=== Ursinha is now known as Ursinha-afk
=== Ursinha-afk is now known as Ursinha
rsalvetiplars: sorry, was off today01:41
rsalvetiplars: you could try to start the console before the container is fully up01:41
rsalvetiand see if it helps at least01:42
didrockssil2100: Mirv: hey! FYI, I've changed quite extensively the status computation on the CI Train spreadsheet to get some new features unblocked I'm working on. If you see anything unusual, do not hesitate to tell me :)07:33
sil2100didrocks: ok ;) Thanks! What are the new features? Is it for the ci bot? ;p07:34
didrockssil2100: it's for the bot, I wanted to finish the data side before getting to the bot side :)07:34
didrocksnothing should be visible in the change (as long as you don't look at the hidden columns)07:35
didrocksbut the formulas have changed07:35
Mirvdidrocks: ok :) I took away "/" from B2 in metadata since Chipaca was annoyed by the double "//":s in the URL:s, but it will need to be added back if there was some URL that then lacks the "/" (all URL:s seemed to work for me)07:38
sil2100didrocks: coolio ;) Was there any news regarding CI Airline lately?07:38
didrockssil2100: I guess ask the CI guys. They are working on it, I see tickets flowing!07:38
didrocksMirv: you have done that in every silos?07:39
Mirvdidrocks: I tried all the button links in two silos, I guess the links should be identical in each silo?07:39
didrocksMirv: ah, metadata, yeah! all is fine, thanks ;)07:40
Mirvyes, metadata, that :)07:40
didrocksMirv: maybe it's what was breaking the redirect?07:40
Mirvdidrocks: hmm, that's possible too07:45
=== psivaa-afk changed the topic of #ubuntu-ci-eng to: ChipacaUbuntu CI Engineering Team | Vanguard: psivaa | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: -
sil2100ChipacaUbuntu - I wonder what that is ;)08:02
=== popey changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: psivaa | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: -
Chipacaooh, i was in the topic?08:20
Chipacamorning, all :)08:20
=== xnox is now known as NoNameYet_xnox
ChipacaNoNameYet_xnox: some day you too will be famous08:20
NoNameYet_xnoxChipaca: lol =)08:20
ogra_once he has a name for sure ;)08:21
Chipacababy steps08:21
NoNameYet_xnoxChipaca: this is a protest for inability to upload things into development series =)08:21
ogra_i was expecting to see a post over the easter weekend ...08:21
ogra_seems i was wrong08:22
Chipacalamp posts, maybe?08:22
Chipacasil2100: Mirv: didrocks: good morning! Could I have a landing for silo 8 / row 25, please?08:23
Chipacaor is that what NoNameYet_xnox is complaining about08:23
sil2100Chipaca: let me see :)08:23
sil2100Chipaca: ah, you mean, you want it landed?08:23
ChipacaNoNameYet_xnox: are you complaining about there not being a dev series yet, or are you complaining about not having package upload rights?08:23
Chipacasil2100: um.... is that silly of me?08:24
ogra_Chipaca, no dev series08:24
sil2100Chipaca: yeah, well, we would recommend waiting for U to open, since the other road is making the upload as an SRU - which might even take more time than waiting for U08:24
ogra_Chipaca, we are completely stuck until mark comes across with a name08:24
sil2100So I would recommend joining NoNameYet_xnox in the protest!08:24
Chipacasil2100: well... the fixes there would be nice to have on a stable image, unfortunately. Is that SRU land, or do we cut stable images from dev series too?08:25
=== Chipaca is now known as UtilitarianUvula
didrockspopey: sil2100: ogra_: Mirv: davmor2: do we need to have a meeting this morning?08:28
popeyhappy to.08:29
ogra_didrocks, we need to talk about python-evdev ... there was a long discussion about us bringing it back on the foundations ML08:29
davmor2didrocks: yes08:29
ogra_but we can do that on IRC08:29
didrocksogra_: IRC is fine for me08:30
* davmor2 slaps didrocks with a kipper for getting him up this early for no good reason ;)08:31
sil2100didrocks: so no meeting? I personally have nothing to report, due to things being blocked landing-wise ;)08:31
ogra_didrocks, so barry seems to have had the expectation that dropping of python-evdev actually removes all of python2 his mail sounds slightly unhappy that we brought it back in08:31
Mirvdavmor2: "this early", 5 hours ago! :)08:31
didrockssil2100: yeah, I think it's useless08:31
didrocksogra_: well, it didn't removed it, right?08:32
didrocksogra_: we saw it only removed python-evdev itself08:32
ogra_didrocks, we did for a few images ...08:32
didrocksbut nothing else08:32
ogra_which shows that it does *not* drop any other packages08:32
didrocksogra_: did he even checked on those images before?08:32
davmor2Mirv: look just cause you live in the Wrong timezone is not our problem ;)08:32
didrocksogra_: exactly08:32
didrocksogra_: did he track/check that?08:32
didrocksor work on why?08:32
ogra_didrocks, no, but he insists it has to go any wants the change reverted in an SRU ...08:33
=== psivaa-afk is now known as psivaa
ogra_seems gallery app keeps py2 around additionally08:33
ogra_another proposed SRU for that case08:33
MirvI'm taking some time to please bzoltan to try to get saucy qt 5.2 backport ~done08:33
didrocksogra_: he needs to talk to the SRU team about it anyway08:34
ogra_didrocks, not sure what to do, i dont think we should break the world with SRUs08:34
bzoltanMirv: that would be awesome :)08:34
didrocksogra_: well, let him lead that and do all the paperwork08:34
ogra_and dropping evdev again will surely cause extra work in the infra08:34
didrocksogra_: I don't care for us, it's easy to bring it back if it's broken08:34
UtilitarianUvulaMy SRU is small, cherry picked to address problems in what got into the stable image08:34
UtilitarianUvulaFWIW :)08:34
didrocksogra_: he better has to fix the packaging and ensure he doesn't regress08:35
didrocksogra_: if he wants that08:35
ogra_didrocks, it still breaks testing08:35
didrocksogra_: so, it's a no08:35
didrocksogra_: it's him who wants to bring that issue, it's him to fix it08:35
ogra_which means infra needs work or we keep it as is08:35
didrocksnot us08:35
didrocksand he needs to coordinate with the infra guys08:35
didrockslet him lead that08:35
didrocksif it's important to him08:35
ogra_well, he noeeds to know about the additional work it causes08:35
didrocksbeing clear we won't accept it if it's not all polished08:35
didrocksyeah, I didn't see that in any ML I'm in08:36
ogra_no, itsthe foundations ML08:36
didrocksso if he wants that, tell him that he needs to do the work08:36
didrocksand coordinate with the needed people08:36
NoNameYet_xnoxdidrocks: coordinate exactly what?08:37
NoNameYet_xnoxdidrocks: the gallery-app merge-proposal is stand alone, porting remaining portions of ap to python3.08:37
ogra_NoNameYet_xnox, changing and fixing all tests08:37
ogra_NoNameYet_xnox, that wont help if the automated tests all fail08:37
didrocksNoNameYet_xnox: apparently ogra_ is telling that dropping python2 from the image will impact the infra08:38
ogra_we rolled back because there were issues08:38
ogra_right the issue was only found because the infra tests went flat08:38
NoNameYet_xnoxdidrocks: gallery-app merge proposal has nothing to do with dropping python2 from the image, as in just merging that will not cause python2 to be dropped from the image.08:38
ogra_NoNameYet_xnox, barry wants python-evdev gone again08:39
didrocksNoNameYet_xnox: were we talking about the gallery-app MP?08:39
ogra_NoNameYet_xnox, whioch causes the automated tests to fail08:39
didrocksNoNameYet_xnox: read the beginning of the discussion08:39
ogra_gallery-app is just a side note here :)08:39
ogra_NoNameYet_xnox, and the gallery-app fix is fine if it passes testing ... its is the breaking of the other tests due to dropping -devdev that causes issues for us since all tests would have to be touched and app devs would always need to test in rw mode)08:41
popeyNoNameYet_xnox: did you autopilot test sudoku before uploading to the store?08:42
NoNameYet_xnoxogra_: that's why i said -evdev should be seeded into ubuntu-touch seed, instead of relying on transitive dependency from one package.08:42
ogra_NoNameYet_xnox, right, i was with you on that one but others insisted to just revert the change08:43
ogra_which was done in the end08:43
NoNameYet_xnoxpopey: sudoku ? sudoku was published at 180 2 weeks ago and i didn't upload that recently.08:43
popeyNoNameYet_xnox: sorry, I meant stock ticker. my bad08:44
ogra_same thing :)08:44
NoNameYet_xnoxpopey: i did a while a go, but not on the latest-greatest trusty image. it will be a while until i can upgrade to retest on recent. (dual-boot is preventing installing upgrades due to "not enough battery")08:45
popeyNoNameYet_xnox: we always autopilot test before uploading core-apps to the store. please can you run it next time before upload?08:46
NoNameYet_xnoxpopey: ok. in this particular case, it can't go worse, since stock-ticket autopilot tests have never passed =)08:46
NoNameYet_xnoxpopey: and they were fixed to finally work in recent commits.08:46
NoNameYet_xnoxpopey: when my phone charges up enough, i'll rerun the tests.08:47
popeyThank you.08:47
UtilitarianUvulaMirv: in answer to your quesiton on the spreadsheet, it can also go to U, yes; I'll be merging trunk (which has these changes) into automatic (which these changes were cherrypicked from) once things are landed08:52
UtilitarianUvulayou know, the closer we get to Zaftig Zebra, the more nervous I am every time a name needs to be announced08:53
pete-woodsdidrocks: hi, is there any information about how the CI train works with SRUs?08:57
didrockspete-woods: just do as usual, and don't diverge too much, like focus on bug fixing08:58
pete-woodsbasically I have a bug fix for something in trusty main, and wondered what the process would be, (being a person who has never had upload rights, that sorta thing)08:58
pete-woodsas usual means use CI train? or means follow the SRU page on the wiki?08:59
ogra_both :)08:59
didrockspete-woods: yeah, both09:00
didrocksjust follow the SRU page09:00
UtilitarianUvuladidrocks: are stable images only cut from stable releases?09:00
didrocksand proposed your fix to trunk09:00
didrocksUtilitarianUvula: for now, yes09:00
didrocksit's still under discussions09:00
UtilitarianUvulaok :)09:00
pete-woodsdidrocks, ogra_: okay, thanks guys, that makes sense :)09:01
didrockspete-woods: yw ;)09:01
MirvUtilitarianUvula: yeah I meant only U or also T. for SRU you need to file a bug with Impact/TestCase/RegressionPotential and also otherwise follow the https://wiki.ubuntu.com/StableReleaseUpdates instructions09:05
UtilitarianUvulaMirv: ah, ok. But to SRU I need it in U first :)09:06
ogra_well, it needs to land in U at the same time at least09:06
ogra_(you can do paperwork and prepare the trusty side even without U existing)09:07
UtilitarianUvulaokie doke09:07
UtilitarianUvulashould I create a new bug for all the things fixed in the branch, or should I tag each bug fixed in the branch separately?09:08
UtilitarianUvulathe former sounds like less hassle09:08
=== UtilitarianUvula is now known as Chipaca
mandelom26er, hello! so you were right and umd was broken!!! sorry for doubting you! mea culpa!09:44
mandelom26er, I have updated the code and now it works perfectly ok (I have tested and retested) can you take a look?09:45
om26ermandel, sure no problem, Since the release is out i guess you wont need my testing onow09:45
mandelom26er, probably not, but you added a comment and it does not feel right if I edit it :)09:46
mandelom26er, is better that you do it09:46
om26ermandel, yes, sure. let me do that.09:46
mandelom26er, awesome, thx09:47
didrocksogra_: I kicked an image to check that the infra is working fine09:53
davmor2Morning all09:58
imgbot=== trainguard: IMAGE 303 building (started: 20140422 10:00) ===09:59
* popey spies a bazillion webapps from ogra_ 10:30
ogra_only 1510:31
* didrocks goes for a run10:46
mardyCan please someone help me in making Jenkins build this: https://code.launchpad.net/~mardy/signon-ui/oxide/+merge/215661 ?10:48
davmor2ogra_: did you poach popey 's create a webapp is one easy script or something10:50
popeynot my script10:50
ogra_i have my own script10:51
ogra_but i just learned that the app name doesnt come from the .desktop file anymore10:51
* ogra_ edits all the app neames :(10:51
popeyyeah, when i saw them appear in the store I thought "he probably doesn't want that"10:52
ogra_and with bild.de we finally have the true yellow press in the store10:54
* ogra_ notices that no brit is brave enogh to upload a sun.co.uk webapp :P10:54
davmor2ogra_: that's cause we'd rather chew our own arms off10:55
NoNameYet_xnoxogra_: haha10:55
ogra_davmor2, well, you dont have to use it :)10:55
imgbot=== trainguard: IMAGE 303 DONE (finished: 20140422 11:15) ===11:14
imgbot=== changelog: http://people.canonical.com/~ogra/touch-image-stats/303.changes ===11:14
ogra_didrocks, seems to still work11:16
popeyanyone else see nothing "available" from the app scope on #302?11:24
sil2100Works fine here11:25
davmor2popey: how many apps do you have open?11:27
ogra_same here11:27
popeydavmor2: 511:27
popeyjust the ones you see on that screenshot11:27
ogra_is your U1 auth messed up perhaps ?11:28
popeyits there11:29
popeydunno how I'd know if it was messed up tbh11:29
ogra_heh, me neither11:29
popeyif I search, i then see available category11:29
davmor2popey: http://davmor2.co.uk/~davmor2/screenshots-phone/device-2014-04-22-122906.png same 5 apps open11:29
popeynowsearching unblocked it11:30
sil2100Notabug, ship it!11:31
popeywe did!11:31
davmor2only click updates right?11:32
=== MacSlow is now known as MacSlow|lunch
=== cjohnston changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cjohnston | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: -
=== retoaded changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cjohnston | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: iSCSI server is down affecting s-jenkins and q-jenkins
=== retoaded changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cjohnston | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: iSCSI server is down affecting s-jenkins and q-jenkins
bfillersil2100: hi, can I have a silo assigned for line 26 please?13:24
=== MacSlow|lunch is now known as MacSlow
sil2100bfiller: sure, give me a moment ;)13:27
Mirvbfiller: sil2100: assigned13:28
bfillerthanks guys13:28
Mirvbfiller: as usual, it targets trusty still so if you want to land it to trusty you'll need SRU bug filed13:29
Mirvand if you want to land it to U series, you need to wait until we have a name for U...13:29
sil2100Well, for experimenting and testing, silos are good anyway13:29
sil2100Since we don't have too many reliable usages now anyway before U is opened13:30
bfillerMirv: are we building anymore trusty phone images?13:30
Mirvbfiller: today we built one, and probably tomorrow too at least if U doesn't open. I don't know the grand plan, though.13:30
=== jhodapp|afk is now known as jhodapp
sergiusensMirv: is there any plan in motion to be able to still release the gallery, camera et.al. as just click? I gess bfiller can continue work on that side13:43
sergiusensa merge and commit without a publishing13:43
ogra_sergiusens, there is  big MP from barry to switch gallery to py313:44
sergiusensogra_: can't go into trunk, which means it can't get to the click which means it is also blocked with our current infra13:45
ogra_yes, it is SRU blocked anyway13:45
ogra_without a U release where it can land first13:45
sergiusensogra_: if it were just a click package it can get to the store13:46
cjwatsonIn the past we've gone ahead with SRUs in advance of the next series opening13:46
sergiusensogra_: no need for SRUs there13:46
cjwatsonI don't see why we should suddenly start hard-blocking on that now13:47
ogra_cjwatson, yeah, i would expect it could land at the same time at least13:47
ogra_so that the SRU work could already be done13:47
cjwatsonWe just need to make sure that we copy things forward as needed13:47
ogra_sergiusens, well, not so sure ... we should actually define how we update core apps for released systems ...13:48
sergiusensogra_: all core apps have been getting updates since saucy was released ;-)13:48
ogra_probably less strict than SRUs but we kind of need to make sure they get enough testing on the old image etc13:48
ogra_did anyone ever test them against saucy ? :)13:49
sergiusensogra_: well that was one of my complaints about autopilot making a change in it's API; apps became untestable13:49
ogra_right ...13:50
ogra_that will bite us at some point i guess13:50
ogra_(i really dont care if they work on saucy but we will have to test backwards compatibility from some point on)13:50
ogra_(once we had our first "stable" release)13:51
sergiusensogra_: anyways, fwiw, I thought the store for clicks would of been treated the same as a partner archive13:52
sergiusensogra_: I would put emphasis to still run the click tests on trusty now (as part of ci)13:53
ogra_well, i think core apps that we actually ship by default need some special treatment wrt testing backwards compatibility13:54
sil2100Damn, thunder close by14:00
ogra_russians ?14:01
rsalvetiplars: would you mind trying the serial console with http://paste.ubuntu.com/7307282/?14:12
rsalvetithat will give you a tty1 right after the kernel is done booting14:12
ogra_rsalveti, why wait on cdmanager ?14:13
plarsrsalveti: will do14:13
rsalvetiogra_: don't need to wait on it specifically, just wanted it to be started together with the container14:13
rsalvetiand not blocked by it14:13
ogra_i would just start on startup :)14:14
ogra_then you can even watch the container starting :)14:14
rsalvetiright, let me try if that would work14:14
rsalvetiplars: ogra_: yeah, startup went fine as well: http://paste.ubuntu.com/7307308/14:16
ogra_right, as long as the device node exists and getty is available it shoudl be fine14:16
ogra_botrh is dealt with in the initrd14:16
plarsrsalveti: that seems to get me a login pretty early14:17
ogra_dont say that !14:18
plarsit's hard to say if adb is available by then anyway14:18
rsalvetioh, it's not14:18
ogra_it isnt14:18
ogra_adb starts after the container14:18
rsalvetiadb needs the container to be up14:18
ogra_or after the container properly failed14:18
plarsthe hope here was that even if the device gets wedged in a state where nothing works (stuck booting but can't get to fastboot, or recover, or adb, or anything) that we could use this to recover from it14:18
ogra_if the container didnt attempt to start at all you are screwed14:18
rsalvetiplars: you can still14:18
plarsok, so this would get us shell access earlier at least14:18
ogra_(on my list for fixes in U)14:18
rsalvetijust use the tty to reboot the phone14:19
plarscould be some corner cases still14:19
ogra_i doubt serial will hellp you with the fastboot case14:19
plarsso now we just need a good way to shove this into the device at install time.. I also added --autologin root so that we get a shell rather than a login14:19
plarsogra_: yeah, I suspect there are lots of cases where we're still screwed14:19
plarsogra_: so the other avenue we're pursuing is instrumenting the buttons with relays14:20
* ogra_ wonders why you dont just deply a hacked up adbd job that always starts regardless14:20
rsalvetisudo reboot -f bootloader should get you back to the bootloader14:20
plarswhich is still complicated I guess since sometimes we don't know what power state the device is in, but in theory you could have it hold the power button down for [? seconds] and also hold the volume up/down for [? seconds] and end up at the bootloader ready to flash a known good image14:21
rsalvetiogra_: you could still get adb in some weird states14:21
rsalvetiuart would always work14:21
ogra_true but i guess thats a very rare corner case14:21
sergiusensplars: keep in mind that the boot into fastboot sequence is different per device ;-)14:22
plarssergiusens: yeah14:22
rsalvetiyeah, and uart would only work for mako atm14:22
plarssergiusens: these devices need a big debug plug on them that gives you full control, so that our lives would be simple :)14:22
ogra_they do have it ... if you buy these devices without the case and call them devboards ...14:23
rsalvetihm, nexus 7 could have the same uart interface, let me check14:23
plarsrsalveti: so these plugs might work on flo also?14:24
rsalvetiwill check14:24
=== alan_g is now known as alan_g|tea
plarsrsalveti: https://plus.google.com/104363289464878469115/posts/AbvyWECLPXo14:24
rsalvetibut this is for the older n714:26
plarsrsalveti: sounds like the n5 might have it also14:26
plarsrsalveti: can't find anything about manta though14:26
ogra_call samsung :)14:26
rsalvetiyeah, I think they all have an easy way to get uart14:27
plarsogra_: yeah, I'll just call their customer support line and ask about that14:28
rsalvetiplars: yeah, so we could make it work for both mako and flo14:31
=== alan_g|tea is now known as alan_g
plarssergiusens: any way with u-d-f to specify an extra/local custom tarball to add at install time? Something like this could go in at install time so it's there from the first boot like that, but I'm guessing this conf file won't be part of the normal image for obvious reasons14:33
ogra_you just have to dump it in the right place before the device reboots i guess14:34
ogra_or even after install and a forced reboot to recovery14:34
rsalvetican't it just be part of the custom tarball?14:35
rsalvetior do we need more than one custom tarball?14:35
sergiusensplars: wrt rsalveti is on the right track14:35
ogra_rsalveti, no, but we need to deploy that tarball14:35
rsalvetiogra_: flashing it with u-d-f14:35
sergiusensogra_: what would it have and how do you guarantee it's not stepping on anythings toes?14:36
=== Ursinha is now known as Ursinha-afk
ogra_does u-d-f have an option to inject that ?14:36
plarssergiusens: but u-d-f doesn't give us an option to do that, and also, what if we are already testing the custom image?14:36
rsalvetiwe could add the support for multiple custom tarballs, which I believe it was discussed before already14:36
sergiusensplars: well I had no requirement for that; was trying to follow the upgrade spec as close as possible14:36
plarswe really just need to deploy one file, and it's not one that would exist already. but we can't wait until the first boot to deploy it14:37
ogra_dumping a signed tarball into /cache/recovery and rebooting into recovery should be enough, no ?14:37
sergiusensI would rather have stgraber's opinion on this14:37
sergiusensogra_: to sign it you would need access to the magic keys14:37
sergiusensI can't sign client side14:37
ogra_wasnt there an option for unsigned ...14:38
* ogra_ thought he saw some MP pass by recently 14:38
sergiusensplars: it doesn't need to be a customization tarball; just an extra tarball in the image server14:38
plarssergiusens: but I can't put this on the image server14:38
sergiusensogra_: that's for the https server itself14:39
plarsthat's the point14:39
ogra_ah, k14:39
sergiusensplars: not even on a different channel?14:39
plarssergiusens: not if we're going to test the same images14:39
ogra_well, if you modify them you taint them in any case14:39
sergiusensplars: well it's sort of the same if you are going to inject it anyways14:39
plarssergiusens: this is to help make the smoke testing devices more resilient so we have another avenue to recover them if we get a busted image14:39
ogra_and if you do that you can as well just adb push ...14:40
plarsogra_: and if the image we just flashed is busted?14:40
ogra_then you have serial to force it back into bootloader at least14:40
sergiusensplars: can't we have this always in and triggered form initrd from a property? ogra_ rsalveti?14:40
plarsogra_: then we wouldn't have the opportunity to adb push14:41
ogra_sergiusens, i have "fix adbd if no container is there" on top of my TODO14:41
sergiusensplars: I can add the option to inject; but it needs to be signed; or you need to discuss this with stgraber at least14:41
ogra_wont need a property14:41
ogra_but yeah, we could ship some code for special cases14:41
rsalvetiwe could have a property for it, but not sure if we want this enabled by default14:42
ogra_properties dont work without container though14:42
ogra_apart from grepping presistent.$prop files14:42
rsalvetiatm we could safely have an upstart job that could start a tty at the right UART14:43
rsalvetiwe're starting adb by default14:43
ogra_but how do you know the UART device ?14:43
rsalvetiit'd just need to parse proc/cmdline to find out the right UART14:43
ogra_assuming you want that job to be generic14:44
ogra_ok, thats trivial14:44
ogra_but only works if the cmdline has a console= entry14:44
rsalvetithat's fine14:44
ogra_(and one that points to an existing device node actually)14:44
=== Ursinha-afk is now known as Ursinha
sergiusensogra_: rsalveti plars I still prefer to have a .*-debug channel; it would also be used to enable adb once we disable it by default14:50
sergiusensand by handling it on the server/image side; we can keep in sync and not play catch up14:50
ogra_well, we want to have a developer mode14:50
ogra_even on images where  its disabled by default14:50
plarssergiusens: that also means we'd need a -debug of things that already have customizations (customized image)14:50
sergiusensogra_: yeah, but you need to manually enable it ;-)14:51
ogra_sergiusens, yeah14:51
ogra_manually or via a script14:51
sergiusensplars: custom images only really on the customization framework; changing the boot cmd goes away from a customization scope14:51
sergiusenscustom images only really stick to the customization framework; changing the boot cmd goes away from a customization scope14:52
plarssergiusens: I wasn't suggesting to change the boot command, just add in that extra file under /etc/init14:53
sergiusensplars: well file placing is supposed to stay in /custom for customization14:53
plarsah, I see14:54
sergiusensplars: just saying, while you say you are customizing the image; it is far from a custom image that we promote as custom images14:54
plarsso it's more than just something a customization tarball can add14:54
sergiusensplars: any tarball can add anything, so we are fine there; it's just not a custom image if you add that mod (market speak)14:56
ogra_lets just ship the needed bits by default and add switches to en/disable them14:57
plarsogra_: as long as that switch can be flipped before the first boot, but then I think we're back to the same problem14:59
ogra_it could be a cmdline option ... but that would mean to need to modify the boot.img15:00
ogra_or it could be a property ... for that you would have to touch some file in /data/property/ from recovery15:00
=== doanac changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: doanac | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: iSCSI server is down affecting s-jenkins and q-jenkins
rsalvetiI'd like the developer mode to be enabled via a property or such15:02
rsalvetiand for that we could use a custom tarball as well15:02
sergiusensyeah, let15:02
ogra_i dont want to force people to custom tarballs for just setting a switch15:02
sergiusenslet's just no call it custom tarball15:02
sergiusensas it seems to confuse people15:02
sergiusensogra_: well it's only for ci15:03
ogra_adb shell touch /data/property/persist.plars.mode15:03
plarsogra_: that doesn't get wiped when you flash?15:03
ogra_if itr does you just touch it again while you flash15:04
plarssure, so we would just  need a --pwned option in u-d-f15:05
rsalvetiogra_: right, the custom is just because we can deploy custom properties with it15:06
ogra_--plarsed :)15:06
rsalvetibut yeah, anyway, we could hack it around15:06
rsalvetilet's just enabled it by default and we can clean everything up once we create the developer mode switch15:06
ogra_right, we just need to put it under the switch later15:07
plarssure, this could probably all be part of the developer mode once we have a real consumer/product mode for it to be distinct15:10
ogra_we should probably aggregate all that into one package15:12
ogra_so that you can just omit that package at build time for images that are not supposed to ever have a dev mode15:12
ogra_didrocks, did you plan to have an evening meeting ?15:24
ogra_(its not like much changed since the morning)15:24
sil2100I wouldn't mind, but I don't have too much to say/discuss landing-wise still15:26
ogra_right, and landings are still getting to -proposed only anyway15:27
ogra_or did anything migrate to -updates yet15:27
didrocksogra_: we do have an image15:33
didrocksso I guess just a 10 minutes one would be find15:33
ogra_not like it has any interesting changes though :)15:34
didrocksogra_: what? popey's changes are uninteresting? :)15:34
ogra_nah, but i trust them to be heavily tested anyway15:34
* popey hugs ogra_ 15:34
popeynice recovery15:34
ogra_and tehz CI infra is off15:35
ogra_so we dont get any automated results15:35
didrocksogra_: hum, why?15:36
didrocksplars: ? ^15:36
ogra_didrocks, see the ML15:36
plarsdidrocks: apparently there's a storage failure in the dc15:36
didrocksah, like 10 minutes ago :p15:37
didrocksdon't ask me to be fresh on the last 10 minutes ;)15:37
=== gatox is now known as gatox_lunch
didrocksrobru: popey: coming?16:02
robrudidrocks, any word when U is going to open?16:24
didrocksrobru: I don't know more than you. We need a name first :p16:25
* ogra_ points robru to sabdfl 16:25
ogra_please ask there16:25
Laneywe should redesign this new name process :(16:26
ogra_robru, dont count on being able to land anything in U this week ... even if we have a name it will take a few days to get everyhitng up and runnning16:26
ogra_Laney, ++16:26
ogra_we should juat use random letter generators16:26
Laneylike the early nexus 7 images?16:27
* Laney sniggers16:27
robrudidrocks, so what are we doing with landings then? should I just take the week off? ;-)16:27
ogra_you can land stuff in trusty-proposed16:28
Laneynot anything though16:28
robrujust fixes right?16:28
ogra_SRU stuff, yeah16:30
=== gatox_lunch is now known as gatox
=== bfiller is now known as bfiller_afk
sil2100robru: it is best to ask everyone if they want their fix to land as an SRU or maybe they prefer waiting for U ;)16:57
robrusil2100, well I doubt anybody will *prefer* to wait an extra week on their landings.16:58
ogra_and make sure they understand how much paperwork an SRU can be :)16:58
robrusil2100, it's more like "how life threatening is this landing?"16:58
sil2100robru: well, SRUs can take a week as well, and a lot of paperwork which we will not do for them16:58
sil2100robru: so sometimes it might not be even worth it16:59
robrusil2100, oh, we're not doing the SRU paperwork? yippee!16:59
ogra_why would you17:03
asacall went fine today?17:03
ogra_asac, what would you expect to have "gone"17:03
ogra_we are all stuck17:03
ogra_new release cant open17:04
asacok i realize it was a bit of a loaded question :)17:04
asacI take, besides that it seems to be all good :)17:04
ogra_nothing moved, so nothing broke :)17:04
ogra_well, not true, popey landed some well tested click packages and didier rolled an image including them17:05
asacthats how it should be; just hold the line by name :P17:05
=== alan_g is now known as alan_g|EOD
robruogra_, cyphermox : I seem to be getting red light of death on my nexus 4. you guys ever seen this? http://forum.xda-developers.com/nexus-4/general/final-fix-nexus-4-red-light-death-t225045418:29
robrui'm not really well equipped for opening the phone18:29
robruguess i gotta go buy some torx wrenches... sigh18:36
pmcgowanrobru, you should not need to18:39
pmcgowanred light is good, let it charge18:39
robrupmcgowan, yeah, I've seen it before where it won't turn on but then i let it charge for a few hours and it's fine... but in this case it was plugged in through the entire easter weekend18:40
pmcgowanrobru, with the wall wart?18:40
pmcgowanit can get in a state where its not actually charging18:40
robrupmcgowan, no, plugged into laptop usb18:40
pmcgowanyeah usb not good enough18:40
pmcgowanwhen its low, needs more amps18:41
robrupmcgowan, ok, it's plugged into the wall now, i'll give it a shot again in a bit18:41
robruhmmm now it's turned itself on after just a couple minutes of charging ;-)18:48
pmcgowanrobru, perfect18:49
pmcgowanthats what it do18:49
robrupmcgowan, thanks, saved me a long trip to the store ;-)18:49
pmcgowanin my experience they can always be revived18:49
pmcgowanplug in, loooong press, what for redlight, let it charge18:50
robrupmcgowan, yeah, I've revived it a couple times before, but usually only after it had been unplugged for a long time. never seen it die after being plugged in all weekend18:50
pmcgowanthat is odd18:50
pmcgowanbut if laptop was off, usb was too18:50
robrupmcgowan, no no, laptop is always on ;-)18:50
kgunnasac: you on?19:14
kgunnwondering if there is like a master view of transitioning touch from trusty to whatever "u" is gonna be19:14
kgunnand sort of a hidden question there, if i think we're gonna land mir soon... e.g.  non-blocking swap buffers which fixes qt gui thd block19:17
kgunnshould that be for both trusty and "u"19:17
robrukgunn, from my perspective, the only value of landing fixes in trusty is if they impact the unity8 preview session for desktop users... if you're talking about phone-only fixes then they should just wait for u.19:18
kgunnrobru: thanks...in that instance it sounds like a "u" target19:20
kgunnsorry to be so out of touch, if i land something today...is it "u" by default?  or does it go to trusty ?19:21
kgunni would assume "u"...19:21
robrukgunn, well it goes to "trusty" because u doesn't exist yet. but it can't get to trusty without going through SRU, which is a lot of paperwork and extra testing19:21
robrukgunn, so for the most part all landings are blocked until u gets created, unless we're talking about some critical desktop fix, then we can SRU it for trusty19:22
pmcgowankgunn, we can all take vacation until mark has a name19:25
robrupmcgowan, yes, that's what ogra told me earlier today ;-)19:25
pmcgowanI see NoNameYet is catching on19:25
* kgunn thinks ...what!?!?!19:25
robruit does seem a little bit silly to me that our entire operation is completely held up on one guy picking a codename. I'm not sure why we couldn't just open an archive called literally "u" and starting landing there, and then just rename it later19:27
cjwatsonSadly that would break a bunch of stuff19:27
cjwatsonI spent some time thinking about it last cycle and thought "nah, Mark will have learned his lesson and this won't happen again"19:27
robrucjwatson, does your IRC client ping you at every mention of "archive"? ;-)19:27
cjwatsonNo, I just happened to be reading ...19:27
* kgunn also considers old will shakespeare...a rose is a rose19:28
cjwatsonI also thought about making the version the primary "name" rather than the codename, although that also has complicated problems associated with it19:28
josharensondoanac, I had a build fail yesterday, but I think I fixed it. How do I manually kick of the jenkins job again?19:28
doanacjosharenson: can you share the MP?19:29
cjwatsonBasically I was kind of not motivated to work on this because it's a really stupid problem to have to spend any cycles at all on fixing, rather than just having a dratted codename in the first place19:29
doanacjosharenson: we are in the midst of an outage so it might not be possible at this very moment19:29
sergiusenscjwatson: more importantly, what happens after z?19:29
robrucjwatson, true19:29
kgunnsomeone needs to file a bug on the next release that's a showstopper that says we can't release w.o a name for next19:29
robrusergiusens, easy, it'll just be zz ;-)19:29
cjwatsonsergiusens: I assume we wrap around.  I've been whack-a-moling all the cases where people assume codenames are sorted alphabetically for some time19:30
josharensondonac: :-p bummer. The MP is https://code.launchpad.net/~josharenson/mir/install_glmark2/+merge/216504 if its possible19:30
cjwatsonkgunn: AFAIK we actually have a codename, we're just blocked on Mark blogging about it so that we're allowed to make it public19:30
cjwatsonI kid you not19:30
robrucjwatson, how do you "fix" code that sorts ubuntu codenames alphabetically? just hard-code the list without sorting?19:30
doanacjosharenson: you've done the right thing. I believe this isn't running because our jenkins servers are affected by the outage19:30
doanaca new commit on the branch should trigger jenkins to re-run19:30
cjwatsonrobru: Well, you fix the data so that that code doesn't need to exist19:31
josharensondoanac, ok I'll just be patient19:31
cjwatsonrobru: e.g. a while back we fixed the versions of packages in -backports to be ubuntu14.04.1 or similar rather than trusty.1, that kind of thing19:31
cjwatsonrobru: And there's distro-info for mapping codenames to versions19:31
robrucjwatson, BUT WHAT HAPPENS IN 2099?! WHAT THEN?!?! ;-)19:32
cjwatsonI see no problem with 1.0~ubuntu100.04.1 :-)19:32
robrugood call19:32
cjwatsonVersions act sensibly with that kind of thing19:32
cjwatsonWe'll all have made our millions from the 2038 bug by then anyway19:32
renatohi guys could you check why jenkins did not build this MR: https://code.launchpad.net/~renatofilho/address-book-app/fix-1306798/+merge/21673219:39
dbarthrobru: good evening, we have a new silo ready to land (silo 005)19:39
robrudbarth, it'll probably have to wait until U opens unless you feel the fixes are important enough for SRU for desktop trusty.19:41
dbarthrobru: what are the options? severity level expected for an SRU?19:42
dbarthrobru: you can see the bug reports, we're talking about links not opening in g+ or facebook for example19:42
dbarthrobru: ie, good to land imo for phone at least, and desktop when updates are considered appropriate19:42
pmcgowandbarth, can we get a blanket SRU for webapps?19:43
pmcgowanwhat ever happened to that19:43
robrudbarth, yeah, we do tend to SRU everything for webapps19:43
robrudbarth, https://wiki.ubuntu.com/StableReleaseUpdates#When19:43
dbarthright, i think that qualifies19:43
dbarthpmcgowan: could do yes, since we'll have some more to close the gap with what we had when using the browser as a container19:44
dbarth(vs the new oxide)19:45
pmcgowandbarth, right, be worth greasing the landings now19:45
pmcgowanbased on the bug set target19:45
robrudbarth, ok i'll hit publish on that silo, I see 2/3 of those bugs are already in SRU format, will need that third one converted into an SRU19:46
dbarthoh sure?19:46
dbarthlet me search the main sru request we had19:47
robrudbarth, this is the one to SRU-ify: https://bugs.launchpad.net/unity-webapps-googleplus/+bug/130878419:47
ubot5Launchpad bug 1308784 in unity-webapps-googleplus (Ubuntu) "External links broken in Google+" [High,Fix committed]19:47
renatorobru, could you check why jenkis did not build this branch? https://code.launchpad.net/~renatofilho/address-book-app/fix-1306798/+merge/21673219:50
robrurenato, well, jenkins is down. but that's more of a fginther question.19:50
dbarthrobru: https://bugs.launchpad.net/unity-webapps-googleplus/+bug/130878419:53
ubot5Launchpad bug 1308784 in unity-webapps-googleplus (Ubuntu) "External links broken in Google+" [High,Fix committed]19:53
dbarthrobru: the description should be better now19:53
robrudbarth, great, thanks19:53
robrudbarth, i'll subscribe ubuntu-sru to those bugs and then it's up to them.19:53
dbarththank you19:54
dbarthi'll m&c tomormrow morning19:54
=== bfiller_afk is now known as bfiller
=== doanac changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: iSCSI server is down affecting s-jenkins and q-jenkins
balloonsfginther, ping20:03
fgintherballoons, pong20:03
balloonsfginther, lol your email answered my question. I was really wanting music to have the real device testing for mp's. The mp I'm proposing requires it20:04
balloonsmusic still has the plugin issue, so, got it :-)20:04
fgintherballoons, how far off is that problem from being solved?20:05
balloonsfginther, let me see20:05
sergiusensfginther: can we add the testing of the clicks pre store landing? as in the ones in the 'click' view?20:06
sergiusensjust a post build job that installs and runs the tests20:06
balloonsfginther, sounds like it's rolled up in media-hub, and is *soon*20:08
fginthersergiusens, I had started some work on what you are describing, but that got redirected by working on a different method with balloons and others20:09
fginthersergiusens, we're trying to basically do autolanding with the click build and test as a part of that20:10
balloonsright.. so there's no confusion.. the click that was generated by the mp that passed everything can go into the store20:12
balloonsless work, less surprises I hope20:12
sergiusensballoons: it's not part of media hub20:12
sergiusensballoons: your music problem is tied to grilo and media scanner20:12
fginthersergiusens, is this something that would directly help you, because converting all the projects to this method may be a little ways off. Adding a test job to the current click build jobs may be a short term solution20:12
balloonssergiusens, well mediascanner 2.0 is supposed to drop the plugin20:13
balloonsI guess I'm misspeaking to say mediahub is a requirement then, eh?20:13
sergiusensfginther: if you do what I ask, we can atest off the bat that a click app passes it tests and upload directly to the store20:13
sergiusensballoons: yeah, media hub was a req for removing qtpowerd, which it already has been20:14
plarsI just heard that q-jenkins has storage again, and confirmed that the image tests on 303 are now running20:14
sergiusensfginther: this would also offload testing autopilot manually from popey and myself20:16
balloonssergiusens, I believe what's happening does that already20:17
balloonsyou remove an MP, it generates a click, tests in on the device and you approve. You can then simply grab that click and place it into the store20:17
balloons*review an MP, hah20:17
sergiusensballoons: what if you need multiple MRs?20:18
fgintherballoons, I do have some concerns about this working right away. we've already noticed that some apps are not passing when testing on the phone20:18
sergiusensballoons: what if we need a new translation pushed? Those don't go through MP/MRs20:19
fginthersergiusens, I'll try to do an interim solution to testing of clicks. it's just a bit of work20:19
balloonssergiusens, yes it doesn't eliminate building a click from trunk20:19
sergiusensfginther: balloons in the following weeks; I'm adding x86/armhf fat click packages; those would also be harder to test from MPs I guess20:20
balloonssergiusens, but I don't think we should force testing on the click build.. that's a parm or separate job20:20
sergiusensballoons: that's why I said post build step ;-)20:20
sergiusensit's different job20:20
balloonssergiusens, any hopes of landing this btw? Or did I miss it? https://code.launchpad.net/~sergiusens/phablet-tools/emu_prov/+merge/20744020:20
fgintherballoons, it would be implemented as two separate jobs, using the existing job to build it and then another job to run the test20:21
balloonssergiusens, click-buddy still needs some love for armhf builds I think.. I'm still using my own chroot.. I'm guessing you too20:21
balloonsfginther, fair enough20:21
sergiusensfginther: balloons to summarize one job builds trunk -> if it passes, it runs the tests in a job (passing the click as a parameter)-> if it passes, a possibility of pushing to store (passing the click as a parameter)20:22
sergiusensballoons: my chroot is currently broken; I can't build anything20:22
balloonssergiusens, you want jenkins to push to the store?20:22
sergiusensballoons: yes; to mitigate the one password/key to rule them all problem20:23
balloonssergiusens, yes I remember your broken chroot :-)20:23
balloonssergiusens, I like that20:23
balloonsuploading has been streamlined, it's completely possible20:23
sergiusensballoons: wrt to click-buddy; I'm going to try and get consensus with zbenjamin during sprint time and just do everything once; we have a google doc rolling with some loose ideas20:28
balloonssergiusens, yes I've been working on the file manager click build issues, and one of the issues is click-tools and qtcreator don't agree on the manifest file layout20:28
balloonsso at the moment I'm not sure of a way to allow building with both to work20:29
balloonssergiusens, do you mind sharing the doc with me?20:32
sergiusensballoons: yeah, I'll have to find it again though :-)20:32
=== jhodapp is now known as jhodapp|afk
josharensonjosepht, I'm trying to get lp:qa-dashboard to run locally, and I'm having some issues. Are you able to help?21:29
bfillerrobru: can I have a silo for line 28 please?22:50
robrubfiller, sure thing, you got 1622:52
bfillerrobru: cheers22:52
bfillerrobru: and one for line 29 too please23:09
robrubfiller, ok, you got 18 ;-)23:10
robruyou're welcome23:11
josharensoncjohnston do you happen to _not_ be eod?23:56

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