2015/08/27/#ubuntu-ci-eng.txt

=== robru changed the topic of #ubuntu-ci-eng to: Train trouble? ping trainguards | CI problems? ping cihelp | Train: http://bit.ly/1hGZsfS | QA Signoffs: http://bit.ly/1qMAKYd | Known Issues: OTA-6 final freeze in effect, but the vivid-overlay landing gates opened
[00:09] <sil2100> Goodnight! o/
[00:09] <sil2100> No issues with the OTA so far so time to sleep
[00:23] <robru> sleeeeep
[02:34] <nuclearbob> Trevinho: I saw test failures in the unity8 suite with the ppa that I didn't see without it. There may have been other packages with failures as well, but I stopped looking once I found some. I'm going to see if we can narrow down which package specifically is causing the problem
[02:35] <Trevinho> nuclearbob: ok, let's see..
[02:36] <Trevinho> nuclearbob: in that ppa, the only thing I think might interact with is autopilot-qt, as the others have no influence at all on u8
[02:36] <nuclearbob> Trevinho: okay, I'll try rolling back to the archive version of that and kick off the tests
[02:36] <nuclearbob> they'll take a while to run, and it's almost bedtime here, so I can collect the results in the morning
[02:37] <Trevinho> nuclearbob: yeah, I know how it goes (AP in u7 takes almost 2hrs)...
[02:37] <Trevinho> (bedtime here was about 5hrs ago :P)
[02:59] <nuclearbob> Trevinho: I'm running into issues trying to get just autopilot-qt4 downgraded and leaving everything else, so I'm going to try to come back to it in the morning when I'm cleverer
[03:00] <Trevinho> nuclearbob: give a try to ppa-purge
[03:01] <nuclearbob> Trevinho: I can get them all cleared out with that, but then I'm having trouble getting just the one upgraded
[03:02] <Trevinho> nuclearbob: mh, I think it's because there are new dependency on new libxpathselect, and this could lead something else
[03:04] <nuclearbob> Trevinho: yeah. I can take a further look at it later. Do you have info on how to run the tests if you want to keep working on it? I can send it to you if not
[03:04] <Trevinho> nuclearbob: no, but feel free to send them... I can look at it tomorrow
[03:12] <nuclearbob> Trevinho: I sent you the instructions from veebers, and the additional stuff I've done. I'll pick this back up in the morning, and talk to you then
[03:13] <Trevinho> nuclearbob: ok, thanks
=== chihchun_afk is now known as chihchun
[04:41] <Mirv> morning
=== _salem is now known as salem_
[08:44] <seb128> kenvandine, jgdx, cihelp, do you have any idea why there is no CI on https://code.launchpad.net/~morphis/ubuntu-system-settings/fix-device-actions/+merge/269188 ?
[08:53] <psivaa> seb128: that has CI pending
[08:53] <seb128> psivaa, yeah, I just added that
[08:53] <seb128> psivaa, what team membership is needed for that to auto happen?
[08:53] <seb128> we suspect that Simon isn't in the right group to get CI builds
[08:55] <psivaa> seb128: that might be the reason, let me check
[09:02] <psivaa> seb128: that was the reason, 'morphis' was not in the allowed users. I've just added them.
[09:02] <psivaa> s/them/him
[09:02] <morphis> psivaa: thanks!
[09:02] <seb128> psivaa, thanks
[09:02] <psivaa> np
[09:02] <seb128> psivaa, what list do we use? aren't all canonical people in there by default
[09:05] <psivaa> seb128: this is a list in a config inside jenkins. this has been being manually appended.  I'm not sure if adding all of canonical employees is the right way, but i can certainly discuss it with our team
[09:06] <seb128> psivaa, well, I was mostly wondering, would probably be easier than having to figure out why things don't append and add people 1 by 1 as we just did
[09:07] <psivaa> seb128: I agree it would be easier for maintenance, just not sure if there was any particular reason to restrict to those only require. I dont have the history
[09:07] <seb128> k; anyway thanks for fixing it ;-)
[09:08] <seb128> for that particular mp, was what I did enough? (subscribing ci-jenkins)
[09:11] <psivaa> seb128: It was not needed to be subscribed, if the user was in the allowed_users list, the MP should automatically trigger the CI
[09:12] <seb128> psivaa, right, but the mp was done before you added Simon
[09:12] <psivaa> seb128: still it will trigger once 'trigger-ci-on-stacks' job runs again, after 15 mins :)
[09:13] <seb128> oh ok
[09:13] <seb128> thanks
[10:33] <Mirv> sil2100: is it useful to track via milestone all bugs that are fixed for OTA-7? I mean to add the milestone to "other bugs than what were especially targetted"?
[10:37] <popey> sil2100: i have no idea how to add click packages to the citrain now it's a site and not s sheet, can you help me?
[10:38] <sil2100> popey: sure :)
[10:38] <sil2100> Mirv: hmm, I personally would like that, yes, since it makes easier to show users what fixes are in the image
[10:38] <Mirv> popey votes for going back to the spreadsheet
[10:39] <sil2100> Mirv: but you would have to double confirm with pmcgowan
[10:39] <sil2100> popey: so!
[10:39]  * popey gets coffee
[10:40] <Mirv> sil2100: yeah, not really sure if we'd remember it for every bug so the commit log would be the only place to surely have everything (but then that's not a list of bugs fixed, but commit messages)
[10:40] <pmcgowan> Mirv, sure you can also affects to the milestone on your own, I care about managing targets
[10:40] <pmcgowan> but it doesnt preclude tracking other bugs that get fixed
[10:40] <Mirv> pmcgowan: ok
[10:41] <sil2100> popey: you log into bileto and prepare a new request, fill in the description, test plan and target ubuntu vivid from the dropdowns
[10:41] <sil2100> popey: then, in the manual download URLs field, include the convinient link to the click package to be reviewed
[10:42] <sil2100> popey: I also recommend to write down the name of the click app in 'Manual Source Packages', but that's not super required
[10:42] <sil2100> popey: the final step is to switch the last drop-down to 'Ready for QA'
[10:42] <popey> sil2100: what is bileto?
[10:42] <sil2100> popey: once those are filled, press 'Create New Request'
[10:42] <sil2100> popey: https://requests.ci-train.ubuntu.com/
[10:43] <popey> ah
[10:43] <popey> codenames ftw
[10:43] <popey> "prepare a new request"? ?
[10:43] <popey> i probably have no access, i am logged in and there is no obvious way to create a new request sil2100
[10:45] <jibel> sil2100, there is a card for silo 037 created less than an hour ago in the trello board but there is no request for it on bileto, any idea where is comes from?
[10:45] <Mirv> pmcgowan: sil2100: right, it seems we can't add a bug to a milestone, only add the System Image project to the bug, so it doesn't scale. maybe it's not worth the trouble unless the milestone managing rights are given to trainguards with a note that they can only use it for milestoning bugs that were already fixed to get them listed.
[10:46] <Mirv> it's only https://launchpad.net/~canonical-devices-products-team/+members that can manage
[10:46] <pmcgowan> Mirv, I am happy to add landers to that team
[10:46] <sil2100> popey: let me check, but you should be...
[10:47] <Mirv> pmcgowan: ok, timo-jyrinki sil2100 robru then
[10:47] <sil2100> popey: ok... you're added now
[10:47] <sil2100> Strange that you weren't!
[10:47] <sil2100> popey: re-log and be sure to check the checkbox with membership in SSO
[10:48] <sil2100> jibel: let me take a look
[10:48] <popey> Thats better! thanks sil2100
[10:48] <Mirv> thanks pmcgowan
[10:48] <jibel> sil2100, apparently it's only in the 'ready for qa' view
[10:48] <pmcgowan> Mirv, ok done
[10:51] <sil2100> jibel: heh, no - it's on the other page
[10:51] <sil2100> jibel: the train now has pagination
[10:51] <sil2100> jibel: it's a good thing to have but I wonder if it's not set for a too small value, it should at least show the first 100 landings
[10:52] <sil2100> pmcgowan: thanks!
[10:54] <jibel> sil2100, yeah the link to navigate is very well hidden at the bottom of the page
[10:54] <jibel> needs design obviously
[11:13] <nik90> popey: Sry, clock-app rev 361 is the new cut off for QA. No more changes.
[11:13] <popey> hey nik90
[11:14] <popey> convenient as I just built a click of that, ready for me to submit to QA?
[11:14] <nik90> popey: yes :D
[11:14] <nik90> popey: you give me more time, I am going to only push more MPs...to hard to resist. so let's freeze here.
[11:15] <popey> hah
[11:15] <popey> okay
[11:17] <sil2100> nik90: ;D
[11:18] <popey> sil2100: created https://requests.ci-train.ubuntu.com/#/ticket/274 (cc nik90 )
[11:18] <nik90> :P
[11:18] <nik90> popey: awesome thnx.
[11:18] <popey> np
[11:19] <sil2100> popey: thanks! Looks goodish, let's see when it pops up at QA's trello
[11:20] <nik90> popey: Can you include a click package of the music-app click in the changelog. QA will need that to test the custom alarm sound feature.
[11:20] <nik90> I am adding a note as well
[11:21] <popey> ah, okay.
[11:23]  * sil2100 off to prepare lunch
[11:23] <popey> added
[11:24] <nik90> cool
[11:29] <davmor2> nik90: will that require the new media hub as that hasn't landed yet
[11:29] <nik90> davmor2: nope
=== salem_ is now known as _salem
=== chihchun is now known as chihchun_afk
=== MacSlow is now known as MacSlow|lunch
=== alan_g is now known as alan_g|lunch
[12:25] <Trevinho> robru: is there a reason why all the CI-train mail are sent to ubuntu-phone ML but not ubuntu-devel (or something similar)? this is not anymore related to phone as well...
[12:25] <pstolowski> jibel, hello, quick question - is silo 44 (wily only landing) in your pipeline yet, can we expect it to land soon?
[12:28] <jibel> pstolowski, if it's wily only, we don't block you can land.
[12:32] <pstolowski> jibel, thanks
[12:49] <popey> sil2100: do you know if theres been some Ubuntu One creds invalidation happen? My U1 account has completely disappeared from my device.
[12:53] <sil2100> popey: hm, not sure, did that happen right after upgrade or just like that suddenly?
[12:54] <popey> not sure when, but today certainly
[12:54] <popey> [M#jHchannel: ubuntu-touch/rc-proposed/bq-aquaris.en
[12:54] <popey> last update: 2015-08-26 09:32:01
[12:54] <popey> the wifi here in canonical seems to trigger network manager to go berzerk too
[12:54] <popey> phone is completely unusable here, wondering why nobody else has this issue
[12:54] <popey> dbus is being spammed by network manager
[12:58] <DanChapman> popey: is it https://bugs.launchpad.net/ubuntu/+source/network-manager/+bug/1480877 your experiencing?
[12:58] <ubot5> Ubuntu bug 1480877 in Canonical System Image "Access points' "PropertiesChanged" dbus signals freeze UI on mobile devices" [High,Confirmed]
[12:58] <popey> hah, yes!
[12:58]  * popey hugs john-mcaleely 
[12:59] <popey> thanks DanChapman
[13:00] <DanChapman> np i recall faenil being plagued by it a while back when at canonical office
[13:02] <john-mcaleely> uh oh
[13:02] <john-mcaleely> what I have I done
[13:02] <jibel> you filed a bug?
[13:02] <sil2100> ;)
[13:02] <jibel> not even that :)
[13:02] <john-mcaleely> it seems it may be assigned to me
[13:03] <john-mcaleely> hugs may be a hopeful way of getting it fixed :-)
[13:03] <popey> there are an insane number of access points around here
[13:03] <john-mcaleely> welcome to the big smoke :-)
[13:11] <brendand> john-mcaleely, you should all come to Bath. i can see exactly 3 access points here
[13:14] <sil2100> mandel: hey! :)
[13:14] <sil2100> mandel: do you know when we would have the trust-store pulseaudio integration landed in wily?
[13:15] <mandel> sil2100, ho! lets go!
=== alan_g|lunch is now known as alan_g
=== _salem is now known as salem_
[13:25] <Mirv> pstolowski: top approvals https://ci-train.ubuntu.com/job/ubuntu-landing-044-2-publish/12/console
[13:25] <pstolowski> Mirv, ah, sorry, working on it
[13:27] <Mirv> ack
[13:28] <kenvandine> mandel, FYI, trust-store in wily has had a major version update which was never merged into trunk
[13:28] <kenvandine> so you'll need to get those in sync to get the fix in wily
[13:33] <mandel> sil2100, ok, we are landing the pulse changes to the git repo, once there we should take a look
[13:33] <mandel> sil2100,  but ken already said the bad news
[13:38] <boiko> rvr: hi, do you need any help with silo 26 testing procedures or is everything clear there?
[13:40] <rvr> boiko: I'm having issues with the keyboard
[13:40] <boiko> rvr: oh :/
[13:41] <boiko> rvr: what kind of issues?
[13:41] <rvr> boiko: After installing the packages keyboard doesnt appear, and removing the .cache/QML/ doesn't fix it
[13:41] <rvr> So I'm stuck in the greeter screen
[13:42] <rvr> boiko: I flashed arale to check, and happens there also
[13:42] <boiko> rvr: did you see any other package being installed besides the ones from the silo?
[13:44] <boiko> kenvandine: was that the issue you were seeing in your tests? (keyboard not showing up)
[13:44] <rvr> boiko: http://paste.ubuntu.com/12205776/
[13:44] <kenvandine> very rarely
[13:44] <kenvandine> but yes
[13:44] <kenvandine> it was in autopilot tests
[13:44] <kenvandine> but jgdx did reproduce it locally
[13:44] <boiko> rvr: unity8-fake-env? hmm, never saw that one
[13:45] <kenvandine> probably not the same problem
[13:45] <rvr> Hmm
[13:45] <rvr> Remove: ubuntu-touch:armhf (1.221vivid7), qtubuntu-android:armhf (0.62+15.04.20150814-0ubuntu1)
[13:45] <boiko> kenvandine: yeah, looks like it is not the same thing
[13:46] <boiko> rvr: it is also installing lots of qt5 packages
[13:47] <rvr> boiko: From the overlay ppa
[13:48] <davmor2> michi: http://paste.ubuntu.com/12205807/
[13:48] <boiko> rvr: yeah, but shouldn't that be in the image already?
[13:48] <rvr> davmor2: Hey
[13:48] <rvr> davmor2: I think we have a problem with the overlay ppa
[13:49] <rvr> davmor2: Maybe silo 29 broke something
[13:49] <davmor2> rvr: keyboard is working here what issue are you seeing
[13:49] <rvr> davmor2: Oh
[13:49] <rvr> davmor2: citrain device-upgrade to 26, and the keyboard doesn't work anymore
[13:50] <davmor2> rvr: let me see if I can confirm
[14:02] <davmor2> rvr: boiko: something bad is happening http://paste.ubuntu.com/12205934/  this didn't happen with silo 11 or silo 27
[14:03] <rvr> Good, not just my imagination
[14:04] <boiko> davmor2: well, I have no idea why that is happening
[14:04] <boiko> davmor2: rvr: we haven't changed anything on packaging/deps on those branches
[14:05] <boiko> robru: do you by chance know what is going on?
[14:05] <rvr> davmor2: boiko: http://paste.ubuntu.com/12205948/
[14:05] <rvr> That's without any silo
[14:05] <boiko> sil2100: maybe you know something about it? ^
[14:06] <davmor2> rvr: so that isn't uninstalling the universe then
[14:06] <rvr> There is some package in the silo that requires unity8-fake-env
=== MacSlow|lunch is now known as MacSlow
[14:19] <rvr> Weird, I don't find any direct dependency on unity8-fake-env
[14:21] <rvr> unity8-fake-env: Replaces: unity8-autopilot (<< 8.02+15.04.20150422-0ubuntu1)
[14:22] <rvr> Not sure what's going on
[14:32] <jibel> rvr, which silo?
[14:32] <rvr> jibel: 26
[14:33] <sil2100> rvr: what's the problem in the end? Something funny happens when you try to install silo 26?
[14:34] <rvr> sil2100: Yes, it pulls unity8-fake-env and removes ubuntu-touch, not funny :)
[14:34] <sil2100> uh
[14:34]  * sil2100 looks
[14:34] <jibel> rvr, looking
[14:34] <jibel> rvr, just checking you're on latest image build from the overlay and pinningg is correct?
[14:35] <rvr> jibel: I think so
[14:35] <jibel> rvr, confirmed
[14:35] <jibel> The following packages will be REMOVED:
[14:35] <jibel>   qtubuntu-android ubuntu-touch
[14:35] <jibel> The following NEW packages will be installed:
[14:35] <jibel>   unity8-fake-env
[14:35] <rvr> A regular apt-get upgrade with the overlay ppa don't remove those packages
[14:36] <rvr> It happens with citrain device-upgrade 26
[14:39] <davmor2> jibel: doesn't happen on other silos either :)
[14:39] <jibel> sil2100, it's probably a problem with qtdeclarative5-qtmir-plugin 0.4.4+15.04.20150513-0ubuntu2~vivid1~test1~qt1 dependencies
[14:41] <davmor2> jibel: no without that ppa it updates with no issues, I'm wondering if it is qtdeclarative5-ubuntu-history0.1 History service to store messages and calls - QML plugin  that is having an effect on things?
[14:43] <jibel> sil2100, output of the resolver http://paste.ubuntu.com/12206225/
[14:43] <rvr> Broken unity8
[14:48] <davmor2> jibel, rvr: Oddly if I reinstall the 2 packages that were uninstalled the phone works and it didn't remove anything
[14:48] <jibel> sil2100, the version of qtdeclarative5-qtmir-plugin in the overlay is higher than the ppa
[14:48] <jibel> s/ppa/silo
[14:49] <rvr> davmor2: Yup, I have continued testing after reinstalling those
[14:49] <sil2100> jibel: silo 26 doesn't have qtdeclarative5-qtmir-plugin now does it?
[14:49] <jibel> sil2100, it does      0.4.4+15.04.20150513-0ubuntu2~vivid1~test1~qt1 0
[14:49] <jibel>        1002 http://ppa.launchpad.net/ci-train-ppa-service/landing-026/ubuntu/ vivid/main armhf Package
[14:49] <sil2100> It shouldn't: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-026
[14:50] <sil2100> There's no such non-deleted or non-deprecated version there, hmm
[14:51] <jibel> sil2100, so with silo 26 a dist-upgrade replaces qtmir-android by qtmir-desktop and removes ubuntu-touch
[14:51] <sil2100> That really shouldn't happen, apt shouldn't try installing packages that are removed
[14:52] <jibel> sil2100, it doesn't look like the content of bileto https://requests.ci-train.ubuntu.com/#/ticket/209
[14:53] <sil2100> That's exactly what's in the PPA
[14:54] <sil2100> Bileto says there's history-service and telepathy-ofono, and the PPA has those 2 packages
[14:54] <sil2100> So I have no idea where the qtdeclarative5-qtmir-plugin comes from in that PPA
[14:54] <jibel> sil2100, and what are all the mir packages in ppa 26 then?
[14:55] <sil2100> LP doesn't see any so I have no idea
[14:55] <sil2100> https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-026 <- this only has those 2 packages
[14:56] <sil2100> I'll dig into what's happening with that PPA after our team meeting
[14:56] <jibel> sil2100, there are 2 qtmir packages too http://ppa.launchpad.net/ci-train-ppa-service/landing-026/ubuntu/dists/vivid/main/binary-armhf/Packages
[14:56] <jibel> 3
[14:57] <jibel> which totally confuse the upgrade
[14:57] <sil2100> jibel: well, those files also show deprecated or removed packages
[14:58] <sil2100> So it's not relevant to the current issue I think
[14:58] <sil2100> For instance: http://ppa.launchpad.net/ci-train-ppa-service/landing-020/ubuntu/dists/vivid/main/binary-armhf/Packages <- while this PPA in fact is empty right now
[14:59] <sil2100> Wait, wrong paste
[14:59] <sil2100> Ignore that
[14:59] <jibel> sil2100, since the silo is pinned it'll try to upgrade them isn't it?
[14:59] <sil2100> Yeah, but still, those are removed - they shouldn't be there
[14:59] <sil2100> Let me re-try removing
[15:00] <sil2100> hmmm
[15:00] <sil2100> I think we'll have to poke the LP team
[15:00] <oSoMoN> ubuntu-qa: silo 14 is ready for QA validation
[15:03] <pstolowski> Mirv, MPs in silo 44 approved
[15:04] <rvr> kgunn: Can you add some words about how to test silo 46? The test plan is empty.
[15:05] <kgunn> rvr: so bregma is looking at it here....and there's a test plan attached to the silo
[15:05] <bregma> rvr, there is a link from the test plan to the u-a-l test plan
[15:05] <rvr> kgunn: Hmm... let me check
[15:06] <sil2100> cjwatson: hey! It seems one of our PPAs didn't remove some packages completely - they're removed from the LP interface, but their binaries are still published in the archive
[15:06] <rvr> kgunn: So correct way to test the silo is to run ubuntu-app-launch's test plan, isn't it?
[15:06] <bregma> rvr, correct
[15:06] <rvr> bregma: Perfect
[15:06] <sil2100> cjwatson: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-026
[15:11] <cjwatson> sil2100: I don't understand why they should have been; https://requests.ci-train.ubuntu.com/#/ticket/209 is only Ready for QA.  Could you be more specific?
[15:11] <cjwatson> sil2100: Oh, I see there's more scrollback, let me see
[15:11] <sil2100> cjwatson: yes, but the silo should only have packages from the telepathy-ofono and history-service
[15:11] <sil2100> cjwatson: while the PPA has stale packages from qtmir
[15:12] <cjwatson> There are some race conditions possible there if it was removed while a build was in progress, IIRC
[15:14] <cjwatson> sil2100: re-removed
[15:14] <cjwatson> <cjwatson@niejwein ~ (master)>$ remove-package -A ppa:ci-train-ppa-service/ubuntu/landing-026 -s vivid -m 'partially removed some time ago' -e 0.4.4+15.04.20150513-0ubuntu2~vivid1~test1~qt1 qtmir
[15:14] <sil2100> cjwatson: \o/
[15:14] <sil2100> jibel, rvr: ^
[15:14] <cjwatson> you'll need to wait for publication of course)
[15:15] <sil2100> I think something like this happened once in the past
[15:15] <sil2100> But it's really bizarre
[15:15] <rvr> sil2100: Great
[15:16] <cjwatson> so the removal started at 2015-06-01 10:45:17 UTC; the armhf build finished at 2015-06-01 10:43:48 UTC
[15:17] <cjwatson> that means it was probably in flight when the removal happened
[16:12] <cjwatson> sil2100: just rechecked, definitely gone now
[16:26] <robru> sil2100: Mirv: btw https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-033 ;-)
[16:29] <ogra_> rvr, tagget the gibberish bug and added another one with the tag for the "always american location info" in Nearby
[16:29] <rvr> ogra_: Thanks
[16:29] <ogra_> *tagged
[16:31] <rvr> ogra_: Link for one of them, please?
[16:31] <rvr> Don't appear in the tag search
[16:31] <sil2100> cjwatson: thanks :)
[16:31] <ogra_> bug 1489536 and bug 1488162
[16:31] <ubot5> bug 1489536 in Canonical System Image "Nearby Scope Location format is always american" [Undecided,New] https://launchpad.net/bugs/1489536
[16:31] <ubot5> bug 1488162 in Canonical System Image "today scope has gibberish for moon phase in german translation ..." [Undecided,New] https://launchpad.net/bugs/1488162
[16:31] <ogra_> https://bugs.launchpad.net/canonical-devices-system-image/+bugs?field.tag=touch-i18n
[16:32] <sil2100> robru: ah ha! Looking awesome :)
[16:32] <rvr> ogra_: touch-l10n...
[16:32] <robru> sil2100: thanks, dunno why I didn't do this sooner, there was a bug report for this from forever ago...
[16:32] <ogra_> rvr, ouch ...
[16:32] <rvr> ogra_: touch-i18n is more correct, but... :)
[16:33] <sil2100> robru: now see? We could have been using LP for all things CI Train instead of Bileto! Silos have all the buttons now + LP bugs and violla!
[16:33] <sil2100> ;)
[16:33] <rvr> ogra_: Can you add touch-l10n to the other bug as well?
[16:33] <ogra_> yeah
[16:33] <rvr> ogra_: Done
[16:33] <rvr> Thanks
[16:34] <robru> sil2100: lol. the thing is, that info in the PPA comes straight from jenkins, which makes it more authoritative than bileto at the moment (eg if you forgot to reconfigure, the PPA description is more correct than what bileto has). so that was one thing I feared about dropping the dashboard, losing that authoritative info. but now with that I'm comfortable
[16:34] <robru> dropping the dashboard, which is one step towards making bileto the authoritative source
[16:38] <robru> brb
=== alan_g is now known as alan_g|EOD
[17:43] <rvr> robru: Bileto is returning errors
[17:43] <rvr> robru: API Error 400
[17:44] <robru> rvr: what are you doing?
[17:44] <rvr> robru: Nothing involving guns and journalists
[17:45] <rvr> robru: https://requests.ci-train.ubuntu.com/#/?req=245
[17:45] <robru> rvr: and why not?
[17:45] <rvr> robru: Just going there returns me error
[17:46] <robru> rvr: huh, bizarre that it shows up in the main list but the permalink is broken
[17:46] <jibel> rvr, the url changed
[17:46] <robru> and permalinks for other requests aren't broken
[17:46] <jibel> rvr, it's now #/ticket/nnn
[17:46] <robru> oh right
[17:47] <rvr> Ah
[17:47] <jibel> rvr, I  fixed the bot but didn't edit existing cards
[17:47] <robru> rvr: https://requests.ci-train.ubuntu.com/#/ticket/245 right, sorry. I mentioned it in an email, but not prominently
[17:47] <rvr> Done
[17:48] <rvr> It's nice, I could approve the silo from the main page
[17:48] <jibel> not prominently is an understatement :)
[17:48] <robru> rvr: it used to be that the frontend had a bunch of special cases in the URL handling, but now it doesn't a direct URL pass-through to the API. so eg https://requests.ci-train.ubuntu.com/#/ticket/245 and https://requests.ci-train.ubuntu.com/v1/ticket/245 return the same ticket, while one is displayed nicely and the other is raw json. simplified a lot of
[17:48] <robru> code, should have been that way from the start
[17:49] <rvr> Cool
[17:50] <robru> jibel: yeah, sorry. I know breaking URLs is a terrible sin, but they were ugly before, the new way is much better. it should have been this way from the start but it was a bit rushed. they're stable now, I promise I won't change them EVER AGAIN. ;-)
[17:51] <rvr> robru: Oh, came on. Then jibel's life would be more boring.
[17:53] <jibel> robru, bots are insensible to the beauty of urls ;)
[17:53] <jibel> robru, BTW I've an issue with click and tarball requests.
[17:54] <jibel> robru, where are landers supposed to says what it is?
[17:54] <dobey> trainguards, cihelp: do any of you have magical powers over the debci stuff for autopkgtests?
[17:55] <jibel> robru, 1 sec I'll give you an example
[17:57] <robru> dobey: no me
[17:57] <robru> jibel: yeah click & tarball are a bit of a special case
[17:58] <jibel> robru, for example for silo 014 I know from the mps that it's the webbrowser
[17:58] <robru> jibel: I mean theoretically they can say what it is in the description. not sure I understand.
[17:58] <jibel> robru, for for the clock app request there is no such information, just the description
[17:58] <jibel> and it's free text
[17:58] <psivaa> dobey: cihelp does not have that power now either, it's now mostly handled by pitti, iiuc
[17:58] <robru> jibel: which request is the "clock app request"?
[17:58] <jibel> robru, yeah that's what people do but they also write tons of other info in this field
[17:59] <jibel> robru, 274
[17:59] <robru> jibel: the problem with click & tarball is that it's not actually touched by jenkins so ALL the forms are free-form, there's no enforcement of any standards for any fields.
[17:59] <robru> jibel: brb, food on burner
[18:00] <jibel> robru, for the clock app for example in the description it's written "Updated Clock click app" and I'd need only the name of the app
[18:00] <robru> jibel: well the "manual download URL" field was given for the .click URL, you can parse the name from the URL?
[18:01] <robru> jibel: "download_links" field
[18:02] <jibel> robru, but then for custom or device tarball it's somewhere else for example https://requests.ci-train.ubuntu.com/#/ticket/237
[18:02] <dobey> robru, psivaa: oh, ok, that sucks :(
[18:02] <jibel> hard to know what is what
[18:02] <jibel> robru, I'll use the description for now and see how it goes
[18:02] <jibel> it'll be better than N/A
[18:02] <dobey> jibel: do you know anything about the debci autopkgtests stuff?
[18:03] <jibel> dobey, not really, it's all pitti's stuff now
[18:04] <jibel> robru, ah previously such requests were prefixed with [non-citrain] and we lost that that's why the bot is lost
[18:04] <dobey> ah ok :-/
[18:04] <jibel> dobey, what is the problem?
[18:05] <dobey> jibel: https://objectstorage.prodstack4-5.canonical.com/v1/AUTH_77e2ada1e7a84929a74ba3b87153c0ac/autopkgtest-wily/wily/amd64/u/ubuntu-system-settings-online-accounts/20150827_145513@/log.gz
[18:05] <dobey> ubuntu-system-settings-online-accounts seems to have regressions not of its own doing, and it's blocking several things in proposed
[18:06] <robru> jibel: yeah that request is really funny, URL field should have URLs in it.
[18:06] <jibel> dobey, ENOMEM
[18:07] <dobey> jibel: great, how do we fix it?
[18:07] <robru> jibel: you need to talk to penk and john-mcaleely and get them on the same page with the format of those fields. once you all agree I can put in some input validation that enforces it.
[18:07] <dobey> and how is this the only package failing in that way
[18:08] <jibel> dobey, is the system running out of memory because of a issue or the tests really require lot of memory?
[18:08] <jibel> infinity, ^ can you help with out of memory issue in autopkgtest? have you seen that before?
[18:08] <dobey> jibel: i don't know, but i guess they probably aren't requiring any more memory now, than they did 3 weeks ago when gcc5 transition was happening
[18:08] <jibel> ok
[18:09] <jibel> dobey, if infinity cannot help I'm afraid it'll have to wait for pitti
[18:10] <jibel> robru, right I'll do that
[18:10] <dobey> i'm a bit surprised that those autopilot tests would have ever passed under X though. *shrug*
[18:41] <jibel> robru, as example I created 597 for a custom tarball and 598 for a click on staging, would it work or it would break something if we use the fields I used to store tarball and clicks info?
[18:43] <robru> jibel: yeah that would work but please don't put anything into the siloname field. Look at what happens with the Build/Publish/Merge links when you do that.
[18:43] <robru> that should really be a read-only field...
[18:44] <jibel> ah 404
[18:44] <robru> jibel: it constructs URLs with the value of that field, it should really be blank if it's not a train silo
[18:44] <jibel> OTOH there is nothing to build/publish
[18:44] <jibel> I can just leave it empty
[18:44] <robru> jibel: exactly, there's nothing to build/publish, so leave that field empty and those links won't appear ;-)
[18:47] <jibel> so, if it's empty I'll use 'manual source package' for the type and 'manual download urls' for the details of the package and people can put what they want in description and test plan
[18:47] <jibel> I'll propose this to penk and john-mcaleely
[18:49] <robru> jibel: i think it would be better if the download_links field had full URLs in it, not sure if that's possible for "rc-proposed/bq-aquaris.en/krillin/492"
[18:49] <robru> like I don't know what that is, if there's a URL that can represent that information
[18:50] <jibel> robru, there is no full url for tarballs
[18:50] <robru> jibel: alright then
[18:50] <jibel> robru, technically there is one but it cannot be used directly and doesn't mean anything
[22:11] <bregma> robru, regarding silo 46 and libertine: do we really need a NEW review?  We were hoping to postpone landing in distro for a while (but if that isn't possible, OK, we have been debating if it was required at this point or not)
[22:13] <robru> bregma: your plan was to land it in the overlay PPA without landing it in wily at all?
[22:14] <bregma> robru, yeah, pretty much ... for the time being
[22:14] <bregma> because of, um, internal project priorities
[22:15] <bregma> if that's not kosher, we'll do the right thing
[22:15] <robru> bregma: that sounds like a horrible idea. Process is pretty clear that you should be landing in wily first and overlay second. I'm not comfortable with the thought of having packages on overlay that don't even exist in wily... Much less being a couple commits behind...
[22:15] <bregma> we blame ted for everything
[22:17] <robru> bregma: maybe raise it with slangasek if you want some kind of temporary exception for that but generally that sounds really wrong to me
[22:22] <bregma> robru, sure thing, thanks
[22:22] <robru> bregma: your welcome!