[01:00] axw: quick one? https://github.com/juju/juju/pull/7249 [01:05] wallyworld: looking [01:17] wallyworld hml: shouldn't we have v2-unstable while we work out the breaking API changes we want to make? [01:17] on goose [01:19] axw: i'm sorta ambivalent about that - pita to change the imports again, plus we are really the only ones using it [01:19] wallyworld: true enough [01:19] ok [01:20] axw: and we *still* have charm.v6-unstable et al in core :-/ [01:20] indeed [01:21] axw: I did add a note to the readme.md that the branch was experimental for now and to use v1 for stable. perhaps unstable would have been better phrasing [01:21] hml: okey dokey [01:22] axw: just have to remember to change the readme.md at some point. :-) [01:22] hml: on other repos we use -unstable suffix to indicate that hte API may break, then get rid of the suffix when we're happy with it [01:22] hml: but since nobody else is using this, it should be fine [01:22] just so long as it doesn't go on for too long [01:23] axw: agreed [02:15] babbageclunk: i know you're having fun with GCE, but here's a review for when you need to give your brain a rest https://github.com/juju/juju/pull/7251 [02:31] wallyworld: looking [02:31] yay, ty [02:31] sorry [04:28] we doing the tech board now? [04:28] thumper: I'm going to show up [04:28] thumper wallyworld jam menn0: I need to eat, so won't be coming [04:28] thumper, jam: my brain feels full but I'm ok to join [04:29] menn0: we're in there, but we'll try to keep it lighter [04:41] wallyworld: lgtm'd [04:42] yay, ty [05:05] wallyworld: ping? Or are you in a call? [05:05] in a call, finished soon [05:28] babbageclunk: hey [05:31] wallyworld: hey, sorry - are IngressRules being sorted somewhere before being hashed to form a key? [05:32] wallyworld: I couldn't see it, but it seems like that would be needed. [05:32] for gce provider? not sure tbh. it's been ages sincei looked at that code [05:32] i think they are yes [05:32] probs in the firewaller [05:32] i'm sure there's a bespoke sort function [05:33] babbageclunk: yep. network.SortIngressRules [05:34] wallyworld: I mean the cidrs within the rukes [05:34] rukes [05:34] gah [05:34] rules [05:35] oh, not sure [05:35] otehr providers don't need that sorting i don't think [05:35] so it would be gce specific if it were there [05:37] wallyworld: Ok, I'll check again and add it if I don't find it. (I just mean so that if there's one rule for tcp:80 from (1.2.3.0/24 and 2.3.4.0/24) and another for tcp:3306 from (2.3.4.0/24 and 1.2.3.0/24) they get combined correctly. I'll add a test for it anyway.) [05:38] babbageclunk: yeah, that scenario should be handled, i though it was. the code to do that is indeed in the gce section somehwere. i recall seeing it [05:39] but add a test for sure if there's not coverage [05:39] there's also a bunch of tests [05:39] for various scenarios [06:50] wallyworld: something dodgy about max-status-history-age, see http://juju-ci.vapour.ws/job/github-check-merge-juju/781/artifact/artifacts/trusty.log/*view*/ [06:51] axw: i fixed that failure locally and once tests passed, pushed before landing [06:51] how did that get through i wonder [06:52] i'll look, see if latest landing still has it === frankban|afk is now known as frankban [09:08] is anyone around that might be able to give this a review, please? It's been waiting for 9 days now. https://github.com/juju/juju/pull/7222 [09:08] wallyworld, axw: ^ [09:09] rogpeppe: not right now sorry, but will tomorrow morning if nobody else gets to it first [09:09] axw: do you know if there are any juju-core devs in non-antipodean timezones any more now? [09:11] rogpeppe: jam, hml, and externalreality [09:12] axw: ok, that's good to know, thanks [09:14] rogpeppe: reviewed +1 [09:15] * menn0 isn't actually working right now but saw rog's plea :) [09:15] rogpeppe: in future, email me if you're having trouble getting a review [09:16] menn0: ❤ [09:16] menn0: i've been away for a week [09:16] rogpeppe: ah right [09:16] menn0: so it wasn't too much of an issue [09:16] rogpeppe: I had started to look at that PR last week but must have gotten distracted [09:17] menn0: what's the magic string to get a test CI run on a PR, BTW? [09:19] rogpeppe: $$merge$$ will run all the tests, !!build!! I believe to request a pre-merge check of the merge tests [09:19] rogpeppe: that's supposed to happen automatically though [09:19] at first submission at least [09:20] menn0: ah, it didn't seem to on the above PR for one [09:22] rogpeppe: it seems unreliable (or maybe user specific?) [09:22] menn0: i think i prefer it to be explicit anyway [09:23] menn0: like the golang.org one === salmankhan1 is now known as salmankhan === salmankhan1 is now known as salmankhan === salmankhan1 is now known as salmankhan [14:34] @team, https://bugs.launchpad.net/juju/+bug/1684143 [14:34] bdx_: Error: "team," is not a valid command. [14:34] Bug #1684143: applications deployed to lxd on aws instances failing === frankban is now known as frankban|afk [16:38] ongoing juju command-level mocking simplifications https://github.com/juju/juju/pull/7254 [16:38] anyone around for a review of this? ^ [17:40] wallyworld: rogpeppe - question. If i'm having an issue with resource-get, and getting zero output on the controller, is just the "how to reproduce" isntructions valid enough? [17:41] i dont want to file a bug that wont help anyone, but at the same time i cannot deploy any of teh older k8s bundles to test an upgrade path scenario as resource-get just hangs indefinitely. [18:01] lazyPower: if it hangs indefinitely, sending SIGQUIT (or doing ctrl-\) can be useful to show where it's hung [18:01] i just filed https://bugs.launchpad.net/juju/+bug/1684242 -- will redeploy and give that a go [18:01] Bug #1684242: resource-get hangs indefinitely on older k8s bundles [18:14] updated, thanks for the detail rogpeppe === frankban|afk is now known as frankban [22:47] babbageclunk: you find the GCE code to handle the CIDR/port aggregation? [22:49] wallyworld: yeah, but I can't see it doing any sorting of CIDRs [22:50] wallyworld: got time for a hangout? [22:50] ok, seems like something that needs to be fixed i guess [22:50] sure [22:50] standup