[00:08] <wallyworld> katco: good evening, not that you should be still be working now, but i've commented on your PR. so you have something to look forward to when you wake up tomorrow :-)
[01:07] <axw> wallyworld: would you be satisfied if I just doc commented syncToolsAPI? i.e. say that it's a subset of state/api/Client?
[01:07] <axw> (thanks for review btw)
[01:07] <wallyworld> axw: yep, sure
[01:07] <wallyworld> just to assist the casual reader
[01:07] <axw> cool
[01:07] <axw> yup
[01:39] <davecheney> thumper: http://paste.ubuntu.com/8174218/
[01:39] <thumper> my fix at the end was very like that
[01:40] <davecheney> ok, let me see if I can make a repro out of it
[01:47] <thumper> wallyworld: that is a long email, will read it fully later, but... good effort
[01:47] <wallyworld> thumper: yeah, it is long, sorry
[01:47] <thumper> don't be sorry
[01:48] <wallyworld> i'm not, well, that's what i say to the girls
[01:48] <davecheney> wallyworld: that's a yello card mate
[01:48] <bodie_> davecheney, any word on pr 617?
[01:48] <bodie_> just when you have a minute
[01:48] <davecheney> lets keep the induendo below the audible level
[01:48] <davecheney> bodie_: i haven't looked at it
[01:48] <davecheney> i'm not the on call reviewer today
[01:49] <wallyworld> it's friday, gotta have a little fun
[01:49] <bodie_> gotcha -- can you leave a comment or something to the effect that it's cool if someone else has a look then?  I don't want this thing to get trapped in limbo
[01:49] <bodie_> I just have a bunch of stuff depending on it https://github.com/juju/juju/pull/617#discussion_r16817826
[01:50] <bodie_> just since you'd left a negative comment
[01:51] <thumper> wallyworld: I have to agree with davecheney on this...
[01:51] <wallyworld> sigh
[01:51] <thumper> sorry dude
[01:51] <bodie_> and since we're on totally different timezones, and nobody else will respond to it since you'd left a comment
[01:51] <thumper> there is a time and place, but this isn't it
[01:52] <wallyworld> thumper: it's not like you haven't been "guilty" before :-P
[01:52] <thumper> wallyworld: doesn't make it right
[01:52] <thumper> I'm happy to be called out on it
[01:53] <bigjools> I have to agree, wallyworld should be more careful
[01:54] <bigjools> I mean, if you're going to try and make someone take offence, do it properly
[01:54]  * bodie_ furiously scribbles notes
[01:54]  * wallyworld makes a mental note who has a thin skin for next time
[01:55] <davecheney> wallyworld: casual sexium isn't about having a thin skin
[01:55]  * bodie_ compares notes with wallyworld and rebases his changelist
[01:55] <wallyworld> it wasn't sexism, but anyway
[01:56] <wallyworld> define sexism: prejudice, stereotyping, or discrimination, typically against women, on the basis of sex.
[01:56] <wallyworld> anyways, let's move on
[01:57] <bodie_> well, where were we.  oh yes.  so, I've noticed that when a core member on a 12 hour offset timezone comments on something, other team members are hesitant to add their two cents to discussion, and thus my review gets trapped in the equivalent of a HTTP syn timeout on a 12 hour latency window
[01:57] <bodie_> I chatted about the PR with TheMue and alexis today and both indicated they wanted davecheney to be the one to respond to that comment on 617
[01:58] <davecheney> thumper: http://paste.ubuntu.com/8174325/
[01:58] <davecheney> repro for the gccgo bug
[01:58] <davecheney> filing an issue upstream now
[01:59] <bodie_> if someone could at least acknowledge what I'm saying, maybe I wouldn't wonder as much whether the cold shoulder has anything to do with wanting to keep us out of the loop
[01:59] <thumper> I'd say more casual sexual harassment rather than sexism... however... moving along
[02:00] <davecheney> https://code.google.com/p/go/issues/detail?id=8612
[02:00] <bodie_> or, you know, you could post a three word reply on that pr, such as "someone else take this" and it would be totally cool
[02:00] <bodie_> oops, that was 4...
[02:00] <thumper> davecheney: cool, ta
[02:01] <perrito666> cool email wallyworld, thanks for taking the time
[02:01] <wallyworld> np
[02:07] <bodie_> so I'm taking that as a "no, your PR is going to sit in limbo until I'm the next on call reviewer, and/or you're on my ignore list now"
[02:07] <perrito666> I really need to go live a few weeks to the countries where you guys are from so I get to learn local english
[02:08]  * perrito666 feels like the hispanic comic relief on an adam sandler movie half the time
[02:08] <bodie_> LMAO
[02:09] <bodie_> perrito666, you're cool by me, I never notice your accent ;)
[02:10] <perrito666> bodie_: trust me I sound terrible when speaking for long periods of time, I am really bad at managing my air and I get really exasperated when trying to speak in english at the same speed than spanish
[02:10] <davecheney> bodie_: nope, that's not it
[02:10] <davecheney> i was pulled off onto another crash this morning
[02:10] <davecheney> the OCR reviewer rotates daily to avoid anyone being on the critical path
[02:11] <bodie_> I just don't like feeling like I'm talking into thin air at timse
[02:12] <bodie_> but I understand you guys are totally slammed
[02:12] <bodie_> just frustrating trying to push our stuff through when nobody will respond to a thread because someone who's only online when it's 10pm where I live owns the review
[02:13] <bodie_> perrito666, english was invented for people who mutter under their breath :P
[02:13] <davecheney> bodie_: i think you need to raise that with alexis
[02:13] <bodie_> yeah, I did this morning
[02:13] <davecheney> bodie_: i've said the same to her in the past
[02:13] <davecheney> i'm not sure what I can do
[02:13] <bodie_> davecheney, she and TheMue both said they wanted the response on that thread to come from you
[02:14] <davecheney> nah, that can't be how it works
[02:14] <bodie_> you could just say something like "I'm content for someone else to look at this" or "bodie_, I'll be able to review it next thursday" or something other than... ya know, nothing
[02:14] <davecheney> he told me that martin is reponsible for landing your stuff
[02:14] <davecheney> so, all i know is eveyrone knows a different story
[02:14] <bodie_> TheMue has replaced martin for working with the actions team
[02:15] <bodie_> I'm just trying to make things work, man :)
[02:15] <bodie_> if we have a blocker, I need to figure out how to get around it
[02:15] <davecheney> bodie_: yup, i hear you
[02:16] <perrito666> ok, lets do this I have a decent overlap with all of you guys, mostly because I have no life and I hardly sleep, bodie_ could you privmsg me with a brief on what you need? davecheney could you tell me what alexis told you so I can try to unlock it tomorrow?
[02:16] <bodie_> maybe you could either respond to that comment, or make a mention to them in #jujuskunkworks, either of which will take no more than 20 seconds
[02:16] <bodie_> that way they won't feel like you own the thread
[02:17] <bodie_> or, I can pass it along tomorrow, or I can wait for your response when you're able
[02:17] <bodie_> I just want to know what to expect so I can be effective here
[02:23] <bodie_> lol
[02:47] <davecheney> thumper: good news
[02:47] <davecheney> upstream have a fix for the gccgo crash
[02:47] <davecheney> now comes the hard part ...
[02:47] <thumper> getting it into ubuntu?
[02:47] <davecheney> :P
[03:08] <davecheney> thumper: https://code.google.com/p/gofrontend/source/detail?r=75739d377426
[03:08] <davecheney> we have a fix
[03:08] <davecheney> what's the best way to report this ?
[03:08] <davecheney> an issue on ubuntu/gccgo ?
[03:09] <thumper> I guess
[03:09] <davecheney> right-o
[03:09] <davecheney> i'll try that
[03:10] <davecheney> might need you and sinzui to ride shotgun on this one
[03:10] <thumper> who is in charge of gccgo?
[03:10] <thumper> inside canonical?
[03:11] <davecheney> dokku maybe ?
[03:12] <davecheney> the lack of an owner probably explains why the last fix hasn't landed for 5 months
[03:32] <mwhudson> thumper, davecheney: i have some gccgo patches i want in ubuntu too
[03:32] <mwhudson> thumper, davecheney: i think the official process is "hound doko relentlessly"
[03:32] <thumper> mwhudson: I suggest we push through alexisb to pat
[03:32] <thumper> mwhudson: who does doku report to?
[03:32] <mwhudson> thumper: slangasek i think
[03:33] <mwhudson> btw, there is already a newer gccgo in trusty-proposed
[03:33] <mwhudson> have you guys tried that?
[03:34] <mwhudson> (it obviously doesn't fix this problem that just got fixed, but doko will probably ask...)
[03:34] <davecheney> mwhudson: thumper `https://bugs.launchpad.net/ubuntu/+source/gccgo-4.9/+bug/1362906
[03:34] <mup> Bug #1362906:  internal compiler error: in comparison, at go/gofrontend/expressions.cc:6508 <gccgo-4.9 (Ubuntu):New> <https://launchpad.net/bugs/1362906>
[03:35] <davecheney> mwhudson: nope, wasn't fixed until 30 minutes ago
[03:35] <davecheney> mwhudson: pls mail thumper and I with your bug numbers
[03:35] <davecheney> i'll add mine
[03:35] <davecheney> and it'll be another thing on alexisb 's plate
[03:35] <mwhudson> https://bugs.launchpad.net/ubuntu/+source/gcc-4.9/+bug/1361940
[03:36] <mup> Bug #1361940: patches for cgo on arm64 <patch> <gcc-4.9 (Ubuntu):New> <https://launchpad.net/bugs/1361940>
[03:36] <mwhudson> davecheney: this is all for cgo though, that probably shouldn't be on alexisb's plate?
[03:36] <davecheney> if it doesn't affect juju or other canonilcal developed products
[03:36] <davecheney> probably not
[03:37] <mwhudson> docker docker docker
[03:39] <davecheney> docker docker docker malkovich docker
[04:08] <ericsnow> axw: do you mind if I add another 10 files and 500 lines to that backups PR <wink>
[04:09] <bodie_> lol
[04:09] <ericsnow> axw: but seriously, thanks (to you, davecheney, and waigani) for the reviews
[04:10] <ericsnow> the rest of my backups patches are *much* smaller
[04:11] <waigani> ericsnow: welcome
[04:14] <ericsnow> who controls the juju account on github?
[04:15] <axw> ericsnow: :)
[04:15] <axw> nps
[04:16] <axw> ericsnow: all the team leads are owners I believe
[04:16] <ericsnow> axw: okay
[04:16] <ericsnow> axw: I'll bug nate about it tomorrow
[04:17] <ericsnow> axw: nps?
[04:26] <axw> ericsnow: no problems
[04:26] <ericsnow> axw: I figured as much
[04:46] <wallyworld> axw: thanks, i was wondering about wording of that message. i wanted to explicitly call out that there could be an instance running; but as you say, i think mentioning destroy --force will be enough
[04:47] <axw> cool
[06:25] <wallyworld> axw: i think testing failure bug 1202039 is either fixed, or will be with your current work?
[06:25] <mup> Bug #1202039: Provider tests are architecture-specific <arm64> <feature> <i386> <ppc64el> <test-failure> <juju-core:Triaged> <https://launchpad.net/bugs/1202039>
[06:27] <axw> wallyworld: I think they're already working on i386 on ppc64 right? so it's fixed already?
[06:28] <axw> CI runs the tests on i386 and ppc64 I'm pretty sure
[06:28] <wallyworld> axw: yeah, that was my thinking too, i'll mark as fix committed
[06:28] <wallyworld> just wanted to double check
[07:36] <mattyw> morning all
[08:00] <TheMue> morning
[08:58] <TheMue> dimitern: ping
[09:00] <dimitern> TheMue, hey
[09:01] <TheMue> dimitern: as I as OCR cannot review myself would you mind to take a look at https://github.com/juju/juju/pull/626 ;)
[09:01] <TheMue> dimitern: and morning btw :D
[09:01] <dimitern> TheMue, sure :)
[09:02] <TheMue> dimitern: thx
[09:04] <axw> wallyworld: just noticed I cannot edit the bugs doc
[09:04] <axw> would like to strike out a line
[09:34] <wwitzel3> fwereade: ping
[09:46] <wwitzel3> wallyworld: got time to chat about a test I want to try and fix?
[09:47] <rogpeppe1> anyone know how to get a list of all charms? i used to use http://manage.jujucharms.com/api/2/charms?text= but it doesn't seem to produce many results now
[09:51] <wwitzel3> rogpeppe1: manage.jujucharms.com is still how I know how to get a list of them
[09:51] <rogpeppe1> wwitzel3: do you have a better URL than that?
[09:51] <rogpeppe1> wwitzel3: i only get about 34 out of that
[09:52] <wwitzel3> rogpeppe1: sounds like you're getting trusty only
[09:52] <rogpeppe1> wwitzel3: actually 20 only
[09:53] <wwitzel3> oh .. hrmm
[09:53] <wwitzel3> there are 38 charms for trusty .. so if you're only getting 20
[09:53] <rogpeppe1> wwitzel3: it looks like there's a limit kicking in, but adding limit=100, for example, doesn't give me any more
[09:54] <rogpeppe1> wwitzel3: have you actually got a current-ish list, by any chance?
[09:54] <wwitzel3> rogpeppe1: https://manage.jujucharms.com/charms/trusty
[09:55] <wwitzel3> rogpeppe1: that's the HTML version if that helps anyway
[09:55] <rogpeppe1> wwitzel3: it's a hassle (i'll have to manually scrape the html) but better than nowt
[09:56] <wwitzel3> rogpeppe1: sadly it doesn't do an ajax request (that I can see) to get the results
[09:56] <wwitzel3> rogpeppe1: was hoping it might be using the API to render that list
[09:56] <rogpeppe1> wwitzel3: yeah
[09:57] <rogpeppe1> wwitzel3: it's annoying because it *did* work once. i even wrote a little program to automate getting all charms.
[10:04] <wwitzel3> rogpeppe1: yeah, according to the lp repo for that API it hasn't changed since 2013-10
[10:05] <wwitzel3> rogpeppe1: but something must have :/
[10:06] <rogpeppe1> wwitzel3: well, the last-modified time on my getcharms.go file is 2013-08, so the 2013-10 change might've broken it i guess
[10:06] <rogpeppe1> wwitzel3: which lp repo is it, BTW?
[10:08] <wwitzel3> rogpeppe1: also doest his give you better results http://manage.jujucharms.com/api/3/search ? that isn't techincally an unstable API, but if 2 stopped working for you
[10:08] <wwitzel3> rogpeppe1: https://code.launchpad.net/charmworld
[10:08] <rogpeppe1> wwitzel3: i only see 41 results there
[10:09] <rogpeppe1> wwitzel3: (assessed by searching for "files": in the JSON :-))
[10:23] <wwitzel3> rogpeppe1: http://bazaar.launchpad.net/~juju-gui-bot/charmworld/trunk/view/head:/charmworld/search.py#L141
[10:24] <wwitzel3> rogpeppe1: I feel like someone else was running in to this 20 limit thing recently
[10:25] <rogpeppe1> wwitzel3: i tried using limit=9999
[10:32] <dimitern> TheMue, reviewed
[10:32] <dimitern> TheMue, ping me if something is not clear please :)
[10:32] <TheMue> dimitern: yeah, seen a lot of it (and commented some back). thanks a lot, will mark it as wip and continue
[10:33] <dimitern> TheMue, cheers!
[10:33] <wwitzel3> rogpeppe1: yeah, sorry I couldn't get it working, the code itself seems to respect limit, but then overrides it somewhere with that default
[10:35] <rogpeppe1> wwitzel3: yeah, i couldn't see where that happens.
[10:35] <rogpeppe1> wwitzel3: and the limit doesn't always appear to be 20 though
[10:35] <dimitern> TheMue, about the dummy provider running "unsafe" networker - it certainly makes sense to use "safe" by default, but there should be a way to test the "unsafe" one if needed - we can isolate what the networker does from the system (config, interfaces), so it's fine
[10:36] <TheMue> dimitern: is it ok to do this in a second, follow-up PR? this one already grows and grows. ;)
[10:37] <wwitzel3> rogpeppe1: yeah, if you leave it off all together, it seems to be 41
[10:38] <dimitern> TheMue, i'm not saying to add more tests, just to tweak the dummy provider a bit so RequiresSafeNetworker can return true or false, just by adding a useSafeNetworker bool in the dummy provider struct and a SetUseSafeNetworker(bool) method on the dummy provider to set that flag
[10:39] <dimitern> TheMue, it's like 5-6 more lines :)
[10:39] <rogpeppe1> wwitzel3: actually no, with a more accurate count, it always seems to be 20
[10:40] <rogpeppe1> wwitzel3: (counted with http://paste.ubuntu.com/8177535/)
[10:41] <TheMue> dimitern: ah, ok, haven't been sure that in this case the "unsafe" (/me still not really happy with that term) doesn't make any harm to the own system
[10:46] <dimitern> TheMue, should we do a quick standup?
[10:46] <TheMue> dimitern: yes, already comming
[12:37] <perrito666> wallyworld: is there any assigning system for the tests that need rework?
[12:38] <wallyworld> perrito666: raise a bug and assign yourself to it. cross out the item in the doc so others can see there that that one is taken
[12:39] <wallyworld> many failing tests already have bugs
[12:39] <wallyworld> so if you go to raise a bug, it should tell you if there is one already
[12:39] <wallyworld> include the test name in the bug description
[12:40] <wallyworld> so that it has a better chance of finding any dupe
[12:41] <perrito666> I guess that created bugs are more important than the ones not yet there
[12:42] <wallyworld> no necessarily
[12:42] <wallyworld> there's failures that people haven't created bugs for
[12:42] <wallyworld> and i found today that bugs i looked at had already been fixed
[12:42] <wallyworld> but the bugs were still open
[12:43] <wallyworld> so pick a failure that you think you can fix - all fixes are good
[12:43] <katco> wallyworld: interesting, i can't add an event to the team calendar =/
[12:44] <wallyworld> katco: that's no good, that means you are not allowed to have the holiday
[12:44] <katco> wallyworld: you are  afunny one ;)
[12:44] <wallyworld> oh, you think i was joking
[12:44] <wallyworld> :-P
[12:44] <katco> wallyworld: we're planning on going to our botanical garden's japanese festival :)
[12:44] <wallyworld> cool, sounds fun
[12:44] <katco> oldest and largest in the country!
[12:45] <wallyworld> well, that's something i didn;t know
[12:45] <katco> http://www.missouribotanicalgarden.org/things-to-do/events/signature-events/japanese-festival.aspx
[12:45] <katco> oops, says "one of" the oldest/largest
[12:46] <wallyworld> ah, you americans always exaggerate
[12:46] <wallyworld> world series baseball = usa plus canada
[12:46] <perrito666> I am amazed on how US people actually measure things in order to be able to have "the largest" for almost anything :p
[12:46] <katco> hey we are THE BEST at exaggerating. no one is better, anywhere.
[12:46] <perrito666> wallyworld: its like in columbus times "the known world"
[12:47] <katco> and me? i'm the most humble person alive.
[12:47] <wallyworld> you too?
[12:47] <wallyworld> we can't both be
[12:47] <katco> hey i'm WAY more humble
[12:47] <wallyworld> of course
[12:47] <katco> lol
[12:47] <wallyworld> you have reason to be
[12:47] <katco> LOL
[12:47] <katco> so nice to work with people who appreciate a dry sense of humor :)
[12:47] <wallyworld> don't know what you mean
[12:48] <perrito666> lets say each of you is the humblest in his/her timezone
[12:48] <katco> what did you ask me yesterday? have i told you today?
[12:48] <wallyworld> lol
[12:48] <wallyworld> no you haven't
[12:48] <katco> fark off! haha
[12:48] <wallyworld> :-( you hate me
[12:49] <perrito666> wallyworld: dont be like that, we all hate you, its because you are a manager
[12:49] <wallyworld> perrito666: oh, that's a low blow
[12:49] <katco> haha
[12:50] <wallyworld> way to hurt a guy
[12:50] <katco> no, wallyworld, i definitely don't hate you. i quite like working with you, axw, and everyone i've had the pleasure to work with so far
[12:50] <katco> with the exception of mgz_, traitor!
[12:50] <wallyworld> katco:  i was joking :-)
[12:51] <perrito666> katco: nahh, dont piss off the guy that can lock your commits
[12:51] <mgz_> ;_;
[12:51] <katco> haha
[12:51] <katco> mgz_: i kid, i kid!
[12:52]  * perrito666 accidentally spills yerba all over his kb
[12:56] <wallyworld> wwitzel3: sorry i missed your ping before, i was at soccer. did you still have a question?
[13:05] <katco> need to cycle my server, brb
[13:12] <perrito666> wallyworld: I cant write the doc either
[13:13] <wallyworld> perrito666: i changed the permissions, have you tried refeshing?
[13:14] <perrito666> tx
[13:33] <natefinch> perrito666: you around?
[13:33] <wwitzel3> wallyworld: yeah, I was going to try to fixup a test from the list
[13:33] <perrito666> natefinch: I am I though you where not here
[13:33] <natefinch> perrito666: yeah, sorry, busy morning.  I'm in the hangout on the calendar event
[13:34] <wallyworld> wwitzel3: ok, or it can be any other one you may have come across that isn't there
[13:34] <wallyworld> or it can be improving the cmd/juju tests to use mocks
[13:35] <mattyw_> dimitern, do you have 10 minutes spare for a quick question, possible via a hangout
[13:36] <dimitern> mattyw_, sure, just give me 15m
[13:44] <mattyw> dimitern, great, just ping when it's a good time
[13:52] <perrito666> natefinch: interesting fact, cold medicine hinders my already low spoken english skills :p
[13:53] <natefinch> perrito666: haha
[13:53] <natefinch> perrito666: btw, it's "beats around the bush"
[13:53] <perrito666> ahh well close enough
[13:53] <natefinch> yeah, I knew what you meant
[13:54] <perrito666> natefinch: it is hard to translate those, the spanish one is nothing like it
[13:54] <natefinch> yeah... those kind of sayings are kind of crazy, because the actual words don't make any sense
[13:54] <perrito666> literal translation of spanish one is "dodges the lump"
[13:54] <perrito666> to give you an idea
[13:54] <natefinch> rofl
[13:54] <perrito666> of how far it is
[13:58] <perrito666> omg, even though its not ergonomic I must say, thinkpad keyboard that emulates the one in the laptop is awesome
[13:58] <perrito666> I wish I had found a newer one, I actually got one ibm branded :p
[13:59] <natefinch> haha
[14:01] <perrito666> I also with it was bt, this one actually has a 2.5M cable :p
[14:14] <katco> is there a place we could put consts representing config key-strings? seems silly and error-prone to repeat them all over the codebase.
[14:16] <dimitern> katco, how about in environs/config ?
[14:17] <katco> dimitern: that seems like a fine place
[14:17] <katco> dimitern: could probably even put it in environs/config/config.go
[14:18] <katco> dimitern: is this a small enough shift that i can just introduce this concept? :)
[14:19] <dimitern> katco, go for it :)
[14:19] <katco> dimitern: awesome, thanks for the input! :)
[14:20] <marcoceppi_> hook execution question
[14:21] <marcoceppi_> do hooks in a service group execute serially or is that only on the unit level?
[14:23] <katco> boy documenting all public values seems silly sometimes
[14:23] <katco> 	// ProvisionerHarvestModeKey stores the key for this setting.
[14:23] <katco> 	ProvisionerHarvestModeKey = "provisioner-harvest-mode"
[14:23] <perrito666> ahh tautological comments
[14:24] <perrito666> katco: you could expand on what exactly does that key means
[14:24] <katco> i know... i understand wanting good documentation, but an absolute rule doesn't make sense
[14:24] <katco> perrito666: that's already covered in the actual setting
[14:24] <katco> perrito666: this is literally the key to get the setting in the map
[14:25] <katco> i especially dislike repeating the name of the thing in the comment. hungarian notation all over again.
[14:47] <bac> hi marcoceppi_
[14:47] <marcoceppi_> hey bac
[14:48] <bac> marcoceppi_: lazyPower_ has filed a bug regarding supporting tags as a constraint for MAAS.  you've seen it i think.
[14:48] <marcoceppi_> bac: yup
[14:49] <bac> marcoceppi_: in the definition tags are a comma-separated list.  until now we've marked bundles with constraints listed as a comma-separated as deprecated.  clean implementation of tags will require us to reject those deprecated bundles.
[14:50] <bac> i.e. constraints: 'k1=v1 k2=v2 tags=a,b,c' is valid bug 'k1=v1,k2=v2' is not
[14:51] <marcoceppi_> bac: the error we're getting is tags is not a valid constraint
[14:51]  * marcoceppi_ checks the bundle file
[14:52] <bac> marcoceppi_: yes.  i'm trying to fix that.
[14:52] <marcoceppi_> oh, I see
[14:52] <marcoceppi_> one sec
[14:52] <bac> marcoceppi_: to properly support tags will require not allowing comma-separation in the other constraint key value pairs
[14:53] <marcoceppi_> bac: maybe constraints are all wrong
[14:53] <bac> is it possible to run charm-proof against the set of all bundles to see if it is even an issue?
[14:53] <marcoceppi_> maybe they should be a YAML list
[14:53] <marcoceppi_> bac: I can download them all and see
[14:53] <bac> marcoceppi_: that would be great
[14:53] <marcoceppi_> constraints:
[14:53] <marcoceppi_>   - key: v1
[14:53] <marcoceppi_>   - key1: v2
[14:53] <marcoceppi_> hazmat: ^^
[14:54] <bac> marcoceppi_: that may be the best long term solution but i thought you needed a fix now.
[14:55] <marcoceppi_> well, I'll look at all the bundles we have
[14:55] <marcoceppi_> and see what each is doing for constraints
[14:55] <bac> changing proof rules already touches way too many parts...
[15:01] <TheMue> oh, bugfix by accident, nice. field of a struct has been prepared but never been set. now I deeded it.
[15:01] <perrito666> TheMue: lol
[15:02] <TheMue> perrito666: yeah, gave me a panic during tests :D
[15:04] <perrito666> TheMue: which one (bug)? just in case
[15:05] <TheMue> perrito666: it never has been filed. it has been a field for checking the authorization of reading a machine
[15:06] <TheMue> perrito666: the function (it's a function type) has been created, but the field never set
[15:06] <TheMue> perrito666: and I now wanted to check if I can read => *bang*
[15:06] <perrito666> ouch, I wonder if its there on the bugs ian added to the doc which are not all reported
[15:09] <TheMue> have to check
[15:57] <natefinch> Anyone up for reviewing my log rotation change?  It's pretty small: https://github.com/juju/juju/pull/512
[16:18] <mattyw> natefinch, I'll take a look
[16:25] <perrito666> mattyw: what is a vanity url?
[16:26] <mattyw> github.com/perrito666/projectname
[16:26] <mattyw> perrito666, A url with someones name in it
[16:26] <mattyw> perrito666, I'm not happy with the name vanity url either :)
[16:27] <mattyw> but that's what they're called in the mailing list
[16:27] <mattyw> and who am I to go against consensus?
[16:27] <perrito666> I would think a vanity url is something like beautifulpopleonly.com/perrito666/thewonderfulwonderfulman/
[16:27] <perrito666> :p
[16:28] <perrito666> now that is a vanity url
[16:32] <jcw4> perrito666: can you get me one like that too?
[16:32] <perrito666> :p
[16:32] <mattyw> perrito666, I'd be happy with that one - at least make it obvious
[16:33] <mattyw> perrito666, what I'm happy and not happy with is inconsistent and changes daily
[16:33] <perrito666> we might want to rething our mentor policy to not be 12 hours distance with our reviewers
[16:34] <jcw4> perrito666: quit complaining... you're on here 24/7 anyway
[16:34] <perrito666> jcw4: Id say around 20
[16:34] <perrito666> but yes
[16:34] <jcw4> :)
[16:35] <perrito666> its the corners of the week which are a problem, my mentor will not be here until sunday (my sunday)
[16:35] <jcw4> yeah, that's a bummer.
[16:35] <mattyw> perrito666, how many corners does a week have? I thought they were circular?
[16:35] <mattyw> they feel circular sometimes
[16:35] <jcw4> I have to say though, working with folks all over the globe make it more than worth it for me :)
[16:36] <perrito666> if weeks where circular in two days it would be last monday
[16:36] <jcw4> spiral then
[16:37] <perrito666> uff, this bug is a heavy thing to pinpoint, I am pretty sure I know what is going on, I am not so sure on the how part
[16:38] <bodie_> my weeks have no corners, only hills and valleys
[16:43] <natefinch> perrito666, mattyw: a vanity url, in the context of Go is a url which is not the same as the url to the repo.  So, for example, the package that gets imported when you do import "labix.org/v2/mgo"  is actually hosted on launchpad, so the real url is launchpad.net/mgo (or something like that).
[16:46] <mattyw> natefinch, I didn't know that - thanks very much
[16:46] <natefinch> mattyw, perrito666: Go supports this through the go tool which, if it doesn't recognize the url of the import path, will fall back to trying to get a web page from that URL, and it'll look for a header that looks something like <meta name="go-import" content="labix.org/v2/mgo bzr https://launchpad.net/mgo">
[16:47] <natefinch> It's pretty nice because it means that if you control the domain, you can move your code from one host to another and not break anyone else's code that imports using the vanity url.
[16:51] <mattyw> natefinch, now I know the proper definition of vanity url my comment probably makes no sense
[16:52] <mattyw> natefinch, I was just wondering if accepting a number of dependencies that had team members names in was ok
[16:53] <mattyw> has anyone else seen this error using the local provider on tip? http://paste.ubuntu.com/8180167/
[16:58] <perrito666> mattyw: are you sure your code is up to date?
[16:59] <mattyw> perrito666, it was tip
[17:00] <perrito666> mattyw: looks a lot like someone made one of those automatic changes on the code and forgot to change a return somewhere
[17:01] <mattyw> perrito666, I was trying to recreate this https://bugs.launchpad.net/juju-core/+bug/1363143 but I can't even get that far
[17:01] <mup> Bug #1363143: local lxc deployments fail to create machines <ci> <local-provider> <lxc> <precise> <regression> <juju-core:Triaged> <https://launchpad.net/bugs/1363143>
[17:01] <perrito666> mattyw: checking
[17:01] <alexisb> perrito666, the mentorship matching is my fault, we did the best we could but most of the mature skills on the team are in +12 timezone, you are one of the few with potential overlap
[17:02] <perrito666> alexisb: sorry didnt meant to make you feel bad
[17:02] <perrito666> mattyw: and this blows running bootstrap right?
[17:02] <mattyw> perrito666, yeah
[17:02]  * jcw4 wonders about relocating to UTC+12
[17:02] <jcw4> ;
[17:02] <jcw4> )
[17:02] <jcw4> ;) even
[17:02] <perrito666> mattyw: rm /pkg/* on your go path and then recompile
[17:03] <alexisb> yes perrito666 I feel terrible, I may never recover from your slam ;)
[17:03] <wwitzel3> lol
[17:03] <natefinch> mattyw: btw, your comment is actually correct - gopkg.in/natefinch/lumberjack.v2 *is* a vanity url.  It redirects to github.com/natefinch/lumberjack at the v2 branch.
[17:03] <perrito666> this will so cost me beer on the next sprint
[17:03] <alexisb> perrito666, you should feel tremendously guilty on my behalf
[17:03] <alexisb> lol totally
[17:03] <alexisb> o wait
[17:04] <alexisb> not thumber buys the beer remember
[17:04] <mattyw> perrito666, same problem
[17:04] <alexisb> no, thumber buys the beer, remember?
[17:04] <natefinch> close
[17:04] <natefinch> one more try?
[17:04] <perrito666> we can levenstein from there
[17:04] <natefinch> lol
[17:04] <mattyw> perrito666, I'm about to EOD to be honest, maybe we can take a look at it on monday?
[17:04] <alexisb> :)
[17:05] <perrito666> mattyw: sure, ping me, honestly looks like a red herring
[17:05] <natefinch> mattyw: btw, the reason it's not under juju is because it's my personal code written during off hours.  Otherwise, yes, absolutely, anything written during hours is Canonical's and belongs under github.com/juju
[17:06] <mattyw> perrito666, ok cool
[17:06] <natefinch> mattyw: in theory, the code is MIT licensed and anyone could just fork my repo to put the code under github.com/juju
[17:06] <natefinch> but some might consider that rude
[17:07] <perrito666> that would remind me the early days of google chrome
[17:07] <perrito666> absolutely all of the deps from the project had been imported into the project svn
[17:07] <perrito666> from here it was a 3 or 4 hs checkout
[17:08] <natefinch> ouch
[17:08] <natefinch> good old SVN
[17:08] <mattyw> perrito666, I like red herrings
[17:08] <natefinch> tasty with a little lemon and butter
[17:08] <mattyw> perrito666, I might be able to work it out for myself once I've slept :)
[17:09] <perrito666> I only ate herring once in my life, in amsterdam, there was this kiosk next to the canal that made raw herring sandwitches
[17:09] <perrito666> mattyw: try coffee, has pretty much the same effect
[17:10] <mattyw> natefinch, my point is probably pretty academic as all the code is on github anyway
[17:10] <natefinch> mattyw: it's a valid point... if it were canonical's code, it should be represented as such, and under canonical's control.  But it's not, so it's not :)
[17:11] <mattyw> natefinch, I just don't want to end up in the situation where core relies on changes landing in your project but you want to take the project in another direction
[17:11] <mattyw> natefinch, but we can cross that bridge if it comes - and there are many ways of fixing it
[17:11] <natefinch> mattyw: that would be a perfectly valid reason to fork the project, and like I said, it's MIT, you can just do that.
[17:11] <natefinch> mattyw: exactly
[17:11] <natefinch> mattyw: especially for a package that's used in exactly one file... it's trivial to change.
[17:17] <mattyw> folks - it's time for me to call it a day
[17:18] <mattyw> have a good weekend all and I'll speak to you on monday
[17:19] <wwitzel3> see ya mattyw
[17:24] <TheMue> so, off for today, have a nice weekend
[17:26] <perrito666> TheMue: likewise
[17:58] <perrito666> I might be wrong here, but does this sounds like a huge isolation bug to anyone else? https://github.com/juju/juju/blob/master/juju/testing/conn.go#L426
[18:01] <natefinch> perrito666: certainly seems like something that should be exposed by MgoServer not something the suite is figuring out.
[18:02] <perrito666> natefinch: I am a bit worried about the shared nature of the mongo server also
[18:02] <perrito666> https://github.com/juju/testing/blob/master/mgo.go#L35
[18:02] <natefinch> perrito666: yeah
[18:02] <natefinch> perrito666: that very well could be why some of our tests are failing... they're running at the same time and then the streams cross and boom
[18:03] <perrito666> The failure I am looking at (only have a few logs) seems to be triggered by the test assuming the server is alive and trying to connect but it gets auth fails, which makes me think its hitting someone else's server
[18:03] <natefinch> hmm could be
[18:06] <natefinch> perrito666: this may actually not be that bad
[18:07] <natefinch> perrito666: each package runs tests as an independent process
[18:07] <ericsnow> anyone have a little time to give a follow-up review?
[18:07] <ericsnow> https://github.com/juju/juju/pull/606
[18:10] <natefinch> ericsnow: is there a reason you need to hide the fields of the archive?
[18:11] <ericsnow> natefinch: I guess not...I've been trying to make things private by default and avoid exposing them if possible
[18:12] <natefinch> ericsnow: for things that are just data... don't bother.... and this is definitely just data
[18:12] <natefinch> ericsnow: I'm used to the same pattern from past experience, too.  But honestly, if someone wants to change the filename after the make a new Archive.... who cares?
[18:13] <natefinch> it simplifies the code a bunch not to hide the values - you don't need a constructor and don't need the getters
[18:14] <ericsnow> natefinch: Archive is one of the types that I actually think might be used outside of juju, so I wanted to be extra careful there to keep the interface clear
[18:14] <ericsnow> natefinch: those fields are just implementation details
[18:15] <ericsnow> natefinch: it's the getters that expose the structure of the archive file
[18:17] <natefinch> ericsnow: it's not really an implementation detail if you're just returning what you got passed in.  It's just data
[18:18] <natefinch> ericsnow: unless you need the functions to fulfill an interface
[18:19] <ericsnow> natefinch: yeah, the fields are not part of the interface I care about here.
[18:20] <ericsnow> natefinch: so you make a good point
[18:23] <natefinch> ericsnow: it's not that the way you have it is wrong, it's just somewhat more complicated that necessary.  Not that it's actually complicated at all, of course - the functions are trivial.
[18:25] <ericsnow> natefinch: yeah, I've run into that several times and it's always tempting to just make fields/types public because it simplifies things to much
[18:25] <ericsnow> natefinch: in some cases I did so because keeping them private wasn't worth it
[18:27] <natefinch> ericsnow: I've come to the conclusion that the only time I need to make fields private is when there are interactions between different fields that would be invalidated if the fields were updated after construction.... like if you were generating the DBDumpDir just one time in the constructor.
[18:27] <ericsnow> natefinch: yeah, the caching case
[18:31] <bac> ping lazyPower_
[18:31] <lazyPower_> hey bac
[18:31] <lazyPower_> sorry i saw teh pings while i was out and about checking on my phone
[18:31] <lazyPower_> you're looking for an example bundle with tags as constraints yes?
[18:31] <bac> yes i am
[18:31] <bac> lazyPower_: any that are blocked
[18:33] <lazyPower_> bac: i edited the bundles for promulgation
[18:34] <bac> lazyPower_: so you took the tags out?
[18:35] <lazyPower_> sure did. I pointed the resource to the charm store, and removed the tags as constraints - and left a note in the bzr log + MP that i had done so for promulgation.
[18:35] <lazyPower_> https://code.launchpad.net/~kirkland/charms/bundles/transcode-cluster/bundle
[18:36] <lazyPower_> this was the hottest culprit, as the orange-box bundle has tags to denote NFS and TRANSCODE should go on hardware / physical nodes
[18:36] <bac> lazyPower_: ok.  i mean it is trivial for me to come up with test bundles but i thought i'd run my charm proof against the real ones.
[18:36] <lazyPower_> bac: well that submission would be a great one to look at
[18:37] <bac> oh, the famous transcoder...
[18:37] <lazyPower_> famous? :D
[18:40] <bac> lazyPower_: minimally.  :)
[18:40] <bac> lazyPower_: that bundle is actually trivial wrt constraints.
[18:41] <lazyPower_> indeed. very minor set of constraints - it was throwing the charmworld proofer through loops though
[18:41] <bac> lazyPower_: i've got a charmworldlib fix that i'll put up for review soon.  it'll require a small code change to charmworld and a redeploy, which cannot happen until monday
[18:41] <lazyPower_> when i run charm proof --offline  it passes.
[18:41] <bac> lazyPower_: you're not currently blocked so monday should be fine, correct?
[18:55] <perrito666> natefinch: yes, but that still leaves a whole package to race against
[18:55] <perrito666> I mean, you cannot guarantee that a mgo server died properly in the last test
[18:57] <natefinch> perrito666: well, the tests are run in series, not in parallel, so there's no concern as long as the setup and teardown do the right thing
[19:08] <perrito666> natefinch: I believe the tear down might be failing
[19:09] <perrito666> this is a heisembug so I might try to make sure this is isolated and hope for the best
[19:29] <bodie_> https://github.com/juju/juju/pull/617 should be done now
[19:31] <bac> hi marcoceppi_, when you get a chance could you look at this charmworldlib branch for the constraints issue: https://codereview.appspot.com/131710044/ ?
[19:39] <marcoceppi_> bac: I wonder if there's a limit to characters that can be in a tag/constraint
[19:39] <marcoceppi_> nvm
[19:39] <marcoceppi_> ignore what I'm saying
[19:39]  * bac ignores you
[19:40]  * bac listens again
[19:50] <natefinch> ericsnow: is backups.create actually used anywhere?
[19:50] <ericsnow> natefinch: not until the next patch
[19:50] <natefinch> ericsnow: ahh ok
[19:51] <ericsnow> natefinch: I linked it in the PR summary
[19:57] <natefinch> ericsnow: aww... reading?! :/
[20:55] <perrito666> aghhh I believe I foundiiiiiittttt
[20:55] <natefinch> huzzah!
[20:59] <perrito666> ish, someone might be re-setting the mgoserver but I cannot find who yet :p