/srv/irclogs.ubuntu.com/2015/12/21/#ubuntu-ci-eng.txt

=== chihchun_afk is now known as chihchun
=== chihchun is now known as chihchun_afk
=== chihchun_afk is now known as chihchun
=== chihchun is now known as chihchun_afk
mardyMirv: hi! Do you know where I can find an older version of the uitk (I'd like to try 1.3.1742+15.04.20151209-0ubuntu1 to verify if a bug I'm seeing is a regression in the uitk)?08:08
mardyMirv: ops, nevermind, got it :-)08:08
=== chihchun_afk is now known as chihchun
Mirvmardy: ok :)08:25
davmor2sil2100: are we having the landing meeting this morning I assume it will be you and me09:20
sil2100davmor2: there might be Mirv as well ;)09:21
sil2100davmor2: let's do a quick one, you know, to say merry christmas and such!09:21
davmor2sil2100, Mirv: there had better be cats is all I can say ;)09:22
sil2100I think our cat is busy around my girl's grandmother now ;)09:23
Mirvdavmor2: sure, two of them09:25
davmor2Mirv \o/09:26
=== chihchun is now known as chihchun_afk
=== plars-off_ is now known as plars-off
sil2100dbarth: ping! Could you top-approve https://code.launchpad.net/~mardy/gnome-control-center-signon/invalid-providers-1287640/+merge/279278 ?09:55
=== john-mca` is now known as john-mcaleely
=== Elleo_ is now known as Elleo
davmor2bfiller: Elleo: was browser app in silo 17 rebuilt, it doesn't look it or it failed judging by the crosses, if you can hit that today we can get that landed then10:20
Elleodavmor2: will trigger a rebuild now10:21
ElleooSoMoN: are the failures for webbrowser-app on ppc arm64 s390x, etc. a normal thing? I notice the train doesn't flag them10:22
Elleodavmor2: oh, looks like silo 45 still hasn't landed yet10:23
Elleodavmor2: it's blocked on an oxide/gcc issue that's preventing it building for xenial10:23
oSoMoNElleo, yes, failures for those arches are expected, because oxide doesn’t build on them10:23
davmor2Elleo: should of sil2100 triggered an new image with it in on friday10:24
ElleooSoMoN: ah, okay; thanks10:24
davmor2sil2100: ^10:24
davmor2ah okay10:24
Elleodavmor2: on the train it's listed as manual published for vivid, while waiting for oxide xenial fix10:24
Elleo(silo 45 that is)10:24
oSoMoNElleo, davmor2: I’m looking into the issue, but it’s like a pandora box, it uncovered a bunch of issues which all need to be fixed for it to work10:24
davmor2Elleo: right okay10:24
davmor2oSoMoN: :'(10:25
dbarthsil2100: ack11:16
dbarthsil2100: hey, what about the oxide compiler issue? is there a possibility to revert to the previously known-to-work compiler?11:17
dbarthsil2100: i know oxide was dictating the gcc build dep a few months ago, for similar reasons11:17
davmor2Mirv: awesome now I can insult you here to.....insult......insult.........insult ;)11:29
Mirvdavmor2: :D11:31
Mirvdavmor2: I can feed your insult machine. We're still not publishing the latest UITK landing to xenial.11:32
Mirv(due to unability to build webbrowser which doesn't build because oxide-qt doesn't build on armhf because of new GCC 5.3)11:32
davmor2Mirv: Shuggin fashin shuggin fashin insult insult random insult pfff fine11:32
davmor2Mirv: right but what does that have to do with uitk11:33
davmor2Mirv: also that will mean no more dual landing fullstop11:34
Mirvdavmor2: nothing, but your insult machine can surely blame SDK team for it anyway! :)11:34
dbarthMirv: on that cascade of issues, see what i was writing to sil2100; can that help ?11:34
dbarthor otherwise let us know if we need to do an emergency call to Chris11:34
sil2100Yeah, still thinking what to do here11:35
Mirvdavmor2: it has that to do that webbrowser needs to land together with UITK (like it did to vivid) but compiler issue now hold the xenialn one11:35
sil2100LP: #152774111:35
sil2100I think reverting to an earlier gcc is not an option11:35
Mirvdbarth: it's probable it's purely GCC ARM problem still, so Chris can't help11:36
sil2100I guess it could be fixed on oxide side as well, probably gcc became more strict or something11:38
sil2100Looks really strange tho, would have to check the source but that would take ages11:40
sil2100;)11:40
dbarthyeah11:41
dbarthand also it feels like a code change that should be backed by upstream11:42
dbarthwhen they want blink or v8 to be gcc 5 compatible11:42
sil2100dbarth: looks like currently we're in a deadlock then... chrisccoulson is on holidays, not really possible to back-out gcc, hm11:52
=== xavigarcia is now known as xavigarcia_lunch
sil2100bfiller: hey! Regarding silo 46 - I can't publishi this silo due to ubuntu-ui-extras12:12
sil2100bfiller: normally manual re-builds like this can be ignored, but there's actually a change missing that's required12:13
sil2100bfiller: ubuntu-ui-extras needs to have the libexiv2-13 hard-coded dependency dropped12:13
sil2100bfiller: (change introduced by 0.2+15.04.20150311-0ubuntu3)12:14
Saviqsil2100, can you publish https://requests.ci-train.ubuntu.com/#/ticket/725 ?12:19
davmor2john-mcaleely: I see no issues with the tarball12:20
john-mcaleelydavmor2, excellent12:21
john-mcaleelysil2100, any reason I should not push the tarballs now?12:21
sil2100john-mcaleely: no, all good12:21
sil2100Saviq: hmm, looks published to me12:21
sil2100Saviq: I published this on Friday12:22
john-mcaleelysil2100, pushed. thank you12:23
sil2100john-mcaleely: thanks :)12:24
Saviqsil2100, oh, sorry, it's not obvious unless you read the status line, sorry :)12:26
sil2100nw ;)12:27
Saviqah but we have regressions in qtmir-gles apparently12:27
Saviqhmm12:28
Saviqok /me confused12:29
Saviqsil2100, can you please upload qtmir-gles with http://paste.ubuntu.com/14127400/ to xenial-proposed (and stable-phone-overlay for good measure...)?12:34
MirvSaviq: which silo, could it be direct upload to the silo + re-publish? anyway, I can help12:36
Saviqsil2100, or shall I better prepare a qtmir silo with those changes12:36
SaviqMirv, https://requests.ci-train.ubuntu.com/#/ticket/72512:36
MirvSaviq: ah so no silo, just stuck?12:36
SaviqMirv, I can do silo if that's better12:36
Saviqno-qa needed12:36
MirvSaviq: broken in overlay too so emulator broken?12:36
SaviqMirv, it's not really broken12:37
MirvSaviq: ok12:37
MirvSaviq: silo is always nicer, one gets the diff from the build job etc that's easier to double-check12:37
SaviqMirv, ok will do12:37
SaviqMirv, problem is new autopkgtest approach seems to have proposed pinned under archive, which IIUC means unless you request a version from proposed explicitly it won't take it12:38
Saviqsil2100, we need to pay more attention to -gles diffs, thus ↑12:38
MirvSaviq: hmm, but, that ticket isn't merge&cleaned? if you get the silo, I can upload manual uploads there and then we can double-check the diff.12:38
SaviqMirv, or we could force merge&clean as qtmir-gles is the last thing stuck and make it unstuck with the new silo asap12:39
MirvSaviq: yeah autopkgtests have something like that, or that they are not dist-upgraded. bug #1524817 is about that when upgrading new Qt version, old libqt5test5 is installed and does not get upgraded to the newer version12:40
SaviqMirv, problem is in qtmir-gles autopkgtests it tries to build against mir 0.1712:40
Saviqit's a somewhat dumb autopkgtest...12:40
MirvSaviq: it sounds pretty similar to the Qt problem - it first setups build dependencies from release pocket, then enables proposed pocket and installs only strict requirements or something like that12:40
Saviqeven though the package was built in silo against 0.1812:40
Saviqbut really it did find an issue with our deps, so it's a valid problem12:41
Saviqs/problem/test/12:41
MirvSaviq: I'm merge & cleaning now12:41
SaviqMirv, ok, I'll have silo asap12:41
MirvSaviq: trunks are now up-to-date12:46
Saviqkdub, need a hand please to unblock Mir from xenial-proposed https://code.launchpad.net/~mir-team/qtmir/gles-sync/+merge/28039612:50
Saviqhttps://code.launchpad.net/~saviq/qtmir/no-change/+merge/28109012:50
Saviqhttps://requests.ci-train.ubuntu.com/#/ticket/80912:50
kdubSaviq, what's needed?12:51
Saviqkdub, review of those two MPs please12:51
Saviqkdub, we might've told you wrong about the -gles sync MP requirements, debian/control needs to be updated manually12:51
sil2100Saviq: well, qtmir-gles did not generate any packaging diffs so those were not reviewed12:51
Saviqsil2100, yeah, which should blink a red light from now on (if there are debian/ changes in non-gles)12:52
kdubSaviq, done, looks good12:52
Saviqsil2100, I try to remember myself when landing anything with -gles, but if it gets through still12:53
Saviqkdub, thanks12:53
kdubSaviq, Saviq, https://docs.google.com/document/d/1VOP_I1dn_9oxwwf209S0ANGOYeOuROmARoOeyth8QaE/edit#heading=h.i57dat7acs4s12:53
kdubstep 13d is the qtmir / gles instructions for the mir team12:53
kdubSaviq, I'm guessing they should be changed to always bump the version number?12:54
Saviqkdub, any changes in non-gles debian/ need to be "ported" to -gles debian/, if applicable12:55
Saviqkdub, I've proposed a change to the step12:57
=== xavigarcia_lunch is now known as xavigarcia
kdubthanks Saviq13:04
sil2100Saviq: true ;)13:05
kdubSaviq, and if i understand, mir is stuck, pending 809 landing?13:05
Saviqkdub, stuck for xenial, yes, should be under an hour13:06
kdubSaviq, ack, thanks for the help. The changes to the doc made sense, thanks for the edit13:09
SaviqMirv, any idea what's with https://requests.ci-train.ubuntu.com/#/ticket/796 ? is it stuck somehow due to the oxide-qt mess?13:11
Saviqit seems to be published to vivid overlay, but not even in proposed for xenial13:11
MirvSaviq: it's purposefully not released to xenial since it would break things since webbrowser-app is not possible to build. so it's better for it not to be there at all than to break xenial images. currently GCC regression causes oxide-qt to fail to build on armhf13:35
SaviqMirv, builds done, train doesn't seem to have caught up yet https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-021 https://requests.ci-train.ubuntu.com/#/ticket/80913:35
SaviqMirv, not a gcc regression any more, according to the bug, but yeah, I understand13:35
MirvSaviq: well it sounds like a change in GCC anyway, any failure, since it built fine just two weeks ago13:36
SaviqMirv, might still be valid issues though, just not shown before13:37
MirvSaviq: and some bounds checking sounds like a gcc change13:37
MirvSaviq: diffs look good https://ci-train.ubuntu.com/job/ubuntu-landing-021-1-build/7/13:38
Mirvjibel: heads up url above ^ no-change rebuild of qtmir going into overlay silo soon13:39
Mirvsmall dependency change for qtmir-gles13:39
bfillersil2100: just seeing your message. little confused about what I should do on silo 46. It looks to have built correctly on both distros in the silo13:50
sil2100bfiller: yes, but you need to include the downstream change13:51
sil2100bfiller: since slangasek didn't release that to the archive for nothing - just remove the dependency and remove13:51
sil2100http://launchpadlibrarian.net/213415209/ubuntu-ui-extras_0.2%2B15.04.20150311-0ubuntu2_0.2%2B15.04.20150311-0ubuntu3.diff.gz13:51
bfillersil2100: oh so he pushed something to the archive and not to trunk you mean?13:51
sil2100s/remove/rebuild13:51
sil2100Yeah... bad practice13:51
bfillersil2100: got it, ok13:52
sil2100But that was during the wily finals, so he's excused partially ;)13:52
sil2100bfiller: thanks!13:52
MirvSaviq: so, should you set 021 to publish without QA and let's publish it?14:07
SaviqMirv, dnoe14:07
Saviqdone, even14:07
MirvSaviq: published, let's hope autopkgtests pass this time14:12
bfillersil2100: what's the best way to get the changelog in sync with what's in archive? getting this error trying to build https://ci-train.ubuntu.com/job/ubuntu-landing-046-1-build/4/console14:21
bfillersil2100: nm, just will copy from version in wily14:22
=== cyphermox_ is now known as cyphermox
davmor2sil2100: silo 58 were you able to deal with that in the end?15:11
davmor2Mirv: so qt-networking silo 5, I assume that it is just double checking that nothing breaks on the phone with networking right then it can be landed?15:12
davmor2Mirv: assume that kenvandine will soon tell you if it doesn't work, iirc the thing that was throwing us was that the phone has no ethernet :)15:13
Mirvdavmor2: that should be it yes.15:13
Mirvdavmor2: ken has confirmed the fix at bug #152397515:14
ubot5bug 1523975 in qtbase-opensource-src (Ubuntu RTM) "[desktop] system-settings crashes on startup with 5.4.1+dfsg-2ubuntu11~vivid1" [High,New] https://launchpad.net/bugs/152397515:14
davmor2Mirv: cool15:14
kdubSaviq, just to check, once https://requests.ci-train.ubuntu.com/#/ticket/809 goes through, mir will become unstuck? I admittedly don't know the process very well15:25
Saviqkdub, yeah, everything will happen automagically, should be any minute now15:26
kdubcool, thanks for the help15:26
Saviqkdub, you can monitor the progress on http://people.canonical.com/~ubuntu-archive/proposed-migration/xenial/update_excuses.html#qtmir and http://autopkgtest.ubuntu.com/running.shtml15:27
kdubah neat, didn't know about those sites15:27
sil2100davmor2: yeah, it's done15:28
Saviqkdub, the first one is linked in the "Status" line in https://requests.ci-train.ubuntu.com/#/ticket/80915:29
davmor2sil2100: great I'll delete the ticket from the qa list then15:29
bfillersil2100: silo 46 has been rebuilt and problems fixed with ui-extras package, should be able to publish now15:54
sil2100bfiller: \o/ thanks, on it now15:55
jhodappMirv, can you publish silo 34 for me please since it needs a core dev to land?17:33
kdubis this silo stuck? https://requests.ci-train.ubuntu.com/#/ticket/80917:53
jhodappcyphermox, around?18:11
cyphermoxyep18:11
jhodappcyphermox, my silo 34 got QA approval, can you please land it since it needs a core dev to do so?18:12
cyphermoxjhodapp: NAK, the changelog there means nothing at all, it doesn't even describe the changes in the patch: https://ci-train.ubuntu.com/job/ubuntu-landing-034-1-build/lastSuccessfulBuild/artifact/qtmultimedia-opensource-src-gles_vivid_packaging_changes.diff18:22
jhodappcyphermox, that's for the -gles package which was not changed, I'm not quite sure what Mirv does for that18:23
jhodappcyphermox, the change is really in the non -gles package18:23
jhodappcyphermox, and actually, that description is accurate18:24
cyphermoxwell then why does it need a patch?18:24
cyphermoxand why not just write what the patch does?18:24
jhodappcyphermox, it does, it's a sync18:25
cyphermoxthat's not a sync, that's copying the same changes again to another package18:25
jhodappcyphermox, that's what the -gles package is18:25
cyphermoxand that brings up the question of why do you need two source packages when it's the same code?18:25
jhodappI don't fully understand what it's used for18:25
jhodappcyphermox, not sure, you'd have to ask Mirv as he's the real maintainer of this for Ubuntu18:26
cyphermoxI'd very much like to see better changelogs even if it's just copying the same thing from one package to another. that's the point of changelogs, making sure people who don't have all the context have an idea of what's going on18:27
jhodappcyphermox, sure ideally I agree...we can get Mirv to update this although I'm pretty sure this is how things have always been done18:28
cyphermoxjhodapp: it's not because it's how it always has been done that it's correct18:29
jhodappcyphermox, I never said that :)18:29
cyphermoxwe've been able to do multiple builds of different kinds with a single source package :)18:29
jhodappcyphermox, I really can't answer your questions, we need to get an explanation from Mirv18:29
cyphermoxindeed.18:29
cyphermoxwell, I don't feel it's necessary that I punish you because of this, but I'll bring up the issues to Timo18:30
jhodappcyphermox, alright, we need a better way of reviewing source only package changes...this kind of discussion should happen before I'm asking a core dev to land it18:30
cyphermoxwell, yes and no18:31
cyphermoxpeople should already know to write meaningful changelogs18:31
cyphermoxin fact, it flows from writing meaningful commit messages :)18:31
jhodappcyphermox, agreed, but this should be caught in a code review process like a normal MR, but this isn't an MR18:32
cyphermoxwhat wasn't a MR?18:32
jhodappcyphermox, this change, it's a source package only change18:32
cyphermoxpeople who do direct uploads should even better know to write changelogs18:32
Elleotrainguards: heya, I've just retargetted silo 17 from dual landing to vivid only, I seem to remember there was some manual clean up needed in the past when doing this, am I remembering right?18:33
jhodappcyphermox, but anyway, thanks for publishing18:33
jhodappand reviewing it18:33
cyphermoxsure18:34
cyphermoxElleo: I'm not sure what trainguards are around for this. did you try to rebuild? I would expect this to be the only thing needed.18:35
Elleocyphermox: I haven't tried yet, I just remembered having to ping someone for something last time I did this :P18:35
Elleocyphermox: I'll give it a go and see what happens18:36
cyphermoxok18:36
jhodapprobru, around?18:51
kdubtrainguards,  https://requests.ci-train.ubuntu.com/#/ticket/809 seems stuck somehow waiting for ubuntu-system-settings-online-accounts19:36
kdubnot quite sure why... is there something that can be done to move the ticket along?19:36
cjwatsonkdub: it appears to be waiting for autopkgtests in general to catch up19:58
cjwatsonkdub: but at least some of the relevant tests are running right now19:59
kdubcjwatson, so its just going slow, not stuck then?19:59
kdubI'm not quite sure the details of this process, kinda looks stuck to me20:00
cjwatsonkdub: slow20:00
cjwatsonkdub: http://autopkgtest.ubuntu.com/running.shtml shows the relevant tests urnning20:00
cjwatson*running20:00
kdubcjwatson, thanks for the help20:00
robrucyphermox: agreed, stupid -gles not being built from the same source package has been a thorn in my side for a long time. IIRC according to slangasek it's because they have mutually conflicting build deps that they have to be separate source packages20:20
robrujhodapp: hey what's up?20:21
cyphermoxrobru: that's be the one good reason to have those separate20:21
robruElleo: you can't just retarget, you need to branch your changelog for vivid as the train won't allow generating a vivid version number on a xenial trunk20:22
robruAlso you'll need me to delete the xenial packages manually20:23
robruGenerally you should fix whatever issue is not working in xenial rather than five up on it and try to release to vivid only20:23
robrucyphermox: yeah I tried making it all coinstallable one time and it was a giant fustercluck20:24
jhodapprobru, hey, can you please dput qtmultimedia 5.4.1 from ppa:jhodapp/ubuntu/ppa to silo 5820:26
pmcgowanrobru, that silo is blocked by a gcc5 issue in xenial such that a lib oxide needs wont compile20:27
bzoltan_robru:  may I ask for a silo to fix that system settings issue - https://requests.ci-train.ubuntu.com/#/ticket/814 ?20:27
robrupmcgowan: OK well you'll need a new branch with a vivid changelog then20:28
pmcgowannot sure if Elleo or oSoMoN  are still around20:28
pmcgowanbfiller, ^20:28
pmcgowanthats for silo 1720:29
bfillerpmcgowan: looking20:30
robrujhodapp: done20:30
oSoMoNpmcgowan, robru: I’m around20:31
robrubzoltan_: hm I would but I'm having trouble with sso...20:32
oSoMoNrobru, we’ve not given up, but any xenial webbrowser-app build is blocked on bug #1528297, and no one will take a look at it until tomorrow at the earliest20:32
ubot5bug 1528297 in oxide-qt (Ubuntu) "libvpx FTBFS with gcc 5.3 on armhf" [Undecided,New] https://launchpad.net/bugs/152829720:32
robruoSoMoN: ok but if you switch your silo from dual to vivid you're essentially giving up on xenial ;-)20:33
bzoltan_robru: right... whenever you can :) i will be back20:33
oSoMoNrobru, temporarily20:33
oSoMoNrobru, although from what you said, if we need to branch we might as well revert to dual landing and do like we did for silo 45, i.e. force merge with a broken package in xenial20:33
robruoSoMoN: if you're planning on switching it back to dual before landing I recommend just leaving it as dual and just letting the xenial builds fail until they get fixed later. because you can't just take a xenial trunk and build it for vivid, you have to branch the trunk, not worth the effort20:34
oSoMoNrobru, okay20:34
robruright20:34
oSoMoNElleo, still around?20:34
oSoMoNI’ll update the landing request to revert to dual landing20:35
robrucyphermox: if you're able to log in to train jenkins can you assign https://requests.ci-train.ubuntu.com/#/ticket/814 ? I can't seem to log in20:36
cyphermoxdone20:36
robrucyphermox: thanks20:37
ElleooSoMoN: back now21:10
oSoMoNElleo, I reverted silo 17 to dual-landing and triggered a rebuild, it’s almost done now, I’ll be testing it just to double check, and we can ask davmor2 to give his ok tomorrow morning21:12
ElleooSoMoN: okay, cool21:13
oSoMoNElleo, feel free to do some quick validation on your side as well, I wouldn’t want to break anything the day I leave for holidays21:14
ElleooSoMoN: is the "diff missing" message anything to be concerned about?21:20
ElleooSoMoN: will install now and run through some quick sanity checks21:20
oSoMoNElleo, I don’t know, but it doesn’t look like the train considers it a blocking error21:20
ElleooSoMoN: okay21:21
alecuHi trainguards, I get "ERROR Assignment failed: Low on silos: Ask a trainguard to assign.", so I guess I need to ask y'all.21:27
alecuhttps://requests.ci-train.ubuntu.com/#/ticket/81621:27
robrucyphermox: can you assign another? I still can't log in21:29
cyphermoxrobru: can you try again? there should be no reason why you can't log in... otherwise it would be good to ask on launchpad-ops.21:30
robrucyphermox: the status in #is-outage mentions sso being unresponsive for some people21:30
alecuoutage again? :-(21:32
robrualecu: train is fine I just can't log in because sso21:42
robrucyphermox: I'm chatting with #is people, if you could go ahead and assign https://requests.ci-train.ubuntu.com/#/ticket/816 that'd be super, th21:44
robruthx21:44
cyphermoxok21:45
cyphermoxSUCCESS21:45
alecuyay, santa got here earlier!21:46
alecuthanks guys :-)21:46
cyphermoxwhat? where is it?21:47
cyphermoxwhat I want for christmas is to be able to breathe normally again21:47
alecucyphermox: really? what happened?21:48
cyphermoxjust a cold21:48
cyphermoxbut then I get annoyed at myself for generating so much mucus, coughing, runny eyes and all.21:48
alecusounds annoying... hope you get well!21:50
cyphermoxin due time! :)21:50
robrucyphermox: thanks21:57
cjwatsonkdub: ah, yeah, maybe it is in fact stuck - not something I have the ability to fix I'm afraid22:00
robruawwwwwwww yissssssss ^ new queuebot won't ping spam nearly so much now ;-)22:20

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