/srv/irclogs.ubuntu.com/2014/09/19/#juju-gui.txt

=== alexpilotti_ is now known as alexpilotti
hatchevening huwshimi01:05
rick_h_huwshimi: <3 01:10
huwshimihatch: Evening!01:10
hatchI am now able to sit straight in my chair again01:11
rick_h_hatch: :p01:11
hatchgot one of these http://en.wikipedia.org/wiki/Myofascial_release01:11
hatchhuwshimi: can you review this so I can land it? https://github.com/juju/juju-gui/pull/56601:20
hatchqa has already been done01:20
huwshimisure01:21
hatchhuwshimi: oh looks like rick_h_ is doing it01:22
huwshimiah ok01:22
rick_h_huwshimi: feel free to look as well01:23
rick_h_I'm just being nosy01:23
huwshimi:)01:23
rick_h_hatch: feedback in, just one theme I'd ask a revisit with and a question or two01:25
hatchrick_h_: ok replies done - one q about the defaults for the model01:27
hatchdefault01:27
rick_h_github bit the farm :/01:34
hatchyeah just saw that :?01:34
rick_h_hatch: Hmm, I see what you mean. Most of the code isn't looking at this value so there's a lot of existing code that would have to know to create models with the right value. Setting a value means this has a chance to be a lie. 01:34
rick_h_The more I think on it the more this might be right. I do think it's odd that the comments mention an assumed default if undefined. 01:34
rick_h_I'll make some other suggestions based on sticking with an undefined default. 01:34
hatchsure ok thanks01:34
rick_h_hatch: so my other comment was that the comments you have around "if the value is undefined we set it to comitted"01:34
hatchoh I think that's a lie01:34
rick_h_but the code there after that isn't doing that01:34
hatchyeah - I think I had it do that before but then it wouldn't fail if something was missed01:35
rick_h_so I'd suggest updating that repeat comment or updating the code to get the value, and clearly set it comitted right there01:35
hatchyou bet01:35
rick_h_that way the comment and the code are in line right together and chances of that being kept up to date are better01:35
hatchyeah definitely01:36
* rick_h_ runs away for the night and checks on the wife-hop-a-long01:36
hatchhaha cya01:37
hatchpass her my best01:37
hatchI just recently did the same01:37
hatchon the way back from London, it still hurts01:37
=== urulama-afk is now known as urulama
=== fabrice is now known as fabrice|lunch
* fabrice|lunch is going for lunch with friends09:47
rick_h_morning10:38
rick_h_party fabrice|lunch 10:38
=== fabrice|lunch is now known as fabrice
fabricerick_h_: morning12:07
kadams54rick_h_: looking for a new card; any suggestions? Was thinking about tackling bug #1371107.12:18
mupBug #1371107: Machine view do not show icons after login <juju-gui:Incomplete> <https://launchpad.net/bugs/1371107>12:18
fabricekadams54: kool the one I entered yesterday !12:22
kadams54guihelp: looking for reviews and QA on https://github.com/juju/juju-gui/pull/568 - review covers a whole four lines of code!12:22
kadams54jcsackett: thanks for the QA!13:18
jcsackettkadams54: yw. 13:19
kadams54Oh oh, wait, no…13:19
kadams54Aarrrr, thanks for the QA, you bilge-sucking bit-pusher!13:20
rick_h_morning13:21
rick_h_kadams54: looking13:21
rick_h_kadams54: that one's a tricky one. jcsackett did some work around that13:22
rick_h_kadams54: basically we don't get data until the page loads from the delta stream and so we'd need to live update the UI, but the user might have interacted so not sure we can13:22
kadams54rick_h_: yeah, I approached it with fear and dread ;-)13:23
rick_h_kadams54: can you investigate https://bugs.launchpad.net/juju-gui/+bug/137111213:23
mupBug #1371112: Unable to delete a service after deployment failed <juju-gui:New> <https://launchpad.net/bugs/1371112>13:23
rick_h_kadams54: and see if there's something there we need to look closer at and update please?13:23
kadams54rick_h_: Sure13:23
urulamarick_h_: that might be a core issue, i remember --force to remove a service in error state13:27
rick_h_urulama: right, just want to check it out13:28
kadams54guihelp: anyone have a suggestion on how to get units to fail on deploy?13:33
rick_h_kadams54: check out http://reports.vapour.ws/charm-tests-by-charm and look for a charm that fails that's not a lint issue13:33
kadams54rick_h_: Cool. That page seems super useful. Bookmarking.13:35
rick_h_kadams54: yea, it's in development still but might help with some of this. It's just a little hard to call out something we know fails to deploy atm13:41
rick_h_vs was it a test failure, or a lint failure, or what13:41
kadams54rick_h_: I updated 1371112 with additional comments. Merits looking into more.13:52
rick_h_kadams54: hmm, yea this gets into showing unit health in MV and such. 13:58
kadams54Yeah, that plays into it, but even if we did display unit health… there are still a bunch of other places that the workflow falls over, IMO.14:00
rick_h_kadams54: right, the workflow is in the inspector14:00
rick_h_kadams54: otp atm, thanks for checking into it. We'll have to think on it. 14:00
kadams54rick_h_: It seems like I ought to be able to destroy a unit in an error state. And the GUI makes it look like I should be able to as well - that is, I can go through the whole "make a change, click commit, click confirm" workflow. But it isn't actually being destroyed, which puts the GUI into a really odd state.14:02
rick_h_kadams54: right, and it's a pain point that juju wants you to resolve first14:03
rick_h_kadams54: so the UX is poor on cli and we're carrying it into webui14:04
kadams54rick_h_: since I don't have much experience with this… in the CLI you have to run 'juju resolved <unit>' before you can run 'juju destroy-unit <unit>'? Even if you do nothing to actually resolve the error?14:06
rick_h_kadams54: yes14:06
rick_h_kadams54: and every user learns that by getting angry with juju14:06
kadams54rick_h_: OK, got it.14:06
kadams54lol14:06
rick_h_so I understand what's happening and agree it's :(14:06
=== luca__ is now known as luca
hatchjujugui call in 1014:50
kadams54guihelp: looking for a second review and QA on https://github.com/juju/juju-gui/pull/56814:52
hatchsure14:53
fabricesure14:53
fabriceI was trying to get into my first review but hatch you can do it14:53
rick_h_fabrice: just do it14:54
rick_h_there's no rule that we can only have two on there14:54
kadams54It would actually be a pretty good PR for a first review14:54
rick_h_fabrice: and it'd be good to try out and do the QA14:54
rick_h_and hatch can just do review14:54
fabricekool I'll do that this evening14:54
kadams54Congrats hatch, you get to review four lines of code :-)14:54
hatchkadams54: code looks good I'm more interested in the qq14:57
hatchqa14:57
kadams54Awesome. More eyeballs.14:58
rick_h_jujugui call in 1 go go go kanban, fight chrome, microphones, and your battery power remaining!14:59
rick_h_fabrice: rogpeppe2 ^15:00
jrwren_rick_h_: I'm missing one of those 4 review comment options, can you fillin the blank? 1. What is this doing?   2. Suggestion: do this too/instead.   3. Urgent: This is a problem.   4. _____15:33
hatchhammertime15:33
Makyojrwren_, This is broken but out of scope, make a follow up carad15:34
rick_h_jrwren_: not in scope for this branch but please adda  follow up card15:34
Makyo*card15:34
jrwren_Makyo: thanks!15:34
rick_h_jrwren_: e.g. I see something up but don't block this on it15:34
jrwren_rick_h_: thanks!15:34
=== fabrice is now known as fabrice|dinner
* rick_h_ goes to find lunch now 16:09
hatchooooo boy17:14
rick_h_hatch: what did you do now?17:19
hatchrick_h_: heh nah its just this conflict resolution UI was not designed to be used outside of conflict resolution17:20
rick_h_hatch: right, but you're not doing that yet right?17:20
* hatch digs through closet for shoehorn 17:20
rick_h_one step at a time?17:20
hatchoh I'm done the other part 17:21
hatchhttps://github.com/hatched/juju-gui/compare/config-default?expand=117:21
hatchI now need to figure out  how to make the UI work17:21
rick_h_hatch: then let's move that forward. The UI part can only be on the summary for now blocking and it's an incremental step forward. 17:22
rick_h_hatch: then we can regroup on the UI part on the inspector17:22
hatchrick_h_: but once the summary says there is something busted....how will they resolve it?17:22
rick_h_this is going to grow with tests/such and doing it in one swoop will go bonkers17:22
rick_h_hatch: it's a baby step. 17:22
rick_h_hatch: we'll do that part next17:22
hatchalrighty17:23
hatchI'll split my card up17:23
rick_h_ty17:23
hatchwhile I listen to the new Weird Al album17:23
hatchhttps://play.google.com/music/m/Bbevcxdfomo3gzvbf3aqiv6dubi17:24
* rick_h_ steps away for a little bit17:56
urulamahatch: i'm living in "there be dragons" land :( /me goes find the album somewhere18:20
hatchhaha18:21
urulamakhm ... FB's sponsored ad for me is ... "Move to America" 18:35
hatchjujugui does comingsoon still update properly>20:43
hatch?20:43
jcsacketthatch: i don't think so.20:43
hatchI think I stumbled across a databinding bug20:44
hatchyup20:45
rick_h_hatch: jcsackett it updates as far as I'm aware20:48
hatchjujugui anyone got a second to confirm this bug? rick_h_https://bugs.launchpad.net/juju-gui/+bug/137178920:48
mupBug #1371789: changing the configuration value of one boolean value marks all as uncommited <juju-gui:New> <https://launchpad.net/bugs/1371789>20:48
rick_h_hatch: jcsackett or do you mean a CSS issue? 20:48
hatchhttps://bugs.launchpad.net/juju-gui/+bug/137178920:48
jcsacketthatch: did huw's branch land? thought it addressed this. 20:50
hatchugh this databinding is a house of cards20:50
hatchwhich branch is that?20:50
hatchthere are no open prs20:51
jcsacketthm. it landed. i did not see this behavior qa-ing his branch, but it might be related. 20:55
jcsacketthatch, it may be CSS based, not data binding. 20:55
jcsackettI am in transit right now, but can look soon. 20:55
hatchjcsackett: yeah I didn't see it either while qa'ing kyles or huws heh20:56
hatchI added a screenshot20:57
rick_h_hatch: rgr, I see it and probably another chunk of work around the css stuff that kadams poked at today21:06
hatchconflict resolution is a tough problem - we probably should have invested the time in writing a more generic solution back when21:14
rick_h_Makyo: how goes the ecs stuff? 21:14
rick_h_hatch: jcsackett you'd have to QA with a charm with multiple boolean to see it, and they'd have to fit into the same space of the sidebar21:14
Makyorick_h_, okay, just hunting down a million edge-cases.  Autoplacing units is giving me the most woe.21:14
rick_h_hatch: jcsackett so not surprised it went unnoticed21:14
rick_h_Makyo: how so?21:15
hatchrick_h_: yeah - true true21:15
rick_h_jcsackett: put a card on you for MV for next week, sorry to pull you back but with that and this other bug want to squash them asap monday21:15
Makyorick_h_, placeUnit is just not very comprehensible, so I think I might've messed up in there somewhere.21:15
rick_h_Makyo: want to push WIP and we can peek at it monday before you start?21:16
hatchMakyo: oh I had to parse it too - if you need some help I may be able to help light the way21:16
Makyorick_h_, sure, one sec.21:16
Makyo(Will likely be making up some slow-hours this weekend, fwiw21:16
Makyo)21:16
rick_h_but why does placeUnit come into play about swapping out the ecs with a fresh one on the env after deploy?21:17
rick_h_Makyo: all good, just checking in.21:17
jcsackettrick_h_: dig. on both cause of bad qa and back to mv. 21:17
Makyorick_h_, I simplistically thought that placeunit would only come into play with ecs.changeSets[0], but I might've been kidding myself.21:17
Makyorick_h_, Just needs some digging.21:18
rick_h_Makyo: rgr, ok21:18
rick_h_Makyo: well lean on the team if you need help. 21:18
Makyorick_h_, rgr - wip branch https://github.com/makyo/juju-gui/tree/multi-ecs21:18
rick_h_or fresh eyeballs :P21:18
hatchhttps://github.com/makyo/juju-gui/compare/multi-ecs21:19
MakyoOr https://github.com/juju/juju-gui/pull/569 yeah, sorry.21:19
hatchMakyo: I think something is borked here....this includes changes that frankban made too21:19
MakyoWhaaat happened with that PR21:19
MakyoYeah, one sec.21:19
rick_h_Makyo: need to rebase with develop21:19
MakyoWhaaat the hell21:20
rick_h_Makyo: have a sec to chat? 21:21
rick_h_hatch: jcsackett free to join?21:21
Makyorick_h_, sure21:21
hatchyup21:21
rick_h_standup21:21
Makyostandup or friday call?21:21
hatchin friday21:21
hatchoh 21:21
MakyoOh, okay21:21
jcsackettComing. 21:23
jcsackettWell that was weird. 21:23
hatchI need to get out more - I'm worried that my brain will become acustomed to mouth's moving and sound being out of sync21:33
hatchlol21:33
* rick_h_ runs away for the day night all!21:34
hatchcya rick_h_21:34
rick_h_lol21:34
jrwren_have a good weekend ya'll.  21:37
* jrwren_ leaves21:37
lazyPowerrick_h_: you know you work for an awesome company when you come across an email thread that mentions toga and powdered wig revenue in a large-scale discussion.21:40
lazyPowerAs I stated in the email, I slow clap for you sir.  well done.21:40
=== alexpilotti_ is now known as alexpilotti
rick_h_lazyPower: glad you appreciated me feeding the trolls22:06
rick_h_I almost closed it while writing it but damn that kind of attitude bugs me22:06
lazyPowerI was thorougly impressed22:09
lazyPowerseriously22:09
lazyPoweran argument like that shoudl be put on showcase in the hall of huehuehue22:09
=== alexpilotti_ is now known as alexpilotti

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