[00:29] wow, trying to find somethign that _wont_ marshall into yaml is more complicated than I thought [00:32] urgh [00:32] thumper: menn0 http://paste.ubuntu.com/13205113/ [00:32] * thumper looks [00:32] yaml.Marshal returns an error [00:32] but the code panics ... [00:32] * menn0 looks [00:33] heh [00:33] because... reasons [00:35] https://github.com/go-yaml/yaml/issues/144 [00:35] so, to test that writeyaml fails on invalid yaml [00:35] i have to [00:35] a. find some invalid yaml [00:36] b. find some invalid yaml that doesn't crash the library [00:36] \o/ [00:36] i've checked yaml.v1 and yaml.v2 [00:37] davecheney: it looks like panics that are yamlErrors will be turned into errors [00:37] and all others will be allowed to paniuc [00:38] the whole yamlError thing is a crappy way of allowing deeply nested encoding calls to abort the whole process immediately [00:38] a way to get around Go not having exceptions [00:38] i'm discovering why this function had not test coverage ... [00:40] davecheney: if that panic were a call to the package's fail() func, an error would be returned [00:49] https://github.com/juju/utils/pull/172 [01:38] thumper: could i beg for a small review to address a windows test failure https://github.com/juju/charmrepo/pull/43 [03:15] thumper: tyvm for the review [03:16] np [03:17] sorry for the delay, was otp for 1:1s [03:19] np, long meeting :-) [03:21] thumper: logging doc updates: http://reviews.vapour.ws/r/3091/ [03:21] wallyworld: no, just four in a row [03:21] allover and done with for the week :-) [04:43] thumper: super quick one: http://reviews.vapour.ws/r/3093/ [04:44] done [04:49] menn0: hmm [04:49] 88 files changed, +1,050 −2,330 [04:49] http://reviews.vapour.ws/r/3094/ [04:50] holy shitballs batman [04:50] I felt it was bad, but didn't realise it was that bad [04:52] thumper: good net line delta though [04:52] hmm... diff is off [04:52] includes previous branch... [04:53] which I though landed already [04:53] this will explain a lot [04:54] thumper: right [04:54] i've got to stop now but might jump back later to review [04:54] not sure why the diff is bad [04:54] don't bother [04:54] I'm stopping too [04:54] i want to get a few other things squared away as well [04:57] menn0: that's better, 43 files, +98 −1,375 [04:57] now I'm done [04:58] laters folks [05:53] http://reviews.vapour.ws/r/3095/ [05:53] anyone for a little one ? [09:25] fwereade: morning to you. did you have a few minutes to spare a sunburnt bread thief? [09:34] wallyworld, heyhey [09:34] sure, joining our hangout [09:35] fwereade: on my way [09:36] anastasiamac: https://plus.google.com/hangouts/_/canonical.com/ian-william [10:02] fwereade, standup? [10:34] voidspace, ping, 1:1? [10:34] frobware: omw === akhavr1 is now known as akhavr [14:15] Bug #1514444 opened: Windows github.com/juju/juju/cmd/jujud/dumplogs/dumplogs.go:65: undefined

=== Ursinha_ is now known as Ursinha [14:51] Bug #1514451 opened: TestUniterConfigChangedHook fails

[14:51] Bug #1514456 opened: Juju is unable to detach root volume

[14:54] Bug #1514451 changed: TestUniterConfigChangedHook fails

[14:54] Bug #1514456 changed: Juju is unable to detach root volume

[14:57] Bug #1514451 opened: TestUniterConfigChangedHook fails

[14:57] Bug #1514456 opened: Juju is unable to detach root volume

[14:57] Bug #1514462 opened: Assertion failure in TestAPI2ResultError

[15:05] ericsnow: ut [15:05] ? [15:37] dimitern: ping, hey do you think you're going to wrap up bug 1483879 soon? [15:37] Bug #1483879: MAAS provider: terminate-machine --force or destroy-environment don't DHCP release container IPs

[15:37]

[15:38] katco, yes [15:38] katco, I need another review on it [15:39] dimitern: ok, cool :) just trying to figure out if bugsquad moonstone needed to pick it up [15:39] and was having lots of maas issues, but managed to verify most of the cases at least manually [15:39] nice [15:47] wwitzel3: hey before you get too far into that bug, can you rebase/merge master into our lxd branch so we can get a bless and prepare for landing? [15:47] wwitzel3: " katco: cimmit 1c86f86 is what you want to merge." [15:47] katco: sure [15:48] wwitzel3: ty [15:49] frobware, voidspace, dooferlad, can you please review http://reviews.vapour.ws/r/3088/ ? [15:49] dimitern: *click* [15:49] dimitern: in a bit if no-one gets to it. Got my head stuck in a bug right now. [15:50] voidspace, sure, np [15:51] fixed the test failure I had, now stuck in a different one *grrr* [15:57] voidspace, do I remember you having familiarity with the rsyslog stuff? [15:58] voidspace, wondering if you know why the worker apparent deletes the rsyslog config file when the worker stops [15:58] !!!!! [15:58] fwereade: no [15:58] fwereade: it's changed a lot since I worked on it apparently [15:59] fwereade: I don't think that was done by wwitzel3 and I when we worked on it [16:00] blame would know though... [16:00] voidspace, yeah, that's all axw's code from about 18 months ago [16:01] voidspace, and he probably did it for a reason, but I'm not clear what that reason might be, thought you might have already gone through the digging process [16:02] voidspace, np, I'll ask him tonight, it's not critical [16:02] cool [16:02] voidspace, and moot soon enough what with db-log \o/ [16:02] hey all, happy monday [16:02] fwereade: that sounds like the code to clean up the agent stuff... there was a bug about that happening when you sigkilled the agent [16:03] cherylj, katco anything urgent I should be aware of? [16:03] crawling through the inbox after being out :) [16:03] alexisb: nothing new that I can think of [16:03] alexisb: you feeling better today? [16:03] natefinch, ooh, if there's a reported bug associated I can be a bit grumpier about it ;p [16:03] cherylj, much! thank you [16:04] fwereade: I would search for it, but I think i'm being bitten by the launchpad "feature" where it's not searching across series [16:05] alexisb: nope [16:07] alexisb: glad you're feeling better :) [16:08] thanks katco [16:08] thanks to you and the team for keeping things moving in my absence [16:12] katco: I rebased that commit in to our branch, but I'm not sure where it goes now [16:12] katco: there is still conflicts against current master that need to be resolved [16:13] wwitzel3: i.e. you're having trouble resolving the conflicts? [16:13] katco: so not sure what to make the PR against [16:13] katco: no, that hash is rebased [16:13] katco: but that hash isn't the current HEAD of master [16:14] wwitzel3: oh right. so you've got 1 pr up against the lxd feature branch right? [16:14] katco: no, I rebased, so it won't merge against our feature branch [16:14] wwitzel3: ah i see now [16:15] wwitzel3: gosh... not actually sure how github handles that? if you propose against the feature branch what does it say? just merge conflicts? [16:15] dimitern, will also take a look in a bit [16:15] katco: yeah, unable to merge [16:15] katco: which makes sense, since I rewrote history [16:16] wwitzel3: right [16:16] wwitzel3: hrm. maybe just use git to push directly into the FB? [16:16] frobware, ta! [16:16] wwitzel3: and do a force? [16:17] ericsnow: natefinch: any opinions on how to tackle this? [16:17] dimitern, I started looking then realised it was quite large and I /think/ I now know why one of my tests fails.. [16:17] frobware, yeah? what was it? [16:17] dimitern, (ha, don't ask). :) [16:18] katco: yeah, I had to force push to my copy of the feature branch to get it up, so the merge exists under my namespace right now [16:18] frobware, ok :) [16:18] wwitzel3: b/c the point is to get a bless of the FB, so skipping the tests might be OK [16:18] dimitern, it was one of those... If only I had taken the dog for a walk earlier... type of bugs. :) [16:18] katco: right, makes sense [16:19] frobware, right :) [16:20] wwitzel3, katco: maybe make a new branch off the feature branch and cherry pick what you need (if it's not too many changes). I've had success with that in the past when I get into a bad spot [16:20] natefinch: we need .. a lot [16:21] katco: I don't have permission to do a force anyway ;) [16:21] wwitzel3: lol i can take care of it if we think that's the way to go [16:25] katco: well, I could just merge it instead of rebasing [16:25] katco, wwitzel3: force push onto the feature branch is fine if no one else has outstanding branches against it [16:25] wwitzel3: yeah we can do that too. rebasing is nice if we have to come back to anything later, but if it's this much of a PITA maybe merging is the way to go [16:26] I would just force push the feature [16:26] wwitzel3: ok. i'll take care of that then [16:26] you brought it up, this is for getting blessed [16:26] wwitzel3: ty for rebasing [16:26] katco: np [16:27] very minor, mostly dep.tsv nags [16:28] fwereade: gota minute to talk? I found another odd bug in my unitassigner code. Machines weren't getting created with sequential ids (i.e. I'd get 0, 1, 3, 8..) Any immediate thoughts in what might cause that? [16:29] natefinch: just add "machine number assignment no longer sequental" to the release notes, feature shipped ;) [16:29] lol [16:33] rogpeppe: did you have a script for automating the import sorting (i.e. project internal vs. project external)? [16:33] natefinch: yeah [16:33] natefinch: github.com/rogpeppe/sortimports [16:33] natefinch, huh [16:34] natefinch, yes, for some reason st.sequence() is getting called too often [16:34] natefinch, ideally when you know you need to create a machine you consume a sequence number just once and pass that into the txn-builder loop [16:35] natefinch, possibly one slipped into a createMachineOps that was hitting failure a it later? or something? [16:36] fwereade: sounds plausible [17:02] dimitern, frobware, voidspace: hangout [17:03] wwitzel3: your rebased branch is here? git@github.com:wwitzel3/juju.git [17:06] katco: yeah, lxd-provider [17:06] wwitzel3: k [17:18] dammit [17:24] frobware: dooferlad: dimitern: if you have time http://reviews.vapour.ws/r/3097/ [17:27] frobware: Did someone from your team get a chance to look at bug 1512371? [17:27] Bug #1512371: Using MAAS 1.9 as provider using DHCP NIC will prevent juju bootstrap

[17:27] cherylj, nope [17:29] cherylj, but we talked about dimiter's wip branch which help fix this: https://github.com/dimitern/juju/tree/maas-1.25-better-bridge-script [17:30] frobware: okay, so that should address that issue as well? [17:31] cherylj, should do / can do, but wasn't specifically started for that bug though. [17:31] cherylj, the bug came after dimitern started work on that [17:31] wwitzel3: can you tal at this: https://github.com/juju/juju/pull/3694 [17:32] wwitzel3: i can't force into FB, so did a merge --theirs of your branch [17:32] wwitzel3: and would like to land that into FB [17:32] wwitzel3: just need a sanity check [17:33] frobware: ah, okay. so that branch was just for improving things, and not for a specific bug? [17:34] cherylj, correct. the issue is that the current script changes the order of what's in /e/n/i. the branch script is more complicated and attempts to change things in-situ [17:35] katco: looking [17:36] frobware: is there a target for landing that branch? We'll probably want to make sure we pull that in when we do 1.25.1 [17:36] cherylj, initially we said master [17:36] katco: yeah, that looks like it did the right thing to me [17:36] wwitzel3: k merging [17:37] frobware: it looks like that branch is based off of 1.25? [17:40] gah, I've somehow disabled the scroll wheel in sublime :/ [17:41] ...nope I've disabled both the scroll wheel and right click [18:15] how ennerving, a place that does one of the best croissants in the city makes also one of the worse coffees [18:17] heh [18:17] this country would really beneffit from discovering donuts :p [18:17] haha [18:55] ....and reboot fixes it. [19:10] anyone here familiar with dumplogs to review http://reviews.vapour.ws/r/3100/ ? [19:10] i'm not familiar with juju-dumplogs but windows seems to feel left out [19:13] Bug #1514555 opened: local provider "machine0" agent fails to start up on wily if "/var/lib" is not on the "/" partition. [19:16] Bug #1514555 changed: local provider "machine0" agent fails to start up on wily if "/var/lib" is not on the "/" partition. [19:22] Bug #1514555 opened: local provider "machine0" agent fails to start up on wily if "/var/lib" is not on the "/" partition. [20:04] Bug #1514570 opened: 'JUJU_DEV_FEATURE_FLAGS=address-allocation' blocks after first 3 ips are allocated [20:07] Bug #1514570 changed: 'JUJU_DEV_FEATURE_FLAGS=address-allocation' blocks after first 3 ips are allocated [20:19] Bug #1514570 opened: 'JUJU_DEV_FEATURE_FLAGS=address-allocation' blocks after first 3 ips are allocated === urulama is now known as urulama__ [21:13] katco: good news, both bugs with unit assignment were from the same source, and I have a fix for it. [21:14] natefinch: nice! === dames is now known as thedac [21:51] uh, how do un-revert a PR? I tried re-PRing with the same old branch with some added fixes, and only the fixes are getting applied, not the whole of the branch... which means the unit assigner code is not getting added back to master [21:52] * natefinch wonders if he should revert the revert PR... [21:55] I got it.. just cherry-picked the original merge === natefinch is now known as natefinch-afk [22:00] katco, wwitzel3, ericsnow_: if you guys have time... the fixes are very small: (easier to see on github) http://reviews.vapour.ws/r/3103/ [22:00] ^ https://github.com/juju/juju/pull/3698 [22:00] natefinch-afk: ok, i'll take a look [22:15] rick_h__: ping [22:19] thumper: pong [22:19] rick_h__: hey there [22:19] rick_h__: wheren't we going to talk this week about CLI stuff? [22:19] I thought you were going to make some meetings [22:20] thumper: yes I was/am [22:20] k [22:20] thumper will get them on tonight sorry for the delay [22:20] that's fine [22:20] it isn't like I've got nothing to do :) [22:20] thumper: thanks fornthe poke [22:20] np [22:21] heh +1 [22:36] natefinch-afk: wwitzel3: ericsnow_: as you EOD, be sure to update any bugs you're working on [22:36] katco: k === ericsnow_ is now known as ericsnow [22:39] thumper: added please adjust if those don't work out [22:40] rick_h__: can we move tomorrows one out 30 minutes? [22:40] rick_h__: I'm normally at the gym between 12 and 1 [22:40] thumper: wfm [22:40] cheers [22:41] alexisb: why is the TL call this week on a different day? [22:41] thumper, becuase weds is a US holiday [22:41] ah [22:42] thumper, hopefully I actually changed this weeks meeting and not some random day in dec ;) [23:32] katco: yep, will do, thanks for the reminded [23:32] reminder