[00:15] plars, oh, do the devices have network access? [00:26] Saviq: yes, you'll need to use phablet-network, and pass the config in for it to use. The config with wifi credentials for that lab get deployed to: /var/lib/jenkins/.ubuntu-ci/wifi.conf [00:45] plars, thanks! === salem_ is now known as _salem === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [11:06] pete-woods: ping [11:08] rvr: hi [11:08] pete-woods: Hi. I'm testing silo 56 [11:08] pete-woods: The VPN test is ok [11:08] rvr: okay, well that's something, at least [11:09] pete-woods: But when I open the VPN settings from the indicator, System Settings crashes [11:09] Or at least it opens and closes [11:09] rvr: yeah, system settings doesn't have the VPN UI in yet [11:09] so it just closes [11:09] you can't add VPN connections without manually installing my click package at the minute, though [11:10] so this isn't something a regular user can see [11:11] pete-woods: Hmm... I see [11:11] pete-woods: So, what's the plan? [11:11] rvr: jgdx is currently building the final VPN UI for USS [11:11] but it's not finished yet [11:12] this is why I hacked up that VPN editor app quickly [11:12] so that QA would have a decent UI to test with [11:12] instead of having to use console commands [11:13] at any rate, the VPN link appears with or without this silo [11:13] this just fixes the password storage [11:13] (which is completely broken beforehand) [11:14] pete-woods: Hmm [11:14] jgdx: The VPN UI for USS, is it planned for OTA10? [11:15] rvr: he's probably not started work yet [11:15] I think he works afternoons [11:15] pete-woods: Oh [11:15] but i *think* that's what he's aiming for, yes [11:18] Ok [11:54] pete-woods: Silo approved [11:54] rvr: awesome, thanks for testing it :) [12:24] * sil2100 off to lunch === tsimonq2alt is now known as tsimonq2 === _salem is now known as salem_ [12:28] rvr, yes, and I had started. Working while the cortisol levels are high [12:30] pete-woods, filed bug 1550283 if you didn't see it already. [12:30] bug 1550283 in indicator-network (Ubuntu) "Connectivity VPN API, hard to change active connection " [Undecided,New] https://launchpad.net/bugs/1550283 [12:30] pete-woods, also, is the route backend planned/wip/shelved? [12:30] jgdx: yep. saw it, thanks [12:30] some vpns are useless without [12:31] I have literally no time to work on it at the moment [12:31] I would definitely like to add the routes feature [12:31] even the check box for "only route this connection's ip ranges" would help a lot [12:31] yeah, it would' [12:32] the bug above becomes superfluous if the connectivity api can detect changes to a live connection and then apply them by reconnecting [12:32] pete-woods, ^ [12:33] jgdx: yeah, that'd be a nice fix to it, and totally feasible [12:34] pete-woods, great, added to bug. [12:35] jgdx: btw, if you did want to re-use any of the QML code from my test UI (or indeed you have replacement stuff), I have no problem at all with a MP for indicator network to add it to the connectivity QML plugin [12:38] pete-woods, yeah, I've used a bunch of it. Most prominently the file selector (since we don't have a certificate manager yet). But, we're adding it to ubuntu-settings-components instead [12:39] pete-woods, that's the current plan, anyway. [12:39] jgdx: okay, that works for me too [12:40] pete-woods, I've made it so it doesn't explicitly depend on indicator-network. Not sure if that makes any sense, because you'll have to guess the vpn types. [12:40] connection.type === 0 // Openvpn, etc [12:41] jgdx: surely you still depend on the connectivity API to get the model / know about the various properties on the connections? [12:43] pete-woods, yes, but that's currently in the consumer of u-s-components. So Settings/Wizard passes along the model. [12:43] ah, okay, well fair enough if you think it makes it more easy to move around [12:44] i'm not sure that it does, and I'm not sure I like the implicit dep. I mean, you don't implement a freedesktop-like spec, right? [12:44] pete-woods, portSet and the like is a pete-woods (tm) concept [12:45] indeed, well the whole thing is a pete-woods (tm) concept, really [12:48] how can I retrigger the autopackage testing? https://requests.ci-train.ubuntu.com/#/ticket/1040 seems to have failed a downstream test due to a passing system image issue [12:55] trainguards, how can I retrigger the auto autopackage test on: https://requests.ci-train.ubuntu.com/#/ticket/1040 ? === alan_g is now known as alan_g|lunch [13:13] kdub: you're meaning that they'd start, it should be automatical and we've no manual triggering for their starting. if you'd have the excuses page we could help with retrying already failed tests. [13:14] Mirv, thanks, guess I should have been more patient before toggling === alan_g|lunch is now known as alan_g [14:00] dbarth__: alexabreu: Silo 45 approved [14:02] dbarth__: alexabreu: Remember that silo 16 needs a rebuild [14:02] rvr, yes I am working on solving the conflicts & repushing [14:03] alexabreu: Ok! [14:19] jibel: 049 already passed automatic signoff, somehow a retry on one flaky unity8 autopkgtest was done. it's now running yet again so I guess that would mean it could create another card on the trello board, but on the other hand the no fields were toggled so I'm not sure if it will work or not. [14:19] (a retry was done on a test that already passed since it got into QA queue) [14:23] Mirv, yes a new card will be created when it's ready again. I'm leaving the card on the board for now and will delete the new one [14:36] kenvandine: Silo 23 approved [14:38] rvr, thx [14:50] jibel, hey, could you please point me at your jenkins-job-builder files? I've started to write yaml for our jobs but having issues wrapping my head around it... because I need to allow others set up some defaults, I'm trying to use includes and anchors, but JJB just complains about everything :( === alexabreu is now known as alex-abreu [14:54] morphis: hey! I noticed that in my android silo britney migration-testing fails on unsatisfiable Depends: ubuntu-emulator-runtime etc. - did you have that with your android silos too? [14:55] sil2100: last time I landed that the package britney wasn't active for citrain :-) [14:55] Bummer ;) [14:55] sil2100: where is ubuntu-emulator-runtime coming from? [14:56] jibel, it's almost like JJB expects all the context for a file to be included in that file, which feels rather limiting :/ [14:56] morphis: from android ;p [14:56] android-emulator/amd64 unsatisfiable Depends: ubuntu-emulator-runtime <- looks strange [14:56] sil2100: wonderful :-) [14:57] a bit [14:57] I think this might be something that's confusing for britney [15:01] can we retrigger the qtmir-gles here?: seems its a known intermittent failure in the downstream https://requests.ci-train.ubuntu.com/static/britney/xenial/landing-032/excuses.html [15:06] sil2100: do you know if the image tests that were happening with arale still do anything useful? I'm guessing no, but that the dev-jenkins-server just never got shut down. I'd like to make those available to some other teams, but wanted to check first that they are not in use [15:07] trainguards: can we retrigger the qtmir-gles here?: seems its a known intermittent failure in the downstream https://requests.ci-train.ubuntu.com/static/britney/xenial/landing-032/excuses.html [15:08] plars: hm, I don't remember those being used, so fine with me at least [15:08] Not sure what other teams could have been using that [15:08] kdub: on it [15:08] thanks sil2100 [15:11] sil2100: those were just the ci.ubuntu.com style jobs, and afaik that part of it is no longer maintained [15:11] sil2100: so it would have just been you :) [15:14] rvr, ok approved https://requests.ci-train.ubuntu.com/#/ticket/955 [15:22] sil2100: that's one of the packages that has a specialised hack for the Ubuntu proposed-migration instance [15:22] http://bazaar.launchpad.net/~ubuntu-release/britney/britney1-ubuntu/view/head:/fauxpkg/FauxPackages#L39 [15:23] I don't think that's ever been stitched into the citrain instance. You may have to force it [15:25] cjwatson: thanks for looking into that [15:25] :) [15:29] jibel, davmor2, rvr: I know your guys queue is pretty long already, but if you could get silo https://requests.ci-train.ubuntu.com/#/ticket/1030 at least forced somewhere into the queue that would be good - there is a britney hack for it in -proposed since otherwise this just fails [15:29] jibel, davmor2, rvr: so the silo is basically good for QA [15:30] sil2100, done [15:39] sil2100, is there any way to tell when that retrigger is done? [15:43] kdub: hm, I guess it happened already [15:44] kdub: I can retry again, but I suppose it was retried once [15:44] yeah, apparently its a been a vexing downstream test with previous releases [15:46] sil2100, lets retrigger, the mir team apparently has had some history with this particular test in that package before during mir releases [15:47] kdub: retriggered again, let's see when it gets picked up [15:47] thanks sil2100 [15:50] sil2100, was checking greyback, apparently its a test that can be affected by heavy load situations (and has a test timeout increase coming to qtmir soon), so might need to force-pass if possible [15:51] test timeout for qtdbustest, not qtmir [15:51] or, we could just retrigger during light load :) [16:01] sil2100, you can now disable unity8 jobs as well, I've enable autopilot testing in our jenkaas, so we're migrated :) [16:01] Saviq: ok, on it [16:43] plars, thanks, everything seems in order :) [16:44] Great, let me know if there's anything else I can do to help! [16:55] sil2100, is there a way to force approval for that aoutpkg test? we're being affected by this https://bugs.launchpad.net/qtmir/+bug/1541854 which is dependent on system load [16:55] Launchpad bug 1541854 in QtMir "SharedWakelock test unstable" [High,In progress] [16:56] kdub: hm, you would have to negotiate with QA, so jibel, davmor2 or rvr - can't this test be written so that it's not affected by load? :) [16:57] Tests shouldn't generally be affected by load at best, so that they can be run in multiple environments [16:58] sil2100, its in a downstream and the fix hasn't gotten in there :/ [16:58] Ouch... [16:58] so, I suppose i'll just retrigger this weekend, hopefully lower load then [16:59] sil2100: one for jibel or rvr more likely [17:05] sil2100: is OTA-10 expected to be the one that ships on the net round of devices from the factory? M10 / MX5 Pro ? [17:08] popey: yes, at least from the M10 POV [17:08] popey: won't want to say anything official regarding the MX5 Pro [17:09] ok [17:13] it's Pro 5, not MX5 Pro :) [17:13] MX series is a different from Pro [17:18] Mirv: thanks :) === tyhicks` is now known as tyhicks [17:41] Mirv: Silo 49 approved === Elleo_ is now known as Elleo === alan_g is now known as alan_g|EOW [18:03] robru: want to check https://ci-train.ubuntu.com/job/ubuntu-landing-049-2-publish/7/console ? [18:14] Elleo, about silo 30 what provides fonts-droid-fallback? [18:15] Elleo, the change makes the keyboard depend on an alternative package that doesn't exist [18:27] Mirv: looks like lp didn't respond to the REST request. Did the copy actually fail? I've seen it a lot where lp performs the copy but doesn't return any value, that's why that error is ignored... Though not sure why the job failed [18:29] Mirv: oh the failure is due to no artifacts, which is expected for a vivid copy. Basically that looks fine to me, as long as you check the overlay ppa actually has all the copied packages in it === mhall119_ is now known as mhall119 [18:52] Mirv: indeed, train found all packages copied successfully and then merged ;-) [19:04] trainguards: the train doesn't understand git branches? "ERROR Failed to understand "https://code.launchpad.net/~mterry/geonames/+git/double-free/+merge/287234". Is it a merge?" -- is that a planned feature with an ETA or is it foolhardy to use git at this point? [19:05] sure doesn't! [19:05] mterry: I'm currently working on a big overhaul that will a) modularize the bzr integration so that git support is at least theoretically possible, but we're talking months and months away [19:05] oh and b) do a bunch of other stuff like use sbuild instead of cowbuilder and c) parallelize source package builds [19:06] robru, gotcha. Canonical own the geonames LP project, so maybe I can convince the maintainers to use bzr [19:06] robru, is there a wiki for how to set up a project for train integration? (like autolanding of merges and all that jazz) [19:07] robru, I know CI/jenkins is the team's responsibility, but I guess I just mean the autolanding/automerging ibts [19:07] *bits [19:09] mterry: autolanding? that hasn't been a thing for years. you just mean you want to use the train, right? [19:09] mterry: err, I guess you mean jenkaas [19:09] robru, is it all jenkaas? I thought there was something the train did to merge into trunk and whatnot once a silo landed [19:10] mterry: yeah but that isn't "autolanding". "autolanding" was the thing from like 3 years ago where everything would just be automatically merged. "having a silo, testing it, hitting publish", not very automatic. but it does automatically merge after the publication has been successful [19:11] robru, OK. Is there a name for that last merge step? :) But a project doesn't need any special setup for that to happen? [19:12] mterry: you can't do "that last merge step" unless you're fully on the train. I'm not really sure what you're asking here. like you just want merges automatically merged without having a silo at all? [19:13] mterry: https://wiki.ubuntu.com/DailyRelease/InlinePackaging this is the train packaging requirements [19:13] robru, no I want to be fully on the train. My situation is that I'm looking at this git project (geonames) and am looking into what it would take to make it compliant with our workflow -- bzrification, what else? [19:13] ok, that's good [19:14] mterry: if you want "that thing that automatically finds new merges and builds them & reports test failures", that's jenkaas. train is "manually put merges into train, manually click build, manually review package, manually click publish, then once it hits archive it automatically merges" [19:15] robru, yeah I want the train. I understand jenkaas CI is a different beast, which can be tackled separately [19:15] mterry: so yeah basically bzr if you're in a hurry. I know some projects do development in git and then just export to a bzr branch for train use, and then merge back at the end... [19:16] mterry: I'm literally working right now on extracting the bzr bits from the train so it can be modularized and a git plugin created. [19:16] robru, I don't think the maintainers (basically desktop team I think?) are super git-oriented. Just the guy that wrote it originally was more git-oriented [19:16] but probably months to get it all together, there were so many huge changes required I decided to just sort of make a clean break, so this change got coupled with a bunch of other stuf [19:17] robru, but he's not working on it anymore and I'm just trying to whip it into shape so I can more easily land a few changes I want :) [19:17] cool [19:18] robru, so I bzr-ify the project and just use silos like normal and it does all the right things automatically? there's no master list of projects using the train or anything? [19:18] mterry: nope, it was a foundational design goal not to have master lists of anything. the precursor to the train had master lists and it was awful [19:19] robru, awesome :) [19:19] mterry: bzr-ify and set a few packaging details to be what the train expects as per that wiki page [19:19] robru, so my question was basically dumb. It just took me a while to figure out why :) But that InlinePackaging page is a useful reference [19:19] mterry: yeah even in that url 'DailyRelease' is a reference to the train-precursor. train inherited all that code though [19:20] robru, yeah and that page uses the phrase "autolanding"... :) [19:20] I should really update that page and move it under /citrain/ [19:31] jibel: it will exist in xenial soon (and fonts-droid will be removed there) [19:32] jibel: this is to allow them to make that transition [19:32] jibel: about to go underground, can answer more later if needrd [19:32] needed* [20:13] robru: oh, I didn't realize the copy did happen, thanks for checking it. [20:14] Mirv: yeah apologies that job is a bit of a mess, and it's not worth to fix right now because I'm literally working to replace all of jenkins entirely. === dpm is now known as dpm-afk === tsimonq2alt is now known as tsimonq2 === salem_ is now known as _salem === _salem is now known as salem_ === salem_ is now known as _salem