[05:40] hello, got a problem with my u 13.04 [05:40] unity [05:40] I can't login with main user [05:40] just with guest [05:40] In tty allows me login [05:41] but I need the gui [05:41] I made freshinstall from upgraded 13.04 with broken unity [05:41] made a backup [05:42] now I need restore all that for yesterday [05:42] and cant login the user [05:42] Helpme! [07:12] veebers: do you know how to fix the maguro tests failing? [07:17] mzanetti: or you? [07:17] or can we just disconnect it for the time being? [07:27] tsdgeos: not yet... I tried quite a while to reproduce it on saturday, but couldn't manage to do so [07:28] tsdgeos: it's on my todo to get it either fixed or temporary disabled for today [07:28] tsdgeos: we also have the QA meeting in 2 minutes. feel free to join [07:30] veebers: ping [07:34] damn i missed it [07:34] mzanetti: you guys still there? [07:50] tsdgeos: just saw a mail from veebers that he can't make it today [07:50] ah [08:15] mzanetti: still happening like crazy :-/ https://code.launchpad.net/~jpakkane/unity8/quickfixes/+merge/182048 [08:23] tsdgeos: yeah... I just disabled it 15 minutes ago [08:23] oka [08:23] have you retriggered all the jobs? [08:23] want me to? [08:26] ahhhh, my mouse wheel broke [08:26] driving my craaaaazy [08:41] tsdgeos: no, don't re-trigger all of them. I want to do it one by one [08:41] ok, sorry [08:41] tsdgeos: at least for the first bunch until I'm sure it works again [08:41] no problem if you did so... [08:46] well i did for ~2 [08:57] mzanetti: hey, yeah sorry as mentioned couldn't make it [08:59] mzanetti: ping [09:04] veebers: hey [09:04] veebers: I only had one important topic for today's meeting [09:04] veebers: I wrote a mail [09:05] to you and Omer [09:05] mzanetti: ah see that. Hmm odd, the failure I see from that job is that NetworkManager isn't running then a whole bunch of apt-get failures (due to network stuff) [09:06] veebers: ok, that's new [09:06] veebers: last week it was all input failing [09:06] mzanetti: hmm. Will ping omer tomorrow [09:06] veebers: anyways, I've disabled them for now. [09:06] mzanetti: ah I had that happen to me today, had to restart the device etc. [09:07] veebers: but obviously keeping them disabled is not the solution. It would be great if you and Omer could try to figure what it is. [09:07] mzanetti: aye, agreed. Will do [09:07] veebers: as I said... I spend most of my Saturday here trying to reproduce but no success :/ [09:07] mzanetti: ugh sucks. You were starting and stopping unity8 in-between runs? [09:08] veebers: yeah... I used the scripts from jenkins [09:09] mzanetti: odd, alright will explore [09:09] veebers: I wen to the last successful jenkins jobs and copied the archive.zip url. passed that to the scripts from omer that he uses in jenkins. [09:09] veebers: and then I went through the history of jenkins jobs and copied the other archive.zip urls trying to reproduce the exact thing as it happened on jenkins. [09:10] veebers: my suspicion is this: we install some version of unity8 (which consists of like 5 packages). [09:10] mzanetti: right thanks, will try that. Today I was just using `phablet-test-run -n etc.` (with the 'stop shell' option) a couple of times and realised that after a couple of runs none of the applications were getting input [09:10] veebers: and the next run we reinstall some packages with the same version number (but different content) [09:10] with the shell started or stopped (but unity8 was, i.e. could reveal the launcher) [09:11] veebers: looking at the apt-cache policy output from the jenkins log, there seems some weirdness, like packages not matching any more [09:11] i.e. unity8 installed from current run, while unity8-fake-env still from the previous run. [09:12] mzanetti: that's really odd. Right will explore further tomorrow. Yeah good point, will check that out [09:12] ok. [09:12] mzanetti: might need to bite the bullet and re-flash between runs or something similar [09:13] veebers: I vote for that... still we should at least understand/be sure what's happening that it breaks [09:13] veebers: but imo reflashing would be a good idea anyways. ideally _after_ a job, not before, to save some time ;) [09:14] * tsdgeos pities poor Saviq [09:14] come to sunny Catalonia [09:14] and get rain D: [09:14] haha [09:14] mzanetti: yes agreed, it would be good if we could have a node have a deconstructor job assigned to it, that gets run after a job uses that node but as a separate run so as not to make the job take ages [09:15] veebers: yeah. fginther might be able to help with that I guess [09:15] veebers: actually, reverting the jenkins VMs already works like this [09:15] veebers: so jenkins might have that capability already [09:16] mzanetti: sweet, will email and cc you in so might get looked at in my night [09:16] cool [09:17] tsdgeos: will dandrader finally come back today? [09:17] or tomorrow? [09:18] i guess today [09:19] Cimi: ping [09:27] paulliu: ping [10:05] mzanetti: whatthe? https://code.launchpad.net/~unity-team/unity8/unity8.new_hud-button_interaction/+merge/181829/comments/412412 [10:06] all success but failed? [10:13] mzanetti: dandrader leave says 1-27 [10:13] so not back today for sure [10:13] either tomorrow or wednesday [10:13] probably wednesday [10:23] probably yeah [10:55] mzanetti: greyback: if you have some time can you read https://code.launchpad.net/~nick-dedekind/unity8/indicator.visibility/+merge/181642/comments/412434 and tell me if i'm crazy (more than usual) or not? [10:55] ok [10:55] tsdgeos: ack [10:56] are the buildiers dieng? [10:56] /usr/lib/pbuilder/pbuilder-modules: line 238: 59996 Alarm clock mount -t devpts /dev/pts "$BUILDPLACE/dev/pts" -onoexec,nosuid,gid=$TTYGRP,mode=$TTYMODE [10:56] https://jenkins.qa.ubuntu.com/job/generic-mediumtests-builder-saucy/2608/console [10:56] greyback: now that I see you, I'd need another chat regarding the app manager's api stuff. let me know when you have time [10:56] wtf is wrong with my computer, I click that link, and have to wait over 30 seconds before firefox opens it [10:56] mzanetti: ok, after lunch would be good [10:56] greyback: I had really bad troubles with DNS lately... similar symptoms [10:58] mzanetti: I don't see how dns is the problem, unless firefox only opens the tab only after dns resolution is made [10:58] greyback: yes, I think that's what happening [10:58] anyhoo, maybe I should just reboot [10:58] weird [11:04] tsdgeos: tbh I don't see why we want to store the visible flag in the model either [11:05] in other words, I agree with you. [11:06] tsdgeos: +1 from me too. [11:07] tsdgeos: also, I don't see why we want to filter tabs in case the Loader is not ready yet. couldn't the SortFilterProxyModel just filter on "enabled"? [11:07] or whatever the role is that causes an item to be "enabled" [11:08] mzanetti: but enabled is also calculated at the QML level [11:08] so he'd still need to inject it to a model somewhere [11:08] tsdgeos: yeah, but isn't enabled calculated based on other data in the model? [11:08] hmmm [11:08] probably [11:08] tsdgeos: so the sortfilter model couls calculate it on its own === hikiko is now known as hikiko|lunch [11:18] * greyback to the post office, back in a while [11:28] tsdgeos: hi [11:29] paulliu: just wanted to say i'ved added a comment to https://code.launchpad.net/~paulliu/unity8/movie-preview/+merge/181856 [11:29] tsdgeos: ok.. thanks. I'll reply it. [11:30] lunch! === hikiko|lunch is now known as hikiko [12:56] mzanetti: afternoon! [12:56] hi kgunn [12:56] mzanetti: hey, maybe someones already been pestering...but [12:56] mzanetti: any ideas on this http://reports.qa.ubuntu.com/smokeng/saucy/image/3758/unity8-autopilot/ [12:57] kgunn: no, this is new to me. however, seems the autopilot tests for our ci fail too. Problem is, we can't reproduce it locally yet and it seems to be related to the way jenkins sets up the device. [12:58] mzanetti: not just the lockscreen tests...but in general...as part of "going green" [12:58] sure [12:58] kgunn: but om26er, veebers and me are aware of it and trying to find out what it is [12:58] mzanetti: thanks.....by can't repro locally...you mean, you run AP local and everything is hunky-dory [12:58] ? [12:58] oh ok...your on it [12:59] kgunn: I've spent most of saturday running the tests on the my phone here. no tests failing. I even used the same scripts as jenkins is using [12:59] mzanetti, there is only one way to run tests on devices these days i.e. to kill the shell first and we do just that. Nothing different [13:01] om26er: still they fail in jenkins but work here on my phone. I've even been using your scripts [13:01] om26er: btw. please move them to some place where more people have write access [13:01] and I've found some other issues too [13:07] ahhhhhhhhh nothing is landing ahhhhhhhhh [13:08] tsdgeos: come on... we've already landed 2 out of 11 branches today :D [13:09] mzanetti: i'll reapprove your branch that just failed again [13:09] tsdgeos: ok... I'm currently monitoring VM's, hoping I find some reason why all this random weirdness is happening [13:10] mzanetti: can you have a look at https://code.launchpad.net/~aacid/unity8/expandedItemPersistant/+merge/182094 ? [13:10] it's 5 lines and fairly easy to repro [13:10] tsdgeos: ack [13:11] Cimi: any joy on testing multimon mir ? [13:12] mzanetti: om26er veebers: heard you have problems reproducing the unity8 failures on maguro? [13:12] asac: yeah, can't reproduce it on make either [13:12] mako [13:14] tsdgeos: I only have "Applications" in the home scope. what else do you have in there? [13:17] Trevinho: any joy testing multimonitor xmir ? [13:19] mzanetti: jsut search [13:19] mzanetti: the magic of scopes will give you more stuff [13:19] tsdgeos: not happening here [13:20] not getting stuff like "Info" and "Reference"? [13:20] tsdgeos: nope. never seen that so far [13:20] tsdgeos: on the phone or on the desktop? [13:20] mzanetti: both [13:20] mhr3: ↑↑↑ What is he missing? [13:20] tsdgeos: I just flashed a device one hour ago [13:21] mzanetti: do you have interwebs? [13:21] searching for "a" in the home scope gives me 4 applications, that's it [13:21] hmm... maybe not [13:22] mzanetti: we dont see it on mako. only on maguro [13:22] mzanetti: do you have that device? [13:22] asac: yes. but in our ci jenkins we see the same on both devices. [13:23] asac: are you setting up devices in the same way? [13:23] asac: only re-flashing when there is a new image, otherwise installing packages and then running tests on them? [13:23] mzanetti: what does ci jenkins do? [13:23] is that x86? [13:23] asac: no. we test on real phones too [13:23] mzanetti: yes, i think thats what is done, yes. [13:23] mzanetti: its not we or them. its us [13:23] :) [13:24] right :D [13:24] tsdgeos, kgunn national holiday in UK today... [13:24] tsdgeos, read your reviews, will fix tomorrow... [13:24] ah.. that explains why I can't reach designers :) [13:24] mzanetti: so only explain that was brought up was that the system is busy [13:24] while running [13:24] Cimi: ah, you didn't add yourself to the leave document ;) [13:24] mzanetti, :) [13:25] mzanetti: now we dont have that and today the autopilot test on maguro fails reliably [13:25] (two times in a row) [13:25] e.g. i believe its a real thing this time [13:25] tsdgeos, I didn't add national holidays :) [13:25] someone needs to confirm [13:25] so yeah whole design team is off today [13:25] and I am hime sick with headache! [13:25] *home [13:25] Cimi: to be honest not sure if that's the plan or not, but helps if i can go there and see if you're off, though i can probably do the same in the directory thing [13:25] in bed [13:25] how good [13:26] holidays and be sick [13:26] asac: oh... I've just seen that the logs are different from the ci jenkins [13:26] asac: UInput: UInputError('"/dev/uinput" cannot be opened for writing',) [13:27] asac: the user running the tests needs to have permissions to write this. autopilot does set up some group magic for this, but not sure if that requires a reboot to work [13:28] asac: ci jenkins manually sets /dev/uinput to 666 iirc [13:28] mzanetti: it worked in the past [13:28] e.g. we had that a month ago [13:28] now its back [13:28] but only on maguro :) [13:28] interesting [13:28] hmm [13:28] it all worked well on main dashboard [13:30] mzanetti: kgunn: standup? [13:30] mzanetti, I have a suggestion for the upstream merger problem, we could try to see if there is an app running in the background already does that remedy the issue. In the past the test suite would fail if unity was killed and no other app was running in the background then the focus issue that we are seeing now will happen [13:31] summary, if there is some app running while unity is killed the focus issue may not happen [13:33] om26er: does the webcam still exist actually? [13:33] om26er: we had a webcam that was giving us a view on the devices [13:33] mzanetti, it does pointed at 3 devices only [13:33] , [13:39] tsdgeos: sorry gonna miss you guys today! [13:39] must be a lonely mumble today :) [13:40] kgunn: it was :D [13:40] like 5 of us [14:06] mzanetti, hello! if you're feeling like reviewing some stuff (who doesn't love that?), my demo and launcher-items branches would love to see the light of day. I can trade reviews :) [14:07] mterry: lol [14:07] mterry: Can you trade for debugging autopilot? [14:08] * mterry slowly backs away [14:12] asac: any findings regarding the permissions on /dev/uinput? [14:46] mzanetti: did you get searching for "a" to give you intenet thingies? [14:46] tsdgeos: yes. currently compiling your branch on the phone (takes ages nowadays) [14:49] :D [14:55] tsdgeos: look at this...http://s-jenkins:8080/job/generic-mediumtests-touch/ [14:55] tsdgeos: the minute I disabled the touch tests in autolanding, they start working again for -ci [14:55] om26er: ^ [14:55] :S [14:56] like they don't get disabled ? [14:56] om26er: no. I disabled them only in unity8-autolanding, but kept them active int unity8-ci [14:57] mzanetti, impossible isn't it ? [14:57] om26er: huh? [14:57] mzanetti, like is there a relation between a CI and autolanding not behaving well ? [14:58] om26er: wild guess: could it be that the device locking between those jobs has an issue and two jobs try to run the tests on the same device? [15:00] mzanetti, we have exactly 1 executor for each device in the jenkins setup [15:00] om26er: right... === _salem is now known as salem_ [15:14] can't reproduce the issue in jenkins any more either... it's like last friday's weather wouldn't have allowed testing :/ [15:25] paulliu: you on? [15:25] paulliu: whether unity-lens-mock should/can go in archive or what's holding it out of it? [15:25] so lool was asking this ^ [15:25] tsdgeos: do you know ? ^ [15:26] not sure about tbh [15:26] but do we need it? [15:27] we're feeding "live" data everywhere now, aren't we? [15:28] kgunn: yeah.. why we want that to be in the archive? I'll talk to lool.. [15:28] paulliu: i actually wondered the same thing when i wrote it :) [15:28] why we want a mock ?...aren't we going to real backends :) [15:29] paulliu: he's in #ubuntu-touch [15:31] Hey, is unity-services needed by unity8? [15:32] Oh, seems it's related to libunity-core-6.0-7 [15:32] thostr_, ^ [15:32] tedg: which means we shouldn't need it... [15:32] So unity8-private deps on libunity-core-6.0-7 which deps on unity-services which deps on libindicator [15:33] thostr_, It probably needs to be broken up if we want to break that chain down. [15:34] tedg: ok, the we better don't touch it for now [15:34] bregma, Is there a reason that libunity-core-6.0 depends on unity-services not just recommends? Then we could have unity depend on the services? [15:36] mterry: can't manage to get your demo branch running [15:39] tsdgeos: I guess MacSlow will have a cigar when he starts working tomorrow. His autopilot tests branch finally merged :D [15:40] mzanetti, :( [15:40] mterry: I executed that dbus command and restarted unity8... nothing [15:40] mzanetti, like it won't run at all, or you don't see the demo? [15:40] mterry: well, I executed the command and rebooted, the screen just stayed black [15:41] let me try again [15:41] mzanetti: yep :D [15:41] mzanetti: he's back on next week though afaik [15:41] mzanetti, you got the 'demo of a broken phone' version [15:41] I see [15:41] looks good then [15:43] mterry: yep, can reproduce it. Could this be related? file:///usr/share/unity8/Components/EdgeDemo.qml:126: TypeError: Object AccountsService(0x17bdea0) has no method 'getUserProperty' [15:45] mzanetti, you installed the debs? [15:45] mterry: yes. the ones I downloaded from jenkins [15:45] mterry: from the last post of jenkins in the merge [15:46] mzanetti, I'll try. Maybe something broke [15:49] mhr3: ping [15:50] tsdgeos: do you know where the home scope code is located? [15:51] tsdgeos: or better, where to modify in order to change the favorite apps in there? [15:54] mzanetti: in the home scope? [15:55] mzanetti: that's probably in some backend [15:55] unless we are overrding it [15:55] that i think we're not [15:55] tsdgeos: no, I don't think we do [15:55] actually we do :D [15:55] oh [15:55] tsdgeos_work@xps:~/phablet/unity8/Dash$ wcgrep -i home [15:55] ./ScopeDelegateMapper.qml:22: "home.scope": "DashHome.qml", [15:56] so DashHome.qml === boiko_ is now known as boiko [15:57] tsdgeos: cool, thanks [15:57] yw [15:57] Dash/Apps/FrequentlyUsedAppsModel.qml actually [15:57] yep [15:59] autopilot tests in VM: 6 min building, 18 min testing [16:00] autopilot tests on phone: 18 min building, 6 min testing === jasoncwarner__ is now known as jasoncwarner [16:09] mzanetti, just installed the debs, they seem to work for me [16:09] (for the demo branch) [16:10] mterry: hmm... strange [16:10] mterry: I can reliably reproduce the issue [16:11] mterry: with those packages: http://jenkins.qa.ubuntu.com/job/generic-mediumtests-builder-saucy-armhf/213/artifact/work/output/*zip*/output.zip [16:11] mterry: installed on an image flashed today [16:11] on mako [16:11] mzanetti, I used http://jenkins.qa.ubuntu.com/job/unity8-saucy-armhf-ci/759/artifact/work/output/*zip*/output.zip [16:11] mzanetti, I didn't flash today though [16:12] mzanetti, I'll try your link. Do you know the difference between those two deb lines that jenkins makes? [16:13] I don't think there's a difference except one builds with coverage enabled, the other without. the mediumtests being the one without [16:14] EOD! [16:14] om26er: I reenabled the touch mediumtests [16:14] mzanetti, they worked for me too. Will try reflashing with today's image [16:14] om26er: can't reproduce the failure any more. jenkins works fine again [16:14] om26er: there must have been a bad image in between or the like [16:21] mterry: I have an idea: [16:21] mterry: I downloaded that .zip, pushed it to the phone, extracted it and did a dpkg -i * [16:22] mterry: which means, I installed unity8-fake-env too [16:22] mterry: could it be, that the accountservice thing has no api in lp:unity-api and the mock implementation does not match with the real one? [16:22] mzanetti, I did *.deb [16:22] as well [16:40] mterry: tried it on a freshly flashed maguro now. it boots, the stays black for longer than normally, then, at some point unity shows up, but without the demo :/ [16:41] mzanetti, yeah I get unity too, but not the demo right now. This used to work, am trying to see what happened [16:42] mterry: ok. I'm calling it a day now. please write some status into the merge so I can review tomorrow morning [16:42] mzanetti, ok [16:42] mzanetti, thanks [17:07] mzanetti, or just that race condition is not happening anymore :) [17:14] * greyback eod [20:49] om26er: still here? [20:50] mzanetti, yep [20:50] om26er: I'm watching jenkins and just cane across this: http://s-jenkins:8080/job/generic-mediumtests-runner-mako/ [20:51] om26er: this was a test run for the addressbook-app but there are artifacts from unity8 in there [20:51] om26er: not a real problem. just noticed it... [20:51] om26er: otherwise its running quite nicely again now [20:52] mzanetti, what's up ? [20:52] om26er: ? [20:53] mzanetti, hm, I see they are for the phone-app [20:53] mzanetti, I could make sure to delete all the .xml in the workspace [20:53] om26er: lol... I messed up with the jenkins links :D [20:54] during each run, so that way we'll always have the right test results attached [20:55] yeah... but I can't find that link any more [20:55] mzanetti, that's magix [20:55] *magic [20:59] mzanetti, btw the address-book suite failed because: address-book-app : Depends: qtdeclarative5-ubuntu-telephony-plugin but it is not installable [20:59] om26er: yeah. its all real failures now. looks really good [21:00] it still annoys me that we have no clue what happened on friday. I'm it'll strike back [21:01] maybe, just maybe there was an update and that resolved the issue somehow ;_) === jhodapp is now known as jhodapp|afk === salem_ is now known as _salem