/srv/irclogs.ubuntu.com/2014/01/20/#ubuntu-ci-eng.txt

=== Ursinha is now known as Ursinha-afk
=== bfiller is now known as bfiller_away
=== Ursinha-afk is now known as Ursinha
=== Ursinha-afk is now known as Ursinha
didrockspopey: hey, once you are around: are you on latest nux/unity (and if so, it's not crashing for you I guess, on the x220?)08:19
popeydidrocks: morning!08:20
popeydidrocks: I'm on saucy on my laptop, trusty on nvidia desktop08:20
popeyso sure, can test latest of whatever on desktop08:20
didrockspopey: that would be excellent if you can confirm :)08:26
popeykk08:27
didrockspopey: or even maybe just a usb stick in live mode for the laptop?08:27
didrockspopey: I think testing on the x220 like that would be helpful :)08:27
popeyI have been resisting moving to 14.0408:29
popeyI will when I migrate to btrfs, probably on the weekend08:29
didrockspopey: yeah, so live usb seems to be the way to go for now :)08:31
popeydidrocks: unity is fine on my trusty desktop08:32
didrocksensure on the live that you have unity 7.1.2+14.04.20131106.1-0ubuntu2 and libnux-4.0-0 4.0.3+13.10.20131011-0ubuntu208:32
popeyah08:32
didrockspopey: I'm interested on the x220 (hence the live usb so that you don't scratch your system)08:32
didrocksglew isn't used on nvidia08:32
popeyah08:32
popeymeh08:33
* popey moves to 14.0408:33
didrockspopey: just use a live session, don't break your laptop!08:37
didrocks:)08:37
popeyI am outraged that you suggest upgrades don't work!08:37
didrocksahah, I'm not. I just don't want you to end up in a session where you can't use your laptop :p08:43
* popey installs gnome-shell ☻08:45
didrockspopey: heh, I fallbacked to it as well!08:47
popeyThe upgrade has completed but there were errors during the upgrade08:53
popeyprocess.08:53
popeybah!08:53
didrocksargh, something else then ;)08:54
didrocksif you can note them :p08:54
popeyhttp://paste.ubuntu.com/6784779/08:54
didrockswaow, more information on those?08:54
didrockslike overwriting files?08:54
popeyapt-get -f install, it wants to pull in another truckload of packages08:55
popeyhttp://paste.ubuntu.com/6784782/08:55
* popey continues with teh configuring08:56
didrocksyep08:56
didrocksI'm pretty sur eyou had an overwrite08:57
didrockswhich blocked08:57
didrocks(would be great to get your apt full logs of the previous update)08:57
didrocksonce you finish your update ;)08:57
popeyfrom dpkg.log?08:57
popeyoh, there's a dist-upgrade log too, handy08:58
didrockspopey: /var/log/apt/term.log rather08:58
didrocksoh ;)08:58
=== ev changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: ev | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
tsdgeosfginther: the inexplicable otto crashes that went away are now back :'( https://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/2150/?09:28
vilatsdgeos: fginther is not up and today is a US holiday IIUC, so better ping the vanguard in the topic line ;)10:19
vilatsdgeos: that said, a quick look in the log file (https://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/2151/consoleFull) reveals :10:19
vilatsdgeos: /home/ubuntu/.xsession-errors: init: unity8 main process (7403) killed by SEGV signal10:19
tsdgeosi know10:19
vilatsdgeos: /var/log/syslog: Jan 20 09:34:02 trusty-amd64-20131210-1707 kernel: [842034.353231] QThread[30955]: segfault at 1cf16 ip 000000000001cf16 sp 00007f50255e0228 error 14 in unity8[400000+c000]10:20
tsdgeosi was not really asking for help10:20
tsdgeosi was just telling fginther10:20
tsdgeosit crashes10:20
tsdgeosjust on those machines and noone else can reproduce10:20
vilatsdgeos: I didn't sync with fginther since Friday but I think it's a different issue10:20
vilaha10:20
tsdgeosthis was happening a few weeks ago10:20
tsdgeosthen it went away10:20
tsdgeosand now is back10:20
vilatsdgeos: unity8 crash ? (Making sure we're indeed talking about the same thing)10:21
tsdgeosyes10:21
vilaok, I didn't heard about that then, sorry10:22
sil2100huh10:23
sil2100After flashing to the latest image, it seems my unity-system-compositor dies10:23
sil2100terminate called after throwing an instance of 'boost::exception_detail::clone_impl<boost::exception_detail::error_info_injector<std::runtime_error> >'10:24
sil2100  what():  could not activate surface with eglMakeCurrent10:24
sil2100I guess it might be the same issue discussed?10:24
tsdgeosnot sure10:25
tsdgeoscan't find that line in any of the logs of the build10:25
tsdgeosbut might be10:25
sil2100tsdgeos: actually, nevermind, misread the discussion just now ;) Carry on!10:26
=== vrruiz_ is now known as rvr
sil2100ogra_: hm, I remember that on the meeting just now there has been some mention of mir session env variable not being set properly, yes?10:28
ogra_sil2100, for adb sessions, yeah10:28
sil2100ogra_: since u-s-c fails on my device after stating: Unable to set active session, unknown client name session-010:29
ogra_that was fixed in 13610:29
sil2100I wonder why it's crashing here10:32
ogra_sil2100, thats an unmodified image 137 ?10:33
ogra_(system image that is ... we know the cdimage install seems to have issues, but thats lower prio)10:34
sil2100ogra_: system image, unmodified, 137 (just downloaded and installed) - maybe it's some user config data?10:35
ogra_shouldnt, but try the fix we are discussing in -touch since a while10:36
ogra_we know there seems to be a race with setting the QPA_PLUGIN var10:36
ogra_err, QPA_PLATFORM10:36
sil2100ogra_: yeah, that fix helped - so it seems MIR_SOCKET thing was the problem10:48
ogra_yep, we see that in the test rrsults between 136 and 13710:49
ogra_thanks for confirming10:50
=== MacSlow is now known as MacSlow|lunch
silDroiddidrocks: had to jump out for a short while, my girlfriends dog is at the vet dying, so had to take her there so she can say goodbyes12:23
silDroiddidrocks: but regarding ubuntu-keyboard12:23
silDroiddidrocks: it seems its broken ;/12:23
didrockssilDroid: argh, just ping directly bfiller_away with those once he's around12:23
didrocks(probably tomorrow)12:24
silDroiddidrocks: the new version just segfaults on start...12:24
didrockssilDroid: thanks for the catchup12:24
rsalvetiogra_: hey, did we already got an image after reverting the nested support?12:24
ogra_rsalveti, building ...12:24
bfiller_awaysilDroid: what are you testing? you need new seed if you're using new keyboard12:24
=== bfiller_away is now known as bfiller
didrockssilDroid: you will be able to do the rest after a while12:25
bfillerogra_: did you have a chance to look at the MR I sent you? https://code.launchpad.net/~bfiller/ubuntu-seeds/ubuntu-touch.trusty-keyboard-english/+merge/20187912:25
didrocksbfiller: isn't today a national holidays for you? :)12:25
rsalvetiogra_: great12:25
ogra_rsalveti, not really12:25
didrockssilDroid: as bfiller suggested, you installed all ubuntu-keyboard-<locale> packages, right?12:25
bfillerdidrocks: yes but I'm working part of it anyway, capetown sprint going on..12:25
ogra_bfiller, sil2100 is on it ...12:26
didrocksbfiller: ah ok12:26
silDroidbfiller: hi! I was testing the ubuntu-keyboard from the ppa12:26
ogra_bfiller, he is testing your package changes first before i make the seed changes12:26
bfillersilDroid: make sure to install ubuntu-keyboard-english12:26
rsalvetiogra_: sure, just wanted a cdimage that could have a working shell at least12:26
silDroidbfiller: hm12:27
rsalvetibut not sure anymore if that was because of the nested merge12:27
ogra_rsalveti, well, we'll land it again ... after the UI ordering issue is fixed i suppose12:27
rsalvetiright12:27
ogra_rsalveti, though by experience that ight take another 8 weeks12:27
silDroidbfiller: will try, but then I guess we need to modify the packaging12:27
rsalvetiyeah =\12:27
ogra_rsalveti, it was ... something is racy in setting MIR_SOCKET apparently12:28
silDroidbfiller: since it leads to a situation that without it installed (which can be done) it crashes on start12:28
ogra_rsalveti, http://ci.ubuntu.com/smokeng/trusty/touch/ ... image 136 and 137 are identical ... no changes at all ... 137 has the same error messages as you reported ... 136 doesnt (and local tests here dont either)12:28
silDroidbfiller: I installed ubuntu-keyboard-polish, set the key language to polish in settings, rebooted and it was the same12:29
rsalvetiogra_: hm, right12:29
bfillersilDroid: you need to set the active-language key12:29
ogra_rsalveti, so it looks like cdimage hits that race condition harder12:29
bfillersilDroid: actually 2 keys need to be set12:30
silDroidbfiller: system settings doesnt do that when switching keyboard layouts?12:30
bfillersilDroid: it's not been updated yet, I talked to attente about it12:30
bfillersilDroid: gsettings set com.canonical.keyboard.maliit enabled-languages "['en','pl']"12:31
bfillersilDroid: gsettings set com.canonical.keyboard.maliit active-language pl12:31
silDroidbfiller: ok, I will try that once back, but I guess we'll have to wait with osk to be released with system-settings then, not to break existing functionality12:32
bfillersilDroid: the existing functionality in system settings doesn't work already12:32
bfillersilDroid: so don't know if we need to wait for it12:32
bfillerbut fine with me12:33
ogra_rsalveti, cdimage part is done12:41
rsalvetiogra_: great, thanks12:42
ogra_didrocks, fyi ... http://people.canonical.com/~ogra/touch-image-stats/20140120.1.changes12:52
ogra_(seems we dragged in new e-d-s and a bunch of other bits12:52
ogra_)12:52
didrocksogra_: ok, so we grabbed latest systemd as well12:52
ogra_yep12:53
didrockslet's see… :p12:53
didrocksthe diff isn't published yet on system-image12:53
didrocksfor mako12:53
ogra_yep, still running12:53
didrocksogra_: I'm anxious now :)12:54
ogra_why ?12:54
didrockswell, I hope the image will be back in shape, e-d-s and systemd… let's see :p12:54
ogra_we built images for years like that :) 90% of them worked12:55
didrocksI think the pass rate is lower for touch though of promotable images :)12:55
davmor2didrocks: why do we have systemd installed? Is it just that half the components we need are integrated into it now?12:55
=== MacSlow|lunch is now known as MacSlow
didrocksdavmor2: we do have for sure udev coming from systemd12:56
didrockssystemd package itself, hum…12:56
ogra_davmor2, systemd swallowed all important plumbing sources12:56
didrocksah, we do use logind12:56
didrocksdavmor2: so that's why ^12:56
davmor2ogra_: yeah that is what I thought12:56
ogra_oh, wait, the binary shouldnt be there !"12:56
ogra_ah, only -services12:57
ogra_that might be for hostname changing12:57
davmor2ogra_: it isn't I just saw about pulling in systemd and wondered why sorry12:57
didrocksogra_: yeah, hostname (and maybe logind is using it too)12:57
ogra_right12:57
davmor2ogra_, didrocks: did you see the announcement that systemd is soon to start running gnome soon too (opt in/out, by all accounts)12:59
* ogra_ only knows the discussion about that from 2 years ago 13:00
davmor2ogra_: https://plus.google.com/+LennartPoetteringTheOneAndOnly/posts/iCNySJKFseo there was a better one somewhere on google plus too I think13:01
ogra_well,let them13:01
* ogra_ doesnt care what gnome uses to be honest 13:02
didrockswell, we do that with upstart as well, we have session support13:02
didrocks(but no kdbus)13:02
=== josepht changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: josepht | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
ogra_there we go13:03
ogra_=== Image 138 DONE ===13:03
didrocksyeah \o/13:05
* didrocks upgrades13:05
didrockshum, should reset to vanilla config first13:05
ogra_nah, that tears out all excitement :P13:06
didrocksahah :)13:06
didrocksI'm still unsure what I shold kill from /userdata13:07
ogra_just flasdh with --wipe/--bootstrap/--whatever-is-the-wiping-option-of-the-day13:07
davmor2ah infact here you go https://plus.google.com/115250422803614415116/posts/apZrj54GYee13:07
didrockswell, I would prefer to start being smarter :)13:07
ogra_pfft, your phones is already smart ...13:08
didrocksheh13:08
didrocksok, trying to rm -rf /usrerdata/*, let's see what happens :p13:08
didrocksat worse, I'll reflash13:08
davmor2ogra_: it was smart with android it has to be smarter now it has ubuntu on right?13:09
rsalvetiogra_: cdimage on maguro still failing with 'Failed to load platform plugin "ubuntu". Available platforms are...'13:10
rsalvetiwith 13813:10
ogra_rsalveti, properly wiped ?13:11
rsalvetiogra_: yup, flashed with -b13:11
rsalvetiogra_: trying to run with SF13:12
rsalvetilet me see with mir now13:12
ogra_stacking issue is gone here on maguro13:12
ogra_account creation still has issues for me though13:12
rsalveticool, worked with mir, let me try what slangasek said on that thread13:13
didrocksflash in progress here with -b13:13
ogra_rsalveti, he reverted that stuff in 135already13:13
ogra_err 13613:13
davmor2ogra_, didrocks: maguro still rebooting, mako bbc news installed13:14
ogra_136 and 137 both have his change ...13:14
ogra_davmor2, did you try removing and re-adding the U1 account ?13:14
didrocksdavmor2: I guess while we are getting a test for wifi, you will need a "wipe data and connect to a wifi hotspot, enter WPA/WEP"13:14
davmor2ogra_: no I'm going to try that on maguro13:14
ogra_hmpf, now system-settings doesnt even get me into accounts anymore13:15
davmor2popey: why is there a big gap at the top of the bbc news app13:15
ogra_thats where popey will place his ads soon ;)13:15
popeythere is?13:15
didrockshum, lot of failures when copying…13:15
ogra_didrocks, copying ?13:16
didrocksfailed to copy '/home/didrocks/Téléchargements/phablet-flash/imageupdates/pool/ubuntu-dbad0141353e07914b8a6fdf462ac9a4e18994eb768c9f4a5c1d0e27e5500c5b.tar.xz' to '/cache/recovery/': Is a directory13:16
ogra_wow13:16
* didrocks retries13:16
* ogra_ reboots again, i definitely cant create a U1 account here 13:17
ogra_yay13:18
didrocksok, passed it seems13:18
ogra_so after 3 reboots it allowed me to13:18
davmor2so I was able to remove the account and add it back again and managed to crash unity8 in the process13:19
* ogra_ installs an app13:19
rsalvetiogra_: right, tried to check the qtubuntu package split, but it fails with both qtubuntu-android and qtubuntu-desktop13:20
rsalvetiSF support is definitely broken13:21
ogra_rsalveti, well, then it is time to call it dead i guess13:22
rsalvetiogra_: I'm not that sure yet13:22
ogra_or do you want to invest time into it13:22
rsalvetiI don't, but I know we had a plan to drop it13:22
* ogra_ doesnt think it matters if we do it now or in two weeks13:22
rsalvetiif so, we should just warn the porters and give them directions for proper mir support13:23
ogra_right13:24
ogra_i think the plan to drop SF was definite ...13:24
ogra_so i would say it doesnt matter if we drop it a few weeks earlier13:24
ogra_but thats an asac decision indeed13:25
rsalvetiright, I know we want to drop it, but we wanted it to be planned :-)13:25
rsalvetinot as a side effect :-)13:25
ogra_rsalveti, well, we would have to find out as side effect of *what* first13:26
ogra_which will eat time13:26
ogra_... and then fix it as well ...13:26
asackeeping SF around was done because of phonedations team think it is a) cheap and b) useful for bringup... once it is a) expensive and b) not really useful anymore, I don't see a reason to keep it around13:27
asac:P13:27
rsalvetiasac: sure, but we should first sync with the oem team13:27
didrocksogra_: ok, both setting up wifi and associating an account works here13:27
rsalvetiif they are ok, then fine13:27
rsalvetiaka achiang13:27
asacrsalveti: yeah, we could check tomorrow on the sim call13:27
ogra_asac, b) is definitely always true13:27
ogra_going step by step via SF to Mir makes porting surely easier13:28
didrocksdavmor2: popey: can you try image #138 and test it on maguro/mako?13:28
ogra_but it doesnt lok like a) is still cheap13:28
asacsure, but if SF doesnt deliver a UX that is anywhere useful, we dont need to support it anymroe in unity etc.13:28
popeydidrocks: ya13:28
asacyou can still bring up ubuntu touch13:29
asacand run an example app that works on SF for instance13:29
ogra_rsalveti, when did you try SF last ? can we even make out a point when it stopped ?13:29
davmor2didrocks: I am13:29
didrocksthanks popey, davmor213:29
rsalvetiogra_: last week or sure, probably before the qtubuntu package split13:29
didrockspsivaa: tests started on #138?13:29
ogra_ok13:29
ogra_so we would only have to bisect a week13:30
popeydavmor2: did you change the permissions on that spreadsheet?13:30
rsalvetiogra_: I know because I was using SF with 4.413:30
asachttp://ci.ubuntu.com/smokeng/trusty/touch_sf4p/13:30
ogra_yeah, just remembered that13:30
asacseems it was last time safelyy green13:30
davmor2popey: nope13:30
rsalvetiogra_: and MIR support for it is still not merged13:30
asacfro build 12113:30
popeydavmor2: i can't access it.. it's "comment only" for me13:30
ogra_asac, we're at 13813:30
ogra_thats quite a few images to check13:31
asacright. you asked if we know when ti was last working13:31
asacdashboard says 121 or maybe 12413:31
ogra_yeah, thannks13:31
asac126 :)13:31
ogra_i had forgotten we still do sf tests13:31
davmor2popey: try it now13:31
popeythanks13:32
davmor2popey: it was on comment only but I hadn't touched it13:32
popeyodd. nvm13:32
asacrsalveti: so what do you think?13:32
asacresurrect SF or good bye?13:33
asaccan it even be resurrected? in particular with this nested stuff coming etc.13:33
asacwhat everyone surely agrees is that we shouldnt start hacking SF itself :)13:33
rsalvetiasac: probably, but as I said, not my call, we should sync with people that are doing bringups13:33
asacrsalveti: ok lets suggest to achiang and folks tomorrow that we kill it and see13:34
rsalvetialright13:34
asacrsalveti: i think you could try answer this question yourself if you ask "would i be screwed if i dont have SF to bringup a N5"13:34
asac(assume you really shoot for MIR and are wiling to invest what it takes anyway)13:35
rsalvetiasac: yeah, personally I'd prefer to invest my time with MIR13:35
ogra_yeah13:36
ogra_its just easier to get some success feedback with SF currently13:36
ogra_since it will just work13:36
ogra_(well, used to)13:36
ogra_while Mir needs extra work added13:37
ogra_but would likely be a good thing to get used to that extra work13:37
ogra_for us as well as for the porters13:37
ogra_(so we can answer their questions)13:37
popeydidrocks: mako seems okay at first pass13:45
didrockspopey: great!13:46
psivaadidrocks: yes, tests are running with 13813:56
ogra_oh, and all 100%13:57
ogra_stop them now !13:57
ogra_:)13:58
didrocksahah ;)13:58
davmor2popey: for you on your bbc new app do you see a grey block at the top of the app where the header should be (like a blank title bar)14:14
ogra_davmor2, hmm, might be the new webapp-container thingie14:15
ogra_did you try another webapp ?14:15
davmor2ogra_: no I'm just running tests now14:15
davmor2g+ is alright but that has it's own config iirc14:16
ogra_davmor2, i see it in all webapps apparently ... just checked14:16
ogra_G+ doesnt use the webbrowser-app webapp mode14:16
ogra_use IMBD14:16
ogra_or dict.cc14:16
ogra_thats definitely broken :(14:16
davmor2ogra_: indeed14:19
ogra_davmor2, see -touch ... talking to osomon14:19
popeydavmor2: yes14:29
=== alan_g is now known as alan_g|tea
=== alan_g|tea is now known as alan_g
* popey afk16:30
davmor2ohhh I want popey 's telepathy hat if he can type and not be at the keyboard16:36
ogra_i guess thats only a telepathic macro ... only says "afk"16:47
=== josepht changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -
sil2100Ok, webbrowser-app is good17:49
ogra_let it in then :)17:50
sil2100Pushing ze button17:51
sil2100\o/17:51
=== alan_g is now known as alan_g|EOD
ogra_=== Image 139 Building ===19:11
asacogra_: did we get all blockers fixed in that?19:26
asace.g. regressions are under control?19:26
davmor2asac: I think it fixes the blockers between this and reverting nested mir iirc19:36
asacdavmor2: not sure i understand... i think you say "no, it doesnt fix all promotion blockers"19:38
asacwhat blocker do we expect to be still in?19:38
davmor2asac: No I think it fixes all the blockers19:39
davmor2asac: between reverting nested mir in 138 and the webbrowser-app fix in 139 I think all the blockers are fixed19:39
davmor2ogra_: do you know if didrocks was going to promote 139 if everything was fixed?19:41
davmor2asac: ^19:41
davmor2or are we looking to tomorrows images now?19:41
popeydavmor2: i asked this and in all likelyhood we won't promote this one because the next one will arrive and we'll be testing in the morning20:05
popeybut if it's good in the morning and the next one 140? is bad, then sure, we could20:06
* popey updates to #13920:25
asacdavmor2: didrocks would decide tomorrow morning based on findings20:40

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