[00:05] hatch: Also, if you have any ideas on how to test the expose option exist in the new "ghost" inspector I'd love some help with that [00:05] https://github.com/juju/juju-gui/pull/461 [00:05] (see the last commit) [00:06] looking [00:06] don't like partials I see :) [00:06] hatch: Oh, no, it's just it only gets included in one place now [00:07] huwshimi well you can only test that the class is there [00:07] not that it's actually doing what it is supposed to be doing [00:12] hatch: My problem is that I don't actually know how to put the inspector into "ghost" mode. Then 'pending' variable doesn't actually do anything: https://github.com/juju/juju-gui/pull/461/files#diff-f05fb78d550f50174f5355a05130fa02R742 [00:15] huwshimi sorry my computer decided it didn't like my keyboard anylonger [00:15] :) [00:16] hmm that should work [00:17] huwshimi oh the mix command is wrong [00:17] huwshimi https://gist.github.com/hatched/b8b2d12bb6d4c6ad90a8 [00:18] this goes around line 90 of test_inspector_overview [00:18] to replace what's there [00:18] hatch: Ah great, I'll try that in a minute! [00:31] Makyo: Are you around? [00:34] Makyo: Just wondering if you actually started your card, I have inadvertently fixed it in the branch I'm working on... [00:35] hatch: This branch is ready for review now too: https://github.com/juju/juju-gui/pull/467 [00:36] cool ill take a look [00:36] maybe....if my internet stays up [00:37] oh huwshimi we now don't rebase until just before landing [00:37] so reviewers can see the changes [00:37] like you can rebase before putting up for a PR [00:37] just don't rebase between PR and landing until just before landing [00:38] hatch: https://twitter.com/rwjblue/status/494992271235100672 [00:38] hatch: Sure, makes sense [00:39] rick_h__ w000000t [00:39] hatch: ember? for a blog? seems a bit extreme [00:39] hatch: and upgrade time [00:39] the admin panel [00:39] hatch: In that pr I just sent you I fixed Matt's card he is in 'Branch start' should I steal that card off him? [00:40] wordpress is extreme for a blog too :) [00:40] I'm not going to upgrade the charm to include the RC [00:40] but this ight be time to provide the option to pull in the source from elsewhere [00:40] huwshimi yes, and send him an email just to be safe :) [00:40] hatch: Sure! [00:44] huwshimi shipppppit [00:44] hatch: Cheers mate [00:46] rick_h__ can you put some input on what should be happening here https://bugs.launchpad.net/juju-gui/+bug/1351112 I think that it should add the unplaced unit back into the unplaced unit list.....buuuuuuut [00:47] hatch: So with that new mix line should I just pass the 'pending' as I have done? [00:47] <_mup_> Bug #1351112: Deleting an ecs'd machine with placed service still shows service in change list [00:47] huwshimi yep [00:47] hatch: looking [00:49] hatch: remind me to look tomorrow [00:49] hatch: brain is slow and trying to qa kadams54's stuff and no more capacity [00:49] haha - too much thinking? [00:49] :) [00:51] updated: https://github.com/CanonicalLtd/charmstore-charm/pull/1 [00:51] its practically all new. [00:52] oops, private. [00:56] jrwren: Hello, welcome! [00:59] jrwren putting in some OT? [00:59] :) [01:02] hi huwshimi, thanks for the welcome. good to be here. [01:02] hatch: yes, I guess so. [01:02] You know how it is when you just really want to get something done :) [01:03] jrwren: Great, things going well so far? [01:03] good enough. [01:03] yup I do [01:03] I'm funding bugs and warts and stuff. :) [01:04] And I'm learning tons. I'm still very new to pretty much everything. [01:04] jrwren: Yeah, lots to learn [01:26] jrwren yeah after a couple months you'll be right in the swing of things [01:31] huwshimi be sure to update the board when your stuff lands [01:31] hatch: Into daily call? [01:31] yep [01:31] np [01:53] hatch: When I set pending to true it doesn't render the UI, any obvious ideas? [02:02] Actually I think it's rendering the old ghost inspector [02:21] oh u need to se the mv flag [02:21] set [02:21] huwshimi [02:26] hatch: Yeah, I have! [02:26] (well, in my local branch) [02:26] hmm [02:27] are you setting it before calling the setupInspector? [02:27] yep [02:27] well I have noooo idea [02:27] you'll have to step through it I guess [02:28] hatch: it's hitting this here: https://github.com/juju/juju-gui/blob/develop/app/views/environment.js#L97 [02:29] OHHHH [02:29] ok that's why it's not working [02:30] sec [02:31] huwshimi so under mv we no longer hit that in the real code flow we now use the dispatcher https://github.com/juju/juju-gui/blob/develop/app/subapps/browser/browser.js#L586 [02:32] hatch: OK, so any ideas how I get it to call that instead? [02:33] huwshimi well I can't believe we haven't tested any of this stuff being passed a ghost model yet [02:34] hatch: Isn't this what you and jcsackett did in London? [02:34] nope this was done pre-london [02:34] I think by jc though [02:34] oh ok [02:39] hatch: pending must be set at a later point than where I'm doing it if this works not in the test [02:39] no that area is never called in a real app [02:39] under mv [02:39] it is instantiated in the browser.js file [02:43] sure, but setting pending is what is making me hit this [02:49] yeah so that's not going to work [02:49] like we are hitting a branch of code that's going to be going bye bye once mv hits [02:51] yeah [03:00] hatch: Any brilliant ideas or should I head to lunch? [03:00] huwshimi the only thing I can think of is to instantiate the view manually [03:00] the overview view [03:01] hatch: OK, I'll try that [03:01] but maybe dig in the closed PR's for when this stuff landed [03:01] maybe use git blame to find it [03:01] because we must have added tests around this stuff [03:06] sure [06:28] urulama: mornin' [06:35] rogpeppe: morning [06:41] rogpeppe, urulama: morning [06:41] huwshimi: hiya [06:48] hey there, huwshimi [06:53] urulama: the reason we don't use errgo.Any everywhere is that we want to have control over the error codes that are being generated (and that can be relied on) [06:53] urulama: the place to use errgo.Any is where there's a callback that we want to give control over the error codes [06:54] urulama: a classical example being when we have a general handler that has more specific handlers plug into it [06:55] urulama: in general, my view is that use of errgo.Any is to be avoided when possible. i am aware that this view has been the source of some contention in the past :-) [06:57] rogpeppe: yeah, that errgo.Any is funny :) [06:58] rogpeppe: ok, could we have a comment then for us non-errgo-experts, to know the reasoning ... just as a reminder not really comment on each Any call [06:59] urulama: the motivation comes from experience with large systems where there can be a lot of "action-at-a-distance" where something high up relies on specific error codes generated by something much lower level. [06:59] urulama: yes, that's a good idea. will do. [07:04] rogpeppe: the comment would be exactly what you've written in the reply on the PR [07:05] urulama: the "see at a glance" comment? [07:06] Because we don't want parseURL setting status codes. By omitting errgo.Any, we can see at a glance which are the possible places that could be returning an error with a Cause (the only kind of error that can end up setting an HTTP status code) [07:10] urulama: oops, just realised there was a file missing from my commit. i've just re-pushed (the file is params/error.go) [07:12] rogpeppe: that's a problem with reviews on github ... it's PITA to see the whole code and then one just assumes that "it's there from before" :( [07:30] rogpeppe: regarding our morning call ... it's way too loud today at home, i'll have to move to another location [07:30] urulama: ok. let us know when. [09:22] * urulama brunches [09:25] trivial PR: https://github.com/juju/charm/pull/32 [09:25] * rogpeppe grabs breakfast [10:09] urulama: i've made all those changes requested, i think (and having a clear policy now made it easy to find a couple more places where I omitted an errgo.Any) https://github.com/juju/charmstore/pull/50 [10:17] rogpeppe: looking [10:17] urulama: it's a pity that github doesn't make it easy to see what changes have been made since the last review comments were made [10:18] indeed [10:22] rogpeppe: thanks for the comments, i think that the code is much easier to read now. [10:22] urulama: great, thanks [10:35] frankban, urulama, jrwren: fairly trivial fix to charm.v3: https://github.com/juju/charm/pull/33 [10:37] review appreciated, because this is blocking the "user charm.v3" branch from landing [10:39] rogpeppe: looking, see that you got lgtm from core for #32 and #31 as well ... nice [10:49] rogpeppe: package version independent code :D [11:05] * frankban lunches [11:16] bac, jrwren, frankban: i need a second review of this, please, before i can land it: https://github.com/juju/charmstore/pull/50 [11:23] morning all [11:23] morning, rick_h__ [12:04] morning all [12:13] morning jrwren [12:18] jrwren: just a quick notice to pay attention on the details of error handling in that PR #50 as we'll use that on all services (until replaced by something more new and shiny) ... if you haven't already that is [12:18] * rick_h__ heads to the coffee shop for morning work [12:18] urulama: ok, thanks. [12:30] jrwren: hiysa [12:30] jrwren: hiya, even [12:30] jrwren: any chance you could review that branch too, please, as i can't land it without 2 reviews [12:30] ? [12:31] I thought I did. [12:31] rogpeppe: review done [12:31] jrwren: ah, perhaps i just didn't refresh the page [12:32] jrwren, frankban: thanks! [12:32] jrwren: make sure to leave a final 'comment' to help indicate the review is complete. (didn't look if you did there but just general fyi) [12:33] guihelp: yesterday I realized that the error message I was seeing was not the same as the one I was supposed to be seeing. I setup a trusty machine and was trying to deploy precise charms. [12:34] I think there have been discussions about this in the past, but are there any plans to filter charms based on series? [12:34] Also: should I setup my ec2 as precise for the time being? [12:35] kadams54: is this in reference to your current card? [12:35] Yeah [12:35] kadams54: frankban landed some work this week that has machine auto match the right 'series' based on what's placed on them [12:35] kadams54: so there might be some other work effecting what you see atm [12:36] kadams54: hmm, so what error are you seeing now? [12:36] kadams54: This might work now with frankban's changes actually [12:36] "Series does not match" [12:36] * rick_h__ goes to get the bug [12:36] I checked the core codebase and it throws that when the unit's series does not match the machines [12:36] So yeah, frankban's work might have helped keep me from shooting my foot off :-) [12:37] kadams54: ok, so you're getting that error out of core [12:37] ? [12:37] Yes [12:37] GUI tries to deploy the unit to the container and gets an error… which it should, according to the card. [12:37] I just didn't realize initially that the error I was getting back was the wrong one. [12:38] So even though I'm working in the GUI, the error is coming from core [12:38] kadams54: ok, so the bug is that the error is machine doesn't exist, but that's "fixed" and what you're seeing is the wrong series error instead now [12:39] kadams54: hangout please? I want to make 100% sure of the steps and such [12:39] I'm not sure if it's "fixed" or if I was just hitting the "wrong series" error first. [12:39] If it is fixed, I didn't do anything to fix it :-) [12:39] Sure [12:40] Standup hangout? [12:41] kadams54: sorry, the friday one [12:43] gah, no rain here for six weeks and the tropical storm is due to arrive about the time i have to go to the airport tomorrow. [12:51] bac: ha, no rain here for ages either, and the downpours are just starting in time for a nice weekend out camping [12:53] rogpeppe: they aren't so good at periodic maintenance here. the slightest rain causes the highway to flood since the stormwater system is clogged up. makes getting to the airport hard. [12:54] bac: well at least we don't have that problem... unless it *really* rains [12:54] yes, because you're not 1 GDP in debt [12:54] :( [12:55] kadams54: pr #171 can land correct? [12:56] bac: Where does it flood? [12:57] jrwren: mainly the roadways. and parking lots. and low lying areas. not our house, though. [12:57] We have some of that here, but only on a few select roads. [12:57] * urulama sees the Sun for the first time since London ... happy to switch with people with no rain ;) [13:11] jcsackett: kadams54 are either of you up for reviewing/qa'ing frankban's branch please? https://github.com/juju/juju-gui/pull/469 [13:12] rick_h__: i thought frankban had another commit he wanted to push up to that branch? [13:12] i've been holding off waiting for changes; did i misunderstand? [13:13] jcsackett: the branch is ready to be reviewd [13:13] dig, well i'm sorry for holding it up. :p [13:13] * jcsackett goes to review it. [13:13] jcsackett: np and thanks [13:18] hatch: did your ac branch land? [13:19] bac last I checked I needed one more review [13:19] oh [13:19] ok [13:19] hatch: I reviewd it this morning [13:19] hatch: one point in there please [13:19] ohh ok [13:19] will look [13:19] sorry there is like 200 emails, haven't gotten through themall yet [13:19] heh [13:19] heh, np. Figured it was early for you yet [13:20] rick_h__ replied [13:21] yeah it's still only 7:20 [13:21] trying to buy a kite on ebay so had to get up early to check on it haha [13:21] er get online early [13:21] hah [13:22] hatch: ty [13:24] anyone here use duolingo? [13:33] hatch: i did for a while. plan to start again [13:33] hatch: i thought about it recently, but didn't [13:33] hatch: it is really good [13:33] hatch: i know some friends that are really into it [13:33] I just started on Spanish, prepping for a winter getaway (if we can ever get our S$%^ together) :) [13:34] so, "e", "ev", "evt" for events? i see all three, and i see many diffs that switch them back and forth. can we just pick one? :p [13:34] after one lesson I've already learnt more than I did before heh [13:34] jcsackett I vote for e (ev is from benji, evt is from frankban) [13:34] lol [13:34] :-) [13:34] "evt" however gets "<3" from the boss man. :p [13:35] hatch: evt is because I want t make you happy: I'd write that as "event" ;-) [13:35] hatch: my pr is updated btw. [13:35] haha [13:36] hatch: duolingo produces some silly things to say: 'the duck drinks milk' [13:36] single letter var names are evil no matter what :P [13:36] evt reads nice to me and love s/e/evt [13:36] how many letters have we saved, if you include this discussion? [13:36] but it's personal preference. I think ev was me caving to hatch. [13:36] bac: 3, it's a fact [13:36] :) [13:36] * jcsackett laughs [13:36] lol [13:37] do we have actual style guidelines written anywhere for jujugui? or is it just "whatever the reviewer and lint will let through"? [13:37] come on, the linter hates on the rest of my party, have to find the points to chat about somewhere [13:37] jcsackett: no, I don't know we'd ever get everything through a process [13:37] lol [13:37] yeah, let's do that some friday i'm gone [13:37] I've still resisted the urge to one weekend s/ / /g [13:37] we still can't agree on doc formatting [13:38] (or at least after lunch) [13:38] lol [13:38] Spanish is a wako language though bac, there is a male/female/adult/child version of everything [13:38] and they say English is hard [13:38] hatch: yeah, uh, i've noticed that. [13:38] haha [13:39] no, English is insane. gendered languages are stupid but English is insane if you're not 3 [13:40] there their they're [13:40] i have a theory computer scientists have a harder time learning a natural language since the inconsistencies are so aesthetically offensive. [13:40] hahaha [13:40] hatch: most eu languages are gender specific, we also have the third option, for "it" :D [13:40] English grammar is simpler than other languages, pronunciation is hard, due to the fact English is not a phonetic language [13:41] rick_h__: pr#171? You mean #471? That's the one that hasn't been QA'd yet… [13:41] hatch: and not just singular and plural forms, but also special dual form ... [13:42] kadams54: ah gotcha [13:42] surely there is a list of things you DO agree on which could be written. [13:42] kadams54: did you find out how you did it locally? [13:42] urulama haha damn.... [13:42] rick_h__: no, I got distracted… looking now. [13:42] jrwren: I thin kit's that code is good [13:42] jrwren: that's about it [13:42] hatch: so, try learning German (or slavic languages), then you'll consider Spanish simple :D [13:42] jrwren: so much history...I'm actually kind of floored that I asked teo join this team. The first 6mo was so rough. [13:43] rick_h__: lol [13:43] urulama well I know some Ukranian does that count? [13:43] rick_h__: how long has it been? [13:43] jrwren: hmm, 1.5yr on this work? [13:43] the GUI stuff [13:43] I remember after christmas working on charmworld/charm browsing in the GUI [13:43] hatch: :) [13:43] so little over [13:44] jrwren no we don't agree, we just tolerate... [13:44] it's like a supersaturate [13:44] so sad. [13:45] one of these days it's gona explode into extra comment styles and single char variables everywhere [13:45] * jcsackett one day will work through the holidays to re format everything, update the linter, land a style guide, and pronounce it the law of the codebase... [13:45] sounded like flake8 was agreed upon. [13:45] at some point you realize there's natural things that are purely opinion and try to do the best you can [13:45] python is easier to agree on. [13:45] javascript is a mess to begin with, so... :p [13:46] the python 'everyone agrees' is a lie [13:46] GTK [13:46] lol [13:46] there's tools that half the people don't follow, the 'I pep8'd everything for you' pull request is a flame war in the making [13:46] haha [13:46] and then there's always the old fallback of "if the stdlib doesn't even follow..." [13:47] rick_h__: never said everyone agrees, just that it's easier. [13:47] * rick_h__ notes that as the guy that asked jrwren to add a make target to the charm linting the python [13:47] but that's for another channel [13:47] i've always liked the policy that new code landing should follow the style guidelines, but you don't ONLY change style for style sake. [13:48] * rick_h__ heads home from coffee shop, afk for a couple of min [13:49] as a side note, the js linter we use sometimes goes crazy, e.g.: wrong indentation, this must be indented at 6, 9, 12.7 or 42 spaces... and I don't remember why we decided not to just indent everything with 2 spaces... [13:50] frankban: +1 on your PR, QA OK. [13:50] jcsackett: cool thanks [13:50] frankban: i think when you can't find a logic to a style decision in GUI, the answer is there was no logic to it. :p [13:51] jujugui: i need one more review for https://github.com/juju/juju-gui/pull/470 (hatch, you had a chance to look at the update?) [13:52] jujugui: also, because my juju install is borked, kadams54 needs a QA on https://github.com/juju/juju-gui/pull/471 (it requires a live env) [13:53] jcsackett: yeah, rick_h__ volunteered but I need to find more QA info for him. [13:54] kadams54: ah, cool. [13:58] jcsackett: out of curiosity, do you know the value you passed in as source when you were trying to QA that PR? [14:00] kadams54: "repo branch-name" [14:02] i've since had juju refuse to connect to any of my deployments for my hackerspace--not sure if they're related, but i've got a todo to reset my juju stuff sometime today. [14:02] i've never cleaned up various ppas etc since i was on qa--figure there's a chance i'm working with bad stuff. [14:03] kadams54: not literally "repo branch-name", of course--whatever copied from github for your repo and branch name. :p [14:04] jcsackett: Yeah, rick_h__ was having problems too… [14:04] huh. [14:04] urulama, frankban, jrwren: here's the stats Store code imported from the old charm store: https://github.com/juju/charmstore/pull/51 [14:04] ...wonder if the two things are unrelated, and the gui charm is having issues... [14:04] * rogpeppe loves it when code is that modular [14:06] its a good sign. [14:12] rick_h__, jcsackett: just figured out why I wasn't finding anything in my shell history… I hacked up fakebackend.js to return errors so I could dev locally. (Which was when I asked about updating it to do some of the same error checking as the real thing.) [14:12] kadams54: ah, cheater [14:12] Yup. [14:13] So… still no closer to understanding why the gui charm isn't getting pointed to the right source. [14:13] kadams54: yea, now that it's morning I'll try again and see what's up [14:14] sorry for the delays [14:15] jcsackett, rick_h__: do we really want to collapse var declarations when the definition is "simple"? [14:15] frankban: it was agreed long ago that if they were simple one liners then it was ok to do [14:16] frankban: and anything multi-line, objects, arrays, etc would be split out [14:16] rick_h__: I know it's ok to do that, I am not sure if it's optional or not [14:16] frankban: I take it as optional [14:16] frankban: the only time I call out is like hatch's when he had more than one var per line [14:16] as that hides potential vars [14:18] rick_h__: cool thanks, are you QAing my branch? [14:18] frankban: no, left that for jcsackett [14:19] rick_h__: cool, he did it, thanks, will submit soon [14:19] frankban: ty much [14:21] frankban: bac didn't we update the charm for the KeyError: u'NetworkScope' issue? [14:22] rick_h__: we updated quickstart, not the charm. Is the charm affected? [14:22] frankban: oh, maybe I've not updated quickstart /me goes to dbl check source [14:30] hatch: buld failed email? [14:32] yup got it re-running [14:34] kadams54: hmm, no idea why this is failing. The apt-get install runs and then it gets to Processing triggers for libc-bin ... [14:34] ldconfig deferred processing now taking place [14:34] and then it hangs and config goes into a bad state [14:34] before it ever hits git [14:57] jujugui call in 3 [14:58] kadams54: oh crap, all my issues were because I missed the s in your github username. [14:58] hahahaha [14:59] found where it logs out to, not the charm log :( but the all log [14:59] anyway, finally getting somewhere with your qa, thanks for your patience [14:59] "I must've put a decimal point in the wrong place or something. shit, I always do that. I always mess up some mundane detail." [15:00] hatch that's a great look. doesn't look dorky at all. [15:01] jcsackett: ^ [15:01] bac? whowhatinthewhatnow? [15:01] hatch: :) [15:12] PR just cut back from 19 to 15 holidays, starting next year. (i don't get them) [15:20] rick_h__: OK, I'm getting the right error now for my card, so venturing forth… [15:20] kadams54: woot [15:20] frankban: i'm off in 40 mins, BTW. 't'would be great if you could perhaps have a look at that PR before then, please. ta! [15:20] ish [15:21] rogpeppe: sure [15:21] frankban: thanks [15:21] rick_h__, rogpeppe, frankban: are we using github for the bug tracker for juju/charmstore ? i see two issues are there. jrwren has documented it as a TBD so i'm curious. [15:21] curses! out of tape [15:21] bac: yes, I think we should. [15:22] bac: since we don't have a LP fallback project [15:22] bac: i think so, yes [15:22] hmm internet isn't being too awesome today [15:22] bac: but I think we did have issues in LP in core we might need to peek at [15:22] bac: i prefer associating the bugs directly with the repo, tbh [15:22] rogpeppe: what do you mean? [15:23] bac: well, since we're hosting on github, it makes sense to track bugs there, i think [15:23] rogpeppe: ok, agreed [15:24] rick_h__: those bugs would be more appropriately raised against github.com/juju/charm though, i think [15:25] rogpeppe: ok [15:25] rogpeppe: so, is your branch just moving stuff from _old? [15:25] frankban: yes [15:25] frankban: with a few tweaks to fit the old stuff with the new [15:26] frankban: (e.g. session becomes db) [15:26] frankban: i haven't changed any of the actual logic [15:26] rogpeppe: ok [15:29] kadams54: I think we fixed the error that you were working around and so QA isn't valid atm [15:29] kadams54: but the change is small, I +1 to land [15:37] rogpeppe: LGTM [15:37] frankban: ta! [15:43] rogpeppe: by "i haven't changed any of the actual logic" you mean that it's original form is good enough to be just easily hooked with handlers? [15:43] s/it's/its [15:43] urulama: yes [15:44] urulama: and there's already a handler in the old code that implements the actual stats/counter serving logic [15:44] urulama: which is what i'm porting over currently [15:44] guihelp: I need reviews/QA for https://github.com/juju/juju-gui/pull/473 ? Anyone? thanks! [15:46] rogpeppe: and the code covers /stats/counter use-cases? does it do more as well? [15:47] urulama: i don't think so, but /stats/counter is fairly general [15:47] frankban on it [15:48] hatch: thanks [15:48] frankban so machines and containers use the same model? [15:49] hatch: yes [15:49] frankban how did you find this out? trial and error? Or is it documented somewhere? [15:49] the MANAGE_ENVIRON [15:50] oh constants.go ? [15:50] hatch: I don't know if it's documented, I saw that while implementing the mega-watcher for machines [15:50] ohhh ok cool [15:50] hatch: so server side [15:51] hatch: I knew that sooner or later that jobs attr would be useful ;-) [15:52] haha truth! [15:56] jujugui have a great weekend and may the gods of CI be on your side next week [15:57] urulama: :-) and you, have a good trip [15:58] rick_h__: What changed to fix the error? [16:00] rick_h__: Maybe I should clarify… which error? As far as I know, there's still an error state that occurs when deploying to a ghost container. You'd mentioned earlier that frankban landed work that would fix deploying the series mismatch between charm and machine… [16:00]