=== akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === frobware_ is now known as frobware === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr [09:01] voidspace: hangout? [09:03] dooferlad: omw [09:27] angleterrettes, someone free to look at my collection of branches changing yaml deps? see bug 1504821 comments [09:27] Bug #1504821: Please switch dependency from gopkg.in/yaml.v1 to gopkg.in/yaml.v2 [09:27] rogpeppe: maybe if you're about? ^ [09:28] mgz: will do [09:28] mgz: we've found that upgrade problem BTW [09:29] rogpeppe: ace, both halves? [09:29] mgz: yeah [09:29] mgz: not easy to fix though [09:30] mgz: we're having a hangout about it now if you want to come and join the fun :) [09:30] ooh, ooh, invite me! [09:31] mgz: https://plus.google.com/hangouts/_/canonical.com/gogogo?authuser=1 [09:31] (or authuser=0) [09:33] mgz, I'm actually semi through that bug (#1504821) for the chicago cubs branch - shall I carry it on or are you happy to carry on? [09:33] Bug #1504821: Please switch dependency from gopkg.in/yaml.v1 to gopkg.in/yaml.v2 [09:33] rogpeppe: I need to create a tempdir for a test and then remove it on test completion [09:33] rogpeppe: I can do that with ioutil [09:33] rogpeppe: but IIRC there is a test suite method to do it, and I can't seem to find it [09:34] rogpeppe: do you know it? [09:34] mattyw: I think our work will be complimentary [09:34] mgz, I'm only doing it in core - not any of the deps [09:35] mgz, if you're happy for me to carry on I'll carry on :) [09:35] mattyw: mgz: ^^ do you know of a test suite method for creating a temp dir for a test? [09:35] mattyw: right, I started at doing all the deps [09:35] I'm sure there is one, I just can't find it [09:35] voidspace: I think we have a few... [09:35] voidspace, I don't remember the name - but it's in gocheck [09:35] mgz: mattyw: heh, thanks [09:35] voidspace, https://godoc.org/gopkg.in/check.v1#C.MkDir [09:35] voidspace, is that what you're after? [09:36] mattyw: awesome, thanks [09:36] voidspace, you're very welcome [09:36] mattyw: just looking [09:36] I expect it's exactly what I'm after [09:36] mattyw: yep [09:36] ah, after the *suite* finishes running [09:36] well, for the suite would be fine [09:37] just need to modify the test [09:38] looking at our test suite, we use that method "per test" anyway [09:38] I guess it doesn't matter [09:40] voidspace, I was sure there was another one - I'm just having a look [10:04] mgz, just to confirm then - migrate to yaml.v2 in master first right? [10:06] mattyw: yes, and your life will be a little easier when some of these other dep change branches land [10:07] mgz, awesome stuff, thanks [10:07] mgz, I'll be starting that in the next few minutes [14:01] katco: ping [15:51] mgz, ping? [15:51] mattyw: hey [15:52] mgz, hey, you might be able to ignore me now actually [15:52] mgz, was looking at this https://github.com/juju/httprequest/pull/35/files [15:52] mgz, the json body is because juju/testing was updated it looks like [15:53] mattyw: and the go 1.2 vs later thing [15:54] mgz, I see [15:54] mgz, LGTM [15:54] I could also have updated testing to always set content-type for post, even if no content has been supplied [15:54] but this seems like it's making the test more realistic and avoiding the golang bug fine [15:54] mgz, any idea what's happened here? https://github.com/juju/cmd/pull/22 [15:55] mgz, again ignore me [15:55] mgz, I thought it had been accepted but not merged [15:55] mattyw: I had to do futzing with the gating jobs [15:55] because I made them hostile to dependency changes [15:56] mgz, 2 days ago - were you doing it saturday? [15:56] in my defense, it was tipping it down outside the hotel all morning [15:56] mgz, ah yes - at the sprint [16:11] rogpeppe: is the intention of godeps -t to get all testing deps, or just those of the given package? [16:12] mgz: all testing deps of the named packages [16:12] mgz: so it won't get testing deps of deps that aren't in the original package list [16:12] mgz: (because I found that's almost never what you want) [16:18] rogpeppe: yup, that sounds sane to me. [16:19] so, with the current stuff landed, plus charm which I'll have a quick look at again, should be able to do the remaining charm* bits [16:19] mgz: tbh I'd prefer it now if -t was the default [16:19] mgz: i don't think i've ever not wanted to update testing deps too [16:19] rogpeppe: yeah, that surprised me when I was updating the other day, forgot about the nuance [16:22] katco: ping [16:22] (^was waiting on natefinch for lumberjack landing, and it's US holiday today, but it's only a dep for the tests, not the project, so other dep updates are fine) [16:22] wwitzel3: was she swapping today? or just holidaying? [16:22] mgz: thought she was online today, but maybe not, yeah [16:27] Bug #1456659 opened: Juju bootstraps a machine but cannot SSH to it [16:40] :q [17:09] mgz: i've proposed https://github.com/juju/juju/pull/3485 which should fix the bug, although i haven't had time to actually try the upgrade for real yet: https://github.com/juju/juju/pull/3485 [17:11] rogpeppe: I'm fine with merging it after review and letting CI find out if it doesn't blow things up any more [17:18] Bug #1505309 opened: apiserver: pinger can access state before upgrade has completed [17:21] Bug #1505309 changed: apiserver: pinger can access state before upgrade has completed [17:27] Bug #1505309 opened: apiserver: pinger can access state before upgrade has completed === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr === akhavr1 is now known as akhavr [23:34] Bug #1505435 opened: option to enable --unsafe-caching for uvt-kvm