[06:48] hey Mirv, happy new year! === ev changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: ev | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: - [09:30] didrocks: no landing task force meeting? It disappeared. [09:30] (morning and HNY btw) [09:31] popey: magically! [09:31] popey: happy new year ;) [09:31] lol [09:31] yeah, Mirv and sil2100 are off today [09:31] ah [09:31] no real need to have one, let's get people catching up emails [09:31] and trolling for ogra_ :p [09:31] haha [09:31] ogra_: nexus 5 done yet? [09:31] how about now? [09:31] and now? [09:31] etc [09:32] so i got calendar notifications from my phone every damned day during my vacations ... but the firs meeting after my vacations doesnt notify me :P [09:32] haha [09:32] popey: apparently, he's rewriting something in perl soon… :p [09:32] lol [09:32] ☻ [09:39] Hmm.. Ubuntu Touch on one of these puppies... http://liliputing.com/2014/01/hp-unveils-slate21-pro-android-powered-one-desktop.html? [09:40] ++ [09:42] o/ [09:42] HNY :) [09:42] same same :) [09:46] hey asac, HNY too! [09:47] didrocks: thanks!\ === john-mca` is now known as john-mcaleely [11:46] popey: that device looks quite cool, but would be way better with ubuntu [11:55] i doubt HP would have liked to wait for 14.10 and a working desktop mode in the converged UI though :) [12:41] didrocks, ogra_: are you looking to promote a new image at all today? or can I test the current? [12:42] davmor2: just test current, we won't promote anything I guess until we are back on shape :) [12:42] * ogra_ hasnt heard anything ... and we have no meeting today [12:42] i guess its not *that* urgent to promote one today [12:42] * didrocks wonders why the launchpad cred message isn't shown === alan_g is now known as alan_g|lunch === josepht changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: josepht | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: - [13:20] wasn't there a QA website with boot startup time for Ubuntu? [13:21] Kaleo, for x86/amd64 there is ... [13:21] ogra_, where is it? I cannot find it [13:21] * ogra_ still has an open task to set something similar up for touch [13:21] Kaleo, https://jenkins.qa.ubuntu.com/view/Bootspeed/ [13:22] ogra_, hmmm [13:22] ogra_, it was a very different website before [13:22] hmm [13:22] ogra_, with even rendering stats per hardware too [13:22] yeah, there must be bootcharts somewhere [13:25] http://ci.ubuntu.com/ [13:25] ... [13:25] ...:) [13:25] that's it :) [13:27] hmm, actually not the one i meant [13:28] mine actually had proper bootchart pngs linked [13:28] but i cant find it ... these bootspeed measurements arent very helpful to track down services delaying the boot [13:31] Kaleo: http://ci.ubuntu.com/bootspeed/arch/amd64/ [13:31] ogra_: ^ [13:31] xnox, no bootchart [13:32] (thats the details page of Kaleo's link above) [13:34] ogra_, it has, click on "detailed view" and then "raw data" on the left [13:35] doesnt load anything for me :( [13:35] it's slow for me too, I had to retry 3 times to get the list [13:40] ogra_: no without pain, one can download get everything from there... there is even a full tarball of logs & full-blow bootchart images available.... somewhere.... [13:40] all i could find looks like a self knotted bootchart impostor thingie [13:41] *knitted [13:41] (through jenkins.qa.u.c) [13:45] ogra_, https://jenkins.qa.ubuntu.com/job/bootspeed-trusty-desktop-i386-acer-veriton-01/9/artifact/11/bootchart.png [13:45] yeah [13:45] it finally loaded here too [13:46] * ogra_ just looks at artifact 13 .... 57sec boot time ... ugh [13:46] seems the system stopped working though, no record since 20131108 [13:47] yeah === alan_g|lunch is now known as alan_g === alex_abreu is now known as alex-abreu [14:23] morning [14:25] hey fginther! happy new year :) [14:26] didrocks, happy new year to you as well! [14:26] thanks ;) [14:26] fginther: before your meeting starts, do you know where you backup the gpg and ssh keys for ps-jenkins? [14:27] I think we'll need them for the citrain and ev needs to put that as part of his incoming juju deploy script [14:27] didrocks, yep, I know where they are, let me dig them out for you [14:27] fginther: yeah, if you can send them to me as well, that would help for my local testing! :) [14:27] didrocks, and get them documented [14:27] perfect! [14:59] didrocks: happy new year ! [15:00] didrocks: am i correct in seeing, you guys just have some backlog to work thru before getting to mir ? [15:00] totally understandable [15:00] kgunn: hey! happy new year :) [15:00] kgunn: yeah, there are some backlogs and Mirv/sil2100 start only tomorrow :) [15:00] hmm, nobody pulled in the session changes it seems [15:00] kgunn: so expect that we try landing Mir on wednesday I would say, is that ok for you? [15:00] ok...np, will be looking to make sure mir is well behaved :) [15:01] kgunn: thanks :) [15:01] ogra_: no, see associated comment, it was pending on mterry, I didn't track then when I was on holidays [15:01] hmm, i thought it was ready when i left [15:03] didrocks, bah, thats long obsolete [15:03] ogra_: ah, well, it was still written :) [15:03] (he commented on the MP) [15:03] (not sure why it's not showing up as red anymore) === charles_ is now known as charles [16:39] fginther, hey, can we circle back with the android builder? === gatox is now known as gatox_lunch === alan_g is now known as alan_g|afk [16:49] sergiusens, I'll be available shortly [16:50] fginther, ack, thanks [17:04] sergiusens, here's the last attempt: http://s-jenkins.ubuntu-ci:8080/job/ubuntu-touch-image-builder/12/console === alan_g|afk is now known as alan_g [17:28] fginther, try and install xz [17:29] sergiusens, xz is there [17:30] hmm [17:31] fginther, is this precise, trusty,..? [17:31] sergiusens, it was raring, I've since updated to saucy and am now trying a fresh build [17:32] fginther, ack, error looks weird to me [17:32] as in, not much information :-) === ev changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: - === gatox_lunch is now known as gatox === alan_g is now known as alan_g|EOD [18:44] ogra_, any idea how to disable MTP on the phone? [18:44] ogra_, it seems to be the probable cause of my adb/ssh shell being disconnected constantly [18:45] just dug up https://code.launchpad.net/~pwlars/ubuntu-test-cases/mtp-is-less-broken/+merge/197942 [18:52] Kaleo, iirc there was a regression that cyphermox and sergiusens were about to fix === fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: fginther | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: - [18:57] Kaleo, ogra_ constantly or just at boot? [18:57] sergiusens, still fails on saucy: http://s-jenkins.ubuntu-ci:8080/job/ubuntu-touch-image-builder/14/console [18:57] sergiusens, same error regarding vmlinuz [18:57] Kaleo, ogra_ there is only one disconnect left and it's when android starts and resets the usb device [18:58] sergiusens, what was that regression davmor2 talked about ? [18:58] ogra_, something about mtp not working on boot when the device was connected [18:58] (i dont think the disconnect on boot is a regression, it was always there) [18:58] although that works for me [18:58] but it's not related to Kaleo's comment at all [18:59] sergiusens, constantly when connecting via ssh over adb [18:59] sergiusens, less frequently with just adb [18:59] Kaleo, that means you change the usb settings? [18:59] ogra_, sergiusens: cyphermox I think was looking at that, the mtp from Friday [19:00] sergiusens, I don't know what that means; I'm just connecting via ssh [19:00] ogra_, sergiusens, Kaleo: This is when you're ssh/adb session gets kicked when mtp connects I guess which is an older issue right? [19:01] your even [19:01] davmor2, that's my guess [19:01] davmor2, yes [19:01] Kaleo: yeah it does there is a bug for that :) [19:01] Kaleo, davmor2 I don't know how adb can reset ssh unless you connect over rndis [19:02] davmor2, linkie? [19:03] sergiusens: I don't know about ssh but adb shell work fine if you attach after mtp has attempted to connect if you attach before you get kicked when mtp attaches [19:03] Kaleo: I'll try and find it [19:03] davmor2, thanks [19:03] davmor2, I know; but the definition of constantly disconnects Kaleo gave is a bit different than disconnects on boot [19:04] https://bugs.launchpad.net/mtp/+bug/1233613/ [19:04] Ubuntu bug 1249162 in android-tools (Ubuntu) "duplicate for #1233613 Devices lose adb connection after phablet-flash loop" [High,Confirmed] [19:04] davmor2, mtp attaching should be the cause of the last disconnect; it's actually setting the usb device from ubuntu, then when android starts have it do the same (hence the disconnect) [19:05] davmor2, and the mtp job doesn't touch the usb config anymore, which waas the 3 disconnect [19:06] sergiusens: ah fair enough [19:07] Kaleo: ^ sorry I didn't add your nick to the bug [19:08] sergiusens: that's good to know, so now in theory it shouldn't kick adb off when mtp connects right? if so I'll run a test on that once mtp is working correctly :) [19:09] davmor2, it shouldn't, no; the usb device is reset only when changing the android property for the usb device [19:09] davmor2, thank you [19:10] / [19:12] sergiusens: thanks for the info I'll run a few tests on it when I know the mtp fix has landed [19:14] is there anything special one needs to do, to get a whole new package landed on the image? [19:14] dobey, needs to be in the touch seed [19:15] https://code.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/ubuntu-touch.trusty [19:17] sergiusens: right. i mean, if the package is set up for daily-release and in CI already, but not in ubuntu. [19:18] afaik, the daily release team can get into ubuntu [19:18] robru, ^^ [19:19] who all is the daily release team? [19:22] dobey, long read https://wiki.ubuntu.com/DailyRelease, short read https://launchpad.net/~ubuntu-unity [19:23] so that's basically cyphermox robru didrocks Mirv sil2100 [19:25] yeah, most of the people on it are in europe [19:26] kenvandine: !! :) [19:27] hey dobey [19:27] hey kenvandine. ready for the freezing tomorrow? [19:28] not really :) [19:28] heh [19:28] my plan is to not leave the house :) [19:28] dobey, what package do you need? [19:28] kenvandine: can you trigger a build of ubuntu-purchase-service into the daily ppa? [19:28] what stack is it in? [19:29] then do the rest of the daily-release work to get it into ubuntu if it builds successfully [19:29] kenvandine: click (is that a stack?) should be in the same stack that unity-scope-click is in [19:30] doesn't look like it's in the click stack [19:30] i put it next to unity-scope-click in the CI config, anyway [19:30] huh [19:31] ok, i see it now [19:31] but not showing in jenkins [19:31] hrmm [19:31] maybe it wasn't redeployed? [19:32] not sure. i saw the landing jobs for it when i looked at jenkins the other day [19:32] or maybe some other bit needs to be deployed for daily-release jobs stuff that i don't know about [19:32] who to ping about that? cihelp? [19:33] dobey, looking [19:34] i'm redeploying the stack [19:34] kenvandine, thanks [19:35] ah ok [19:38] dobey, I also deploy the webcred stack to finish ubuntuone-client-data. sorry for overlooking this portion of the deploy [19:38] thanks fginther [19:38] (2 deploys are needed on 2 different jenkins, I only did the first one) [19:40] http://q-jenkins.ubuntu-ci:8080/job/cu2d-click-package-head-1.1prepare-ubuntu-purchase-service/2/console [19:40] fginther, what's wrong with the version? [19:41] kenvandine, looking [19:41] oh [19:41] it isn't UNRELEASED maybe? [19:43] kenvandine, trusty should be ok [19:45] hrmm, one second [19:46] does it require an ubuntu revision? [19:46] * kenvandine doesn't remember what's expected there [19:46] it might. [19:46] kenvandine: can you try to do builds of ubuntuone-client-data or ubuntuone-credentials? [19:47] if those fail on the version it's probably the broken expectation of a non-native version [19:47] so, blame didrocks :) [19:48] not sure what code to fix, in order to fix that, though [19:48] i started ubuntuone-client-data [19:48] same error [19:51] ok, so almost certainly it doesn't like native versions [19:51] hrmm [19:51] let me see if i can fix it [19:53] oh no, that's not the problem [19:53] seems it's parsing the date wrong [19:54] ah, becasue the regexp is wrong [19:54] dobey, kenvandine, looks like it expects 20131204 and not 013.12.19 [19:54] err 2013.12.19 [19:55] oh [19:55] yeah [19:57] where did i get the one with dots from [20:02] huh [20:02] robru: why was this done? https://bazaar.launchpad.net/~dobey/ubuntu-purchase-service/fix-version/revision/7 [20:04] and apparently without an MP for it? [20:17] kenvandine, fginther: https://code.launchpad.net/~dobey/ubuntu-purchase-service/fix-version/+merge/200578 [20:19] ugh... you reverted the packaging cleanup... [20:19] robru would have done that to follow our packaging standards [20:20] robru should have made a merge proposal [20:20] agreed [20:20] and explained why those changes make sense here [20:21] https://wiki.ubuntu.com/DailyRelease/InlinePackaging#Package_inlining_checklist [20:22] not sure why their wasn't a MP... maybe it didn't have CI setup yet? [20:22] it had ci [20:22] or it should have [20:22] at the very least it had people working on the code and reviewing branches [20:23] "jenkins isn't running on it yet" is not a valid reason to not have an MP [20:25] dobey, agreed [20:25] the changes to debian/control made no sense anyway. it didn't make it easier to read [20:25] all of the daily release packages we do that, didrocks set a standard for that [20:25] trailing comma [20:25] etc [20:25] sorting [20:29] robru, can you comment on dobey's MP? [20:30] i have no problem discussing such proposed changes, but that's what MPs are for. you don't just outright commit to the branch [20:32] remvoing debian/source and adding .bzr-builddeb were not necessary, and wrap-and-sort -a -t makes the control file less readable and more annoying to manage. the added comment about Vcs-Bar is also confusing and unnecessary [20:40] fml [20:40] qmltestrunner == pain. [20:42] qmltestrunner: could not find a Qt installation of '' [20:42] how am i even supposed to debug that? [20:42] does that support qtchooser? [20:42] dobey, QT_SELECT=qt5 probably missing [20:42] what sergiusens said :) [20:43] and why would i have to specify that? [20:43] i certainly don't have to when building locally [20:44] to support qt4 and qt5 at the same time [20:44] dobey, you have qt5-defaults package installed [20:44] i don't support qt4 at all [20:44] to support it on the same system that is [20:45] dobey, you could build dep on qt5-defaults [20:46] but i think someone was discouraging that [20:47] it's a clean environment and i build-dep on qt5base-dev; that should be enough [20:48] "should" [20:48] but without qt5-defaults, it can't find qt [20:48] well, software has been known to be buggy [20:49] it would be nice if it was smart about falling back [20:49] why doesn't the qt5 stuff depends on qt5-defaults? [20:49] because then it wouldn't work for qt4 [20:50] kenvandine, dobey: sorry was at lunch. I sometimes don't do MPs for such simple packaging changes since I follow didrock's packaging guidelines strictly, they should be uncontrovercial changes. Also it's possible that there was some rush to get a package out at the time (eg, jenkins was delayed by several hours but we wanted to get the package published for an image build). [20:50] on the same system [20:50] robru: this package isn't in ubuntu yet. it's a brand new project [20:50] robru: and i do find those guidelines controversial :) [20:51] dobey, you are the first upstream to complain about them [20:51] dobey, yeah, I thought we were in a hurry to get it into ubuntu at the time. [20:51] dobey, that is sooo "dobey" of you :) [20:51] kenvandine: i'm probably also the only one who cares enough about package quality to complain about them :) [20:52] dobey, yeah, I find your standards bizarre. 'wrap-and-sort -a -t' makes it way nicer, trailing commas especially make future diffs a lot easier to read. [20:52] robru: i don't care if diffs are easier to read or not. i'm not reading diffs when i'm looking at debian/control :) [20:52] and i don't think wrap-and-sort makes it nicer [20:52] didrocks set strick guidance on package consistency for everything landing in daily release [20:53] dobey, um... lots of people are reading diffs of debian/control? like the release team? [20:53] having 1 space at the beginning of the line for EVERY dependency item, makes it much nicer for me [20:53] robru: really? debian/control shouldn't be changing so often [20:54] dobey, part of this is so the cu2d system can decide if the packaging needs packaging review before publishing [20:54] dobey, it may not change often, but when it does, clean diffs are nice. also one-space indents are literally unheard of, it's a lot easier to read with a nice big indent that sets it apart from the rest of the file. one-space indent is nearly invisible, hard to see where the dep list ends etc [20:54] and gives a packaging only diff for us to review [20:54] kenvandine: they can't both be guidelines, and be strict. it's one or the other. and i've already bugged didrocks about the split/native thing, and i'm trying to make it better by not having it, here [20:55] robru: the problem with wrap-and-sort is that the indent is not consistent across the entire file [20:55] what i care about is consistency... i really don't want some packages split and some native [20:55] kenvandine: i don't want any packages split [20:55] dobey: it is consistent: each following line is indented to the position of the first line ;-) [20:55] kenvandine: things that include the debian/ dir upstream are native packages [20:55] i don't want to go back and change all of them [20:55] kenvandine: using bzr-builddeb split mode is a red herring [20:56] robru: no, it is not consistent [20:56] robru: it is indented according to the value it is filling [20:56] robru: "Build-Depends:" and "Depends:" are not the same width [20:56] dobey, yes, each stanza is consistently indented according to the value it's filling ;-) [20:56] robru: which means copying a dep from one to the other is a pain [20:57] dobey, it's not a pain, just run 'wrap-and-sort -a -t' when you're done, it will indent it nicely for you. or get a real editor ;-) [20:57] and considering they are often the same, especially with packages that are runtime-interpreted code [20:57] whatever [20:57] if you have a problem with the way it exists in the tree, file a bug and a merge proposal and we can argue about it there [20:59] i don't want to waste time arguing about it in irc with multiple people about different aspects of the same thing [20:59] dobey, i don't have a problem, I am following guidelines as I have done for literally hundreds of packages ;-). you're the one that has a problem. [20:59] i have no problem. i am following the policy as i have done for literally hundreds of packages :) [21:00] a) i shouldn't have to run wrap-and-sort, ever. b) my editor is just fine. "get a real editor" is not an argument, it's an insult [21:00] dobey, what policy? I have never seen a policy that says "use one space indents and make sure you leave off trailing commas in a diff-ugly way." [21:00] sigh [21:01] dobey, the thing is, with the move to daily release this is the standard we've been given to follow [21:01] and it has been so for quite a while [21:02] so lots of packaging already following this [21:02] everything in daily release is [21:02] kenvandine, have you ever seen a package with one-space indents? I have literally never seen that before. if you're going to argue that there should be a fixed number of spaces, at least pick a sane number like 4 or 8. [21:02] robru, i have... lots of them [21:02] every package i've ever put in ubuntu has one space indents [21:03] every Description: line has one space indents [21:03] dobey, in the description line, it's a hack to tell the parser that it's a continuation line instead of starting a new field.... for dep lists i've never seen it that way [21:03] dobey, this isn't about difference of opinions... it's about standardizing all the packages we have auto landing for the distro [21:04] kenvandine: no, this is about commiting chages to a tree without a merge proposal [21:04] that would have helped understand, sure [21:05] but that isn't what we're talking about now... it's about packaiging preferences [21:05] we're only talking about that because i'm reverting the changes that were committed without an MP [21:06] and they happen to be packaging preferences [21:06] dobey, ok, so you'd take the changes if robru proposed them? [21:06] no [21:06] but i'd discuss them in the MP [21:06] and with didrocks [21:06] i guess that's the next step [21:07] and you can try to convince didrocks :) [21:07] if the commit i was reverting were style changes to c++, we wouldn't be having this conversation [21:07] and all these are, are style changes [21:11] dobey, you're absolutely right, they are style changes... that's why they should be uncontrovercial and left alone. [21:12] robru: by that argument, you should never had made the change then, because the existing style should not be controversial and it should be left alone :) [21:13] dobey, no... i'm following a standard as I was told to do by my team lead. [21:14] dobey, these debates with you are so much more fun with rum [21:15] i can go get some [21:15] it's about that time anyway [21:15] it is... and getting cold enough that it would be nice for warmth [21:16] kenvandine: https://ubuntuone.com/08D8LG5Pg8wG5Nx6i7bOKe [21:17] that looks much better than this banana stuff i have left from the virgin islands... [21:19] it is very good indeed [21:30] oh man, i should have bought some bananas, caramel, and ice cream, when i went to the grocery for lunch. could totally be having some tasty bananas foster right now if i had. [21:30] might have to go back after work [21:32] kenvandine: how many packages are using the split mode for daily release anyway? [21:34] tons [21:34] gotta be well over a hundred [21:34] everything in cu2d [21:35] so nothing a 30 line python script couldn't handle for turning them into proper native packages? [21:36] probably [21:37] it would still be a pain [21:37] pfft [21:38] only because you'd have to review all the MPs that took only 10 minutes to create with the 30 line python script :) [21:38] yup :) [21:38] and they'd all have to get in the distro too [21:38] and you'd know i'd do it just to annoy you and kill the "there are too many packages using split mode already" argument [21:39] * kenvandine debates a drive to visit dobey [21:39] :-D [21:39] heh [21:39] i have rum :) [21:39] :-D [21:40] well the Kraken is almost gone, but the Flor de Cana is still almost a full bottle i think [21:41] of course there's a packey that's only 1 mile away, so not really an issue :) [21:42] :p [21:42] aka "the man" in virginia, since you can't buy it anywhere else [21:42] would need lots of rum to convince me switching to native is a good idea [21:43] i don't need to convince you. you already know i'm right, and native format is right, because the packages are already native, and split mode is a hack to probably workaround some issue we wouldn't have anyway if we properly used native packages :) [21:44] you might be right... [21:44] heh [21:45] i really don't know... but i'd hate to change everything just because [21:45] well yeah, i wouldn't change it "just because" === fginther changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp | Landing instructions: http://paste.ubuntu.com/6292280/ | Known issues: -