=== Zhenech_ is now known as Zhenech
=== TheMuso` is now known as TheMuso
=== _salem is now known as salem_
=== salem_ is now known as _salem
veebersMacSlow: ping07:17
MacSlowveebers, pong07:19
veebersMacSlow: morning, this is the branch containing the autopilot additions, but I stumbled when it came to the packaging: lp:~veebers/unity-notifications/adding_autopilot_tests07:20
jibeldidrocks, mhr3 I definitely disabled recordmydesktop for unity testsuite (the whole testsuite only it is still enabled for the others) as it is clearly one of the cause of memory fragmentation and OOM issues we have. We'll see how it goes without it.07:24
tsdgeosSaviq: mzanetti: any idea why the sudden surge of failing tests?07:28
Saviqtsdgeos, not yet, will investigate after the QA sync07:28
Saviqah, there's actually a PASS?07:30
Saviqtsdgeos, can you have a look at http://s-jenkins:8080/job/unity-phablet-qmluitests-saucy/354/testReport/junit/%28root%29/ListViewWithPageHeaderTestSection/testInsertItemsBeforeValidIndex/07:55
Saviqtsdgeos, to see if we maybe need to improve that test's reliability07:56
Saviqthe others stopped failing suddenly...07:56
tsdgeoslet me see07:57
* Saviq is going to be back in 20, gotta clean my fans...08:00
MacSlowstill needing some review-love... -> https://code.launchpad.net/~macslow/unity8/snap-decision-expansion/+merge/17209108:05
MacSlowany takers?08:05
fginthermzanetti, morning08:11
mzanettifginther: hi08:12
mzanettiMacSlow: can do08:13
fginthermzanetti, all of the VMs should have the right set of packages on them now08:13
MacSlowmzanetti, that would be great... thanks!08:13
mzanettifginther: yep, I've seen you enabled all 3 VMs08:13
mzanettifginther: we had some weird test failures yesterday evening where we couldn't figure the reason yet. I don't think its related to the VM setup tho08:14
mzanettifginther: but just to let you know. We'll keep an eye on them08:14
mzanettifginther: thanks for fixing the VMs08:15
fginthermzanetti, no problem, please let me know if anything else is needed (or if we need more).08:16
mzanettifginther: yep. I will.08:16
dandraderare autopilot tests working for you guys when run locally?08:24
dandradertsdgeos, ^08:24
dandraderfor me, there are long periods of "nothing happening".08:24
MacSlowdandrader, any test in particular?08:25
dandradersometimes the mouse pointer moves but there's not even a unity8 window open08:25
dandraderMacSlow, no, I guess I big bunch of them08:25
mzanettiMacSlow: is there any way to generate notifications when running unity8 with ./run?08:28
mzanettidandrader: hmm.... I've seen some weirdness too yesterday. sometimes it was clicking outside the window for me08:28
mzanettidandrader: but I haven't seen the long periods nothing happening08:29
MacSlowmzanetti, I use the exmaples from lp:unity-notifications...08:29
mzanettiMacSlow: tell me more08:30
Saviqmzanetti, pkill notify-osd08:30
Saviqmzanetti, and run any of http://bazaar.launchpad.net/~unity-api-team/unity-notifications/trunk/files/head:/examples/08:30
MacSlowmzanetti, and speaking of that... I didn't push the "more-than-2-actions" snap-decisions to lp:unity-notifications... but I'll push a branch with updated examples... one sec08:30
nic-doffaySaviq, it appears in the gallery preview that the components heights are being restrained, which is the cause of the bug I'm running into. Any ideas?08:30
Saviqnic-doffay, wtym restrained?08:31
dandradermzanetti, hmm... then maybe there might be something wrong/missing in my new setup. I installed saucy from scratch (usb stick) a week or so ago08:31
nic-doffaySaviq, because on my components expansion it's height is modified by the code. This works as it should in a qmlscene, however in the gallery the height is never changed for some reason, even though the code is called.08:31
nic-doffayonHeightChanged doesn't log anything.08:32
nic-doffayOnly at startup.08:32
dandradermzanetti, ideas on how to debug, what to check, are welcome :)08:33
mzanettidandrader: hmm... I still have a quantal installation upgraded to raring, then to saucy. But tbh I don't think that should make a difference08:33
mzanettidandrader: well, it depends on what's happening. I would probably start hammering some prints into the autopilot tests to understand what exactly is hanging08:33
tsdgeosdandrader: works here08:33
tsdgeosdandrader: is your user part of the autopilot group?08:33
dandradertsdgeos, yes08:34
dandradertsdgeos, some tests do work. I see the unity8 window there and mouse moves things around08:34
tsdgeosok, then no clue what may be wrong .-/08:34
dandradertsdgeos, but then there are some long periods of nothing showing up at all in between those08:34
tsdgeosthat happened to me when my user wasn't part of the autopilot group08:35
MacSlowmzanetti, bzr branch lp:~macslow/unity-notifications/snap-decision-with-many-actions08:35
tsdgeosSaviq: had a look at that test and tbh unless it is running in an ultra slow machine, i can't see how it'd be failing  :-/08:35
mzanettitsdgeos: :(08:36
Saviqtsdgeos, k, let's see what's gonna happen08:36
MacSlowmzanetti, there try unity-notifications/examples/sd-example-incoming-call.py08:36
tsdgeosSaviq: i mean the only thing that was done before it failing was "scroll down", and then i check that we are where we should08:36
mzanettibtw. jenkins machine is currently at 90% idle with 3 - 6% IO wait.08:37
MacSlowmzanetti, of course the stand-alone and interactive qmltest work too08:37
Saviqtsdgeos, k08:37
Saviqmzanetti, http://s-jenkins:8080/job/unity-phablet-qmluitests-saucy/ looks like we're back in business...08:38
mzanettiSaviq:  a bit unsatisfying tho that we have no clue what actually happened08:38
mzanettibut at least it works again08:38
Saviqmzanetti, I put it down to the fact the machine was just hogged overnight08:39
tsdgeosbut yeah something may be wrong :-/08:39
tsdgeoshttps://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-saucy/345/console failed too08:39
tsdgeosand that's not even moving08:39
tsdgeosit's just inside the "verify position after showing"08:39
* Saviq goes on and restarts all the failed MPs08:39
Saviqnic-doffay, the gallery is just a Column { Repeater { } }, really, no height restriction I can see08:41
nic-doffaySaviq, I can't think of any other reason this would work as a child in another column (in my qmlscene) but not in the gallery.08:42
nic-doffayMouse clicks are detected etc.08:43
Saviqnic-doffay, I can see you disabled some of the behaviours in your branch now, can you bring it back to a state where it exposes the issue you're seeing so that I can have a look?08:44
nic-doffaySaviq, the issue is still there. If you run modules/Ubuntu/Components/ListItems/test.qml you'll see it working correctly in a qmlscene. If you run the gallery you'll see it broken.08:45
Saviqnic-doffay, there's no expansion going on in the gallery, though (it doesn't react to mouse clicks) - is that your issue?08:46
Saviqmzanetti, https://code.launchpad.net/~mzanetti/unity8/launcher-follow-unity-api/+merge/172557/comments/38595608:46
nic-doffaySaviq, yeah but the mouse clicks are detected and the selectors.expanded bool changes accordingly.08:47
mzanettiSaviq: d'oh08:47
Saviqnic-doffay, ok, will have a look08:47
MacSlowmzanetti, updated the branch... and also commented with some additional info.09:15
mzanettiMacSlow: approved09:15
MacSlowmzanetti, thanks09:17
Saviqmzanetti, yeah, looking good http://s-jenkins:8080/job/unity-phablet-qmluitests-saucy/09:17
mzanettiSaviq: \o/09:18
mzanettiSaviq: btw. when porting the launcher code to the unity-api i realized that we generate .pc files for each plugin09:18
mzanettiSaviq: thing is, all of them set include_dir to /usr/include which makes them somehow useless09:19
Saviqmzanetti, but Version will be different09:19
mzanettiSaviq: wouldn't it make sense to actually point directly to /usr/include/unity/shell/launcher/ ?09:19
Saviqmzanetti, we need per-API Version09:19
mzanettiah, right...09:19
Saviqmzanetti, we could point there indeed09:19
Saviqmzanetti, I don't think I have any objections to that09:20
Saviqmzanetti, except maybe that #includes will be less expressive09:20
mzanettihmm... also true...09:20
=== dandrader is now known as dandrader|afk
Saviqnic-doffay, dude, you got me chasing there...10:10
Saviqnic-doffay, you only have a single value for the OptionSelector in ListItems.qml10:11
nic-doffaySaviq, you mean in test.qml?10:13
Saviqnic-doffay, no10:13
Saviqnic-doffay, in ListItems.qml10:13
Saviqnic-doffay, where you added the OptionSelector10:13
Saviqnic-doffay, you have values: [i18n.tr("Value 1")]10:13
Saviqnic-doffay, so there's no expansion going on, 'cause there's only one value to select from!10:14
nic-doffaySaviq, yeah, but I've been trying with four too.10:14
Saviqnic-doffay, anyway10:14
Saviqnic-doffay, I have a bunch of other changes, just preparing the diff10:14
nic-doffaySaviq, but I just remembered that it was before I made some other changes to the Selector >_<10:15
Saviqnic-doffay, http://pastebin.ubuntu.com/5839891/10:17
nic-doffaySaviq, line 24 in the diff?10:18
Saviqnic-doffay, Header is ambiguous10:18
=== dandrader|afk is now known as dandrader
Saviqnic-doffay, so to use it we need to import "." (current dir) under a namespace10:19
dandraderdednick, seems good to go: https://code.launchpad.net/~dandrader/unity8/panelDragHandle/+merge/17214210:19
Saviqnic-doffay, obviously visually  and behaviourally there's still work to do10:20
Saviqnic-doffay, like the fact that the other items should get dimmed10:20
nic-doffaySaviq, yeah def.10:20
Saviqnic-doffay, and that the whole list should move into view10:20
Saviqnic-doffay, and the OptionSelector's height needs to be limited10:21
Saviqnic-doffay, btw, you should not have removed ValueSelector10:21
Saviqnic-doffay, it's in use by apps, we'll need to phase it out10:21
Saviqnic-doffay, (or replace it with the OptionSelector - but that's a decision the SDK team needs to make)10:21
nic-doffaySaviq, I'll move it onto another branch when I'm done with it.10:22
nic-doffaySo it exists with the deprecated ValueSelector.10:22
Saviqnic-doffay, yup10:23
Saviqnic-doffay, if you have any questions about the diff, let me know10:23
nic-doffaySaviq, seems pretty straight forward aside from that ambiguous header.10:23
Saviqmhr3, https://code.launchpad.net/~mhr3/unity8/use-dee-filtermodel/+merge/171846/comments/38600310:24
dednickdandrader: taking a look now10:27
mhr3Saviq, thx, pushed fix10:30
mhr3and made a note to run make test next time as well :)10:31
Saviqmhr3, :)10:33
pstolowskiTrevinho: ping10:33
Saviqmhr3, you could also install our shiny commit hook :D10:33
mhr3Saviq, oh? where is it?10:33
Saviqmhr3, .bazaar10:33
mhr3hmm, now i'll have to commit from inside jhbuild shell10:35
mhr3ah, actually it won't work in jhbuild setup :/10:36
nic-doffaySaviq, one question. Selected is set twice, what are the reasons for that? Even if they are separate components...10:40
Saviqnic-doffay, might've been an omission on my side10:40
nic-doffaySaviq, nm10:40
Saviqdidrocks, https://code.launchpad.net/~unity-team/unity8/packaging-cleanup/+merge/172578/comments/38601610:40
nic-doffaymy mistake.10:40
nic-doffayI see one was removed further down.10:40
didrocksSaviq: ah, I think there is a missing dep, one sec10:41
mhr3Saviq, i wish the ci bots were doing what you are :)10:42
Saviqmhr3, ;)10:42
nic-doffaySaviq, what are your thoughts on the OptionSelector being repeated four times in the gallery? Is that correct behaviour?10:49
nic-doffayThe contained values I can understand.10:49
Saviqnic-doffay, it's just a gallery, it's supposed to show you the interaction10:49
nic-doffayBut the entire thing?10:49
nic-doffayYeah but four version with labels Saviq ?10:49
Saviqnic-doffay, you wouldn't be able to show the dimming otherwise, for example10:49
nic-doffaySurely one would be enough?10:49
Saviqnic-doffay, because the dimming is only supposed to affect the immediate neighbours10:49
Saviqnic-doffay, it's the same as with the others, it just shows four of them to show how they interact10:50
Saviqnic-doffay, having just one would be somewhat lonely :)10:50
nic-doffaySaviq, fair enough.10:50
nic-doffaySaviq, soo, further behaviour mods.10:51
Saviqdandrader, as for your autopilot issues, I sometimes find that the unity8 window isn't created10:51
Saviqdandrader, even when using ./run10:51
nic-doffayWhat do you think still should be added to this component?10:51
nic-doffayDimmed items.10:51
didrocksSaviq: I'll look in a bit, it seems that it was expanded to nothing in my trial, will debug it10:52
Saviqnic-doffay, add a set of contained: false to the gallery10:52
nic-doffaySaviq, waiting on assets too. I'll chase that up now.10:53
Saviqnic-doffay, yup, and make sure you talk to the guys about the visuals for the SDK10:53
Saviqnic-doffay, and there's the "non-full-width" case10:54
Saviqnic-doffay, btw, that's what "contained" stands for, no?10:54
nic-doffaySaviq, yeah the ubuntu shape.10:54
nic-doffaySaviq, contained is whether it's always expanded or not.10:54
Saviqnic-doffay, I beg to differ :) http://design.ubuntu.com/apps/building-blocks/option-selector10:55
Saviqnic-doffay, or at least ckpringle does ;)10:55
nic-doffaySaviq, I assumed that was set manually by the person using the component.10:56
Saviqnic-doffay, it is10:56
nic-doffaySaviq, I was working off this: https://docs.google.com/a/canonical.com/document/d/1cg9ruHiEpRynEr_Ipf39Z0MQINEWoChIN-d111yx-Qw/edit#heading=h.cpfn81n0c0rr10:57
Saviqnic-doffay, I just mean that the "property bool contained" is used wrong10:57
nic-doffaySee page 6710:57
nic-doffayOn the left...10:57
Saviqnic-doffay, it should be "property bool expanded" instead10:57
Saviqnic-doffay, which says that it is supposed to be expanded all the time10:57
nic-doffaySaviq, there's already an expanded boolean on the selector which deals with temporary expansion.10:58
nic-doffayWhich is why I made another.10:58
Saviqnic-doffay, simply rename the internal one, then10:59
nic-doffaySaviq, yeah that's what I've done now.10:59
Saviqnic-doffay, what we want to expose as the API needs to be as close to the design docs as possible10:59
Saviqnic-doffay, so that we don't confuse people (and ourselves)10:59
nic-doffaySaviq, gotcha11:00
Saviqdandrader, also, what you might be seeing is (I'm seeing it now after libusermetrics support was merged)11:02
Saviqdandrader, is that the libusermetrics plugin blocks for a long time on startup...11:02
Saviqwe need to fix that11:02
Saviqdandrader, UserMetricsOutput::UserMetrics::getInstance(); takes a good 15s on startup (probably waiting for something on DBus that's never showing up)11:03
dandraderSaviq, well, that behavior I explained was without the libusermetrics patch11:03
Saviqdandrader, k11:04
Cimimzanetti, https://code.launchpad.net/~cimi/ubuntu-settings-components/calendar-newmodel/+merge/17259811:07
mzanettiCimi: ack11:07
nic-doffaySaviq, fixed this up now. Throwing in the assets then I plan on getting a review from design.11:07
Saviqnic-doffay, there's one thing you might want to investigate11:08
nic-doffayThanks for taking a look. Btw I couldn't find any information on the "interactive" property on the ListView.11:08
Cimimzanetti, I should not clear and repopulate the model onMinimumDateChanged and onMaximumDateChanged11:08
didrocksSaviq: ok, the main issue is in libunity, let me fix it11:08
Cimimzanetti, something smarter would be better11:08
Saviqnic-doffay, http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-flickable.html#interactive-prop11:08
Saviqnic-doffay, when you switch from Value 1 to Value 411:08
Cimimzanetti, however, I do something else now because I'm sick of this calendar for a while :)11:08
Cimididn't find bugs so far and tests are better11:09
Saviqnic-doffay, the list contracts faster than the item gets into position11:09
=== MacSlow is now known as MacSlow|lunch
nic-doffaySaviq, yeah I noticed that.11:09
Saviqnic-doffay, you might want to positionViewAtIndex first11:09
Saviqnic-doffay, http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-listview.html#positionViewAtIndex-method11:09
Saviqtsdgeos, do you know what speed is used when using the ListView::position* methods?11:10
Saviqtsdgeos, or when the ListView takes care of it itself by positioning the current item in view?11:11
tsdgeosnot really, i can check11:11
Saviqtsdgeos, you're our ListView expert, so you're going to be the fastest to do so :)11:12
nic-doffaySaviq, it looks good.11:13
Saviqnic-doffay, actually11:13
Saviqnic-doffay, it could be enough to just use http://qt-project.org/doc/qt-5.0/qtquick/qml-qtquick2-listview.html#highlightMoveDuration-prop11:13
Saviqnic-doffay, and set it to SnapDuration11:13
Saviqtsdgeos, ↑ found it (I think11:13
Saviqhmm or maybe not11:13
tsdgeosSaviq: that may work for current index changes11:14
tsdgeosnot sure if positionAtViewBeginning is the same11:14
CimiSaviq, I start having a look at unity theming now..11:14
Saviqtsdgeos, and anyway that concerns the highlight, not the view itself11:14
Saviqtsdgeos, so it probably doesn't anyway11:14
didrockspstolowski: hey, mind having a look? https://code.launchpad.net/~didrocks/libunity/rename-while-installing/+merge/17278811:15
SaviqCimi, thanks11:16
nic-doffaySaviq, highlightMove didn't do the job.11:18
Saviqnic-doffay, yeah, thought as much11:18
nic-doffayStayed with the positionViewAtIndex.11:18
Saviqnic-doffay, yup11:18
pstolowskididrocks: hi, looking11:18
Saviqdidrocks, why that? didn't it work via the .install file?11:19
tsdgeosSaviq: i'd have to do a test, but my guess right now is 200ms11:20
tsdgeosSaviq: do you have a test file i can trace where the code goes?11:20
Saviqtsdgeos, will have one in 211:20
=== alan_g is now known as alan_g|afk
didrocksSaviq: no, it's creating a directory with the name11:21
didrocksand put the file in it11:21
didrocks(even without trailing /)11:21
Saviqdidrocks, uh11:21
Saviqtsdgeos, http://pastebin.ubuntu.com/5840020/11:24
Saviqdidrocks, so we'll have to wait for another release of libunity... :/11:25
nic-doffaySaviq, the issue with positionViewAtIndex is that it always positions it at the first index for some reason.11:25
Saviqnic-doffay, meaning?11:26
tsdgeosSaviq: ok, may take a while, just realized i had a weird mixup in my qt selfcompiled sources11:26
tsdgeosgit submodules suck11:26
Saviqtsdgeos, ignore11:26
Saviqtsdgeos, leave it11:26
tsdgeosor more likely, my knowledge of git submodules suck :D11:26
tsdgeosSaviq: ok11:27
nic-doffaySaviq, it shows the first index the whole time when positionViewAtIndex is used after it collapses.11:27
Saviqnic-doffay, can you push? I'm not sure I get what you mean11:27
nic-doffaySaviq, pushed.11:28
=== alan_g|afk is now known as alan_g
pete-woodsdidrocks: hi!11:33
Saviqnic-doffay, right, so it actually doesn't do anything useful11:33
Saviqnic-doffay, interesting...11:33
Saviqnic-doffay, it might be because the list is able to contain all the items at this point11:34
Saviqnic-doffay, just leave it for now11:34
Saviqpete-woods, hey, it seems that UserMetricsOutput::UserMetrics::getInstance() blocks for quite some time11:37
Saviqpete-woods, like it's maybe waiting for the DBus service that never shows up?11:37
pete-woodsSaviq: it does block on the service11:38
Saviqpete-woods, could it not?11:38
pete-woodsSaviq: but the service should be there? (Id have thought)11:38
Saviqpete-woods, it should never block on any service11:38
Saviqpete-woods, even if it's there, or coming up11:38
didrockshey pete-woods :)11:38
pete-woodsSaviq: okay, should it return null or something when there's no service?11:38
didrocksSaviq: as soon as pstolowski ack it, I'll republish libunity11:38
didrocks(well, once merged of course :p)11:39
Saviqpete-woods, ideally it should return straight away11:39
Saviqpete-woods, and wait for the service in another thread, if it needs to block11:39
Saviqpete-woods, and just update itself when the service comes up11:39
pete-woodsdidrocks: I was looking to make a new release of the libusermetrics11:39
Saviqor, if possible in the same thread, just wait for the service to come up and update then11:39
pstolowskididrocks: I approved your MP 10 mins ago11:40
pete-woodsSaviq: I'll have a look at it11:40
Saviqpete-woods, thanks11:40
didrockspete-woods: it's in the indicator stack, check with cyphermox who is responsible for it. It seems though that indicator-client is failing to biuld rejecting the whole stack then11:42
didrockspete-woods: it was published yesterday FYI11:43
pete-woodsdidrocks: yes, but I fixed an important bug just after release :/11:44
pete-woodsdidrocks: I will speak to the correct person, thanks! :)11:44
didrockspete-woods: yw :)11:44
Saviqpete-woods, ah, but usermetricsservice only runs in the LightDM session, does it?11:50
Saviqpete-woods, btw, http://pastebin.ubuntu.com/5840105/11:53
Saviqpete-woods, sorry for the PL in there, let me know if I can translate something :D11:53
pete-woodsSaviq: whoops! looks like I need to fix the removal script!11:57
pete-woods(I like I get the idea)11:58
mhr3didrocks, dee-qt is not part of unity stack releases?11:58
didrocksmhr3: I think it was put on the platform stack11:59
didrockshum, no sdk stack11:59
=== MacSlow|lunch is now known as MacSlow
pete-woodsSaviq: I can't seem to reproduce that package purging error (just tried purging my install), what was the exact command you ran?12:15
Saviqpstolowski, https://code.launchpad.net/~stolowski/unity8/activation-and-previews/+merge/172102/comments/38608612:15
Saviqpete-woods, the usual - apt-get purge12:15
Saviqpete-woods, but the service was running under user 11912:15
pstolowskiSaviq: ok, fixing12:15
Saviqpete-woods, and that resulted in the error, I believe12:16
Saviqpete-woods, right, 119 is the usermetrics user12:17
Saviqpete-woods, http://pastebin.ubuntu.com/5840153/12:17
=== dandrader is now known as dandrader|lunch
Saviqpete-woods, so it somehow needs to stop the service first12:18
pete-woodsSaviq: there's a prerm script that is supposed to be doing that12:21
didrockspstolowski: I'm puzzled, this pass both builders and failed in the ppa: mv debian/libunity-scopes-json-def-phone/usr/share/unity/client-scopes-phone.json debian/libunity-scopes-json-def-phone/usr/share/unity/client-scopes.json12:22
didrocksmv: cannot stat 'debian/libunity-scopes-json-def-phone/usr/share/unity/client-scopes-phone.json': No such file or directory12:22
didrocks(the local pbuilder worked as well12:22
didrockswth is happening?12:22
pstolowskididrocks: weird... I built it locally with no problems..12:22
pstolowskididrocks: I built with dpkg-buildpackage (not pbuilder)12:22
=== ChrisTownsend1 is now known as ChrisTownsend
didrockspstolowski: I retried trunk and it's fine as well :/12:24
pstolowskiSaviq: fixed12:27
pstolowskididrocks: just guessing.. is it possible that it builds e.g. twice but unpackas it only once, and the file is no longer there on 2nd build?12:34
didrockspstolowski: I checked the logs and it doesn't seem so…12:34
didrockspstolowski: and anyway dh_install is run just before, so we would have it in debian/tmp/12:35
didrockshum, pbuilder fails now on the tests, grrr12:36
didrocksit's probably a race, it's passsing on i38612:36
didrocksbut what?12:36
=== greyback is now known as greyback|lunch
Saviqpstolowski, cheers12:46
=== greyback|lunch is now known as greyback
MacSlowSaviq, did this show up before yet... https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-saucy/381/testReport/junit/(root)/ListViewWithPageHeaderTestSection/testCreationDeletion12:59
SaviqMacSlow, nope, I've restarted, let's see what happens13:02
tsdgeospstolowski: you have a conflict now, can you fix it?13:03
tsdgeosah, see Saviq already said so13:03
Saviqtsdgeos, it should be fixed already13:03
pstolowskitsdgeos: yes, it's fixed13:03
tsdgeosi'm confused by https://code.launchpad.net/~aacid/unity8/remove28403workarounds/+merge/17260813:04
tsdgeoswyh everything is success but the topone is failure?13:04
paulliuI have a MouseArea, and with onClicked and onDoubleClicked. But when I double clicked it, it brings up onClicked first, and then onDoubleClicked. Is that correct?13:04
tsdgeosSaviq: answered your question about the clip13:06
tsdgeospaulliu: think that's the default behaviour yes13:06
tsdgeosotherwise all the clicks would have to be delayed a noticeable number of ms13:06
mzanettihey ho. here's a MP ready for review: https://code.launchpad.net/~mzanetti/unity8/launcher-improve-flicking-behavior/+merge/17264813:07
paulliutsdgeos: hmm, so how to implement two different actions. I mean I don't want the onClicked triggered if it is a double-click.13:07
tsdgeospaulliu: well, usually the click and double click are "not exclusive"13:09
tsdgeosi.e. click selects and double click opens13:09
tsdgeosso it's not a problem doing click and then double click13:09
tsdgeospaulliu: what does click and double click do?13:09
paulliuIn App lens, it seems to me that the click opens the preview, and double click "activates" it.13:09
mzanettiand this one too: https://code.launchpad.net/~mzanetti/unity8/launcher-follow-unity-api/+merge/17255713:10
paulliuok.. then I have to use onPress and onRelease to implement that by hand.13:11
tsdgeospaulliu: afaics it "can't be done", other than making the click wait a bit to see if double click is going to happen or not13:11
tsdgeoswhich is baaaaaaaad imho13:11
tsdgeosand will give us a weird usability13:11
mzanetti+1 for what tsdgeos said13:12
mzanettipaulliu: tsdgeos: well.. what I guess could be done is this:13:13
paulliutsdgeos, mzanetti: oh? but what's so weird? I mean isn't double-click always have a delay threshold?13:13
tsdgeoswooooo, Qt 5.1 is out13:14
tsdgeospaulliu: the wierd is that i do a single click13:14
mzanettiactually... true... tsdgeos: if you double, tap, I'm not sure if the clicked signal is even emitted13:14
tsdgeosand then you wait 300ms to see if it is really a single click or a double one13:14
paulliutsdgeos: yeah, that's it.13:14
Saviqpaulliu, that's not right anymore13:14
tsdgeosthat 300ms will drive me crazy13:14
Saviqtsdgeos, there's no dbl-click for apps13:15
paulliuSaviq: https://docs.google.com/a/canonical.com/document/d/1qjnUirxkswgCvWRT5ervfQoR03zs-7HIIMI2e3Lvuqs/edit13:15
tsdgeosmzanetti: it has to, otherwise everything is delayed by the "valid double click interval", no?13:15
paulliuSaviq: in Installing the apps, it is double click, isn't it?13:15
tsdgeosor maybe it is and the "valid double click interval" is too short my mind doesn't realize?13:16
mzanettitsdgeos: yeah true, but I'm not sure if we aren't used to that delay yet... And I think its less than 300ms13:16
* mzanetti tries13:16
Saviqpstolowski, where was sabdfl's comment about the dbl click in apps scope?13:17
Saviqpaulliu, I know, but I think that's outdated13:17
paulliuSaviq: ok...13:17
mzanettitsdgeos: nah... you get both... clicked on the first one, and doubleClicked on the second13:18
tsdgeosmakes sense13:18
pstolowskiSaviq: comment on dbl click? not sure, I only know of this - https://bugs.launchpad.net/unity/+bug/1189088 - comment #20; so we won't be showing previews for installed apps on single left-click inunity713:20
ubot5`Launchpad bug 1189088 in unity (Ubuntu) "Unity Dash, left mouse button starts preview instead of directly the application" [High,In progress]13:21
Saviqpstolowski, yeah, I meant https://bugs.launchpad.net/unity/+bug/1189088/comments/2013:21
Saviqpaulliu, ↑13:21
=== dandrader|lunch is now known as dandrader
=== _salem is now known as salem_
greybackmterry: http://s-jenkins:8080/job/ubuntu-touch-phablet-image-saucy-mir/ <- grab the preinstall from there and sideload it on to your flipped saucy device13:44
mterrygreyback, thanks, will try13:45
greybackmterry: I'm using yesterday's image, which works fine. I'm a tiny bit worried the MESA problems that hit me yesterday are in that image, but hopefully not13:45
greybackmterry: don't "apt-get update" :)13:45
fginthermzanetti, I see the qmluitests are running better now, may I asked what led to the improvement?13:46
dandradergreyback, can you point me to that unity8+mir image you talked about?13:55
greybackdandrader: http://s-jenkins:8080/job/ubuntu-touch-phablet-image-saucy-mir/ grab the preinstall image13:56
dandradergreyback, thanks13:57
mzanettifginther: good news: yeah, they seem ok now14:10
mzanettifginther: bad new: we have no clue what was wrong14:10
mzanettifginther: we decided to blame it on machine load14:10
fginthermzanetti, yeah, that always gives me a bad feeling too14:10
fginthermzanetti, thanks for the update. glad it's working again14:11
pstolowskiSaviq: can you top-approve my MP?14:14
Saviqpstolowski, oh, didn't I already?14:14
Saviqpstolowski, sorry14:14
mzanettitsdgeos: may I ask you for this one? https://code.launchpad.net/~mzanetti/unity8/launcher-improve-flicking-behavior/+merge/17264814:14
Saviqpstolowski, done14:14
tsdgeosmzanetti: okidoki14:14
pstolowskiSaviq: ty14:15
mzanettiSaviq: and I fear you're the right person for this one: https://code.launchpad.net/~mzanetti/unity8/launcher-follow-unity-api/+merge/17255714:15
Saviqmzanetti, yeah, now I'm past indicators-client I'll definitely do more :)14:15
Saviqmzanetti, one thing I didn't like in the WIP one for launcher backend14:16
* mzanetti tries to avoid asking Saviq for reviews as he does way too many of them already14:16
Saviqmzanetti, was the fact that you hardcoded /usr/share/applications14:16
mzanettiSaviq: gone already14:16
Saviqmzanetti, cool14:16
Saviqmzanetti, does the backend look at XDG_DATA_DIRS or does the .desktop file parser?14:16
mzanettiSaviq: dunno... I defined a method called QString desktopFile(appId) which I expect Wellark to fill :P14:17
Saviqmzanetti, re: me doing too many reviews, all of you guys have been very helpful in that field recently14:18
Saviqmzanetti, so I can't complain14:18
Saviqmzanetti, :D14:18
dandraderSaviq, yeah. now it takes a long time for the unity8 window to finally show up :(14:20
Saviqdandrader, -f14:20
Saviqdandrader, pete-woods is already noticed we can't be doing that (blocking on a service to show up)14:20
pete-woodsSaviq: I have a fix in for the service actually starting now - haven't done the blocking fix yet14:21
Saviqpete-woods, that's fine14:21
dandraderok, "./run --fake"  bring back the old "./run" behavior14:21
Saviqpete-woods, mterry, btw, we have a separate, private qml import path for the shell, feels like a separate one for lightdm (or $greeter) would be good to have, too?14:22
mterrySaviq, maybe...  the greeter might just want to share the shell's though14:22
mterrydepends on what's in there14:22
Saviqmterry, yup14:23
mterryI imagine my split branch shares the same import path unless we explicitly do something different14:23
mterry(just because it uses the same main.cpp with a few minor DEFINE changes)14:23
Saviqmterry, otoh most of the interfaces will need to be proxied through the accounts service or whatever forwards things from the user session to the greeter session14:23
mterrySaviq, true, but does that change logic about the private import path?14:25
tsdgeosMirv: do we have a timeframe for Qt 5.1 in saucy now that it's officially released?14:25
Saviqmterry, if there's different plugins for greeter (talking to accounts service) and shell (directly talking to whatever they want to talk to)14:32
Saviqmterry, we could have separate plugins for shell and greeter implementing the exact same API14:32
Saviqmterry, reducing the need to adapt between the shell and greeter, where applicable14:33
tsdgeosmzanetti: why are dragLauncherIntoView and waitUntilLauncherDisappears in a different UnityTestCase?14:33
mterrySaviq, oh hmm, sure, alternatively we might call them different names.  But regardless, we can solve that when we come to it14:33
mzanettitsdgeos: because I need it for multiple different test cases14:33
mzanettitsdgeos: and it needs to be inside a testcase because it uses compare()14:34
mterrySaviq, truly, even for things like accountsservice, accountsservice can be considered the source data, and even the session would pull data from it14:34
mzanettitsdgeos: could have left it in one of the others, but I decided its better structured this way14:34
Saviqmterry, sure, that would be possible as well14:35
mzanettitsdgeos: I see this quite frequently: https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-saucy/390/testReport/junit/%28root%29/ListViewWithPageHeaderTestSection/growWindowAtBottom/14:35
Saviqmterry, that doesn't preclude a separate greeter-only set of plugins14:35
mterrySaviq, yup14:35
Saviqmterry, that override those that would normally be used by the shell14:36
mterrySaviq, we may indeed want our own set of plugins.  But not yet I don't think14:36
tsdgeosmzanetti: that's bad14:36
Saviqmterry, sure, I wasn't saying right now14:36
tsdgeosmzanetti: let me finish your review and then see if i can loop it and see how it goes14:37
Saviqmterry, although the fact that we need --fake now to get the shell to run did trigger me thinking about this :)14:37
mzanettitsdgeos: ok14:37
Saviqmterry, are we supposed to be able to just consume the usermetricsservice from a user session, btw?14:37
=== alan_g is now known as alan_g|tea
mterrySaviq, I believe that's what the libusermetricsinput1 library is for (vs the libusermetricsoutput1 library that the greeter consumes)14:38
Saviqpete-woods, maybe a "you" question - should we be able to use libusermetricsoutput1 from a user session?14:39
Saviqpete-woods, i.e. is that the reason why we're failing to connect to the service? is it only accessible from the greeter session?14:39
pete-woodsSaviq: definitely14:40
pete-woodsSaviq: no, the reason you can't connect is to do with dbus being very strict, I think14:40
Saviqpete-woods, ok, so for some reason we can't (it's blocking for like 15s)14:40
Saviqpete-woods, let me know if we can help debugging that14:40
pete-woodsSaviq: basically I say "only this user can own the service", but when the config files get put in /etc/dbus-1/... the user doesn't exist yet14:41
tsdgeosmzanetti: what's the expected startup position for the Laucnher?14:41
tsdgeosat bottom?14:41
pete-woodsSaviq: so you have to "touch /etc/dbus-1/system.d/com.canonical.UserMetrics.conf"14:41
mzanettithe side where the Home button is should be unfolded14:41
pete-woodsSaviq: I have added it to the packaging rules in 1.0.314:41
Saviqpete-woods, indeed, so fixed already, thanks!14:42
Saviqmzanetti, are you sure you saw https://jenkins.qa.ubuntu.com/job/unity-phablet-qmluitests-saucy/390/testReport/junit/%28root%29/ListViewWithPageHeaderTestSection/growWindowAtBottom/ more than once?14:43
Saviqmzanetti, first time I saw it14:43
Saviqmzanetti, tsdgeos there's two other failures in there, though14:43
Saviqbut all three of them I only saw once14:43
tsdgeosmzanetti: that timer looks spooky14:43
mzanettitsdgeos: yeah... any better idea?14:44
mzanettitsdgeos: actually its sort of invokeMethod(Qt::QueuedConnection)14:44
mzanettitsdgeos: as the interval 1 makes it run the next time the event loop idles. Its not that I wait for a random time until hopefully everything is done14:45
mzanettitsdgeos: also, there is a test which tests this. It doesn't seem to fail in jenkins so its kinda a confirmation that its not a problem on slow systems14:46
tsdgeosmzanetti: is there?14:47
tsdgeosi commented the code and test still passed :D14:47
* mzanetti fixes14:47
tsdgeosmzanetti: http://paste.kde.org/~tsdgeos/788318/14:48
mzanettitsdgeos: ah yeah... you never set it to snapped now14:49
mzanettitsdgeos: in that case it starts up as it should, but later snapping is not working14:49
mzanettitsdgeos: but try to set snapping directly in the beginning, or in onCompleted. that will make the test fail14:49
tsdgeosis there a way we can make the tests fail if no snapping is set?14:50
tsdgeosi.e. flick a bit and check x,y coords of something14:50
mzanettitsdgeos: hmm... well, I guess I could compare(listView.snapping, true14:52
tsdgeosthat'd be cheating :D14:52
mzanettitsdgeos: by the time the test runs that init logic should have happened already14:52
mzanettitsdgeos: no, in combination with the initUnfolded test its not I'd say14:52
tsdgeosmzanetti: so why is snapping from the beginning a problem exactly?14:53
mzanettitsdgeos: set it and you'll see14:53
=== alan_g|tea is now known as alan_g
Saviqyay! we're flipping!14:54
mzanettitsdgeos:  also the comment on the timer should explain it, I hope. If not, I need to improve it14:54
tsdgeoswell, it says "if i snap to the beginning it fails"14:55
tsdgeosdoesn't tell me why14:55
tsdgeosanyway i did http://paste.kde.org/~tsdgeos/788324/ and tests are still passing :-S14:55
tsdgeosmzanetti: ↑↑14:55
* mzanetti -> pacepalm14:55
mzanettitsdgeos: ok... will check again in a minute...14:56
rsalvetididrocks: well, noticed you already merged the qtubuntu one :-)14:59
mzanettitsdgeos: regarding the comment... I think it says quite more than "if i snap to the beginning it fails"14:59
mzanettitsdgeos: I mean line 45 in the diff14:59
rsalvetididrocks: any other pending mr?14:59
tsdgeosmzanetti: sure it does say a lot more, but tbh the jump from "snapping from the beginning" to "might cause the first item to be half folded at the beginning" is a bit of a "leap of faith" to me15:01
mzanettitsdgeos: ok. I'll improve it15:01
didrocksrsalveti: yeah, I merged it once ogra has sent his email15:01
didrocksrsalveti: I'm just building it as part of dailies and then will update the seed + upload15:02
tsdgeosmzanetti: thanks :-) Also having the test fail would be cool :-)15:02
didrocksthis will free Saviq (libunity is already testing)15:02
Saviqdidrocks, I'm fixing the branch, too15:02
rsalvetididrocks: cool15:06
mzanettitsdgeos: ok... the reason why the test wasn't failing is because I resized the window after writing the test...15:26
mzanettitsdgeos: unluckily the new size did not trigger the snapping15:26
mzanettitsdgeos: so I'll add a verify() that the launchers height is in fact one that causes the issue15:27
mzanettitsdgeos: bad news is, I'm hitting a bug somewhere and I guess it'll take me a bit to figure it out15:27
tsdgeosok, no worries15:28
mzanettitsdgeos: problem is, if I set the snapping some inner logic in QML is updated to fold the icon while the painting part is not15:28
mzanettitsdgeos: try this: do a ./run and reveal the launcher15:28
mzanettitsdgeos: you will see the launcher looks as it should15:28
mzanettitsdgeos: but then try to click on an icon if you're careful enough you'll notice the whole view jumping a bit15:29
tsdgeosi'm not that fast :D15:29
tsdgeosneed a faster set of eyes15:29
tsdgeoslauncher hides too quickly15:30
mzanettitsdgeos: if you click on the phone app you will notice that the list is flicked upwards instead of launching the phone-app15:30
mzanettitsdgeos: thats because the inner logic says, yep, it is folded for more than 12 degrees15:30
mzanettitsdgeos: while the UI still painted it as non-folded15:30
mzanettitsdgeos: tbh I'm a bit lost15:31
mzanettitsdgeos: how the heck can I make sure this thing is unfolded in the beginning while still keeping the highlight area in the middle of the view and have snapping to that enabled?15:32
tsdgeosmzanetti: why the highlight in are is in the middle?15:33
mzanettitsdgeos: because design requirement is that folding is symmetrical15:33
Saviqmzanetti, hmm hanging ?15:42
mzanettiSaviq: maybe... but if yes, I'd say because of an endless loop in the test15:43
tsdgeosphew, not me :D15:43
Saviqmzanetti, it feels like it's not using the fake Unity plugin15:44
mzanettiSaviq: right... that might be15:45
Saviqmzanetti, otoh I'm getting the same errors locally, so maybe not in the end15:45
Saviqor not, the scopes are not installed at all15:46
tsdgeosmzanetti: i'm confused :D So the phone item (the one at the bottom) is index 0 but i can't find where you tell the list you'll be bottomup instead topbottom :-S15:46
mzanettitsdgeos: hehe15:46
mzanettitsdgeos: LauncherPanel.qml15:47
mzanettitsdgeos: inverted: true15:47
tsdgeosmzanetti: but that only affects the delegate not the list15:48
tsdgeosare you basically painting the delegate somewhere else it thinks it is?15:48
mzanettitsdgeos: no... the listView is rotated15:48
tsdgeoswhy that instead ListView.BottomToTop?15:49
mzanettitsdgeos: Saviq: what's also really weird. I have a MouseArea and onClicked I call listView.flick(). That _always_ works and flicks 4 items15:49
tsdgeosah because we don't want bottomtotop15:49
mzanettitsdgeos: but when I call mouseArea.clicked() programmatically, which should do the same, it's not strong enough to flick the list over the snapping15:49
mzanettiSaviq: ^15:49
Saviqmzanetti, "not strong enough"? wth?15:50
mzanettiyeah exactly15:50
tsdgeosmzanetti: both under qmltestrunner?15:50
Saviqmzanetti, shit, it's hanging here, too
tsdgeosmzanetti: weird :S15:51
Saviqor does it15:51
mzanettitsdgeos: Saviq: lp:~mzanetti/unity8/dafuq-is-happening15:52
Saviqmzanetti, lol15:52
mzanettireveal the launcher and make the phone-app icon half-folded15:52
Saviqpstolowski, merged!15:54
Saviqpstolowski, jeez that took ages15:54
pstolowskiSaviq: yeah, seen that. finally :)15:55
Saviqmzanetti, now it hangs at the finish here ?15:57
tsdgeosmzanetti: at least i can tell you why clicking on the phone makes it do weird stuff15:57
Saviqmzanetti, awesome!15:57
Saviqmzanetti, it's alive :D15:57
* mzanetti is listening15:57
mzanettiSaviq: huh?15:57
tsdgeosmzanetti: priv.distanceFromEdge is -22 and that's why you get a -30 as angle15:58
tsdgeosnow, why  priv.distanceFromEdge is -22 that's something you need to find out :D15:58
* Saviq makes a video15:58
mzanettiSaviq: ah, you mean the flicks in the launcher?15:58
mzanettiSaviq: yeah, thats a timer that clicks on the mousearea every few seconds15:58
mzanettiSaviq: to demonstrate that if you click it manually it works, but if the timer does it, its not strong enough15:59
Saviqmzanetti, we wanted a heartbeat, we have a heartbeat :)15:59
Saviqmzanetti, actually it does, I think15:59
Saviqmzanetti, it flattens out completely here15:59
mzanettiSaviq: all the time?15:59
mzanettiSaviq: here it fails to flatten in most of the cases16:00
Saviqmzanetti, it takes two clicks, but yeah16:00
mzanettihmm... doesn't work here :/16:01
Saviqmzanetti, ok got it to fail when swipied16:02
Saviqmzanetti, if I click to move the list, it works16:03
Saviqmzanetti, but not if I swipe16:03
Saviqmzanetti, so it feels like the overshot influences that16:03
Saviqmzanetti, wth :/16:04
Saviqmzanetti, it's like that for 10 minutes now16:04
mzanettiSaviq: yeah... already investigating in the qmltests16:04
mzanettiSaviq: does that hang all the time?16:05
mzanettior just sometimes?16:06
Saviqmzanetti, first two times I noticed16:06
Saviqmzanetti, and it hangs in a different place now16:06
Saviqmzanetti, the second and third jobs here v16:07
Saviqdednick, there's a bunch of conflicts in indicators-client I'm afraid16:07
tsdgeosmzanetti: something that i think will help you debug your problem with the angle, just comment out the whole onClicked in the delegate16:08
tsdgeosand click on the phone icon16:08
tsdgeossee that it gets angled16:08
mzanettiSaviq: oh, now I see what you mean with 394... thats weird16:08
mzanettifginther: help please!16:09
fginthermzanetti, you called?16:09
mzanettifginther: seen this before?
mzanettitsdgeos: yeah, seems like qml is out of sync somehow16:09
tsdgeosmzanetti: my understanding is that clicking on it makes it the currentitem and some bad things happen with the contentY that break your angle calculations16:09
tsdgeosi.e. the list "moves" when clicking in it16:10
tsdgeosand that is what makes your angle change16:10
mzanettitsdgeos: no... I don't think it makes it the current item16:11
tsdgeosmzanetti: well, the contentY of the list is changing16:11
tsdgeossomething makes it change16:11
fginthermzanetti, does this appear stuck to you as well?16:11
mzanettitsdgeos: as I said, I think qml is somehow out of sync because the snapping wants the list somewhere else then it actually is16:12
mzanettifginther: yeah... sits there for like 10 mins16:12
mzanettifginther: and we have some other jobs that are hanging too16:12
fginthermzanetti, *maybe* it's coping files to archive, but that's just a guess16:12
fginthermzanetti, I'll hope on the VM and see if ps doesn't help16:12
mzanettifginther: the other jobs are hanging somewhere else16:12
mzanettifginther: seems like the VM's are frozen and jenkins still keeping the pipe open waiting for more data16:13
mzanettiok. someone killed it now16:13
fginthermzanetti, I didn't kill it, Saviq did you kill it?16:14
Saviqfginther, nope16:14
mzanettimaybe timeout16:15
mzanettiIIRC they are killed after 1.5 hours or so16:15
mzanettioh yeah: Build timed out (after 60 minutes). Marking the build as failed.16:15
fginther393 only ran for 1 hour16:15
fgintherok, that's it then16:15
Saviqyeah, it timed out16:15
Saviqthe other one SUCCESSed16:16
Saviqafter having hung (hanged?) for 15 mins or so at "closing log"...16:16
mzanettihmm... maybe related to the other VM going down16:16
mzanettifreed up some resource that made this one continue16:16
tsdgeosmzanetti: right the index is not changing16:17
tsdgeosmzanetti: but contentY is16:17
mzanettitsdgeos: yeah... but I actually think the issue is this16:17
mzanettitsdgeos: I set contentY somehow (with that timer for example)16:17
tsdgeosmzanetti: i can try to debug it tomorrow with my compiled 5.1 and see who is changing contentY if you want16:17
mzanettitsdgeos: that does not work well with the defined snapping16:17
mzanettiso what happens is that contentY is set, but not all the other states updated accordingly16:18
SaviqI triggered a rebuild of the one that timed out16:18
Saviqlet's see16:18
Saviqmzanetti, fginther ↑16:18
mzanettitsdgeos: once there is interaction from the mouse, contentY is adjusted to what it should be according to the other states16:18
fginthermzanetti, after looking at the files archived, I don't think that was the problem (there was nothing huge). Perhaps the VM host was overwhelmed, it is close to capacity16:18
mzanettitsdgeos: does that make somehow sense to you?16:18
fgintherSaviq, will watch16:18
mzanettifginther: btw... I told this mmrazik some time ago already but I fear someone else needs to catch up with this:16:19
Saviqfginther, /me too16:19
mzanettifginther: some time back I monitored naartjie for its usage because we were struggling with having 4 VM's there16:20
fginthermzanetti, and now we have a bazillion?16:20
mzanettifginther: turns out the machine's CPU is 95% idle all the time, the memory consumption is like 10% of it16:20
mzanettifginther: BUT: disk IO waits are more than 10%16:20
tsdgeosmzanetti: not sure tbh, bit tired atm, talk tomorrow, ok?16:21
mzanettifginther: so imho we are wasting 95% of naartjies resources waiting for its HDD16:22
mzanettitsdgeos: ok, sure16:22
mzanettitsdgeos: thanks a bunch already16:22
fginthermzanetti, the sys admin has already mentioned the need to rework the hardware resources... I'll ping him again and see if the solution is already in the works.16:22
fginthermzanetti, thanks for the reminder16:23
fginthermzanetti, what tool were you using to collect your data?16:23
mzanettifginther: just sitting there and watching top for a while16:23
fginthermzanetti, ack16:24
mzanettifginther: try to do anything on your machines, you won't be able to get disk IO's > 1% unless you really badly mess with the system16:24
mzanettifginther: watch top on naartjie and there is everything close to 0 except big IO waits16:24
mzanettifginther: and you feel the VM's getting lots slower whenever the IO waits are bigger than like 3%16:25
fginthermzanetti, I'll check it out16:26
dednickmzanetti: ping16:40
mterrypete-woods, can you propose https://pastebin.canonical.com/93700/ ? I don't remember why that was a proposed change, besides to "fix things"16:44
nic-doffaySaviq, would the parent view of the OptionSelector handle scrolling if the OptionSelector expands below the view bounds?16:47
mzanettidednick: pong16:51
dednickmzanetti: hey. was just wondering how CI runs the autopilot tests. where does it get the commands from?16:52
dednickeg autopilot run unity816:52
mzanettidednick: job parameters16:52
mzanettidednick: actually... I think jenkins doesn't support multiple test suites (yet)16:52
dednickmzanetti: ok. i htink i'll just leave it as a single target for now then16:53
mzanettidednick: I could create a test job for you16:53
mzanettidednick: shouldn't be too hard to hack support for multiple targets in16:54
mzanettidednick: do you have a branch that would do everything already?16:54
dednickmzanetti: or the other option is to create multiple targets + a global one,16:54
mzanettidednick: jenkins doesn't run make autopilot16:55
dednickmzanetti: ah right16:55
mzanettidednick: it installs the autopilot package into the system and runs "autopilot run $test_suite"16:55
dednickso it would be running 'autopilot run unity8'?16:55
dednicki c16:55
mzanettiyeah. thats what it does16:55
mzanettiso to have a single target we would need one python autopilot test suite that calls the others16:56
dednickmzanetti: my indicators-client branch has a indicators_client ap suite.16:56
dednickmzanetti: doesnt REALLY matter right now. it is a very simple siute, but i want to get started on some more fairly soon16:57
mzanettidednick: ok. I'll have a test run16:59
mterrypete-woods, I filed https://code.launchpad.net/~unity-team/unity8/older-months/+merge/172860 -- please fill out with your infographic smarts16:59
mhr3pete-woods, hilfe - unity/source/libusermetrics/src/libusermetricsinput/MetricImpl.cpp:56:50: error: ‘logic_error’ was not declared in this scope17:00
mhr3pete-woods, where is that supposed to come from?17:01
mhr3hm, stdexcept, who would have expected that17:02
=== alan_g is now known as alan_g|EOD
mzanettidednick: do you have a jenkins run on your branch that created packages for the multiple autopilot test suites?17:14
mzanettidednick: I think I should have a job now that can do that17:15
dednickthere's only a single package17:15
dednickbut 2 suites17:15
mzanettidednick: doesn't matter. should work17:15
mzanettidednick: do you have a link perhaps?17:19
dednickmzanetti: https://code.launchpad.net/~unity-team/unity8/indicators-client17:19
mzanettidednick: it has conflicts17:23
mzanettidednick: and no recent ci run that would have built the packages17:23
mzanettidednick: please fix the conflicts and ping me when CI as run over it. then I can grab the package and run through my testing job17:24
dednickmzanetti: ah. sorry, i pushed latest to wrong branch17:24
Saviqveebers, ping17:24
Saviqright, won't answer on 5am ;)17:25
Saviqmzanetti, ideas about traceback-1 from https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-saucy/597/testReport/junit/unity8.tests.testlockscreen/TestLockscreens/test_unlock_wrong_Pinlock_/ ?17:32
dednicksigh.. bring back the days when it took 3 seconds to build unity817:34
mzanettiSaviq: code-wise: lightdm didn't respond to the authenticate() call17:34
mzanettiSaviq: did anything change when loading the libs?17:35
Saviqmzanetti, no, I meant the python traceback17:35
Saviqmzanetti, not the test failure17:35
Saviqmzanetti, TypeError: on_test_end() takes no arguments (2 given)17:35
mzanettioh... traceback-117:35
Saviqmzanetti, re: the failure, not that I know of, restarted and will check again (it might be)17:35
Saviqwe've moved all the libs around enough for this to break...17:36
mhr3how do i get some useful info from make test?17:36
mzanettiSaviq: hmm...the on_test_end(). no, haven't seen that before17:36
Saviqmzanetti, it's everywhere here https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-saucy/597/testReport/?17:37
mhr3Saviq, dzenkuje17:37
Saviqmhr3, ę17:37
Saviqmhr3, but proszem17:37
mhr3don't have that here :)17:37
SaviqNení zač17:38
mhr3meh, google translate cheating :P17:40
Saviq-translate ;)17:40
Saviqjust google17:40
mhr3you guys are making it harder and harder to run this on precise17:42
sergiusensmzanetti: what do you mean by multiple test suites in jenkins?17:42
mzanettisergiusens: hi. how come you know of this? :P17:43
mzanettisergiusens: unity8-autopilot installs multiple test suites now. unity8, indicators_client17:43
sergiusensmzanetti: read the back log ;-)17:43
mzanettisergiusens: I just modded the job to run a for-loop over $test_suites instead of just calling it on $test_suite17:44
sergiusensmzanetti: so from a jenkins perspective you should be able to collect them17:44
mzanettisergiusens: so a space separated list should do17:44
sergiusensmzanetti: ah, great17:44
Saviqmhr3, hum? ;)17:47
mhr3Saviq, https://code.launchpad.net/~mhr3/unity8/use-dee-filtermodel/+merge/171846 you know you want to ack it, right? :)18:02
Saviqmhr3, I do, assuming I will review it first :)18:03
Saviqmterry, shite, I missed the powerd meeting (my calendar hates me, as usual)18:03
mhr3Saviq, oh you do actual reviews? i thought you just copy hidden bot messages around ;)18:04
Saviqmhr3, when there's no more hidden bot messages, yes!18:04
mterrySaviq, you saw the notes?18:05
Saviqmterry, yeah, reading now18:06
Saviqmterry, anything else I should know? or maybe provide input for?18:06
mterrySaviq, from my end, no.  It sounds like powerd will be able to provide the interface for blanking, so I'll wait on that for the final 'show greeter on blank' fix18:07
Saviqmterry, k18:08
mzanettidednick: http://s-jenkins:8080/job/generic-mediumtests-runner-testing/18:10
mzanettidednick: this will run your branch now with 2 test suites18:10
* mzanetti crosses fingers18:10
Saviqmterry, would those failures suggest LD_LIBRARY_PATH is not set correctly ?18:12
Saviqyes it does...18:13
Saviqseb128, if you're still around... any idea how to find the multiarch libdir in a python script?18:15
Saviqah sysconfig!18:16
mterrySaviq, yeah, could be pointing at 'single' instead of 'single-pin' or some such18:17
Saviqmterry, yeah, it points at a non-existing dir after we've moved stuff around18:17
mterrySaviq, or even maybe QML_IMPORT_PATH is wrong and we're using real LightDM plugin18:17
mterrySaviq, ah18:17
pete-woodsmterry: sorry, just got back from re-attaching the exhaust pipe to my car :/18:20
pete-woodsmterry: have added details to that commit18:24
mterrypete-woods, thanks man18:25
mterrypete-woods, you can't use wild animals as exhaust pipes, they'll just keep running off on you18:25
pete-woodsmterry: :p18:26
pete-woodsmhr3: sorry for not getting back sooner, have you managed to compile libusermetrics now?18:26
pete-woodsI'm a bit puzzled why built-in C++ classes wouldn't be visible18:26
mzanettidednick: looks good. just one small mistake but I think the next run should succeed18:43
=== seb128_ is now known as seb128
=== bschaefer_ is now known as bschaefer
=== jhodapp is now known as jhodapp|holiday
mhall119so when I have a lot of apps open on my Nexus 7, the unity8 process consistently runs at about 10% - 30% CPU utilization when sitting on the Apps lens21:47
mhall119since all of those apps should have their process suspended, I would expect more CPU idle than this21:48
mhall119it doesn't appear that any of the app processes are active, so something just with unity821:49
mhall119Saviq: ^^ what can I do to help determine why this is?21:50
Saviqmhall119, we had a report like this some days ago, just didn't mention that apps need to be running21:50
Saviqmhall119, if we can reproduce we'll have a look21:51
Saviqmhall119, obviously it should be idle21:51
mhall119Saviq: I have 17 apps open, all of them QML apps using qmlscene21:51
Saviqmhall119, 17!? and they're not killed?21:51
Saviqmhall119, nice one ;D21:51
mhall119which, granted, multi-tasking 17 apps on a nexus 7 is pretty damn impressive on it's own21:51
Saviqmhall119, regardless, should be idle21:52
Saviqmhall119, we'll have a look tomorrow21:52
mhall119ok, let me know if there's some extra data I can get you21:52
Saviqmhall119, will try and reproduce now21:54
mhall119Saviq: looks like a combination of unity8 and surfaceflinger22:01
mhall119if that helps any22:01
Saviqmhall119, yeah, I expect something to be pinging the shell unnecessarily22:02
Saviqmhall119, resulting in UI updates22:02
=== salem_ is now known as _salem
Saviqmhall119, can't reproduce on neither manta or maguro :/22:38
Saviqmhall119, several apps, unity8 goes down to 0.naught CPU22:38
Saviqto the point where htop is at the top with 5%22:38
Saviqmhall119, will ask the guys tomorrow to try on all the devices we have around22:42
mhall119Saviq: like I said, I'm happy to provide whatever data I can from my N723:10

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!