/srv/irclogs.ubuntu.com/2014/06/23/#juju-gui.txt

rick_h_morn huwshimi 00:03
huwshimirick_h_: Hey :)00:03
hatchhey huwshimi 01:06
huwshimihatch: Hey01:07
hatchglad to start the week off without an active PR? heh01:07
huwshimihatch: Yes!01:09
huwshimihatch: Thanks :)01:09
hatchhaha np01:09
hatchthere was one small cleanup I wanted but i think kyle is going to add it to his branch01:10
hatchhe found a bug with his branch that he said he was going to fix over the weekend but I haven't seen it land yet01:10
huwshimiyeah01:12
hatchI just updated the PR with the bug he found for others....I guess he didn't add the comment afterall01:12
rick_h_huwshimi: getting the next set of cards lined up in the backlog/on deck for the next 2wk cycle01:13
rick_h_huwshimi: let me know if anything seems odd or there's anything missing we don't have cards for that we should01:13
huwshimirick_h_: Ah great, I'll have a look01:14
rick_h_huwshimi: I think this next 2wk cycle is mainly around the updated scale up journey and the 'change version' update and really we're finishing/polishing it seems. 01:14
huwshimiyay!01:14
rick_h_huwshimi: so shatter my dreams of having this become a scaled back 'polish' task after this 2wk iteration if you can think of anything :) 01:15
rick_h_but better now than later01:15
huwshimiok :)01:15
hatchhaha01:15
hatchrick_h_ have you ever used `setcap` http://linux.die.net/man/8/setcap ?01:15
rick_h_hatch: nope01:16
hatchrick_h_ darn, I'm having a hard time trying to figure out how it'll help me use port 80 in the ghost charm https://bugs.launchpad.net/charms/+bug/1229377/comments/701:16
_mup_Bug #1229377: Charm needed: Ghost blogging platform <Juju Charms Collection:Fix Committed by hatch> <https://launchpad.net/bugs/1229377>01:16
rick_h_hatch: why not just have apache/proxy requirement for port 80?01:17
hatchrick_h_ currently, that's what it is - but until bundles become first class citizens in juju it's not a perfect experience if the user doesn't fully read the readme01:17
hatchand I really don't want to build in nginx or the sort01:18
rick_h_hatch: hmm, ok guess. Most frameworks default to a non-80 port though, but yea it'd be cool to support01:18
rick_h_hatch: so since hooks run as root why not just have root bind and run with it01:18
rick_h_?01:18
hatcha previous comment said that ghost shouldn't run as root01:19
hatchit's been quite confusing01:19
hatchdon't use root, use port 80...ahhhhh01:19
hatchlol01:19
rick_h_well, if it's run in a container/etc then what's the issue?01:19
rick_h_I'd push back and ask for an example charm that does this without running as root. What does wordpress use? 01:20
hatchI think it has apache running inside or something01:20
hatchI have pushed back on the port 80 requirement01:20
huwshimirick_h_: There are a few things in Luca's QA doc that don't have cards: https://docs.google.com/a/canonical.com/document/d/1p-wPlgVqZoYlMRuChLpAjkJAKZcw-wkSZfs6deIWmc8/edit01:20
hatchso hopefully they let it in01:20
rick_h_huwshimi: onboarding is saved for last01:21
rick_h_huwshimi: and some of these I think are fixed released. Will go back through it01:21
huwshimirick_h_: Things like the popover size controls01:22
huwshimirick_h_: close controls on the summary01:23
huwshimirick_h_: I can add cards for the ones I know are missing if you like?01:23
rick_h_huwshimi: hmm, thought those were the cards in backlog01:23
rick_h_huwshimi: yea, those are in the backlog on deck01:23
huwshimioh yeah that's there01:24
rick_h_huwshimi: but do feel free to add anything you see missing that's not fixed already01:24
huwshimiok sure01:24
huwshimirick_h_: I'll start picking up some of those backlog cards today as I'm nearly done here.01:26
rick_h_huwshimi: oh ok, I was going to mention it but you had a card in 'branch start' so figured you were set01:26
rick_h_huwshimi: but definitely, grab any on deck if you're free01:27
huwshimigreat01:28
=== jcsackett_ is now known as jcsackett
hatchhuwshimi did the review on your branch03:41
hatchyou seem to be getting the hand of this js thing :)03:41
huwshimihatch: Ah brilliant thanks.03:41
hatchs/hand/hang03:41
huwshimihatch: Would you prefer a switch?03:42
hatch6:1 6:12 of another :)03:43
hatchI think a switch would minify smaller03:43
hatchI would probably go with a switch if I was doing it, but the benefit would be negligible 03:44
hatch:)03:44
huwshimihatch: Sure, I'll change it :)03:44
hatchhah ok cool03:46
huwshimihatch: Ugh, 'this' isn't passed into switch cases? Yay!03:56
hatch:) if it's gona be too much of a pita you don't have to do it 03:58
huwshimihatch: Ah, no, I forgot to pass this to the new forEach04:00
huwshimihatch: Changes are up if you want to take a look: https://github.com/juju/juju-gui/pull/39804:04
hatchcool looking04:06
hatchlooks good04:07
huwshimigreat04:07
huwshimithanks!04:07
hatchhuwshimi I set up to re-run your merge04:49
huwshimihatch: Ah thanks04:49
hatchit failed with a intermittent error04:49
huwshimihatch: We seem to have a lot of those04:50
hatchdont I know it!04:50
hatchhuwshimi the other branch failed with the same error04:56
huwshimihatch: Yay!04:56
hatchit looks like saucelabs might be borked04:57
huwshimiYay!04:57
hatchI can try and land them in the morning04:57
huwshimithanks05:03
hatchannnnd I'm out, have a good one05:08
rogpeppe1mornin' all07:42
rick_h_morning party people11:35
* rogpeppe1 lunches12:28
hazmatwhy is charm config read only during deploy?13:19
hazmatalso why do interfaces/relations not show for a  deployed charm?13:19
hazmatoh.. you have to explicitly select not default configuration..13:20
hazmatodd13:20
rick_h_hazmat: charm config is getting updated and we're looking at pulling that read-only bit13:23
rick_h_hazmat: the thought was that the defaults should be sane and to make it as clean/few things to click on as possible to get deployed/going13:23
rick_h_hazmat: but we've got work this 2wk cycle to update that13:24
rick_h_hazmat: and for interfaces/relations, no idea. I guess by clicking the charm you can see what to relate to. Constructed relations show in the tab for them13:24
rick_h_but never really had relation info, not sure if we get that from the charm data from the env or not. 13:25
hazmatrick_h_, ie. i've deployed nova-compute.. charm details on it, doesn't show the related/interface tab13:25
hazmatcharm browser does..13:25
rick_h_hazmat: yea, because we don't have the data from the charm model data from the environment so we scope down the tabs to match the data we have13:26
hazmatrick_h_, we do get interfaces/relations from env.. (not related though)13:26
rick_h_hazmat: with the work in progress to add the data/update the apis to the juju store we'll be able to update that13:26
rick_h_hazmat: right, we don't have the info to build that tab. We didn't have any request/requirement to build an alt UI to that tab13:27
hazmatrick_h_, this is part of the charm definition.. doesn't need the store to display named rels and interfaces13:27
hazmatits available in the charm info api off the env13:27
rick_h_hazmat: right, sorry. You're right that we have the end points, but the api doesn't match the info to build the UI. It needs updating to work for deployed charms. 13:27
rick_h_hazmat: and just never been asked for or thought about to be honest. 13:28
hazmatseems like a basic usability bit.. ie. i've got x deployed, what can i do with it (what else do i need to deploy)13:29
rick_h_hazmat: right, there's work in the machine view stuff to update the inspector to show 'popular relatable charms'13:29
rick_h_and moving config to it's own tab. I'll file a bug on the tab though for deployed charms 13:30
rick_h_hazmat: https://bugs.launchpad.net/juju-gui/+bug/133325813:32
_mup_Bug #1333258: gui does not show related charms tab for deployed charms <juju-gui:Triaged> <https://launchpad.net/bugs/1333258>13:32
hatchmorning all13:53
rick_h_party13:53
hatchso party13:54
bacrick_h_: i submitted a vacation day request14:01
rick_h_bac: k thanks. hatch sent me an email as well will get those in14:01
hatchw00t14:02
rick_h_jujugui any time off the next 2 weeks please submit so we can plan cards for the 2wk sprint14:02
* hatch isn't sure if new system is better than old system14:02
hatch:P14:02
bachatch: yeah.14:02
bacbe aware, if you click on the big calendar to start a new request, it pre-populates not with the date you clicked on but with today's date.  sneaky.14:02
hatchI suppose this new one actually allows you to go back and modify things14:06
hatchthat almost makes up for the scheduled downtime that seems to be a feature in the salesforce world lol14:07
bacjujugui: does anyone know if the new HR system allows you to see a calendar for colleague's work schedule? i can't find it.14:07
hatchyep14:07
hatchgo to Team Members14:07
hatchPaid Time Off14:07
hatchthen click on the team members face14:07
hatchyou can also change the level of PTO.....I always change it to level 10.....( I don't think this does anything but I feel special at level 10 days off ) :D14:08
hazmatrick_h_, thanks14:09
bacrick_h_: i'm going to be traveling on the thursday before the sprint, the 17th. mark that as a swap day so you'll see i'm unavailable?14:14
rick_h_bac: sure, works for me. 14:15
bachatch: that works but you can't see schedule for people not on the team14:16
hatchohh, I suppose - could you with the old one? I never tried14:18
hatchtbh14:18
hatchjujugui in ecs mode do we really need to confirm removing relations? Or can we just do it and then they can add it back in before hitting commit if they did it by accident (somehow)14:27
rick_h_hatch: you mean do we need to ecs remove relation at all?14:28
hatchno, there is a dialogue that pops up asking you to confirm that you want to remove the relation14:28
rick_h_hatch: ah, no. I'd assume that no we'd not need to14:29
rick_h_we should ping UX though on how to represent a removed relation in the canvas. I feel like that's a whole in the UX14:29
hatchgood point, I'll fire off an email14:29
rick_h_bac: pm14:39
Makyojujugui call in 1014:50
rick_h_get your planning poker buzzers ready!14:51
hatch314:56
hatch214:56
hatch414:56
hatch414:56
hatch414:56
hatch114:56
hatch214:56
hatch314:56
hatch*phew*14:56
hatchok warmed up14:56
rick_h_jujugui call in 115:00
rick_h_antdillon: ^15:00
hatch__kadams54 http://fromanegg.com/post/67170468715/intro-to-go-from-a-javascript-developer-barcamp-yxe15:12
=== hatch__ is now known as hatch
hatchkadams54 of course there was a whole spoken track which outlined the differences while showing it in go15:50
kadams5481334515:50
kadams54hah!15:50
kadams54hatch: Thanks, looks like some good material.15:50
hatchyeah the talk got good ratings in the reviews 15:51
hatchugh Ci15:52
hatchrick_h_ I wonder if we used a bigger instance for our CI if we would end up with the same issues15:52
hatchkadams54 I re-triggered the merge run for #39915:53
kadams5485961915:53
hatch36871864181814381315:53
hatchI win15:53
kadams54Man, I gotta quit doing that.15:53
hatchlol!15:53
kadams54hatch: thanks15:53
hatchI'm thinking of painting the edge of mine with some clear nail polish and only leaving a small sliver open15:54
hatchthen it can't be hit as easy15:54
hatchgoing to try it first with some tape15:54
rick_h_hatch: it's a big instance currently15:58
hatchrick_h_ interesting - I'm just wondering why the http requests are so.....darn......slow......15:58
kadams54dilrklkdecijlvjntbluutlkffgkbebjfbfvlr15:59
hatchrofl16:00
rick_h_hatch: it's 4 cores and 8gb of ram16:00
hatchrick_h_ lol what the heck16:00
hatchblame java?16:00
hatch:D16:00
rick_h_hatch: :)16:00
hatcheveryone, in your local laws do you have a legal limit of .04 BAC for driving? 16:31
rick_h_hatch: .0816:33
hatchours are changing to .04 and the penalties are bonkers imho16:33
rick_h_hatch: with .02 for < 21yrs old and .17 is bad bad times16:33
hatchI've tweeted at the govt asking for statistics 16:33
hatchhttp://www.sgi.sk.ca/about/articles/2014/trafficsafetychanges.html16:33
hatchhttp://www.sgi.sk.ca/images/0-content-pages/2014_consequences_660.jpg16:34
rick_h_well, I'm always for locking that down imo. Too many people, my family especially, going about :/16:34
hatchI'm very very skeptical that it will help anything16:34
hatch.04 is one beer after work16:34
rick_h_http://www.michigan.gov/msp/0,4643,7-123-64773_22774-49577--,00.html16:34
hatchright, but what those stats dont' say is the BAC of the person16:35
hatchjust that they were alcohol related16:35
hatchthat's the problem with all of these stats they are missing the details16:36
hatchrick_h_ lol 289,000 crashes.....in a year???16:39
rick_h_hatch: for our one steate16:39
rick_h_state16:39
hatchjeebus you guys like your booze 16:39
hatchlol16:39
hatchmaybe you should drop yours to .04 :P16:39
rick_h_:)16:39
* hatch says that while he waits for his own traffic stats16:39
hatch""Michigan law enforcement agencies arrest an average of 103 motorists a day for drunk and impaired driving.""16:40
hatchwow!16:40
hatchI am kind of guessing that those stats as a % are not outliers and each state/prov probably follows similar stats16:41
rick_h_well I think it does depend on states16:41
rick_h_driving, the types of driving, conditions, etc16:41
=== rogpeppe1 is now known as rogpeppe
rick_h_michigan with the snow/ice seasons, the overall rush/poor driving, really poor roads. I'd expect us to be towards the upper 10%16:42
hatchI suppose worse conditions would increase the probability that an impaired driver would crash16:42
rick_h_right16:43
rick_h_and get noticed/pulled over/etc16:43
hatchtrue true - I drove around a lot this weekend, people don't need any help driving like idiots 16:43
hatchlol16:43
rick_h_jujugui heads up the board is ready for the cycle. We almost hit planned counts and had a little over 20% unsched work. 16:52
hatchreally? awesome - I thought we were sunk :)16:52
hatchgo us!16:52
rick_h_well last cycle I over sub'd the board a bit to get a feel for the bigger team size and the blues is a lot of unknowns16:53
rick_h_so the board wasn't a good indicator, but all good16:53
rick_h_this time we're right on, so should be good and if the new starters help with any of it we should be golden. 16:53
hatchgreat16:54
hatchjujugui fyi 'make devel' almost works 100% on osx :)16:57
rick_h_lol16:57
hatchrick_h_ just reviewed Huw's latest branch, needs a touch of work so I'm sure he'll get it landed early his morning17:19
rick_h_hatch: rgr, ty much17:19
rick_h_bac: rogpeppe can you guys take a peek at https://github.com/juju/charmstore/pull/6 when you get a few mins and make sure that makes sense?17:29
rogpepperick_h_: i'm pretty much at eod, but i'll try to take a look17:29
rick_h_rogpeppe: tomorrow is ok17:29
rick_h_rogpeppe: just on the radar. with francesco out and bac getting up to speed going to lean on you for review help17:30
bacrick_h_: will do17:31
hatchjujugui - if I tell juju-core to delete a machine which has a number of containers in it (via the api) will it auto tear down the containers or will it throw an error?17:35
rogpeppehatch: i'm not sure17:36
rogpeppehatch: there has been some debate about the desired semantics there17:36
hatchok - at the moment for the ecs I'm going to assume that any delete operation will 'just work' and we can re-evaluate later I suppose. 17:36
rick_h_hatch: throw up a branch on ec2 and let's see?17:36
rogpeppehatch: personally i would prefer the former, but some people want things to be more explicit because you might be tearing down important stuff17:36
hatchrick_h_ I'm not sure how to easily simulate an api call through the gui17:37
rogpeppehatch: it shouldn't be hard to try it out17:37
hatch'easily' being the key word there17:37
rick_h_hatch: we have the api calls, just need to console.log to find and build the right call?17:37
bacrick_h_: i looked at cmars PR.  it may be good as a first start but i think webops is leaning towards storing dependencies in an S3 bucket for prodstack deploys rather than bundling.17:37
* bac wishes jcsackett were around17:37
hatchrick_h_ rogpeppe I think I found my answer in the gui source 17:38
hatch      @param {Boolean} force Whether to force machines removal even if they17:38
hatch        host units or containers.17:38
rick_h_bac: oh didn't realize that. Good to know I guess. So for deploying we need a step to get deps into the private space s3 bucket?17:38
rogpeppehatch: cool17:38
bacrick_h_: yes.  i'll check with sinzui to see how they've set things up17:41
rick_h_bac: ok cool. 17:42
* rogpeppe is done for the day17:42
rogpeppeg'night all17:42
bacbye rogpeppe17:43
hatchheh state bites again...oops https://bugs.launchpad.net/juju-gui/+bug/133337618:23
_mup_Bug #1333376: Clicking links in relation dialogue shows inspector under charmbrowser headers <juju-gui:New> <https://launchpad.net/bugs/1333376>18:23
hatchhmm, we need ghost-service-name to service-name utilities 18:31
rick_h_hatch: yea, we dance around that and probably needs to be more formal18:32
hatchatm removing a ghost relation between two ghost services fails because of that18:33
hatchI'm investigating but it might be a follow-up18:33
rick_h_hatch: rgr18:34
hatchyep follow-up Ill create a bug18:37
kadams54hatch: https://github.com/juju/juju-gui/pull/393 should be QA-able again. Also addressed review issues.18:39
hatchok looking18:42
hatchugh github 'expand' buttons aren't working18:46
hatchthis is makign this review very difficult18:47
hatchwell I totally borked my develop branch...19:01
rick_h_stop doing that hatch 19:01
hatchbut it's so funnnnn19:02
hatchugh git19:07
hatchok NOW I can qa the branch19:10
hatchkadams54 (Note: should this link be hidden when a machine is not selected?) <--- you should send an email to design and create a bug/card to track the response so we don't forget about it19:17
kadams54Will do19:17
hatchis clicking create supposed to close the create-machine box?19:18
kadams54hatch: Yes. Close the form, create the machine.19:24
kadams54The mocks don't have the form sticking around after machine creation… unless I'm misunderstanding?19:25
hatchsee the qa issue in your PR19:26
hatchkadams54 if I create a machine, then drag a unit to that machine my only options are lxc or kvm, so I can't put it on bare metal....is this by design because it feels like a bug?19:29
hatchI'll add it as a qa issue in the PR 19:31
hatchkadams54 ok qa done19:32
bachi rick_h_, do you have some time to talk?19:42
rick_h_bac: in 10?19:42
bacsure. ping when free19:42
rick_h_bac: ping freee! told ramm I had important calls to make :)19:44
bacha.19:44
bacstandup19:44
rick_h_rgr19:44
hatchI wonder if i could hire an intern to write tests for me20:15
kadams54testern20:16
hatchlol20:16
hatchthat20:16
hatchit would actually be a great education20:16
rick_h_hatch: but you've gotten so much better at writing them :P20:16
hatchhaha20:16
hatchI could then pass this knowledge on20:16
hatch!20:16
rick_h_there you go, hatch the test teacher20:16
rick_h_here's your 'homework' and make sure it's done by tomorrow so I can get my code review in20:17
hatchlol20:17
hatchexactly20:17
hatchrick_h_ I think that #1333395 should be my next card after this branch lands because the relation removal story is incomplete/broken until then20:28
_mup_Bug #1333395: Cannot delete ghost relation between two ghost services <juju-gui:Triaged> <https://launchpad.net/bugs/1333395>20:28
bacrick_h_: is there a plan for local charm support in bundles? i know we've talked about it but i'm unclear about the decision/roadmap20:28
rick_h_bac: we've talked about a lot of them to be honest. No one is settled on at the moment. 20:32
rick_h_hatch: ok20:32
rick_h_hatch: makes sense20:32
rick_h_bac: if there's a store in every environment out of the box, then you'd just publish to your environment. 20:33
rick_h_bac: or there's some bugs to work out around juju-core and versioning to make bundles work in the current form 20:34
rick_h_bac: and then there's resources/fat charms/bundles that could be used to solve the problems20:34
rick_h_but that's not on the current todo list20:34
bacrick_h_: gotcha.20:34
* rick_h_ runs away for the day20:42
rick_h_evening all20:42
hatchlata!20:53
hatchMakyo any idea where the add relation tests are? There are the add_relation tests for non mv in test_env_go but I don't see any which test the mv path21:19
Makyoone sec.21:19
hatchthanks21:19
Makyotest_env_change_set.js:462 (I think that's what you're looking for, correct me if not)21:20
hatchwell that tests the change set stuff - but nothing that the actual call to the go env ends up in the ecs21:21
hatchI was looking for something to copy....lol21:21
hatchmaybe it was just overlooked?21:27
MakyoNot sure..21:42
hatchoh well no problem22:39
huwshimiMorning23:07
rick_h_morning huwshimi 23:11
huwshimirick_h_: Hey23:11

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!