/srv/irclogs.ubuntu.com/2015/08/17/#ubuntu-ci-eng.txt

=== chihchun_afk is now known as chihchun
Mirvcjwatson: qtdeclarative5-ubuntu-ui-toolkit-plugin-gles Provides: qtdeclarative5-ubuntu-ui-toolkit-plugin and that used to work, otherwise eg vivid emulator wouldn't work, but now last week sil2100 was hitting similar problems on wily and was wondering what had changed.05:30
Mirvbut if that's the only fix needed to have apt happy then it's good. sometimes package dependencies just become more complex, and we've had trouble guiding apt with the -gles packages before too.05:31
=== _morphis is now known as morphis
dbarthhey08:15
dbarthpinging cihelp, cause i have a build / dependency problem in silo 32, ie the oxide-qt rebuild08:15
dbarthi didn't manage to get linked to libmedia-hub-client4 despite it being available in the overlay ppa08:16
dbarthand so when we use the oxide build on the phone the media-hub audio integration is broken08:16
dbarthbecause of an interface binary compatibility issue (libmedia-hub-client3 and 4 are not compatible)08:16
sil2100dbarth: that's more like a trainguard thing then :)08:17
sil2100hmmm08:17
dbarthah08:17
dbarthhi08:17
dbarthif we were to add a dependency on libm-h-c4 strictly, then it would break the desktop builds were 3 only is availble08:17
sil2100That's strange - are all the dependecies correct for it to pull in libmedia-hub-client4 during build?08:17
dbarthso what's the trick?08:18
dbarthsee http://bazaar.launchpad.net/~oxide-developers/oxide/packaging.vivid/view/head:/debian/control08:18
dbarthit depends on     libmedia-hub-dev [armhf],08:18
sil2100This should dep on the latest media-hub, so 4 in case of the overlay I suppose08:19
dbarthah, so as a special armhf thing it could be specific to libm-h-c4 then08:19
dbarthbut we took 3 apparently08:19
dbarthhttps://launchpadlibrarian.net/214500930/buildlog_ubuntu-vivid-armhf.oxide-qt_1.8.4-0ubuntu0.15.04.1~overlay1_BUILDING.txt.gz08:19
sil2100Well, yeah, but this should work nevertheless, let me check what's up08:20
sil2100dbarth: the overlay has media-hub 308:21
sil2100https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/stable-phone-overlay/+sourcepub/5101143/+listing-archive-extra08:22
dbarthhmm, so that's the issue08:22
sil2100This is the latest one08:22
dbarthhmm08:22
sil2100Is it required for us to have a newer media-hub, or are we ok with this here?08:22
dbarthsorry, i was wrong it's not 3 vs 4 it's 3.1 vs 3.008:24
dbarth(re-reading justinmcp's email)08:24
dbarthbut that's still something to do with dependencies it seems08:24
dbarthso the overlay has libmedia-hub-client 3.1.0+15.04.20150527.1-0ubuntu108:25
sil21003.1.0+15.04.20150527.1-0ubuntu1 <- it's using the latest as expected during the build08:25
dbarthyes08:25
dbarthand so that should work08:26
dbarthok sorry, i'll get back to testing the issue for now08:26
Mirvcihelp: what's the status of eg qtubuntu-gles boottest failing because at least in the past it tried to install that x86-only package in arm environment? http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#qtubuntu-gles08:28
sil2100dbarth: make sure you're testing the right oxide version, if you have correct pinning configured then you should fetch the ~overlay1 one08:28
Mirvcihelp that was reported some weeks ago and was said to be looked into, so I'm wondering if that's still the same issue or something else08:28
dbarthsil2100: yup, i forced that with exact version numbers to do the install08:29
sil2100Mirv, jibel, davmor2: be right there on the meeting08:30
psivaaMirv: let me take a look at this08:34
ogra_oops08:36
cjwatsonMirv: "Provides: qtdeclarative5-ubuntu-ui-toolkit-plugin" only satisfies *unversioned* dependencies on qtdeclarative5-ubuntu-ui-toolkit-plugin; it used to work in this case because the dependency in webbrowser-app's binary packages was unversioned until recently08:46
cjwatsonMirv: though as a recent feature in dpkg/apt it's possible to say "Provides: qtdeclarative5-ubuntu-ui-toolkit-plugin (= some-version)" - I haven't tried that personally for anything, so don't know if the full stack works with it08:47
cjwatsonoSoMoN: thanks for the webbrowser-app approval - would you mind sorting out a landing for it to unbreak installability?08:48
oSoMoNcjwatson, sure, I’ll file a landing request08:48
cjwatsongreat08:48
Mirvcjwatson: thanks for the clarification, makes sense08:56
Mirvpsivaa: thanks!08:56
sil2100cjwatson: thanks! We were wondering about this failure, since it suddenly stopped working08:57
sil2100cjwatson: do you know if fixing only the webbrowser-app is enough to satisfy apt?08:57
sil2100cjwatson: since there's a LOT of qtdeclarative5-ubuntu-ui-toolkit-plugin dependencies around there, not sure if versioned ones though...08:57
Mirvsil2100: maybe only webbrowser started to use a feature only available in a newer UITK?08:58
Mirvalthough that wouldn't stay like that for too long even if it would be the case08:58
sil2100But from what I recall (checking that now) there was no recent change of the UITK deps to become versioned09:00
psivaaMirv: The issue hasn't been permanently fixed. I just created a fake result for this, i'll for it to propagate09:00
sil2100hm, no, actually it got bumped recently09:02
sil2100cjwatson: anyway, thanks for explaining, at least I wasn't aware that Provides was working only for unversioned deps and that's why I didn't know what was wrong there09:03
sil2100Apparently this merge should in this case fix the world09:03
cjwatsonsil2100: There are several other versioned dependencies, but they already have explicit | ...-gles09:04
psivaaMirv: http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#qtubuntu-gles is now green09:04
Mirvpsivaa: ok, thank you a lot!09:04
cjwatson(inappropriately written in some cases without the versioning on the -gles part, but that won't actually break things other than partial upgrades)09:04
cjwatsonat some point it might be worth trying to make use of versioned Provides here09:05
cjwatsonI'm not sure what our constraints on that are, whether we need to avoid them in the archive until trusty's apt no longer needs to parse them or something09:05
sil2100cjwatson: could you setup automatic weekly translation exports in the 15.04 series on Tuesdays evenings?09:28
cjwatsonsil2100: That's already been done on the LP side, except Tuesday mornings not evenings (10:30 UTC like all the other exports).  Is that OK?09:30
cjwatsonsil2100: There's something that consumes the output of that that also needs to run, but I think you need to ask pitti for that.09:30
sil2100cjwatson: oh, the auto-exports are done? Didn't know, I requested an explicit export last Friday - good... I think we would rather prefer to have it a bit later in the day to give people more time09:38
sil2100I already talked to pitti regarding langpack-o-matic09:38
sil2100cjwatson: is it a big problem to change the hour of the exports? Don't want to force anything09:38
cjwatsonsil2100: I'm wary of that, because it's a big job that could end up accidentally conflicting with something else like backups09:44
cjwatsonThe current time is known to work09:44
sil2100cjwatson: ok, I suppose we can adjust our schedules accordingly in that case09:44
oSoMoNcjwatson, is there a bug report to track the regression my branch introduced wrt installability of ubuntu-touch on amd64 and i386 ?09:53
cjwatsonoSoMoN: no09:54
oSoMoNok09:54
cjwatsonit was ~midnight, I just proposed a merge :)09:54
oSoMoNthat’s alright, just wanted to check, in case there was one I would have included the link in the landing request09:55
* cjwatson nods09:55
oSoMoNcjwatson, building in silo 809:58
cjwatsonoSoMoN: ta10:01
sil2100oSoMoN: dual landing?10:02
sil2100oSoMoN: since it's the case for both wily and overlay10:02
oSoMoNyes10:02
sil2100Thanks :)10:02
oSoMoN(the silo also includes two unrelated minor bug fixes)10:03
Mirvrobru: almost forgot to reply, but I created bug #1484795 already on Fri10:26
ubot5bug 1484795 in CI Train [cu2d] "Can't handle a package with "-" in upstream version number" [Undecided,New] https://launchpad.net/bugs/148479510:26
Mirvmardy: the unity8 has rebuilt in https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-029 - https://wiki.ubuntu.com/Process/Merges/TestPlans/Unity8 should be executed now additionally to the earlier testing, and of course the MP should be approved before tomorrow's OTA-6 deadline (if not ok with postponing to OTA-7)10:46
* sil2100 off to lunch o/11:16
mardyMirv: thanks, I see that tsdgeos approved the branch11:19
tsdgeosmardy: Mirv: are you going to land that? because mzanetti is planning to land another unity8 silo that conflcits with that one11:19
mardytsdgeos: ah, I see, we have the same branch in silo 14 and 2911:23
Mirvtsdgeos: well yes if the unity8 change seems ok to you, since otherwise no problems were found. but that'd sit then in the QA queue so I'm not sure.. it's not OTA-6 targetted so of lesser priority11:23
tsdgeosmardy: no no, the branch is going away from 1411:23
tsdgeosmardy: what i mean is that your branch conflicts with https://code.launchpad.net/~aacid/unity8/previews_in_order11:23
mardytsdgeos: ok, will Mirv take care of removing it? ^11:23
tsdgeosMirv: ok, then let us get https://code.launchpad.net/~aacid/unity8/previews_in_order in before which is a nice improvement for OTA611:24
mardytsdgeos: maybe since my branch is tiny, you could merge it into yours?11:24
tsdgeosmardy: care to rebase it on mine?11:24
mardytsdgeos: sure will11:24
mardyexpect a ping soon11:24
Mirvmardy: tsdgeos: if the branch can be landed also without the Qt change, then by all means merge and land it! then I can put the Qt alone into queue again.11:25
mardyMirv: yes, it can land without the Qt change11:25
Mirvthanks then, I'll remove the branch from 02911:26
mardytsdgeos: I pushed an updated branch; should I resubmit the MP setting your branch as diff origin, or it doesn't matter?11:31
tsdgeosmardy: makes it easier, so please do11:31
=== _salem is now known as salem_
Mirvogra_: can you ack that https://ci-train.ubuntu.com/job/ubuntu-landing-016-2-publish/lastSuccessfulBuild/artifact/network-manager_packaging_changes.diff adds two patches under debian/...11:37
Mirvthey have headers and all11:37
alan_gfginther:11:41
=== om26er_ is now known as om26er
mzanettimardy, thanks for rebasing that branch, added it to silo 1411:55
mardymzanetti: yw :-)11:56
=== alan_g is now known as alan_g|lunch
=== alan_g|lunch is now known as alan_g
=== om26er_ is now known as om26er
=== charles_ is now known as chalres
=== chalres is now known as charles
=== kyrofa_ is now known as kyrofa
awe_trainguards, can someone vet the packaging changes in silo-016?14:30
infinitysil2100: Please don't sponsor kernel uploads. :)14:31
infinitysil2100: Or, help usher those patches into git instead.14:32
sil2100infinity: it was an universe package, thought it would be alright :)14:32
sil2100Ok, didn't know rules for those14:32
infinitysil2100: All the Android kernels are (sort of) maintained by our kernel team.14:32
sil2100Understood, next time I'll simply direct those to git then, sorry about that14:33
sil2100Hope I didn't cause too much chaos ;)14:33
ogra_someone from the kernel team just needs to re-merge that now14:33
=== chihchun is now known as chihchun_afk
sil2100It was just a minor low-priority description fix anyway14:34
sil2100So I suppose no one would have anything against merging that in14:35
ogra_nope14:35
ogra_but someone needs to do it :)14:35
=== chihchun_afk is now known as chihchun
Mirvawe_: I tried asking ogra for that14:50
awe_I just asked cyphermox14:50
Mirvawe_: ok14:50
Mirvawe_: let me know when he acks it14:51
awe_ack14:51
Mirvwas that it? ;D14:51
awe_no... sorry14:53
awe_sorry for the confusion there14:53
ogra_Mirv, oh, sorry, i'm a bit swamped, havent gotten to it yet14:59
ogra_(and these packaging changes arent really packaging changes beyond the changelog :) )15:00
ogra_so form a "packaging changes" POV this is indeed an ACK15:00
sil2100robru: ouch15:13
sil2100robru: ok, we have a possible breach in the train15:15
sil2100robru: it's a bit serious, I'll fill in a bug for it in a moment not to double type15:16
Mirvogra_: right, it was just adding patches and no actual packaging changes. thanks.15:20
Mirvcyphermox now publish it himself15:21
davmor2sil2100: is it me or are there silos missing from the dashboard?  for example silo 01115:24
cyphermoxMirv: sorry, did I break something? awe asked me to review and publish.15:24
davmor2Mirv: ^15:24
Mirvcyphermox: no, you did the correct thing, I was just doing the same but needed core dev ack15:25
cyphermoxMirv: oh ok15:25
cyphermoxMirv: when you need reviews feel free to ask, I'm there ;)15:25
=== chihchun is now known as chihchun_afk
Mirvcyphermox: sure thanks! earlier it wasn't your TZ yet.15:30
cyphermoxI'm in germany for the week15:31
rvrbfiller: Silo 47 approved15:39
sil2100davmor2: hmmm15:39
sil2100davmor2: let me check15:39
sil2100bfiller: hey! Silo 47 has unbuilt revisions in https://code.launchpad.net/~tiagosh/messaging-app/fix-1394971/+merge/26493615:43
sil2100;/15:43
sil2100davmor2, rvr: if you guys could take a look at silo 8, it has a fix for our emulator builds15:45
rvrsil2100: I'm on it15:45
sil2100Thanks :)15:45
bfillersil2100: let me look15:46
sil2100bfiller: we'll either have to rebuild or revert the missing revisions, but I think they might be needed15:46
bfillersalem_: ^^^ can you take alook at silo 4715:47
bfilleror boiko ^^^15:48
bfillersil2100: should I just rebuild?15:48
ogra_sil2100, i assume you still dont have any magic tool that can re-write the distro series in a debian/changelog of a published package ? seems the android package in the overlay PPA is a wily one ...15:48
sil2100ogra_: that would require a re-publish anyway - I can re-publish it if needed after rewriting15:50
ogra_sil2100, talk to morphis ... if there is a shortcut he can take vs a full rebuild thats probably helpful15:51
morphissil2100: I can rebrand the changelog quickly for wily15:52
morphispackage is identical otherwise15:52
sil2100morphis: yeah, I think we can then publish it then ASAP15:52
morphissil2100: without QA?15:53
sil2100Since rewriting and hacking the PPA is a bad idea in overall, as this android package was built against wily15:53
ogra_well15:53
ogra_the android package ships its own toolchain inside15:53
ogra_i doubt it matters on what distro it was build ... but yeah, do a rebuild15:53
sil2100Ok... anyway, did the package get QA last time? How did they test it?15:54
john-mcaleelysil2100, jibel I understand silo 30 is marked 'ready for qa'15:56
john-mcaleelyI happen to know it's busted on krillin/vegetahd15:56
davmor2john-mcaleely: no jibel15:56
john-mcaleelyand there will be a new device tarball shortly15:56
john-mcaleelyaha15:56
john-mcaleelydavmor2, ^ :-)15:56
morphissil2100: see https://wiki.ubuntu.com/Process/Merges/TestPlans/Android15:56
morphissil2100: and yes, it was QA approved last time by davmor215:57
sil2100john-mcaleely: silo 30 landed already ;)15:57
davmor2john-mcaleely: so new device and then silo030 can be tested right which is the requirement for chinese arale as a minimum right?15:57
sil2100john-mcaleely: it was the pulseaudio trust-store one, right?15:57
davmor2sil2100: see no silo 030 listed either :(15:58
john-mcaleelysil2100, really? how did it pass QA on krillin & vegetahd15:58
john-mcaleelyand (actually) arale15:58
john-mcaleelydavmor2, yes, I thought it needed to land with an arale device tarball, to get chinese image15:59
sil2100john-mcaleely: what's in silo 30? Since I don't see it on request or the dashboard15:59
sil2100https://requests.ci-train.ubuntu.com/#/?search=landing-030 <- all silo 30's got published15:59
sil2100No silo 30 assigned right now15:59
davmor2sil2100: so empty silos don't show on the dashboard any more is that correct?16:00
john-mcaleelyhrm. well, it was pulseaudio trust store16:00
sil2100davmor2: yes16:00
john-mcaleelyand camera trust store16:00
sil2100john-mcaleely: then it landed, not sure how they tested it16:00
sil2100But it was ready for QA so they picked it u16:00
sil2100up16:00
john-mcaleelysil2100, this is the one I thought it was: https://requests.ci-train.ubuntu.com/#/?req=12616:00
sil2100I have the seed change ready16:00
john-mcaleelywell, you have a busted camera then16:00
john-mcaleelyon krillin, vegeta & arale16:00
sil2100davmor2, rvr: ^16:01
boikosil2100: bfiller: yep, just rebuild it, the missing revisions are only adding a QML test, no extra code changes16:01
john-mcaleely(or I think you should have)16:01
davmor2sil2100: meeting16:01
sil2100Google issues16:01
boikosil2100: that's regarding silo 47, I can rebuild it if you want16:02
rvrI tested silo 3016:03
john-mcaleelyI think the camera is busted on it16:03
sil2100boiko: is it required? SInce we can basically try skipping that by reverting those from your branch16:04
rvrjohn-mcaleely: How can pulseaudio mess the camera?16:07
john-mcaleelyrvr, the trust store update within also included a camera service update16:07
john-mcaleelynot a really big deal. the fix is incoming :-)16:08
boikosil2100: well, I asked salem_ to write those, messaging-app is already building again on the silo16:08
boikosil2100: sorry for the trouble anyway16:08
davmor2john-mcaleely: silo 030 testplan on lists pulseaudio no mention of the camera app in there16:09
rvrjohn-mcaleely: Well, there is some deal, that should have been tested, but no word about that on the request :(16:09
john-mcaleelyinteresting16:09
rvrhttps://trello.com/c/3XaS4chY/2182-126-ubuntu-landing-030-pulseaudio-ubuntu-touch-session-abeato-diwic-mandel16:09
davmor2john-mcaleely: https://trello.com/c/3XaS4chY/2182-126-ubuntu-landing-030-pulseaudio-ubuntu-touch-session-abeato-diwic-mandel check the testplan16:09
john-mcaleelytwo people who are on holiday, and one fresh back in16:10
john-mcaleelyit seems the error is ours then16:10
john-mcaleelygiven the new device tarball is incoming, probably best to just press onward16:10
john-mcaleelynot sure if you have other options in the interim?16:11
greyback_cihelp: hey, qtmir stuck in wily-proposed, http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html#qtmir - does boottest even work there?16:11
john-mcaleelybrb16:12
davmor2john-mcaleely: yeap16:12
fginthergreyback_, nope, I'll need to manually pass that one, one moment16:15
greyback_fginther: cheers16:15
rvrsil2100: Silo 8 approved17:04
rvrEOD here, see you tomorrow!17:05
=== alan_g is now known as alan_g|EOD
boikosil2100: messaging-app rebuilt on silo 47, ready to land now17:13
sil2100o/17:19
robrumorphis: uh, you sure silo 11 doesn't need QA? it's targetted at vivid overlay17:30
robrusil2100: ^17:30
sil2100Yeah17:30
sil2100I changed that17:30
sil2100It was basically tested already and it's a binary copy, no sign-off needed as it was already tested in the same state it was before17:31
robruhuh ok17:31
john-mcaleelydavmor2, sil2100 device tarballs to fix the busted camera on krillin/vegeta:17:40
john-mcaleelyhttps://requests.ci-train.ubuntu.com/#/?req=18017:40
davmor2john-mcaleely: will we need one for arale too do you know?  I assume we would17:44
boikokenvandine: could you please check and (possibly) ack the packaging changes on silo 47?17:54
boikokenvandine: it is basically requiring some QML plugins so that the QML tests pass17:55
kenvandineboiko, i'll look17:55
kenvandineadded as build deps then?17:55
boikokenvandine: yep17:55
kenvandineboiko, +1 from me17:58
sil2100davmor2: I think there'll be another one for arale, yes18:00
boikorobru: sil2100: ^ (silo 47)18:00
davmor2sil2100: I assumed there would be18:00
sil2100boiko, robru: reviewing that18:01
sil2100boiko: ok, I think it will require a rebuild - but with an exception of not requiring QA sign-off - the ubuntu archive admins requested that all dependency changes are documented18:04
sil2100boiko: could we get the merge that modifies the deps include information about the deps being changed in the commit message?18:04
kenvandinesil2100, it's a build dep change, probably not an issue18:04
sil2100Quoting the request: "Please enforce the requirement that changelogs for train packages include an explanation of the changes to the packaging."18:06
sil2100ogra_: btw. could you merge in https://code.launchpad.net/~sil2100/ubuntu-seeds/ubuntu-touch.wily-add-iw/+merge/267370 once you have a moment?18:27
sil2100I'll upload that to the vivid seeds in the meantime18:44
robrusil2100: kenvandine: so did we come to a consensus on silo 47 then? we'll wait for boiko to rebuild with a proper changelog?18:52
kenvandineyeah18:52
kenvandinewe should18:52
kenvandineboiko, ^^18:52
sil2100I wouldn't publish that, but maybe someone with power can override this ;)18:52
=== karni is now known as karni-away
greyback_fginther: hey, dobey tested qtmir from wily-proposed and it allows unity8 to come up, adds more evidence it's good!18:55
dobey"tested"18:56
dobeyit fixes the unity8 crashing anyway :)18:56
davmor2video or it never happened18:56
fginthergreyback_, I verified the same and pushed a passing result. It just takes a bit for it to be seen on the excuses page18:57
greyback_fginther: ah ok, thanks for that!18:57
fginthergreyback_, now it's passed :)18:57
greyback_\o/18:58
dobeyyay18:59
pedronistrainguards: hi, you can publish  ubuntu/landing-017 (it's the wily landing matching a recent vivid one that was QAed and is already published there)19:01
robrupedronis: sure one sec19:02
pedronisthx19:02
robrupedronis: right so the reason this wasn't published is that it's not in 'Publish without QA' state (that is, nobody knew it was ready for publishing). for future reference19:03
robrupedronis: also this merge needs to be top-approved: https://code.launchpad.net/~pedronis/ubuntu-push/fixes-to-wily/+merge/26806719:04
pedronisrobru: sorry, done now19:05
nikwenpedronis: Thanks.19:06
robrupedronis: no worries, all good now19:07
pedronisrobru: thx19:07
davmor2john-mcaleely: ^19:55
boikorobru: kenvandine: sorry, I had a doctor appointment, I will change the commit message and rebuild it20:10
boikorobru: kenvandine: rebuilding...20:12
robruboiko: which MP did you change?20:16
boikorobru: https://code.launchpad.net/~tiagosh/messaging-app/fix-1394971/+merge/26493620:42
boikorobru: build finished, btw20:59
robruboiko: heh, ok thanks21:03
robruboiko: https://ci-train.ubuntu.com/job/ubuntu-landing-047-1-build/18/artifact/messaging-app_packaging_changes.diff/*view*/ uhhhh the new words aren't actually in the diff.21:05
robruboiko: I think the changelog code uses git-like commit message parsing, eg the first line is the "title" and subsequent lines are snipped. Can you make the commit message be just one line and rebuild again?21:06
boikorobru: ok21:07
robruboiko: sorry for the trouble, this changelog code has a lot of issues and is definitely on the chopping block21:07
boikorobru: yeah, maybe those policy changes (like requiring the commit message to explicitly mention the packaging changes) should be better advertised?21:08
robruboiko: yeah I'm sure that's written in a policy somewhere but there could definitely be better communication about that.21:09
boikorobru: anyway, one more thing learnt, next time I will make sure the MRs are correct in that regard21:10
robruboiko: thanks, I just sent an email to ubuntu-phone as a reminder for others.21:15
boikorobru: thanks!21:15
robruboiko: you're welcome21:16
boikorobru: now the changes contains the new commit message on silo 4721:59
=== karni-away is now known as karni
robruboiko: thanks22:12
boikorobru: np22:12
=== salem_ is now known as _salem

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