[00:03] cmars: thanks, it has been fun, hopefully the finishing touches won't take too long [00:19] davecheney: hey sorry, wifey and kiddo are home [00:19] davecheney: but rq, i was wondering if you had time to look at my PR again, i was wondering why the name feature_tests was dangerous? [00:24] katco: _test as a package name is reserved for external tests [00:24] the tool will probably let you get away with it [00:24] but it'll bite us (just like testing in package main did) [00:24] davecheney: these are external tests though, correct? [00:24] also, package names should be alphas only [00:24] pacakge foo_test is synthesised by the test runner [00:25] so it's a good idea to stay away from that special cased behaviour [00:25] davecheney: feature_testS is as well? [00:25] katco: please don't use underscores in package names [00:25] davecheney: do you have a suggestion? [00:25] the name of the package should describe what it does, not its contents [00:26] as a stop gap [00:26] package featuretests [00:26] davecheney: so, something that describes the fact that it runs functional/feature tests [00:26] davecheney: i honestly can't think of anything better than featuretests if you don't want an underscore/hyphen [00:27] i don't have any comment apart from please don't use underscores, or any other punctuation in pacakge names [00:27] and don't call a package _test [00:27] 'cos the go tool will end up barfing and we'll hve to back it out [00:28] davecheney: so if i were to rename the package, you would lgtm? [00:28] katco: i guess so [00:28] does this need to live in the juju repository ? [00:29] davecheney: that depends; i think the intent was for developers to have the option of running it, but by virtue of the fact that it is intended for external tests only, it could certainly be external [00:30] davecheney: sinzui, wallyworld, and axw all gave verbal agreement on name/location [00:30] davecheney: perhaps they could weigh in further [00:33] davecheney: as a point of intellectual curiosity, i still don't understand why the go tool will at some future point reject _test packages [00:38] katco: what about "featuretests". i really don't care about the _ [00:38] katco: _test is a majic package [00:38] it never exists on disk [00:39] is is only syntehsised for external tests by the test runner [00:39] this magic will bite us [00:39] please don't anger it [00:53] davecheney: wallyworld: sorry, brushing the little one's teeth for the first time :) [00:53] \o/ [00:53] davecheney: interesting, look at how it laid it out on disk: > pwd [00:53] /tmp/go-build388828020/github.com/juju/juju/feature_tests/_test/github.com/juju/juju [00:54] feature_tests.test binary is in /tmp/go-build388828020/github.com/juju/juju/feature_tests/_test [00:54] so it looks like it generated a blank _test directory within feature_tests [00:54] katco: i'm not sure what you want me to say [00:54] i gave you my advice [00:55] davecheney: sorry, i thought you'd be interested. i'm not looking for you to say anything. thanks for the input. [00:56] i just know this will blow up [00:56] remembver we have to support Go all the way back to 1.2 shipped in a 3 year old operating system [00:56] the possiblilities of backporting fixes to those old distros, after two bitter years of trying, is exactly 0% [01:02] morning all [01:02] hope everyone had a good holiday [01:02] mattyw: happy new years to you sir! [01:02] mattyw: morning, happy new year [01:13] mattyw: morning! and a Happy 2015 :) [01:14] katco, perrito666 anastasiamac happy new year all [01:16] katco: standup? [01:16] wallyworld: shoot brt [01:18] ohno dying pixel [01:19] perrito666: #nopixelleftbehind [01:19] katco: I discovered today that dell announced my next machine :p now I just need a way to get one of those [01:20] perrito666: sweet :) [01:20] none of you want to visit .ar and save me of customs hell? [01:40] axw: package name has been updated; ready for review whenever you're ready [01:40] katco: cool, looking now === kadams54 is now known as kadams54-away [02:01] katco: sorry to be a pain in the butt, but I have a question before I can LGTM [02:01] axw: not at all, what's up? [02:01] katco: published [02:01] axw: k looking [02:10] bleh i fail at review board [02:17] katco: so my concern is there's nothing stopping unit X from requesting leadership for unit Y, but from your response that's what William wants? what are the access controls *meant* to be? I don't think there's any precedent for a free-for-all [02:18] axw: it's an excellent point. again he brought up several examples wherein the unit getting leadership might not be the unit requesting it. i believe they were hypotheticals. [02:19] axw: i think in general, he is trying to be sensitive to bulk-first architecture, but the cross section of that and leadership is interesting to say the least [02:19] menn0: yep, there was a problem with the creation of the environments, test and patch coming [02:19] menn0: just running all the tests to catch bad tests [02:19] axw: and just for fun i'll throw the fact in that i could have misinterpreted :) [02:20] katco: can that particular change be reverted, and moved to a separate PR that fwereade can take a look at? [02:20] katco: I'm not comfortable with it, but I don't want to block the rest [02:21] axw: i'm due to talk with him tomorrow about leadership, why don't i just bring it up there and see what he has to say [02:21] katco: okey dokey [02:21] thanks [02:21] axw: thanks for the review :) [02:21] nps [02:21] axw: 10 points! [02:25] wwitzel3: FYI, I've updated the branch with the gceapi commit [02:25] wwitzel3: I've broken some tests but wanted to get up what I had before calling it a night [02:26] wwitzel3: we can change the name too [02:26] wwitzel3: I may check back in later, but maybe not :) [02:36] menn0: ok. want me to review when it's ready? [02:37] thumper: ^^ :) [02:37] ericsnow: sounds good, thanks :) [02:38] menn0: http://reviews.vapour.ws/r/690/ [02:38] menn0: full test run is still going [02:38] but I expect it to finish shortly [02:38] (running again) [02:38] first run found one other place in the code where we were doing it wrong [02:39] just finished, all tests pass locally [02:40] * menn0 looks [02:44] thumper: is that factory change supposed to be there? [02:45] thumper: it appears like you intend the type of factory.EnvParams.Owner to change but I don't see that [02:45] thumper: never mind [02:45] thumper: i'm being stoopid [02:45] thumper: ship it [02:45] :) [02:46] ta [02:46] thumper: just a typo in that test name [02:46] fixing [02:48] menn0: and thanks [02:55] axw: about facade version numbering... [02:56] axw: should they start at 0 or 1? [02:56] axw: i have seen a few 0s when I was adding mine... [02:58] ah fark..... [02:58] * thumper headdesks [02:59] thumper: again?.. [02:59] anastasiamac: sorry, was afk [02:59] anastasiamac: new ones should be 1 [02:59] anastasiamac: nothing to do with you [03:00] anastasiamac: 0 is for the facades that existed before we had versioning [03:00] or something like that [03:02] axw: oh! thnx - that makes sense ;) [03:02] thumper: i was not expecting it to do with me [03:03] thumper: but ur desk must be very sturdy with all this stress relief :p === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away [03:24] axw: I'll add this explanation to api/facadeversions.go for future references :0 [03:24] anastasiamac: SGTM, thanks [03:25] anastasiamac: I think jam would be able to clarify the rationale, maybe ask him to review [03:26] jam, ping :) [03:26] axw: thnx ;-) [03:26] a bit early I think [03:26] axw: anastasiamac: I'm up, but currently in leads meeting [03:27] no rest for the wicked/leads :) [03:29] axw: :) or OCRs [03:30] axw: could u plz cast ur eyes on my annotations PR when u get a chance? [03:30] anastasiamac: yup, just finishing up with another [03:31] axw: tyvm :p [03:51] m trying to update dependencies.tv with newer version of names... i get the commit hash, I get the date but what is the date suffixed with? for eg. what's T02:50:22Z? [03:51] axw: ^^? [03:52] anastasiamac: just use godeps to get the correct values [03:52] how? [03:52] anastasiamac: godeps ./... | grep names >> dependencies.tsv [03:52] then fix up by hand [03:52] that's how I do it anyway. there's probably a smarter way [03:54] axw: this (from Nate) works for me: "GOOS=windows godeps -t ./... > dependencies.tsv" [03:54] axw: i know what names commit i want dependencies.tsv to contain [03:55] axw: ericsnow: i know what date to put as well, i just don't get the last charas?.. [03:55] chars even [03:55] anastasiamac: not sure what date you're talking about? [03:55] anastasiamac: just set your names repo to that commit and then let godeps sort it out for you [03:55] anastasiamac: ah, the date at the end of the git lines - that's from a newer version of godeps I believe [03:55] axw: newer versions of godeps put a date along with some of the revisions [03:55] ericsnow: axw: I have updated names repo and need juju dependencies.tsv to have my commit in it [03:56] anastasiamac: with the names repo set to the revision you want, run the command I listed and see what it writes to dependencies.tsv [03:58] ericsnow: this removed all dates [03:58] anastasiamac: I'm guessing your godeps is older then [03:59] anastasiamac: try "go get -u launchpad.net/godeps && go install launchpad.net/godeps" or something like that [04:01] * anastasiamac updating godeps [04:01] ericsnow: thnx :) got it ;-0 [04:01] anastasiamac: :) [04:06] anastasiamac: so pong now, in case there is something you still need clarity on [04:07] jam: just to confirm - axw's understnading of facade versioning: [04:08] jam: initial numbering - 1 for new facades, 0 for facades that existed before versioning [04:08] jam: ? [04:09] night all [04:09] * thumper goes to walk dogs [04:09] night [04:09] anastasiamac: the "0" version is for things that were in 1.18 (pre-versioning), yes [04:10] jam: axw: thnx for clarification. I'll add it as a comment to facadeversions on my brnach :-) [04:11] branch* [04:16] anastasiamac: did you just rename annotator.go to annotations.go, or are there other changes in that file? [04:17] oh it's not renamed at all, just extracted [04:17] axw: there r other changes... like GlobalEnitty [04:17] anastasiamac: I meant annotator.go specifically [04:17] axw: m going to remove annotator.go in the next PR [04:17] oh I see [04:18] axw: also state is now passed in [04:18] axw: current annottor is written with embedding into entities in mind [04:19] axw: my implementation ensures that the new annotations does not need to b embedded [04:19] right [04:19] so I see [04:19] sgtm [04:20] axw: :) [04:22] anastasiamac: reviewed [04:22] * axw disappears for lunch [06:01] fwereade: ping [07:15] anastasiamac, pong [07:15] anastasiamac, getting rid of that embedding sounds like a good thing to me, that always felt like a timebomb [07:16] fwereade: about bulk annotation calls, i've sent u an email. Just thought i'd give u a heads-up ;-) [07:16] anastasiamac, ty [07:16] fwereade: m glad u feel this way about embedding ;) [07:30] axw: r u k to drop annotation bulk call review comments until I get confirmation from intrested parties? I'd love to land what I have so far and address the rest in separate PR... [07:30] axw: pretty plz? :) [07:31] fwereade: ping for status chat [07:31] anastasiamac: not really, because 1.22 is about to go out and then this is set in stone... [07:32] anastasiamac: if you go with my suggestion then everybody should be happy because it can handle any combination in one bulk call [08:30] morning [08:30] morning [08:39] wallyworld_, still around? [08:45] axw: addressed :) when u get a chance, could u plz inspect? :P [08:52] anastasiamac: looking [08:59] anastasiamac: reviewed [09:24] axw: thnx! [09:25] nps [09:58] dimitern: TheMue: I'll be a couple of minutes late to standup [09:58] see you soon, sorry! [09:58] voidspace, ok, no [09:58] np even :) [10:02] dimitern: omw [11:03] dimitern: hi, was away for a bit [11:04] voidspace: i think you're ocr? if you get a chance, i'd love to be able to land http://reviews.vapour.ws/r/692/ [11:04] wallyworld_, hey, I wanted to ask are you willing to help now and then with goamz reviews/maintenance on github, as you've contributed in the past [11:04] sure [11:04] i'm not an expert but happy to help [11:04] wallyworld_, ok, you have an invite already :) [11:05] ok, will look, probably tomorrow, am tired tonight [11:05] wallyworld_, no worries, we'll get community members to help as well [11:05] wallyworld_, sure, have a good one :) [11:05] will do [12:03] wallyworld_: nope, ericsnow and axw [12:03] wallyworld_: I can look in a bit if they don't get to it [12:05] voidspace: ah, oh, sorry, i'm a day out. i can wait till tomorrow [12:09] wallyworld_: looking at it now [12:10] i owe you one, thanks [12:15] katco, wallyworld_ : thank you both for the reviews! I replied to some of your comments and I have some questions, (however I am not sure I understand how review board works with comments). could you please take another look when you have time? [12:15] frankban: will do [12:15] ty [12:28] wallyworld_: so imagemanager and the Client, with DeleteImage method, is pre-existing [12:29] wallyworld_: this is just the command to expose it [12:29] voidspace: yeah, all the api and apiserver stuff landed in a previous branch. this is just to add a juju command it use it [12:29] yep [12:29] wallyworld_: LGTM [12:29] voidspace: you rock, tyvm [12:30] wallyworld_: a disturbing amount of boilerplate required just for TestDeleteImage though [12:30] wallyworld_: the mocking, the patching, the fakes [12:30] wallyworld_: not that I see a better way [12:30] yeah. i find i complain about biolerplate in general with go [12:30] lack of generics etc makes it unavoidable [12:30] wallyworld_: and as a result of it, you don't actually test that the delete command calls the DeleteImage client method [12:31] right [12:31] you check the api method is called - but that is mocked out, so it's not tested [12:31] true, i think you are right, i'm more or less following established patterns [12:32] yeah, effectively we test the layers but don't test through all the boundaries [12:32] we are introducing a featuretests package for integration tests [12:32] cool [12:32] anyway, I added my ShipIt [12:32] the first cab off the rank is katherine's leadership service tests [12:33] so with featuretests, we can add a tests to check the end-end wiring [12:33] great [12:33] so we have a mix of fast *unit* tests and fewer slower intergration tests [12:34] work in progress [12:34] thanks for lgtm, i can now land this to meet the 1.22 deadline [12:42] jam: still around? [13:23] * perrito666 reads a little wtf in a review answer and gets a small brain freeze like sensation [13:48] rick_h_: hey [13:49] jam: howdy, got a few min I can steal sorry for the post-hours? [13:49] rick_h_: calling you now https://plus.google.com/hangouts/_/guyjqtpuoh55ecfsilhje5smdia === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [14:58] wwitzel3: ping [15:02] oh right [15:07] ericsnow: no one is here .. [15:10] fwereade: ping [16:44] jw4: ping === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 [16:57] fwereade: you still around? === kadams54 is now known as kadams54-away [17:21] wallyworld_: I updated my branch to reuse the code in envcmd for the default environment. I'll wait for Tim about naming issues === wwitzel3_ is now known as wwitzel3 [17:46] hello folks [17:46] I was doing a deploy of juju 1.21 beta5 on MaaS and noticed that the --constraints argument is ignored [17:47] looking at: https://github.com/juju/juju/blob/1.21/provider/maas/environ.go#L781 , it seams that the Constraints field is missing [17:47] is this by design? [18:03] ericsnow: wwitzel3 ? [18:03] weekly random time meeting? [18:03] perrito666: omw [18:04] rebooting and then coming (weird unity freeze thing I need to get rid of to use the browser) [18:05] TheMue, are you around for the team meeting? [18:05] cmars: ? [18:08] dimitern: as you're still around... [18:08] dimitern: https://code.launchpad.net/~mfoord/gomaasapi/nodegroupsinterfaces/+merge/245891 [18:08] voidspace, sure, will have a look [18:41] g'night all [18:41] nn voidspace [18:41] voidspace, i'm reviewing your MP and will send a review [18:41] wwitzel3: o/ [18:41] dimitern: cool, thanks [18:41] voidspace, have a good one ;) [18:41] o yeah, there is totally uber in tokyo: [18:41] http://www.bloomberg.com/news/2014-08-05/uber-expands-tokyo-service-to-include-taxi-booking.html [18:42] i forgot to mention for anyone who was interested in the spec: it also generates the interfaces for all of the code to ensure parity [18:42] so what you see in the spec is directly in the codebase [18:42] katco, have you and jam talked about your spec format before? [18:42] if not we should make that happen [18:43] alexisb: i have not [18:44] alexisb: i have something regarding leadership to discuss with him as well. it's low-priority, but we could definitely fill a meeting with the 2 topics [18:45] katco, you should get on his calendar [18:45] alexisb: will do, i'll send him an email [18:45] I would offer to set if up but with your earlier timezone you have a better chance of getting on the calendar solo [18:45] what tz is he? [18:46] why am i asking, it's right here lol [18:47] so 10h difference [18:47] alexisb, yeah, uber started in sofia recently as well [18:48] dimitern, cool [18:48] katco, keep in mind he has his son in the afternoon after school [18:49] alexisb: ok ty for the info [18:49] katco, i can't wait for that blog post :) [18:50] dimitern: i'm about halfway done, i will let you know :) [18:50] katco, my org mode skills so far include only storing my emacs conf in org babel source blocks and some todo lists :) [18:50] awesome [18:50] dimitern: this is even easier; it's basically a plain-text outline that you just export [18:50] dimitern: if you can write markdown, you can do this just as easily [18:51] katco, cool! i suppose it goes via latex to generate the pdf [18:51] dimitern: yes, exactly [18:51] dimitern: i also got it to export and odf, but it's not nearly as nice [18:52] dimitern: but i suppose the important thing is that you don't need to know latex or anything to produce one, just how to write an outline in org mode [18:53] katco, that's way better! [18:53] dimitern: yeah, essentially these things just fall out of the notes i take when i'm gathering requirements. it's so easy [18:54] dimitern: I wonder, do you know the answer to gsamfira_ question?? [18:54] well, for me. but i don't think it should be hard for the uninitiated either [18:54] * katco goes to grab lunch [18:54] katco, +1 [18:54] perrito666, hmmm - what's that? [18:54] hello folks [18:54] I was doing a deploy of juju 1.21 beta5 on MaaS and noticed that the --constraints argument is ignored [18:54] looking at: https://github.com/juju/juju/blob/1.21/provider/maas/environ.go#L781 , it seams that the Constraints field is missing [18:54] is this by design? [18:56] perrito666, gsamfira_ it is by design because constraints are processed earlier [18:56] dimitern: then something happens, because acquireNode never sends tags to maas [18:56] so you get a random node [18:56] gsamfira_, ah, wait .. that's total bs [18:56] if you have a maas install handy, give it a shot [18:57] gsamfira_, it seems they need to be there as an arg [18:57] gsamfira_, that has changed recently, I guess with the AZ distribution logic [18:57] gsamfira_, *please* file a bug about it [18:58] sure thing [18:59] gsamfira_, thanks! I'll have a look in the morning [18:59] sure. The fix is easy. Just add Constraints field :). Not sure if its correct, but it works :) [19:01] dimitern: https://bugs.launchpad.net/juju-core/+bug/1408762 [19:01] Bug #1408762: --constraints option is ignored on MaaS provider [19:02] gsamfira_, cheers! [19:43] sometimes I feel we use the word status a bit too much [19:45] need a status update on the status of health status [19:45] :) [19:49] wwitzel3: I am dealing with the status of the entity, which is analog to going to a toolshop asking the thing to put the thiguie in there [20:25] hi marcoceppi, recently you fixed the install hook on squid-reverseproxy for precise. could you make that fix for the trusty version too === kadams54 is now known as kadams54-away === kadams54-away is now known as kadams54 === kadams54 is now known as kadams54-away