[01:05] robru, looks like it merged about an hour ago. The build machines are still under load [01:05] fginther, ok, thanks === _bjf is now known as bjf [04:12] balloons have you noticed that calendar seems to just fail on maguro for the most part? Not saying that makes any sense, just an observation [04:13] balloons: I mean, it does fail on mako too, but not nearly as much. On maguro it's really hard to make it pass though. maybe something timing sensitive? [06:43] Mirv: hey, can you try the camera_app AP tests? (as we have those crashers) to see if you can reproduce with latest image [06:43] and then latest image + camera-app in distro? [06:43] that will help to have first hints for the meeting [06:43] (now that you published libunity, thanks again! ;)) [06:48] didrocks: ok, I ran them once today before publishing the new camera-app, but I'll try some complete reflash + rety. [06:48] Mirv: great, thanks! [06:50] didrocks: so, the DNS has been migrated [06:50] vila: great news! Let's cross fingers now :) [06:50] * retoaded announces that he has switched all DHCP services over to the new DNS/DHCP server(s). Networking will need to be restarted on just about everything on the 10.97.2.0/24 network for the system to pick up their new DHCP leases and DNS servers unless you want to wait for the 24 hour DHCP renewal to take care of it. [06:50] vila: next topic was those autopilot machines for which the nodes are going down randomly [06:51] ^ that was yesterday evening [06:51] hehe [06:51] (again today, the nvidia machine) [06:51] retoaded: I guess I have the next topic (we already discussed that) for you ^ [06:51] something to monitor those machines [06:51] reboot (electrically) if we can't ssh to them for 10 minutes [06:51] and restart the jenkins node if down [06:51] otherwise, rocking on the DHCP! :) [06:52] didrocks: http://10.97.2.10:8080/job/webapps-autopilot-next-daily/306/label=quantal/testReport/webapps.tests.test_hud/HudTests/test_webappicon_firefox_/ [06:53] this is a selenium failure I fixed for sst, the way webapps-tests defines browser launch is way too optimistic, not sure who I should talk to about that [06:53] rsalveti: great work! you should go to bed now :) [06:53] vila: I think this is dbarth's team [06:53] (vvruiz should be the one in charge of those tests AFAIK) [06:54] vrruiz_: ping [06:54] vila: pong [06:54] vrruiz_: http://10.97.2.10:8080/job/webapps-autopilot-next-daily/306/label=quantal/testReport/webapps.tests.test_hud/HudTests/test_webappicon_firefox_/ [06:55] vrruiz_: this is a spurious selenium failure [06:56] * vila searches relevant revision in sst [06:57] vrruiz_: http://bazaar.launchpad.net/~canonical-isd-qa/selenium-simple-test/trunk/revision/425 [06:58] vrruiz_: in a nutshell, building a selenium driver can (and will) fail at times, you have to retry [06:59] Hmm [06:59] vrruiz_: main point of interest in this huge revision is in src/sst/browsers.py (but src/sst/cases.py is also relevant) [07:00] vrruiz_: I realize you can't easily reuse that code but the feature you need is there :-/ [07:00] vrruiz_: mostly I wanted you to be aware that the failure in the url above won't magically disappear [07:00] vila: One question [07:00] until *you* do something about it [07:00] sure [07:01] vila: Do you know which Launchpad project/branch is that job using? [07:02] vrruiz_: lp:webapps-tests ? [07:02] vrruiz_: at least that's where I found the overoptimistic line [07:03] vrruiz_: http://bazaar.launchpad.net/~vrruiz/webapps-tests/trunk/view/head:/autopilot/webapps/emulators/browser.py#L18 [07:03] vrruiz_: i.e. just building a driver without catching the exception and retrying as in src/sst/cases.py roughly (but sst uses a modified driver too) [07:04] Ah, I see [07:04] vrruiz_: oh, indeed, the trunk for that is yours ;) [07:05] vrruiz_: I didn't dig from there so may be you can just use sst and be done [07:06] Someone should package sst and put it in universe :) [07:06] (or main ;) [07:06] vrruiz_: it is packaged but not put in universe (my packaging fu is weak :-/) [07:06] Yeah, elopio told me there is a PPA [07:07] https://launchpad.net/~vila/+archive/selenium/ [07:07] vrruiz_: and (basic) instructions on how to update it in sst: doc/packaging.rst [07:08] vrruiz_: and (basic) instructions on how to update it in sst: docs/packaging.rst [07:09] Thanks, I will take a look [07:09] vrruiz_: ping the qa team about that maybe, elopio and cgoldberg may want to address that now (I pushed for that but didn't go further than the ppa) [07:10] Yup [07:20] Mirv: I tried here and I don't have any failure, is it the same for you? [07:20] (camera-app) [07:27] didrocks: http://10.97.2.10:8080/job/gallery-app-saucy-i386-ci/352/console rings a bell ? [07:27] vila: I would say no more loop device available on the machine [07:28] which makes pbuilder going crazy and can't unmount others [07:28] wow, lucky it rang a bell, would never have suspected a lack of device leading to unmount failing 8-) [07:29] vila: it's a crazy chain of cause/consequence when I looked at it ;) [07:29] didrocks: /var/cache/pbuilder/build also contains stuff as old as March 2013 [07:29] didrocks: on kinnara that is [07:29] * ogra_ doesnt understand the camera-app issues [07:30] fginther: can use your help ^ [07:30] ogra_: me neither, I asked Mirv to look at it (be he doesn't seem around) [07:30] ogra_: I tried locally with image 73, all good [07:30] (multiple times) [07:30] didrocks: hi [07:30] hey Mirv! [07:30] http://people.canonical.com/~ogra/touch-image-stats/20130930.changes [07:30] didrocks: yep, no failures, have tried twice with both older and newer camera-app [07:30] it wasnt updated either [07:30] fresh -b flashing [07:31] Mirv: upgrade here, but same :/ [07:31] vila, woah [07:31] ogra_: yeah, nothing that I can spot on [07:31] fginther: errk, you're here ! OMG, shouldn't you be sleeping ??? [07:31] hmm, pulse was updated [07:31] lots of landed over night :-) [07:31] lool, lots of broke over night :( [07:31] vila, yes [07:31] ogra_: well, we do have that one locally as well… [07:32] fginther: http://10.97.2.10:8080/job/gallery-app-saucy-i386-ci/352/console can't we just kill it ? I suspect it will never get out of that loop [07:32] fginther: but I'm still hesitant to kill jobs without someone confirming ;) [07:33] fginther: http://10.97.2.10:8080/computer/kinnara/ seems healthy otherwise (I've just a couple of other jobs and they progress) [07:33] ogra_: can you update as well and tell us? [07:33] ogra_: if we are 3 with all tests passing… [07:33] vila, yes, it's in a bad place, the job will be timed out after two hours, but no reason to wait that long here [07:34] fginther: ha great, where can I see that timeout ? [07:34] didrocks, takes a while, i only have 2MBit here ... gimme 30min [07:34] fginther: no, stop, go to bed, I don't want to retain you here ;) [07:35] ogra_: still 300 MBit here (well, less as I'm connected through wifi to my router, but I won't complain :p) [07:35] didrocks, there is a .crash file in the tests [07:35] did you check /var/crash ? === vrruiz_ is now known as rvr [07:35] vila, alight, I go back to sleep now :-) [07:36] fginther: job killed, will monitor kinnara a bit, have nice dreams ;) [07:36] ogra_: yeah, I don't have any crash here [07:36] ogra_: but let's give that to upstream [07:36] ogra_: image 74? do you know why? [07:36] with no packaging changes… [07:36] why what ? [07:37] why an image 74? [07:37] 71, 72, 73 I can understand [07:37] dunno, i didnt build it [07:37] but why this 74? (just appeared) [07:37] lool, yours ? ^^^ [07:37] ogra_: this is not a bug, right? http://people.canonical.com/~ogra/touch-image-stats/20131001.2.changes [07:37] it's really like, there is nothing… [07:38] didrocks, yeah, there are only 2h between the two builds [07:38] ok, /me sed -i s/74/75/ on the spreadsheet [07:39] annoying we don't know who do builds [07:39] 2 or 3 for system-image ok [07:39] but we shouldn't do more :p [07:39] heh, we should do a lot more imho :) [07:39] * ogra_ would like hourly builds [07:39] automated ones [07:40] ogra_: well, it's harder to track, I spend my time updating the spreadsheet with build numbers [07:40] so if we do that, we need more automation for all the tracking [07:40] but that would mean that the build *and* the full tests would have to finish in under 1h [07:40] oh, indeed [07:40] the spreadsheet doesnt scale [07:41] the point is, if you could automate all this at such a fine grained manner, you will be ablet to have one image per change automatically [07:41] that means you will spot any regression immediately and will know exactly where it come from [07:41] *comes [07:49] didrocks, i would replace the spreadsheet with an LP form where people put in everything for one changeset to create a ticket and the "lander" has a button called "do testbuild" that runs the whole process of building, publishing, image creation and a full testsuite to show you the results ... if they are good, the system publishes them with one click from you [07:50] argh, sil2100 won't be able to connect :/ [07:50] ogra_: yeah, we are more or less inline on that [07:51] having feature landing [07:51] instead of trunk landing [07:51] anyway, let's see with what we have ;) [07:51] Mirv: ok, so as sil2100 isn't around, do you mind taking the Mir request? [07:52] Mirv: landing #52 === psivaa-afk is now known as psivaa [08:04] nope, I didn't build that image [08:04] weird [08:05] did someone launch in the build in the morning? [08:05] apparently [08:05] Cause Steve L still had one running not too long ago, but he finished testing ~1h ago, however if someone had launched another build, it probably started just while Steve's finished [08:05] i see rsalveti spun 72 [08:06] oh why did steve run a build ? [08:06] Steve did 73 and 74 [08:06] he wanted to test upgrades with system-image [08:06] ah [08:06] well, then we have identified them all [08:06] it's what I suggested yesterday evening actually, except it took the whole night to complete landing and testing of this [08:07] yeah [08:07] (this timezone thing is awesome0 [08:07] i went to bed then it FTBFS the first time [08:07] lool, does it work for you ? [08:07] didrocks claims it doesnt [08:07] (i'm still flashing here, maguro sucks) [08:07] what doesn't work? [08:07] TZ selection [08:07] haha [08:08] oh [08:08] I just meant that it was great for US folks to be in another TZ to finish the work while we sleep [08:08] yeah, got it now [08:08] but I haven't tested the touch TZ support :-) [08:08] to much timezines in my head today [08:08] *zones [08:08] so what do we know about the camera-app crashes? [08:08] exactly nothing [08:08] http://people.canonical.com/~ogra/touch-image-stats/20130930.changes [08:09] thats the change set after which it started [08:09] lool: I sent an email to olivier, ugo and bill [08:09] discussing right now on ubuntu-touch [08:11] didrocks, hmm are you sure we're not seeing an UI issue wrt timezone selection ? i cant even pick it here [08:11] seb128, ^^^ [08:11] is that supposed to work ? [08:11] ogra_, if you type in the text entry you should get locations [08:11] I get location [08:11] I click on one [08:11] seb128, i do, but tapping a location does nothinng [08:11] and nothing [08:11] didrocks: ok, there seems to be a lot to test, maybe sil2100 can then take some of the AP:s if he arrives and I've still tests to run [08:11] didrocks, ogra_: btw, whoever landed that content-hub update without updating setting probably regressed background selection again :/ [08:12] Mirv: right, but this is prio #1 for us, so can you start on those? maybe then sil2100 can pick up when he arrives [08:12] seb128: it's ken I guess [08:12] didrocks, ogra_: that's likely https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1227520/comments/44 [08:12] Ubuntu bug 1227520 in apparmor (Ubuntu) "Timezone changes are not working due to ro /etc and bind mounts" [High,In progress] [08:12] seb128: he didn't ask [08:12] nothing written [08:12] so please check with him [08:12] didrocks: yep [08:12] didrocks, no, ken and I agreed yesterday that I would do the organize the landings with you today [08:12] seb128: someone published the stack without any requests [08:13] didrocks, he said he would land the content-hub commit in trunk and a gallery-app rebuild, but those shouldn't have been uploaded [08:13] didrocks, could be robru? [08:13] maybe he miscommunicated with robru and he Robert landed it? [08:13] I'm pretty confident Ken didn't do it [08:13] we had a 15 min discussion on how to land settings content-hub and gallery-app synced [08:14] yeah, but what I told is that he maybe discussed about it with robru [08:14] seb128, well, i know that apparmor is broken, but is the UI supposed to work ? this zero feedback is weird [08:14] and Robert misunderstood [08:14] ogra_, it's supposed to show as selected the timezone, try on your desktop [08:15] heh, my desktop runs precise :) [08:15] i belive you ;) [08:17] ogra_, https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1232114 [08:17] Ubuntu bug 1232114 in ubuntu-system-settings (Ubuntu) "Adding timezone behviour is odd" [Low,New] [08:17] ogra_, that's a bug about the lack of feedback/current behaviour [08:18] yeah [08:20] oh, wait ! [08:20] it sets the timezone right [08:20] so? [08:20] * ogra_ sees "Europe/Berlin UTC+0" [08:20] haha [08:20] \o/ [08:21] and nothing in any logs :( [08:21] ogra_, it should show Berlin in orange in the list [08:21] in the UI [08:21] well, i used my hometown [08:21] not berlin [08:22] and it shows it highlighted in orange now [08:22] ogra_, it should show as orange all the entries matching your tz [08:22] that doesnt fix the UTC offset though [08:22] it does [08:22] ? [08:22] what offset? [08:22] but my clock is still wrong indeed [08:22] UTC+0 [08:23] was there any time change? [08:23] unless berlin moved to london recently, thats wrong :) [08:23] we had bugs with clock in the past where the new time would only be picked on refresh [08:23] no i freshly flashed [08:23] e.g the new minute [08:23] with a complete re-bootstrapping [08:23] that's not what I meant [08:24] i flashed freshly, and then went to the timezone selection [08:24] I meant, did you wait a minute, in case the indicator is just too stupid to refresh the time during the minute (where the UI is not supposed to changed) [08:24] the TZ selector says UTC+0 [08:24] i doubt it will adjust anything [08:24] what does "time" says in adb? [08:24] root@ubuntu-phablet:/# date [08:24] Tue Oct 1 08:24:45 UTC 2013 [08:25] shrug [08:25] (assuming you meant date) [08:25] yes [08:25] is /etc/timezone having the right content? [08:25] so I'm going to revert to 73 and test an upgrade to 74 [08:25] pitti posted a comment in the bug saying there were apparmor problems still [08:25] root@ubuntu-phablet:/# cat /etc/timezone [08:25] cat: /etc/timezone: No such file or directory [08:25] aha [08:25] :-( [08:25] Laney, yeah, I mentioned that before [08:26] didrocks, ogra_: that's likely https://bugs.launchpad.net/ubuntu/+source/ubuntu-system-settings/+bug/1227520/comments/44 [08:26] oh ok [08:26] Ubuntu bug 1227520 in apparmor (Ubuntu) "Timezone changes are not working due to ro /etc and bind mounts" [High,In progress] [08:26] root@ubuntu-phablet:/# ls /etc/writable/ [08:26] root@ubuntu-phablet:/# [08:26] doh [08:26] :( [08:27] let me update here/try [08:27] is it right in the image? [08:27] http://people.canonical.com/~ubuntu-archive/livefs-build-logs/saucy/ubuntu-touch/latest/livecd-20131001.2-armhf.out says good things [08:28] Laney, it does good things ... [08:29] * ogra_ checks livecd-rootfs code [08:29] ok, let's move that to -touch with pitti [08:29] on other news, robru's landing regressed the background selection [08:29] we need to land [08:29] - a rebuild of gallery-arpp [08:29] - one commit to ubuntu-system-settings [08:30] - the new ubuntu-system-settings to saucy then [08:31] Laney, i suspect the code doesnt cd into the right place when linking and moving [08:48] psivaa: ok so if you have time to also run mir testing, I tried to do mir testing instructions (comments welcome) at http://pastebin.ubuntu.com/6178812/ [08:48] after, those running the actual tests [08:48] Mirv: ack, will do [08:48] ogra_: does that pastebin look correct to you? ^ [08:48] psivaa: thanks! [08:48] Mirv: yw [08:49] Mirv, touch .display-mir needs to happen in /home/phablet [08:50] beyond that, looks ok to me ... make sure to test unity8 last, else the tests behave weird [08:53] ogra_: thanks [09:03] psivaa: hmm, first things first, can you check with top if you also get 100% CPU usage of unity8 under mir? [09:04] Mirv: just doing a fresh flashing, ill check once that's done :) [09:05] psivaa: yeah, there's quite a log to flash/install/etc. I just noted that myself. [09:07] hmm now after another reboot it calmed down, weird [09:09] didrocks: autopilot does not seem to work for me under mir. nothing happens, the autopilot process is there but doesn't do anything. [09:10] psivaa: same for you? ^ [09:10] I've tried -n unity8, camera_app etc, multiple reboots [09:10] didrocks, so with 74 i cant get the camera-app to start at all on maguro ... just FYI [09:10] didrocks: just flashing.. not yet finished :) [09:10] ogra_: hum, we can on mako here [09:10] so, how do I get a gallery-app update to land? [09:10] that seems to be enough to make the background selection works again [09:11] (it needs to be rebuilt with the new content-hub) [09:11] Mirv: ok, but there is no regression visible? [09:11] didrocks, with a fresh --no-backup flash ? [09:11] Mirv: if you play with it? [09:11] ogra_: yeah [09:11] weird [09:11] ogra_: but on majo [09:11] mako* [09:11] right [09:11] same for lool [09:11] ogra_: can you try again (after waiting the session to start, like 1 min) [09:12] yeah, let me reboot [09:12] it seems to be linked to usermetrics [09:12] seb128: you need to check with bfiller, he told me to not release apps without his green flag [09:13] seb128: and I see no landing ask from him on this [09:13] (it seems they don't have an always releasable trunk) [09:13] didrocks, what about doing a direct "no change rebuild" of gallery-app to saucy? [09:13] seb128: once you get his ack, I'll get that in priority [09:13] seb128: fine with me if that's really enough and you backport the changelog upstream [09:13] didrocks, nope ... reboot and 1min wait before doing anything didnt change it [09:13] urgh [09:13] didrocks, let me double check, I get the gallery-app from the daily-ppa, but ken told me a rebuild was the key there [09:14] i get a white screen [09:14] thats all [09:14] didrocks, coming back to you about that in a bit [09:14] ogra_: let me reflash [09:14] seb128: thanks! [09:14] didrocks, yw ;-) [09:15] didrocks: I don't have a reference to compare, but it seems to work, yeah. no flickering for example. [09:16] Mirv: let's wait for psivaa's result, and please email with those results kgunn [09:16] didrocks: but I'm seeing this 100% CPU usage of the unity8 process (not always, but for example on this boot) [09:16] yeah, when the screen is blank [09:16] this is a known thing [09:16] ok [09:17] expected to still exist [09:17] didrocks: also otherwise [09:17] image visible, idling on eg Applications screen, 100% CPU [09:17] interesting [09:17] Mirv: can you detail those in the email? [09:17] (please CC me and olli) [09:18] didrocks: ok. after psivaa has some additions to the report. [09:18] yep [09:19] tested with Mir, doesnt work either [09:19] (though i get a black screen now ... ) [09:19] ogra_: AP or camera app? [09:19] the app [09:19] i wanted to check wallpaper settings [09:19] not even the cam [09:19] :) [09:20] ahah [09:24] ogra_, wallpaper settings are not working again (thanks to robru's landing of content-hub) [09:24] seb128, yeah, i wanted to verifyx that ... but am not able to take a picture :) [09:25] ogra_, just copy stuff to ~/Images over mtp [09:25] that's what I do [09:25] ups [09:25] well, i usually take a closeup pic of something :) [09:25] ~/Pictures I mean [09:26] that way you test both, camera and wallpaper settings at the same time :) [09:26] right [09:27] and its funny trying to find structures to make nice wallpapers from [09:27] ogra_: verifyx, nice one, you fix stuff as you verify it works, you just rock ;) [09:27] :D [09:39] Mirv: 1. i do not get 100% CPU for unity8 under Mir on maguro [09:39] Mirv: 2. unity8 AP test does not pregress.. like you said above [09:40] Mirv: the screen was blank when that happened and there was no response to the power button press [09:51] Mirv: the same thing even after rebooting (removing the battery etc) for other AP tests [09:51] didrocks: Mirv: i see _usr_lib_arm-linux-gnueabihf_hud_hud-service.32011.crash and _usr_bin_maliit-server.32011.crash on reboots [09:52] psivaa: ok, anyway, as long as we don't have AP tests running [09:52] let's not go further [09:52] we can publish Mir, it's not enabled by default [09:52] Mirv: mind doing (that + email?) [09:53] didrocks: ack [09:53] psivaa: sounds reasonable to you? [09:53] ok ;) [09:55] didrocks, the mediaplayer breakage was expected btw (just seeing your mail) [09:56] didrocks, the tests were supposed to be off though [09:56] didrocks: Mirv: just to make sure, i disabled Mir (by deleting .display-mir) and then run AP tests and that's working fine [09:57] ogra_: no, it's another issue [09:57] ogra_: like the app can't start [09:57] oh [09:57] (see, crashing in setup()) [09:58] ogra_: normally, just 2 tests are disabled [09:58] not all :) [09:58] psivaa: \o/ [09:58] psivaa: thanks for confirming [09:58] well, everything that uses the scene selector and thumbnails afaik [09:58] psivaa: can you install Mir on your desktop and reboot as well? [09:58] psivaa: so that we are all sured [09:58] ogra_: right, so 2 ;) [09:58] ogra_: but it's crashing in setup(), when trying to launch the app [09:58] didrocks: ack, will do [09:58] I bet it's the same issue than camera-app [09:58] psivaa: thanks! [09:59] didrocks: yw [10:02] didrocks, https://launchpad.net/ubuntu/+source/gallery-app/0.0.67+13.10.20130924.1-0ubuntu2 and https://code.launchpad.net/~seb128/gallery-app/backport-saucy-upload/+merge/188539 [10:02] seb128: rocking! [10:03] ;-) [10:04] ogra_, ^ that should give us back working background selector [10:04] awesome ! [10:04] now if i could take a pic to use as background that would be even more awesome *g* [10:05] use one of the lovely stock wallpapers we ship [10:06] psivaa: ok, thanks for the report and the fact that you have maguro, maybe the 100% CPU happens only on mako [10:06] Laney, can't do that, the picker only picks from the gallery ... we should probably add those to the gallery by default though [10:06] if you copy them then they appear in the gallery [10:06] Mirv: yea possibly [10:06] right [10:06] but it would be nice to have some content in the gallery by default ;-) [10:07] yes [10:07] I don't think the wallpapers are even installed yet though [10:07] didrocks: doh, the new tick just started, I'm uncertain whether cancelling is ok when prepare job is already running? [10:07] it won't pick up any new changes though, as there are none [10:08] seb128: I agree, like on desktop the wallpapers package [10:09] Mirv: is it wait on stack? [10:09] didrocks: no the mir is the first one to build so it's doing that [10:09]