu) "Add patch to fix loading of ListView items in some corner cases" [Undecided,New]
[14:33] <vila> Saviq: fginther explained, I can answer now.
[14:33] <vila> Saviq: on the otto job pages, there is an 'artifact' url
[14:34] <vila> Saviq: err, 'Build Artifacts'
[14:34] <vila> Saviq: follow 'results/sysinfo' and you'll find dpkg-l.boot and dpkg-l.possetup
[14:35] <vila> Saviq: that should give you the versions you're searching for
[14:35] <Saviq> vila, yeah, fginther pointed us there
[14:35] <vila> Saviq: ha, good ;)
[14:35] <tsdgeos> pstolowski: ok, on yout bug, need to leave the other one for a while
[14:35] <vila> Saviq: I was making sure the info reached you ;)
[14:35] <tsdgeos> pstolowski: so what do you search in the home scope, random stuff?
[14:35] <pstolowski> tsdgeos: k
[14:36] <tsdgeos> pstolowski: ok, got it
[14:36] <tsdgeos> this is really weird
[14:36] <tsdgeos> and obviously i lost the modelchecker
[14:37] <tsdgeos> pstolowski: you won't have it around by chance, no?
[14:37] <Cimi> mzanetti, finally works for me...
[14:37] <Cimi> mzanetti, pushed
[14:37] <pstolowski> tsdgeos: yes, just slowly type >=3 random letters
[14:38] <mzanetti> Cimi: cool
[14:41] <Cimi> dednick, I'm back hopefully
[14:41] <Cimi> dednick, where are the changes?
[14:43] <dednick> Cimi: ah, forgot to push. it's up now. https://code.launchpad.net/~nick-dedekind/ubuntu-settings-components/pre-archive-fixes/+merge/188283
[14:43] <dednick> r58/59 have the changes i made
[14:43] <dednick> Cimi: ^
[14:43] <Cimi> dednick, ok
[14:49] <pstolowski> tsdgeos: I should have model checker somewhere, let me try to find it
[14:50] <Cimi> dednick, quick look seems nice
[14:50] <Cimi> dednick, does it work with infinite loop?
[14:51] <dednick> Cimi: you mean keep scrolling?
[14:51] <mzanetti> pstolowski: calling Scope::cancelActiviation() when I close the preview should be enough, right?
[14:51] <Cimi> dednick, minimum and maximum date could be empty on purpose
[14:52] <dednick> Cimi: ah, it does, but should probably add a test case for it
[14:53] <pstolowski> mzanetti: Scope::cancelActivation will cancel preview request; Preview::cancelAction will cancel any actions; you should call both when closing preview
[14:54] <dednick> Cimi: done and pushed.
[14:56] <mzanetti> pstolowski: what does cancel any actions mean? If I start installing an app and close the preview, will that cancel the isntallation too?
[14:56] <nic-doffay> Anyone got time for a review? https://code.launchpad.net/~nicolas-doffay/unity8/infographic-colour-fix
[14:57] <pstolowski> tsdgeos: your pastebin - http://paste.ubuntu.com/6044489/
[14:59] <pstolowski> mzanetti: no it won't
[15:01] <mhr3> mzanetti, paulliu, i'm thinking about removing imageSourceUri prop from previews, we'll internally handle it and just return a thumbnailer uri for preview.image
[15:01] <mhr3> ok with that?
[15:01] <mzanetti> fine with me
[15:02] <mzanetti> mhr3: what does that mean for our branch?
[15:02] <mhr3> mzanetti, are you using it?
[15:02] <mhr3> cause you shouldn't really
[15:02] <tsdgeos> pstolowski: great
[15:02] <mzanetti> mhr3: yeah, paulliu just added that
[15:03] <mhr3> mzanetti, remove it then :)
[15:03] <mhr3> my original review was already saying that the playable property shouldn't be there
[15:03] <mzanetti> mhr3: ah no... I was wrong... its just a similar name
=== dandrader is now known as dandrader|lunch
[15:04] <mzanetti> mhr3: yeah... that's gone
[15:04] <mhr3> mzanetti, not pushed?
[15:04] <mhr3> still see it in https://code.launchpad.net/~unity-team/unity8/fix-genericpreview/+merge/188355
[15:04] <mzanetti> mhr3: ah right... there it is still. but we need to know that, no?
[15:05] <mzanetti> mhr3: we still need to place the play button there
[15:05] <mhr3> the play button should be there always
[15:06] <greyback> mzanetti: ever seen this error in AP: http://pastebin.ubuntu.com/6179755/
[15:06] <mzanetti> mhr3: ok. fixing
[15:06] <mzanetti> greyback: uh... no, not really
[15:07] <greyback> mzanetti: I've seen similar when Qt wasn't finding the path where the QPA plugins are put.
[15:07] <mhr3> mzanetti, https://code.launchpad.net/~paulliu/unity8/movie-preview/+merge/181856/comments/426394
[15:07] <mzanetti> mhr3: cheers. will fix
[15:07] <mhr3> thanks
[15:11] <mzanetti> mhr3: is line 771 here ok? https://code.launchpad.net/~unity-team/unity8/fix-genericpreview/+merge/188355
[15:12] <mhr3> mzanetti, the entire block should be just one openUrlExternally()
[15:12] <mhr3> no need for those ifs
[15:13] <mzanetti> ack... sounds sane to me
[15:14] <mzanetti> mhr3: ok. all fixed
[15:15] <mhr3> mzanetti, \o/
[15:31] <mzanetti> Cimi: added some comments
[15:31] <Cimi> mzanetti, ok thx
[15:38] <tsdgeos> oh man
[15:39] <tsdgeos> guess what http://paste.kde.org/~tsdgeos/p8a902715/ returns
[15:39] <Cimi> mzanetti, so what's the type for pixel value?
[15:40] <Cimi> mzanetti, thought was double
[15:40] <mzanetti> Cimi: afaik he didn't invent half pixels yet ;)
[15:40] <mzanetti> Cimi: I'd say int
[15:40] <Cimi> mzanetti, well but x and y are real
[15:40] <tsdgeos> well you can use half pixels in x, y and width
[15:40] <Cimi> in qml
[15:40] <Cimi> iirc
[15:41] <Cimi> not int
[15:41] <mzanetti> are they?
[15:41] <mzanetti> hmm.. could be
[15:41] <tsdgeos> it just introduces slowlyness and "antialiasing"
[15:41] <tsdgeos> to give you half pixel
[15:42] <mzanetti> tsdgeos: ture, false ?
[15:43] <tsdgeos> mzanetti: 'xctily
[15:43] <mzanetti> tsdgeos: a.value<QIcon>() == b.value<QIcon>() should return true again
[15:43] <tsdgeos> sure
[15:44] <tsdgeos> but that's the silliness in it
[15:44] <mzanetti> heh
[15:44] <tsdgeos> i've a variant and i want to compare it
[15:44] <mzanetti> well, it can't automatically convert QIcon
[15:44] <tsdgeos> sure
[15:44] <tsdgeos> that's why they have that sofisticate method to registerhandlers
[15:44] <tsdgeos> just they decided not to use it
[15:44] <mzanetti> hah
[15:44] <tsdgeos> because "who cares"
[15:50] <mterry> greyback, I'm looking at a bug where the greeter pauses in its animation after blanking the screen.  Does Mir or unity-mir put unity8 to sleep somehow when the screen is off?
[15:51] <greyback> mterry: I think if screen is blanked, the frame swapper in Mir halts, preventing Qt from rendering new frames
[15:52] <mterry> greyback, ok, makes sense.  I think I can work around the issue by not animating at all
[15:52] <mzanetti> Cimi: can you help me out with a nice screenshot of QtCreator with the ubuntu integration?
[15:52] <greyback> mterry: I'd try the same, yeah
[15:52] <Saviq> greyback, mterry, but that shouldn't impact the animation, should it... the only artifact would be that you might see the old frame and then the new one straight away
[15:53] <Saviq> mterry, greyback, the animation itself should pause, should it...
[15:53] <Saviq> +not
[15:53] <Cimi> mzanetti, what you need?
[15:53] <Saviq> mterry, but yeah, it should never animate when locking
[15:53] <mzanetti> Cimi: a few screenshots that highlight the Ubuntu integration in QtCreator
[15:54] <mzanetti> Cimi: maybe the welcome page, then the device page, or the doc page...
[15:54] <Cimi> mzanetti, when you need?
[15:54] <mzanetti> Cimi: asap :/
[15:54] <mzanetti> read: now
[15:54] <Cimi> mzanetti, I can take it in 1366x768
[15:55] <mzanetti> Cimi: hmm... don't mind the size... make them smaller if possible
[15:55] <mzanetti> Cimi: they should fit on a presentation
[15:55] <mzanetti> so we need to scale them down if they are too large. better resizing the window to only show the relevant content
[15:56] <mzanetti> QtCreator looks like crap on my retina screen, that's why I can't make them myself
[15:56] <Cimi> mzanetti, I am taking them through parallels
[15:56] <Cimi> mzanetti, so I can decide all the res I want
[15:56] <mzanetti> Cimi: ah, cool.
[15:56] <greyback> racarr: can you confirm? If screen blanked, then all clients are blocked from swapping frames - so if a client was animating, screen blanks, pause, screen unblanks, the client resumes animation from before the blank
[15:56] <mzanetti> Cimi: well, I'm sure you got the right eye for this ;)
[15:57] <Cimi> mzanetti, well I'm simply taking screen hot of the out of the box experience
[15:57] <Cimi> mzanetti, guest
[16:01] <fginther> Saviq, Here's a unity8 test run after dist-upgrade: http://10.97.2.10:8080/job/autopilot-testrunner-otto-saucy-fjg/71/
[16:02] <fginther> Saviq, looks like something is still missing
[16:02] <fginther> Saviq, possibly a unity 7 restart?
[16:02] <Saviq> fginther, lookin'
[16:02] <fginther> Saviq, this was a one-off test
[16:03] <Saviq> fginther, huh!
[16:03] <Saviq> fginther, ah, that's a live dist-upgrade is it?
[16:03] <fginther> Saviq, it was also tested from an MP in progress, not trunk
[16:04] <fginther> Saviq, I did the dist-upgrade as a separate task prior to the test
[16:04] <Saviq> fginther, 'cause http://10.97.2.10:8080/job/autopilot-testrunner-otto-saucy-fjg/71/artifact/results/sysinfo/dpkg-l.postsetup/*view*/ says libunity and friends are still at versions from August?
[16:04] <Saviq> fginther, then yeah, possible we need a restart of unity7, as the lib is probably not loaded into memory again after being upgraded
[16:04] <Saviq> fginther, and then unity8 uses the old version, preloaded by unity7
[16:04] <fginther> Saviq, ack, I'll keep working on it
[16:04] <Saviq> MacSlow, fginther is on them ↑
[16:04] <mhr3> that doesn't sound very likely
[16:05] <Saviq> mhr3, what else? still getting "no key hidden-scopes"
[16:05] <mhr3> if the pkg was updated, the schema would be as well, it has nothing to do with the libunity .so
[16:05] <MacSlow> Saviq, fginther: I see
[16:06] <Saviq> mhr3, right...
[16:06] <Saviq> fginther, hmm but then... the artifacts still mention August version...
[16:06] <Saviq> fginther, where is the dist-upgrade log for that job?
[16:06] <fginther> Saviq, sorry, it's lost
[16:06] <mzanetti> Saviq: how is your review load? still high?
[16:06] <MacSlow> Saviq, fginther: any well-educated guess how long it'll take until it works again?
[16:06] <mterry> greyback, so do you know what's happening in Qt behind the scenes?  Like, it can't put any more frames into Mir, so it knows to pause all the timers?  I'm testing not doing any animation and when we come back, I still see a frame of the old content, which means Mir is shutting down frames sooner than unity8 responds to the event  :(
[16:07] <Saviq> mzanetti, how do you think ;)
[16:07] <Saviq> fginther, but if it's *before* that job
[16:07] <Saviq> fginther, shouldn't the dpkg log show the status after dist-upgrade?
[16:07] <mzanetti> Saviq: dunno... are quite a lot "approved, ready to land" and many are "WIP" still
[16:07] <Saviq> mzanetti, wassup?
[16:07] <fginther> Saviq, ah, the dist-upgrade wasn't done as part of the job itself. I can insert one in
[16:08] <mzanetti> just wanted to say that the preview branches are both ready for review now.
[16:08] <Saviq> fginther, wherever it was made, it doesn't seem to have worked
[16:08] <Saviq> mzanetti, yeah, saw that and they're on my queue
[16:08] <fginther> Saviq, hmm, ok I'll see what I can find
[16:09] <greyback> mterry: you'll need to talk to some of the mir team for details. I think when screen blanked, the Mir compositor stops. When that happens, clients frame swapping is blocked, as the compositor is not requesting new frames from the client
[16:10] <Saviq> fginther, what I mean - http://10.97.2.10:8080/job/autopilot-testrunner-otto-saucy-fjg/71/artifact/results/sysinfo/dpkg-l.postsetup/*view*/ shows versions of libunity from August
[16:10] <kgunn> mterry: share that in #ubuntu-mir
[16:10] <mterry> kgunn, yup
[16:13] <fginther> Saviq, sorry, I'm not sure what happened, looking more carefully this time
[16:14] <Cimi> mzanetti, I sent something
[16:14] <Cimi> mzanetti, not sure you received one or two mails, pick the one with the cropped imaged in case
[16:14] <Saviq> mhr3, re: use-thumbnailer - you really think we should be doing this at this level? shouldn't either a) the renderer be able to do its thing or b) the scope be responsible for it?
[16:16] <Saviq> mhr3, also, did you decide with Jussi that there's no need for url encoding the path etc? just take file:/// and replace with image://thumbnail/?
[16:17] <mhr3> Saviq, no strong preference here, jussi wanted it at this level, i don't see why not, having it this central is quite nice
[16:18] <mhr3> Saviq, i have my doubts about the encoding, but apparently i didn't convince him why is it necessary
[16:20] <fginther> MacSlow, I'm shooting for a fix by tomorrow. The fallback to the VM test should work, but I haven't seen a run make it all the way through yet. will keep monitoring
[16:21] <MacSlow> fginther, ok thanks
[16:23] <Saviq> mhr3, I think there's one ~simple reason - it should (could) support other schemes than just file://
[16:23] <nic-doffay> pete-woods, mind reviewing that fix for me?
[16:23] <mhr3> Saviq, think i didn't mention that? :P
[16:23] <mhr3> Saviq, he's no easy to convince
[16:23] <Saviq> mhr3, right
[16:28] <pete-woods> nic-doffay: sure
[16:33] <Cimi> mzanetti, received the mail?
[16:34] <Cimi> mzanetti, I quickly tried putting the code of the drag in __init__.py, but it's a lil bit more complicated to me since I never used python and classes are different. I think its simpler to copy the code for now and wait for a rewrite in qml
[16:43] <nic-doffay> mzanetti mind top approving this dude? https://code.launchpad.net/~nicolas-doffay/unity8/infographic-colour-fix/+merge/188605
[16:50] <mzanetti> Cimi: thanks for the screenshots!
[16:51] <karni> Hey guys, this is my first MR into unity8. Am I missing anything? Not familiar with potential project specific workflows. https://code.launchpad.net/~unity-team/unity8/fix-preview-aspect-ratio/+merge/188657
[16:51] <mzanetti> Cimi: you can put it in emulators/__init__.py and then access it with self._drag()
[16:52] <mzanetti> karni: this is quite an unfortunate time for this merge
[16:52] <karni> mzanetti: Oh, why is that?
[16:52] <mzanetti> karni: we're reworking that code: https://code.launchpad.net/~unity-team/unity8/fix-genericpreview/+merge/188355
[16:52] <karni> aha
[16:52]  * karni looks
[16:53] <Cimi> mzanetti, ah ok
=== dandrader|lunch is now known as dandrader
[16:57] <Cimi> mzanetti, but test_hud.py doesn't subclass the emulator
[16:57] <karni> mzanetti: Does MoviePreview still use the movie cover for the Image { id: previewImage } ?
[16:57] <Cimi> mzanetti, not sure I can access it
[16:59] <mzanetti> karni: yes
[16:59] <mzanetti> Cimi: hmm... odd.
[16:59] <Cimi> mzanetti, what? https://code.launchpad.net/~cimi/unity8/unity8.hud-2_hint-reveal-commit/+merge/187488/comments/431448
[16:59] <mzanetti> Cimi: just checked the code... it seems that they do inherit emulatorbase
[17:00] <Cimi> ok
[17:00] <Cimi> I'll try then
[17:00] <mzanetti> Cimi: yeah... that's the broken jenkins... I think its fixed by now. just wait for the next run
[17:01] <karni> mzanetti: In that case I'm afraid this code will suffer from the same minor problem, but I could propose my changes against your branch, and make your branch prerequisite. How does that sound? (Have you maybe looked at the screenshots I included in my MR?)
[17:01] <karni> mfisch: FYI ↑
=== alan_g is now known as alan_g|EOD
[17:07] <Cimi> mzanetti, ok works in __init__.py
[17:08] <mzanetti> Cimi: yay :)
[17:08]  * Cimi --> GYM
[17:08] <Cimi> I pushed
[17:08] <Cimi> let's hope jenkins will not complain
[17:08] <Cimi> all tests pass on my device
[17:09] <mzanetti> karni: yeah, would be great if you could rebase it on my branch
[17:10] <karni> yep
[17:14] <mhr3> mzanetti, testing the previews branch, i don't see titles for scope previews
[17:15] <mzanetti> mhr3: good catch... fixing
[17:15] <mhr3> mzanetti, shouldn't also all the button be some kind of dropdown?
[17:15] <mhr3> buttons*
[17:16] <mzanetti> mhr3: we don't have those buttons yet. not sure if I should use the OptionSelector
[17:16] <mzanetti> mhr3: but that's another bigger change I think. making that work with the progressbar and everything
[17:16] <mhr3> mzanetti, could you at least make it consistent in app preview?
[17:16]  * mzanetti checks
[17:16] <mhr3> mzanetti, cause apps now do two buttons per row
[17:17] <mzanetti> oh
[17:17] <mhr3> mzanetti, i mean generic preview does it right
[17:18] <mzanetti> mhr3: yeah... got it... make tryAppPreview shows the issue
[17:18] <mhr3> i test on the real thing ;P
[17:19] <mzanetti> mhr3: are you testing it on the phone?
[17:19] <mhr3> nah, not that much real :)
[17:20] <mhr3> mzanetti,
[17:20] <mhr3> file:///tmp/fix-genericpreview/Dash/Generic/GenericPreview.qml:57: TypeError: Cannot read property 'numRatings' of undefined
[17:20] <mhr3> file:///tmp/fix-genericpreview/Dash/Generic/GenericPreview.qml:55: TypeError: Cannot read property 'rating' of undefined
[17:21] <mhr3> GenericPreview shouldn't touch ratings, it's specific for other preview subclasses
[17:21] <mhr3> mzanetti, btw are you ok with this irc review? :)
[17:21] <mzanetti> mhr3: can you please put this stuff into the merge
[17:21] <mzanetti> heh
[17:21] <mhr3> yea...
[17:24] <mzanetti> mhr3: how do you get multiple buttons on the desktop?
[17:25] <mzanetti> mhr3: installing apps doesn't work for me
[17:25] <mhr3> mzanetti, files previews, that uses generic preview
[17:30] <mhr3> mzanetti, weird now i get again two buttons per row
[17:30] <mhr3> mzanetti, i could swear i was getting just one a while ago
[17:30] <mzanetti> mhr3: yeah... there was a mess
[17:31] <mzanetti> mhr3: try pulling now. should be better
[17:33] <mhr3> mzanetti, i'm also not getting title on the generic preview from time to time
[17:33] <mzanetti> mhr3: is there anything special I need to do?
[17:34] <mhr3> mzanetti, i don't see any buttons now
[17:34] <mhr3> after puling
[17:34] <mzanetti> huh?
[17:34] <mhr3> no buttons at all
[17:35] <mhr3> mzanetti, they do work in apps
[17:35] <mhr3> s/work/show up/
[17:36] <mhr3> mzanetti,
[17:36] <mhr3> file:///tmp/fix-genericpreview/Dash/Generic/GenericPreview.qml:59: ReferenceError: root is not defined
[17:36] <mzanetti> I see
=== jhodapp is now known as jhodapp|lunch
[17:51] <mhr3> mzanetti, i added a few more to the review
[17:51] <mzanetti> mhr3: cool, thanks. will try to fix everything today still
[17:51] <mhr3> anyway, time for me to go
[17:59] <karni> mzanetti: there's no longer the ^ tick at the top to close the preview?
[18:00] <mzanetti> karni: nope. that's gone
[18:00] <karni> mzanetti: Too bad. I'm quite sure it was design decision, but I personally thing this is much worse UX. I can imagine the transluent "dismiss" area is not immediately noticable for a new user.
[18:01] <karni> But, no change, no progress!
[18:01] <mzanetti> karni: well... it's tapping outside the preview basically which I think is a pretty common pattern
[18:03] <karni> Yes, I just consider it a bit small (perhaps the fact that it is only at the bottom).
[18:03] <karni> mzanetti: I'll actually consult now, as my change may not be applicable anymore, as the preview image is no longer cropped :)
[18:04] <karni> mzanetti: Do we have ETA when this will land?
[18:04] <mzanetti> karni: I'm hoping for tomorrow
[18:04] <karni> mzanetti: That would be slick.
[18:07] <karni> mzanetti: Ok, cool. I'm deleting my brach, as it no longer applies. The header is gone and preview image of a "poster ratio" fits great by default. Nicely done.
[18:08] <mzanetti> karni: thanks
[18:33] <fginther> Saviq, this one passed: http://10.97.2.10:8080/job/autopilot-testrunner-otto-saucy-fjg/74/?
[18:34] <fginther> Saviq, I figured out the problem with the old libunity9 (update was getting lost when the container shutdown)
[18:40]  * greyback offline for a while
=== jhodapp|lunch is now known as jhodapp
[18:57] <Saviq> fginther, yeah, thought so
[18:57] <Saviq> fginther, so, you switching us back to hardware?
[18:59] <fginther> Saviq, that was my plan.
[18:59] <Saviq> fginther, yay, we're backed up pretty badly again...
[18:59] <fginther> Saviq, the big general problem right now is x86 build capacity. thankfully we're getting some help soon there
[19:00] <Saviq> fginther, right, it's pretty interesting that armhf overtook x86 ;)
[19:11] <thomi> kgunn: https://code.launchpad.net/~gerboland/python-ubuntu-platform-api/add-mir-pkg/+merge/188695
[19:11] <thomi> kgunn: once that merges, We need someone Qt shaped to look into the qt issue Gerry saw. I'll try and look into it on my phone, but I'm no Qt expert
[19:12] <kgunn> thomi: crap
[19:12] <thomi> kgunn: now's not the time, but I really thing someone in your dept should take over the python-ubuntu-platform-api project :)
[19:13] <kgunn> Saviq: assuming your still on any recommendations for a more americano time zone helper on Qt ?
[19:14] <Saviq> kgunn, some guys on bfiller/Chicken team - boiko, renato, tiagosh
[19:16] <Saviq> kgunn, and dandrader (assuming he's not climbing) and mterry of course
[19:16] <dandrader> Saviq, I'm here
[19:16] <kgunn> so for those guys http://pastebin.ubuntu.com/6179755/  this is what is seen when launching AP on unity8 .... suspect library paths set causing Qt to fail to find where the QPA platform plugins are saved (usually /usr/lib/*/qt5/plugins/platforms)
[19:16] <boiko> Saviq: I'm here too
[19:16] <Saviq> dandrader, yeah, I'm just not sure you still will when they need you :)
[19:16]  * dandrader reads backlog
[19:18] <kgunn> thomi verify this...but, dandrader you should be able to take the latest image, rebuild & adb push the python-ubuntu-platform-api using this branch https://code.launchpad.net/~gerboland/python-ubuntu-platform-api/add-mir-pkg
[19:18] <kgunn> then run AP tests....
[19:18] <kgunn> and should get the failure described above
[19:18] <thomi> kgunn: dandrader: correct. you probably already have this set up, but you'll need to cross-compile, since it's armhf only
[19:19] <thomi> dandrader: I'm essentially doing the same thing, but I'm not a Qt expert. If you find anything, please let me know, and I'll do likewise.
[19:23] <thomi> kgunn: dandrader: in case you don't want to build it yourself, get the new hotness here: http://people.canonical.com/~thomir/python-ubuntu-platform-api_1.1daily13.06.13-0ubuntu1_armhf.deb
[19:24] <dandrader> thomi, should I flash cdimage-touch or ubuntu-system?
[19:24] <thomi> dandrader: shouldn't matter. I have ubuntu-system but with a RW root
[19:24] <kgunn> agree it shouldn't matter
[19:30] <kgunn> mterry: ping
[19:32] <dandrader> thomi, is it unity8-mir or unity8-SF?
[19:34] <mzanetti> mterry: ping
[19:36] <thomi> dandrader: it should now work on top of both
[19:36] <thomi> dandrader: the bug was that previously, under unity8-mir, tests would hang
=== dandrader is now known as dandrader|afk
[19:44] <kgunn> dandrader|afk: sorry yeah...this is all about AP not running on unity-mir, which is blocking mir to be default
=== popey_ is now known as popey
[19:46] <mterry> kgunn, mzanetti: hi
[19:47] <mzanetti> mterry: can you think of anything in the greeter that could cause the cpu spinning?
[19:47] <mterry> mzanetti, the edge demo...
[19:47] <mzanetti> mterry: interesting thing is that it stops when you swipe away the greeter
[19:47] <mzanetti> when you lock the device again it starts again spinning
[19:48] <mzanetti> mterry: but the edge demo wouldn't do that when its not shown?
[19:48] <kgunn> mterry: i was curious if you could share the latest story on lightdm/greeter landing in touch
[19:49] <mterry> mzanetti, it shouldn't.  It's not even created when not in use
[19:49] <kgunn> strehl was asking if we were planning on lightdm being in touch (i think its a must from security)
=== dandrader|afk is now known as dandrader
[19:49] <mterry> kgunn, not for v1
[19:49] <mterry> kgunn, well, to be clear, lightdm is in touch v1
[19:49] <mterry> kgunn, just not the greeter
[19:50] <kgunn> mterry: ack
[19:50] <mterry> kgunn, so you're talking post v1?
[19:50] <dandrader> kgunn, ah, ok
[19:51] <kgunn> strehl was thinking it makes work for him  on the indicator side, basically channelling those through account service to make them work again
[19:51] <mterry> mzanetti, the main culprit would be animations, right?
[19:51] <mzanetti> mterry: not sure... but yeah, probably
[19:52] <mterry> mzanetti, or maybe something in a plugin spinning?  Like lots of dbus traffic or something?
[19:52] <mzanetti> mterry: no... QtRenderThread
=== kdub is now known as kdublunch
[19:52] <mterry> mzanetti, you could try to disable parts of it, like the lockscreen or the infographic
[19:52] <mzanetti> mterry: yeah, that's what I did now. problem is I need to wait for half a day until it happens again
[19:52] <mterry> mzanetti, ?
[19:52] <mzanetti> that makes commenting in/out stuff very tricky
[19:52] <mterry> mzanetti, hard to reproduce?
[19:53] <mzanetti> mterry: yeah. I haven't found a way to reproduce
[19:53] <mzanetti> it just starts after a couple of hours
[19:54] <greyback> mzanetti: the 100% cpu thing has been fixed by ricmm and kdub
[19:54] <mzanetti> greyback: thats something else
[19:54] <mterry> kgunn, true, it will require some work in indicator side.  I'd have to check, but last thing I remember is that we didn't have final designs from Design for how indicators look in greeter vs session
[19:54] <mzanetti> greyback: I have a 40% cpu thing
[19:54] <greyback> mzanetti: oh dear
[19:54] <kgunn> mterry: i think i'll throw something on the calendar to synch with strehl...get me on the same page also :)
[19:55] <mterry> kgunn, OK
[19:59] <mterry> mzanetti, that's tough, because if you don't see it, you don't know if you've fixed it or not
[19:59] <kgunn> thomi: no problem...just verifying if this is what i'm supposed to see, so i loaded your deb on my nexus 4 (i haven't switched to mir still using sf)
[19:59] <mzanetti> mterry: exactly
[19:59] <mzanetti> mterry: so... by now I'm sure that it's the QtRenderThread and it happens only when the greeter is created
[19:59] <kgunn> then i run "phablet-click-test-setup", then "phablt-test-run autopilot"
[20:00] <mzanetti> mterry: the amount or around 40% cpu kinda indicates an animation too
[20:00] <mterry> mzanetti, agreed
[20:00] <mzanetti> mterry: but I don't see anything running and have put debug prints in all timer's onTriggered and everything
[20:00] <kgunn> and i see "ran 0 tests in 0 seconds, OK"....surely this is not ok and i did something wrong ?
[20:01] <thomi> kgunn: 'autopilot' should be the test id you want to run.. like 'unity8' or 'dropping_letters'
[20:02] <kgunn> thomi: thanks...i was just guessing at that...cool...now i got ran 15 tests in 1.7 secs and all failed :)
[20:02] <thomi> kgunn: yeah, I'm looking into that.
[20:04] <kgunn> thomi: so we should tell dandrader if he gets it to that point its a success ?
[20:04] <dandrader> kgunn, I'm still downloading&flashing my phone
[20:04] <thomi> kgunn: I'm not sure of the cause yet, sorry
[20:05] <thomi> veebers: perhaps you can lend us your brain for a moment?
[20:05] <veebers> thomi: sure thing
[20:05] <mterry> mzanetti, I'm assuming you've already racked your brain for things that were different or possible triggers the few times you did see the CPU jump?
[20:05] <thomi> veebers: we all get 100% failure rates on the unity8 test suite, using autopilot that's on the imahge
[20:05] <mzanetti> mterry: yeah... it seems total random. sometimes without apps running, sometimes with many apps running
[20:06] <veebers> thomi: jeez, wth.
[20:06] <veebers> thomi: flashing now
[20:07] <thomi> veebers: yeah - it's making it hard to verify if this python-upa thing is working or not :-/
[20:07] <veebers> thomi: ugh I can imagine
[20:07] <thomi> veebers: it's not unity8 specific either - I can't get any of the AP tests to pass
[20:07] <thomi> seems like the proxy objects returned by launch_test_application are always None
[20:07] <thomi> w.t.f?
[20:08] <veebers> thomi: I'm just flashing now, should be able to poke around in a couple of minutes
[20:10] <mzanetti> mterry: well, if you could try to watch "top" on your device every once in a while and maybe try to find some pattern too, that would be great
[20:12] <mterry> mzanetti, I will play with it, yeah.  Stock image?  mir enabled?  writable?
[20:12] <mzanetti> mterry: I'm using SF mostly
[20:12] <mzanetti> mterry: not sure if its happening with mir. probably it does
[20:12] <mzanetti> mterry: writable or not shouldn't matter
[20:12] <greyback> mzanetti: is the renderer causing it?
[20:13] <mzanetti> greyback: yeah
[20:13] <mterry> mzanetti, shouldn't, but who knows  :)
[20:13] <mzanetti> mterry: well, I always make it writable to be able to work with
[20:13] <greyback> mzanetti: so QML_RENDERER_TIMING=1 shows output all the time. In that case some rogue animation must be running
[20:14] <mzanetti> greyback: interesting
[20:14] <mterry> mzanetti, yeah OK.  If it wasn't writable, I could imagine some code having a problem with it, but with writable and seeing it, I'm sure that's not the problem
[20:14] <mzanetti> greyback: do you know where to put it for the upstart user job?
[20:14] <mzanetti> mterry: yeah, I agree
[20:14] <greyback> mzanetti: 1 sec
[20:16] <greyback> mzanetti: /usr/share/upstart/sessions/unity8.conf
[20:16] <mzanetti> greyback: awesome, thanks
[20:16] <greyback> mzanetti: you could also "stop unity8" and then run it in your shell
[20:16] <mzanetti> greyback: as it happens only after a couple of hours I don't want to rely on an attached shell
[20:16] <mzanetti> I could install screen tho
[20:16] <greyback> mzanetti: oh that's annoying
[20:17] <greyback> mzanetti: unfortunately QML_RENDERER_TIMING=1 won't show you what animation is causing the problem :(
[20:17] <mzanetti> :/
[20:18] <mterry> mzanetti, which bug number is this again?
[20:18] <kgunn> thomi: veebers does the ui have to be in a particular state ?
[20:18] <kgunn> if i run unity8 AP locally
[20:18] <mzanetti> mterry: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1219871
[20:19] <ubot5`> Ubuntu bug 1219871 in unity8 (Ubuntu) "when idle with screen blank, unity8 generates hundreds of context switches per second" [High,Confirmed]
[20:20] <veebers> kgunn: we stop unity8 as the tests launch unity with -testability, I guess also that way the shell is in a known state (locked)
[20:23] <kgunn> veebers: ok...will let you debug...like thomi said...this is pretty hot, linked to making mir default now
[20:23] <olli> veebers, no pressure...
[20:23] <olli> not
[20:24] <dandrader> thomi, ok, I got the latest image flashed and your python-... package installed. what's the next step?
[20:24]  * dandrader has never run autopilot tests on his device before
[20:25] <thomi> otp, one sec
[20:25] <kgunn> run "phablet-click-test-setup", then "phablt-test-run unity8"
[20:25] <kgunn> dandrader: ^
[20:26] <kgunn> dandrader: however, AP running locally just seems flat out broken atm...
[20:26] <kgunn> dandrader: so veebers is debugging
[20:26] <kgunn> dandrader: when were you leaving for the day ?
[20:26] <dandrader> kgunn, shortly
[20:27] <dandrader> kgunn, by "locally" you mean on your desktop/laptop?
[20:27] <kgunn> dandrader: any freedom to join a bit later?
[20:27] <kgunn> dandrader: yes locally = run manually on the phone
[20:27] <kgunn> via those phablet commands which are just pumping adb commands to the phone
[20:28] <dandrader> kgunn, ah... so I run "phablet-click-test-setup"  and "phablet-test-run unity8"  from my desktop.. ok
[20:29] <kgunn> dandrader: right with the phone connect
[20:32] <dandrader> kgunn, yeah, "phablet-test-run unity8" fails completely here as well
[20:33] <dandrader> thomi, "ImportError: No module named gi.repository"
[20:33] <dandrader> thomi, are you getting this as well?
=== kdublunch is now known as kdub
[20:34]  * dandrader installs python-git
[20:35] <thomi> dandrader: yeah - that's one for veebers. veebers - any ideas?
[20:37] <veebers> dandrader, thomi: looking
[20:38] <thomi> veebers: that's in the unity8 test suite
[20:38] <veebers> thomi: aye, I guessed that much :-) I'm looking where it is etc.
[20:39] <veebers> right so it's been there for a while
[20:40] <veebers> so it appears that the gi.repo stuff isn't included in the image
[20:41] <thomi> veebers: ok, lets file a bug and move on - it's clearly not the systematic failure we're looking for
[20:42]  * dandrader installs python3-gi and python-gi
[20:43] <veebers> thomi: ok, if I flash using:  `phablet-flash cdimage-touch` the unity8 tests run. But that's not what you are experiencing
[20:43] <thomi> veebers: huh, I'm using the ubuntu-system image
[20:43] <veebers> thomi: right, I'm just reflashing now
[20:43] <kgunn> veebers: i flashed cdimage...it didn't work for me
[20:44] <kgunn> altho...i always wipe
[20:44] <veebers> kgunn: oh? oh with -b?
[20:44] <kgunn> both in real life and in flashing phones :)
[20:44] <veebers> kgunn: ^_^
[20:44] <kgunn> not -b but --wipe
[20:44] <kgunn> also, i use pending
[20:45] <veebers> kgunn: ah ok, trying that now too
[20:46] <kgunn> dandrader: so when i installed python3-gi it said...already latest, but python-gi it installed python-gi & python-gi-cairo
[20:46] <dandrader> ok, now I'm past that gi.repo failure
[20:46] <mterry> mzanetti, I reproduced!  Following cking's suggestion of 2 photos + gallery + timeout.  After swiping greeter away, the high CPU lasted for quite a few seconds, but then went away
[20:46] <dandrader> so now instead of 15 tests run and 15 failed I get 24 run, 24 failed...
[20:47] <kgunn> dandrader: me too
[20:47] <mzanetti> mterry: you don't want to know how often I tried that
[20:47] <mterry> mzanetti, heh
[20:47] <mzanetti> mterry: ah no... if it's just there for a few seconds, its not the issue
[20:47] <mzanetti> mterry: this stays forever
[20:47] <mterry> mzanetti, well, it stayed while blank
[20:48] <mterry> mzanetti, but after waking and swiping, it stayed for a few seconds
[20:48] <veebers> dandrader: could you please run this command and pastebin the output: autopilot run -v unity8.shell.tests.test_hud.TestHud.test_show_hud_appears
[20:48] <mterry> mzanetti, I thought you said it went away after swiping too
[20:48] <mzanetti> mterry: yeah... that's correct, sorry
[20:48] <mzanetti> why the heck doesn't it happen for me?
[20:48] <mterry> mzanetti, just tried it again by just opening the same gallery and waiting
[20:49] <mterry> mzanetti, you've got bad luck
[20:49] <dandrader> veebers, it's the same error for all other tests as well http://paste.ubuntu.com/6181240/
[20:49]  * mterry will try to confirm percent reliability before he changes anything for testing
[20:49] <mzanetti> mterry: do you close or minimize the gallery or keep it in foreground?
[20:50] <mterry> mzanetti, keep it in foreground (though I note that unity did not have a thumbnail for it in "running apps")
[20:50] <mzanetti> hmmm
[20:50] <mzanetti> all thumbnails here for me
[20:50] <dandrader> hmmm but by running autopilot directly instead of through that script I do get that extra line of info
[20:50] <mzanetti> (I just tried it again with 2 pictures + gallery - not happening)
[20:51] <dandrader> "RuntimeError: Expected library path does not exists: /usr/lib/arm-linux-gnueabihf/unity8/qml/mocks"
[20:51] <veebers> mzanetti: sorry to interrupt, has the location of the qml mocks moved?
[20:51] <dandrader> thomi, where do I get it from? ^^^
[20:51] <mzanetti> veebers: not that I know of, no
[20:51] <veebers> dandrader: yeah that's a little odd, ah
[20:51] <thomi> dandrader: that's a question for veebers
[20:51] <veebers> dandrader: how did you install the unity8 tests?
[20:51] <mzanetti> mterry: managed to make the screenshots disappear. waiting...
[20:51] <dandrader> veebers, running  phablet-click-test-setup from my desktop
[20:52] <veebers> dandrader: ok cheers
[20:53] <mzanetti> mterry: nope. got the black screenshots now but it's still not happening
[20:53] <veebers> thomi: would you happen to know how/if phablet-click-test-setup resolves deps.? I.e. unity8-autopilot requires unity8-fake-env
[20:53] <veebers> dandrader: to answer your previous question, unity8-autopilot requires unity8-fake-env too
[20:53] <veebers> that's the package that provides the mocks etc.
[20:53] <thomi> veebers: it's supposed to. sergiusens ? ^^
[20:54] <mterry> mzanetti, have you ever reproduced it when manually trying to blank?  (I'm wondering if I really need to wait)
[20:54] <veebers> thomi: ah ok, I'll bother sergiusens :-)
[20:54] <sergiusens> veebers, thomi no, no deps
[20:54] <mzanetti> mterry: nope, I haven't so far
[20:54] <mterry> mzanetti, ah well
[20:54] <sergiusens> veebers, is that new?
[20:54] <mzanetti> mterry: but well, once it happens you can retrigger it with pressing the power button
[20:54] <veebers> sergiusens: the dependency? no
[20:54] <mzanetti> so might work
[20:54] <thomi> OK, so the problem seems to be that any application launched with autopilot on the mir stack is never displayed
[20:54] <thomi> kgunn: ^^^
[20:55] <sergiusens> veebers, is this for running unity8 tests themselves?
[20:55] <veebers> sergiusens: correct
[20:55] <thomi> kgunn: who can we get to help diagnose this problem?
[20:55] <sergiusens> veebers, and what is unity8-fake-env ?
[20:56] <kgunn> thomi: wait...are you trying to run an application out of shell ??
[20:56] <thomi> kgunn: yes
[20:56] <sergiusens> veebers, I've only worried of click apps now... I don't plan on setting up anything for things that require breaking into rw
[20:56] <veebers> sergiusens: (and mzanetti can correct me if I'm wrong) it provides some fake env stuff/mocks so we can test the lock screen etc.
[20:56] <kgunn> i don't think that's possible at all.... racarr ^
[20:57] <kgunn> thomi: you'd have to launch mir just like shell
[20:57] <mzanetti> veebers: yep, all correct
[20:57] <kgunn> from the app or AP or something....
[20:57] <kgunn> you need the shell to do all that mir stuff
[20:57] <veebers> mzanetti: for on the device, I guess the only needed mock would be for the lock screen?
[20:57] <thomi> kgunn: but it should work if I launch it via upstart, right?
[20:57] <mzanetti> veebers, sergiusens: it's mostly mocking the stuff that's not working on desktop yet
[20:58] <mzanetti> veebers: yes, that would be the best
[20:58] <sergiusens> mzanetti, so we don't use it for desktop testing?
[20:58] <dandrader> hmm, so I suppose using phablet-click-test-setup is not the way to go for running unity8 autopilot tests as it misses several needed packages (python-gi, unity8-fake-env)
[20:58] <sergiusens> oh, mostly
[20:59] <veebers> ok, but to test the lock screen will still need the mocks
[20:59] <mzanetti> sergiusens: ideally this should go away in the long run. for now it mocks stuff where the lower system is not set up yet as we'd need it
[20:59] <mzanetti> sergiusens: i.e. the greeter split into a separate process talking with lightdm
[20:59] <sergiusens> mzanetti, veebers well from the looks of it, I can't support it easily with ro images
[21:00] <kgunn> thomi: i don't think that would be enough...but i might be wrong
[21:00] <mzanetti> sergiusens: don't you have to install the autopilot package anyways?
[21:00] <thomi> kgunn: in that case, (excuse my language), we're kind of fucked
[21:00] <sergiusens> mzanetti, it's preinstalled
[21:00] <mzanetti> oh... do we have that much space?
[21:01] <sergiusens> mzanetti, as it is today, the image is perfectly capable of running tests without breaking into rw
[21:01] <thomi> we need to be able to launch applications... otherwise, how are the autopilot tests supposed to launch the application under test?
[21:01] <veebers> sergiusens: another related question. the unity8 tests use gi.repository and ubuntu-system image doesn't have that apparently. Where would I file the bug for that?
[21:01] <sergiusens> mzanetti, we don't install the tests, just autopilot
[21:01] <mzanetti> sergiusens: yeah, and how do you install the tests?
[21:01] <sergiusens> mzanetti, phablet-click-test-setup
[21:02] <mzanetti> sergiusens: are all apps now click apps?
[21:02] <mzanetti> dialer-app etc
[21:02] <dandrader> veebers, so that's what I'm getting now
=== davmor2_ is now known as davmor2
[21:02] <dandrader> http://paste.ubuntu.com/6181279/
[21:02] <dandrader> veebers, it seems the --testability is not getting to the place it's intended
[21:03] <sergiusens> mzanetti, missing are webbrowser (not going to happen), dialer, messaging and addressbook
[21:03] <sergiusens> they are there, just not installed
[21:03] <mzanetti> sergiusens: and unity
[21:03] <kgunn> thomi: let's make sure we're really f'd first... greyback ^ will we be able to launch apps w/o shell ?
[21:03] <mzanetti> sergiusens: I really don't think we can package unity stuff in click packages
[21:03] <sergiusens> mzanetti, need to wait for usermetrics to get it's apparmor push and camera would be ready
[21:03] <mterry> mzanetti, I rebooted and can't reproduce.  I even deleted all photos and tried again.  Have you noticed if it's always after the first boot or not?
[21:03] <sergiusens> mzanetti, well, unity8 isn't supposed to be a click package
[21:04] <mzanetti> sergiusens: but we have to install the autopilot tests package
[21:04] <sergiusens> mzanetti, that's a special case for which you need to break into rw
[21:04] <mzanetti> sergiusens: ok. and you can't include unity8-fake-env in there too?
[21:04] <sergiusens> mzanetti, so veebers would have to do the usual phablet-test-run -n unity8-autopilot and that should get him going
[21:05] <dandrader> veebers, so "'-testability" is being processed, and rejected, by Mir instead of staying with unity8's main.cpp as it's intended to.
[21:05] <dandrader> thomi, ^
[21:06] <dandrader> or is it a harmless warning that's been always there
[21:07] <veebers> dandrader: no that's a new output/warning
[21:07] <sergiusens> mzanetti, you could, or you can  pull-lp-bin it and dpkg-deb -x it and set QML2_IMPORT_PATH (or DIR... I forget) and it should be picked up without breaking confinement
[21:07] <sergiusens> veebers, ^^
[21:08] <sergiusens> mzanetti, but it seems veebers needs more stuff installed for testing unity8
[21:08] <veebers> kgunn, mzanetti: this looks like a mir/unity8 issue (-testability argument being rejected )
[21:09] <veebers> sergiusens: heh it's not that _I_ need it, it's that the unity8 tests need it ;-)
[21:09] <thomi> hang on
[21:09] <thomi> veebers: on my phone, I see "Loading testability driver."
[21:09] <veebers> thomi: dandrader sees something different, which is what I'm referring to
[21:09] <thomi> erp.
[21:09] <veebers> I should have been more specific
[21:10] <mzanetti> mterry: what do you mean with "after the first boot"?
[21:11] <kgunn> thomi: veebers racarr dandrader sergiusens mzanetti ....would a quick hangout be useful ?
[21:12] <greyback> kgunn: yes, apps launched with upstart, and apps with the desktop_file_hint will launch and appear in shell. The former is preferred
[21:12] <dandrader> so it seems "-testability" should be removed from the args list before it's passed on to createQMirServerApplication....
[21:12] <mterry> mzanetti, like, I'm assuming it's not something that just happens after a fresh install
[21:12] <thomi> kgunn: probably, yes
[21:12] <mzanetti> mterry: no no. happens all the time
[21:12] <sergiusens> kgunn, if you want sure
[21:13] <sergiusens> but need to get some coffee first
[21:13] <thomi> oooh, can we have a coffee break? hangout in 5 minutes time?
[21:13]  * thomi is jonesing for more caffeine
[21:16] <kgunn> thomi: mind hosting the hangout ??...i'm on the mir external engineering call atm
[21:18]  * Saviq feels like there's too many people still here...
[21:19] <mhr3> mzanetti, tried your fixes, works pretty great now.. gj ;)
[21:19] <mhr3> mzanetti, the only thing i miss is some kind of indication that an image is being loaded :)
[21:20] <Saviq> greyback, mzanetti, veebers, thomi, sergiusens, if you guys need me for anything - I'm a ping away
[21:20] <Saviq> mhr3, http://bazaar.launchpad.net/~unity-team/unity8/trunk/revision/356
[21:20] <thomi> sure
[21:21] <Saviq> mhr3, only mzanetti is too lazy to enable it in his rework ;P
[21:21] <Saviq> jeez, 8 branches in queue for merging... and each autolanding takes 3 hrs
[21:21] <veebers> thomi: I can here the hangout ringing noise, I cant see the popup though
[21:21] <thomi> https://plus.google.com/hangouts/_/16ed3d4ac7bbdd50e109aaabbce63c149b8fb526?authuser=1&hl=en
[21:22] <veebers> thomi: cheers
[21:22] <thomi> veebers: sergiusens, mzanetti, dandrader ^^
[21:22] <mhr3> Saviq, you already made him faith from not sleeping and just writing code :P
[21:22] <Saviq> mhr3, I did no such thing - he's just too lazy to get up from his desk
[21:23] <mhr3> Saviq, exactly, you need to task him with something super boring so he'd want to run away :)
[21:24] <thomi> Saviq: can you join us please?
[21:24] <thomi> on that link? ^
[21:26] <Saviq> thomi, coming
[21:42] <veebers> sergiusens: so, just to consolidate re: unity9-fake-env and gi.repository. Are you the guy to bother about that?
[21:45] <mterry> mzanetti, odd.  I just got a 7%-10% consistent surfaceflinger, but no unity8
=== salem_ is now known as _salem
[21:46] <sergiusens> veebers, really depends on what your goal is
[21:46] <sergiusens> veebers, if you plan to use apt and make the image rw, then you need to add those as deps to unity8-autopilot
[21:47] <veebers> sergiusens: to get the unity tests running on the device flashed with ubuntu-system :-)
[21:48] <sergiusens> veebers, ubuntu system can run in rw and ro ;-)
[21:49] <sergiusens> veebers, but if ro is the path, I can look into it, but I need to finish the click stuff first (which is sort of a dep for you to run in ro too)
[21:51] <veebers> sergiusens: ok, I feel that there should be a consistency for running autopilot tests on the ubuntu-system images (i.e. using phablet-click-test-setup)
[21:56] <Saviq> thomi, ok, false alarm
[21:56] <sergiusens> veebers, give the the exact name of all the packages you need and I'll look into it (packages plus depends)
[21:56] <Saviq> thomi, I can launch:
[21:56] <Saviq> $ /usr/lib/arm-linux-gnueabihf/qt5/bin/qmlscene /usr/share/calendar-app/calendar.qml --desktop_file_hint=/usr/share/applications/calendar-app.desktop
[21:56] <Saviq> thomi, and it works fine
[21:57] <thomi> Saviq: hmm, let me try
[21:59] <Saviq> thomi, there's some focus weirdness that I believe greyback was fighting
[21:59] <thomi> Saviq: yeah, for me too, so the issue appears to be with the test suite
[21:59] <Saviq> thomi, yeah, the desktop file needs to be valid
[21:59] <Saviq> thomi, and the path correct
[21:59] <thomi> specifically, the ubuntuuitoolkit test suite doesn't install all it's dependencies
[21:59] <thomi> right
[21:59]  * thomi asks the CI team
[22:01] <veebers> sergiusens: awesome, will do. thanks
[22:07] <Saviq> note to self: never look for "large" in images.google.com just to find a large image...
[22:08] <thomi> wooooOOOOoooo!
[22:08] <thomi> got tests running :)
[22:08] <Saviq> thomi, good!
[22:08] <thomi> just running the ubuntuuitoolkit test suite for a start
[22:13] <sergiusens> Saviq, lol
[22:13] <sergiusens> Saviq, not late at night in a dark room either
[22:13] <Saviq> sergiusens, never
[22:13] <Saviq> sergiusens, things cannot be unseen
[22:22] <olli> :)
[22:23] <olli> so, thomi, Saviq, conclusion: issue in the actual test suite but not in the execution framework?
[22:23] <Saviq> olli, yeah
[22:23] <thomi> olli: there are still problems though, we're not out of the woods yet
[22:23] <Saviq> olli, we're just more strict now
[22:23] <olli> who is taking the lead to fix the suite and other suites so we can finally get a regression view
[22:24] <olli> Saviq, understood
[22:24] <olli> thomi, ack
[22:25] <olli> kgunn, I guess that's a q for you ^
[22:25] <olli> even though while the suites might not be "ours" (if such a term exists at all) it is in our interest to get them moving asap as they are blocking us
[22:26]  * Saviq waits for a phoronix article...
[22:27] <Saviq> 5 mins past the announcement, it should be right around the corner
[22:30] <thomi> olli: My goal is to make sure that all the test suites run by EOD. I'll propose MPs for any changes I have to make
[22:30] <olli> thx thomi
[22:31] <veebers> Saviq: hey, following up a comment that may have gotten lost, based on logs from dandrader (http://paste.ubuntu.com/6181279/) it appears the -testability arg is ignored/handled incorrectly
[22:31] <thomi> nw
[22:31] <olli> Saviq, yeah, I have 20$ on 10min
[22:33] <Saviq> veebers, yeah, bug #1226227
[22:33] <ubot5`> bug 1226227 in unity8 (Ubuntu) "libmirserver parses arguments and fails if it's not something it understands" [Medium,Fix committed] https://launchpad.net/bugs/1226227
[22:33] <veebers> Saviq: yeah just found that :-) sweet, thanks
[22:33] <greyback> veebers: there's a new release of mir which should fix that. Please dist-upgrade and try again
[22:34] <veebers> greyback: ack, wasn't actually my device. Will attempt to re-flash and reproduce
[22:34] <veebers> sergiusens: still around?
[22:39] <kgunn> olli: so i see thomi has volunteered for that... thomi do you need any other help ?
[22:40] <sergiusens> veebers, leaving for a bit, will be back in 2h
[22:42] <veebers> sergiusens: Ack will email :-)
[22:42] <thomi> kgunn: I'll know more in about 60 seconds...
[22:43] <thomi> kgunn: so, my problem right now is that when I boot the phone I don't get a shell
[22:43] <thomi> is this a known problem?
[22:43] <thomi> upstart says unity8 is already running
[22:43] <thomi> but I have a blank screen
[22:44] <thomi> oh wait, it just took *ages* to come up :-/
[22:44] <thomi> like... a whole minute or so
[22:44] <kgunn> thomi: never had that happen weird
[22:45] <thomi> kgunn: I'm just flashing latest image to the phone
[22:45] <thomi> I should know more in a few minutes time
[22:46]  * kgunn hmmm went from 60 seconds to a few minutes
[22:48] <thomi> yeah
[22:48] <thomi> software development (especially debugging) is like the windows file copy dialog
[22:49] <thomi> 60 seconds, 2 minutes, 10 seconds, three days....
[22:50] <thomi> it would be faster if my Internet was not so terrible
[22:53] <thomi> kgunn: OK
[22:53] <thomi> kgunn: the blocker at the moment is that the -testability option doesn't get passed to Qt correctly
[22:54] <veebers> thomi: is that related to bug #1226227? If so greyback mentioned there is a mir release that fixed that, a dist-upgrade should sort you out
[22:54] <ubot5`> bug 1226227 in unity8 (Ubuntu) "libmirserver parses arguments and fails if it's not something it understands" [Medium,Fix committed] https://launchpad.net/bugs/1226227
[22:54] <thomi> possibly, let me try that
[22:55] <kgunn> thomi: that mir fix should be in the latest image...
[22:56] <kgunn> thomi: also...mir literally would crash if you passed it the wrong arg
[22:56] <thomi> kgunn: hmmm, I don't think this is the same. I don't see how libmirserver would be involved, since i'm just running qmlscene
[22:56] <kgunn> right
[22:57] <kgunn> shit!
[22:58] <thomi> yeah
[22:58] <kgunn> thomi: it was this one
[22:58] <kgunn> https://code.launchpad.net/~mterry/unity8/load-testability/+merge/188064
[22:58] <kgunn> it was approved...and should've gone thru...but we got jenkins boned
[22:59] <thomi> kgunn: that might fix it for unity8, but this is jkust running qmlscene
[23:00] <thomi> kgunn: essentially, you should be able to run this:
[23:00] <thomi> /usr/lib/arm-linux-gnueabihf/qt5/bin/qmlscene -testability -I/home/modules /usr/lib/ubuntu-ui-toolkit/examples/ubuntu-ui-toolkit-gallery/ubuntu-ui-toolkit-gallery.qml --desktop_file_hint=/usr/lib/ubuntu-ui-toolkit/examples/ubuntu-ui-toolkit-gallery/ubuntu-ui-toolkit-gallery.desktop
[23:00] <thomi> kgunn: but instead you get this: http://pastebin.ubuntu.com/6181663/
[23:00] <kgunn> thomi: so you're saying thats an issue for ui-toolkit ?
[23:01] <thomi> no, I think it's lower-level than that. Maybe Qt itself? I don't understand how all these things are put together...
[23:01] <thomi> let me try some things..
[23:02] <thomi> ugh. on latest proposed image, unity8 takes *ages* to load... every time
[23:02] <thomi> kgunn: looks mir related - unity8 log shows this:
[23:03] <thomi> terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> >'
[23:03] <thomi>   what():  Could not unblank display
[23:03] <thomi> and it seems now I have no shell at all
[23:03] <thomi> Saviq: still awake?
[23:04] <kgunn> kdub: ^ would that make sense with the discussion we had earlier with ricardo
[23:04] <kdub> unsure
[23:05] <kdub> definitely not the cpu usage issue
[23:05]  * thomi reboots the phone
[23:07] <thomi> kgunn: nope, same thing after a reboot as well.
[23:07] <thomi> so now I'm going backwards - no shell at all
[23:09] <kgunn> thomi: i don't see how you could ever get a shell with that error...that's in the constructor of the thing in mir that talks to hwc
[23:09] <kgunn> which would mean...you ain't got no display
[23:10] <kgunn> thomi: you on system or cdimage
[23:10] <thomi> kgunn: system
[23:10] <thomi> kgunn: I rebooted the phone, and the same thing happens when it reboots
[23:12] <kgunn> thomi: ok...gonna flash system here....
[23:12] <kgunn> thomi: channel ? default or devel-proposed?
[23:12] <thomi> devel-proposed
[23:12] <kgunn> excuse me...meant saucy-proposed
[23:12] <kgunn> oh ok
[23:12] <thomi> phablet-flash ubuntu-system --channel devel-proposed --no-backup
[23:12] <thomi> is what I use
[23:14] <sergiusens> kgunn, devel-proposed and saucy-proposed should be the same thing still (aliases)
[23:14] <kgunn> right
[23:18] <kgunn> robert_ancell: kdub racarr ...while i'm flashing, have any of you built from trunk (or dev branch) and run on the phone today?
[23:19] <kgunn> just wonder if anyone else saw/sees what thomi sees
[23:22] <kdub> have not, on yesterday's image
[23:23] <greyback> kgunn: what does thomi see/
[23:23] <thomi> greyback: https://bugs.launchpad.net/ubuntu/+source/unity8/+bug/1233901
[23:23] <ubot5`> Ubuntu bug 1233901 in unity8 (Ubuntu) "unity8 shell does not start" [Critical,New]
[23:24] <thomi> just uploading the logs to that bug now
[23:24] <thomi> had hoped that apport would do that for me, I guess not :)
[23:24] <greyback> thomi: it's a Mir issue with powerd. powerd blanks the display, Mir can't unblank it. Workaround is to hit the power button and try running shell again
[23:25] <greyback> else unity-mir will need to try something. Agree it's something we need to fix
[23:25] <thomi> greyback: I've hit the power button a number of times. does it not retry?
[23:25] <greyback> thomi: nope. You might be better off disabling powerd entirely
[23:25] <thomi> ugh. why does it keep writing to the log file then!??
[23:25] <thomi> that makes me think it's trying...
[23:25] <thomi> ok
[23:26] <greyback> thomi: oh, upstart will probably be making it retry
[23:26] <greyback> but unless you press the power button at the right time, it'll keep failing
[23:26] <thomi> well, even with hitting the power button, making sure the screen isn't blanked, I still can't get it to start
[23:27] <thomi> greyback: even when I reboot the phone this fails for me
[23:28] <greyback> thomi: mind trying something for me? run "stop unity8", and try "QT_QPA_PLATFORM=ubuntumirserver unity8"
[23:28] <Saviq> thomi, if it gets prevented to unblank once, it gets into a "Address already in use" thing
[23:28] <Saviq> thomi, that only a reboot gets it out of
[23:29] <greyback> Saviq: rm /tmp/mir_socket is faster
[23:29] <Saviq> greyback, right, though it'd be something like that ;)
[23:29] <Saviq> thomi, but it seemed fine for me when I make sure to have the screen unblanked
[23:29] <thomi> ahhhhh
[23:29] <thomi> so I gotta delete that socket file huh
[23:29] <Saviq> thomi, if it fails one, the socket is stale
[23:30] <Saviq> thomi, and it will fail subsequently
[23:30] <Saviq> *once
[23:30] <thomi> greyback: OK, so once I remove the socket file, I can get it to load again
[23:30] <greyback> thomi: glad to hear
[23:30] <Saviq> /afk
[23:31] <thomi> brb
[23:32]  * greyback -> bed
[23:32] <greyback> night all
[23:33] <kgunn> robru: basically....AP on unity-mir still not working....but thomi is trying to help solve that
[23:33]  * thomi is getting increasingly frustrated with the number of things that don't work
[23:34] <thomi> kgunn: any idea how to disable powerd permenantly?
[23:34] <thomi> I'm not spending all my time trying to just start the freaking shell, instead of actually test autopilot :-/
[23:34] <thomi> so I'm thinking of the shotgun approach
[23:34] <kgunn> thomi: no idea... ricmm_ might have an idea
[23:34] <robru> thomi, ok, so your most recent commit on lp:python-ubuntu-platform-api, is that something that is releasable as is, or should I wait for other commits to land on other projects?
[23:35] <thomi> robru: that's releasable as-is. It's a nasty nasty hack, but it'll work, and kgunn promises me his team will work on fixing it properly soon ;)
[23:35] <robru> thomi, ok, it's in daily ppa now, but I will release it just for you ;-)
[23:36] <kgunn> kdub: racarr rsalveti any of you know how to disable powerd to get around the display unblank prob from hwc^
[23:36] <thomi> ps aux | grep unity
[23:36] <thomi> oops
[23:36] <thomi> wrong window
[23:36] <robru> phablet   1416  0.3  6.1 262156 61212 ?        Ssl  21:57   0:18 unity8
[23:36] <robru> phablet   1649  0.0  1.0  99920 10744 ?        Sl   21:58   0:01 /usr/lib/arm-linux-gnueabihf/unity-scope-home/unity-scope-home
[23:36] <robru> phablet   1678  0.0  0.9 108172  9860 ?        Sl   21:58   0:03 /usr/bin/unity-scope-loader applications/applications.scope applications/scopes.scope commands.scope applications/runningapps.scope
[23:36] <robru> root      2963  0.0  0.0   1460   508 pts/1    S+   23:36   0:00 grep --color=auto unity
[23:36] <robru> ;-)
[23:37] <thomi> sudo rm -rf /
[23:37]  * thomi crosses fingers
[23:37] <robru> hey!
[23:38] <robru> actually that probably wouldn't be such a big deal on the phablet... could just fix it with phablet-flash ;-)
[23:38] <thomi> hooray! I got the shell started... it only took an hour
[23:39] <bschaefer> thomi, congrats! (im very out of context)
[23:39] <thomi> kgunn: also, the -testability option earlier is caused by not havign the shell running