2014/08/20/#juju-gui.txt

[00:18] <rick_h__> morning huwshimi 
[00:18] <huwshimi> rick_h__: Hi :)
[02:39] <hatch> huwshimi:  hey
[02:40] <huwshimi> hatch: Hello!
[02:42] <hatch> so did you end up figuring out your issue or do u want me to take a look?
[02:43] <huwshimi> hatch: I haven't figured it out. It appears that when I attach the event it then gets triggered by the event's method that created it!
[02:44] <hatch> clickoutside is a complicated and resource intensive event because it is a 'fake' event
[02:44] <hatch> lemme pull down your code and take a look
[02:45] <huwshimi> thanks!
[02:51] <hatch> huwshimi:  do you guys get the show 'The Dome' there?
[02:51] <huwshimi> hatch: Not that I know of
[02:51] <hatch> count yourself lucky, it's horrible 
[02:52] <hatch> some tv version of a steven king book - somehow it keeps getting renewed 
[02:52] <huwshimi> :)
[03:10] <hatch> huwshimi:  quite the issue we have here heh
[03:10] <huwshimi> hatch: Yeah, fun times!
[03:12] <hatch> so I have a fix, but it sucks
[03:12] <hatch> if you wrap the clickoutside binding in a settimeout then it's ok 
[03:13] <hatch> ohh I think I know how to fix it
[03:25] <hatch> ok I was wrong lol
[03:26] <huwshimi> hatch: oh :(
[03:26] <hatch> the problem is how the clickoutside event handler works....
[03:27] <hatch> it's a synthetic event 
[03:31] <huwshimi> hatch: This is only an issue because the button that triggers the open is outside the menu
[03:35] <huwshimi> (That's not something that we can change, however)
[03:37] <huwshimi> hatch: Any other ideas?
[03:38] <hatch> huwshimi:  yeah so I'm thinking that's what will have to be done
[03:38] <huwshimi> hatch: The timeout?
[03:38] <hatch> the clickoutside handler is just not written properly 
[03:38] <hatch> well that was the wrong thing
[03:38] <hatch> it's not written to be used as we are using it
[03:39] <huwshimi> hatch: OK, well thanks for taking a look.
[03:39] <hatch> huwshimi:  I guess right now in the interest of getting this landed is to render the menu for each token 
[03:40] <hatch> er
[03:40] <hatch> render the ... in the menu 
[03:40] <hatch> then when it gets clicked, show the real menu
[03:40] <hatch> it's a big performance smack but the real fix is going to take too much time
[03:40] <huwshimi> hatch: swap out button with the real one?
[03:40] <huwshimi> the
[03:40] <hatch> no I mean the ... is part of the menu
[03:40] <hatch> widget
[03:41] <huwshimi> hatch: Oh, so just render everything all the time?
[03:41] <hatch> when it's clicked then show the tooltip portion of the menu
[03:41] <hatch> yeah I think that's going to be the only way this will work, because the clickoutside will capture everything on the various up, down, bubble event phases
[03:41] <hatch> it doesn't filter them to only catch one way unfortunately
[03:42] <huwshimi> :(
[03:44] <hatch> it can be written to work as intended
[03:44] <hatch> unfortunately it'll take too long and I think we'd like to get this landed heh
[03:46] <huwshimi> hatch: Oh, I have a solution.
[03:46] <huwshimi> let me push this up
[03:48] <huwshimi> hatch: https://github.com/huwshimi/juju-gui/commit/370e1510bb44d483da21d36f82e9e477d09a578f
[03:49] <hatch> looking
[03:50] <huwshimi> actually that 'all' could be 'one'
[03:50] <hatch> so I was thinking something like that, but it's touching the DOM x number of times 'just incase' 
[03:51] <hatch> hmm
[03:52] <hatch> huwshimi:  you know that's probably an alright idea if it uses a .one
[03:53] <huwshimi> hatch: Yeah, so this: https://github.com/huwshimi/juju-gui/commit/afd157fe01fc39488f0727335cb69b457bbb4d53
[03:53] <hatch> but it's a little bandaidie
[03:53] <huwshimi> yeah
[03:53] <huwshimi> hatch: Is it better than rendering all the widgets?
[03:53] <hatch> much :)
[03:54] <hatch> like orders of magnitude haha
[03:54] <hatch> good work
[03:55] <huwshimi> hatch: OK, I'm going to have some lunch and then add tests. This could probably do with another review tomorrow
[03:55] <hatch> yep for sure I'll review it in the am
[03:56] <hatch> thanks for staying on this :)
[03:56] <huwshimi> hatch: Thanks for that!
[03:56] <huwshimi> np
[03:56] <huwshimi> thank you
[04:07] <hatch> no problem
=== uru-afk is now known as urulama
[08:13] <rogpeppe1> urulama: morning!
[08:16] <urulama> rogpeppe1: heya
[08:17] <urulama> rogpeppe1: how are we doing today, sir?
[08:17] <rogpeppe1> urulama: good thanks
[08:17] <rogpeppe1> urulama: and you?
[08:19] <urulama> rogpeppe1: a bit sleepy, but nothing that proper indian tea could not solve :) 
[08:20] <rogpeppe1> urulama: :-)
[08:20] <urulama> rogpeppe1: could you take a look at Jay's PR, please ... https://github.com/juju/charmstore/pull/77
[08:20] <rogpeppe1> urulama: i'm just looking at it already
[08:44] <rogpeppe1> urulama: reviewed
[08:45] <urulama> rogpeppe1: thanks
[08:46] <urulama> rogpeppe1: i've set him towards the baseURL yesterday, but it was too late in the night to actually review the code
[09:13]  * urulama goes make some tea
[09:29] <urulama> rogpeppe1: ping, when you are ready
[09:29] <rogpeppe1> urulama: ok, one mo
[09:30] <rogpeppe1> urulama: ping :-)
[09:30] <urulama> https://plus.google.com/hangouts/_/canonical.com/roger-uros
[09:59] <urulama> frankban, rogpeppe1: give me 5min please, there's a post-man at the door
[10:00] <rogpeppe1> urulama: ok
[10:03] <urulama> frankban: wonna join for a blobstore talk?
[10:03] <frankban> urulama: sure
[10:04] <frankban> urulama: link?
[10:04] <urulama> https://plus.google.com/hangouts/_/canonical.com/roger-uros?hceid=dXJvcy5qb3Zhbm92aWNAY2Fub25pY2FsLmNvbQ.r0oegfsr6b313oqhn3br43ltqs
[10:05] <urulama> oh, sorry
[10:05] <urulama> frankban: https://plus.google.com/hangouts/_/canonical.com/blobstore?hceid=dXJvcy5qb3Zhbm92aWNAY2Fub25pY2FsLmNvbQ.i8ktli6o0tbmsbe5qmtk62eaik
[10:43] <urulama> frankban, rogpeppe1: google does not like us :)
[10:44] <rogpeppe1> urulama: we're back in...
[11:01] <frankban> rogpeppe1: is it better to compare two *charm.Reference by value (e.g. *id1 == *id2) or by str repr (id1.String() == id2.Strng())?
[11:02] <rogpeppe1> frankban: value
[11:02] <frankban> rogpeppe1: cool thanks
[11:04] <rick_h__> morning
[11:06] <frankban> morning rick_h__ 
[11:06]  * frankban lunches
[11:15] <urulama> morning rick_h
[11:49] <bac> morning
[11:50] <rick_h__> morning bac
[11:51] <urulama> rogpeppe1, frankban: wrote minutes about blobstore. please take a look if anything is missing, lots of topics were covered :)
[11:51] <urulama> rick_h__: you have a minute to chat?
[11:51] <rick_h__> bac: this was funny last night. https://twitter.com/chcholman/status/501906093136969729
[11:51] <rick_h__> urulama: sure thing
[11:52] <rogpeppe1> urulama: looking
[11:52] <urulama> rick_h__: daily standup? https://plus.google.com/hangouts/_/canonical.com/daily-standup?hceid=cmljay5oYXJkaW5nQGNhbm9uaWNhbC5jb20.j0rk5d371ph8331ijtf48t2uj0
[11:52] <rick_h__> loading
[11:52] <urulama> rick_h__: https://plus.google.com/hangouts/_/g4sy3treomztkk367qiounpbg4a
[11:53] <bac> rick_h__: so you know that person on twitter?
[11:53] <urulama> rick_h__: the first one was "borked", hope the second one works
[11:54] <rick_h__> urulama: trying
[11:54] <rick_h__> bac: so she's in a city nearby and we 'know' each other by knowing some of the same tech people
[11:54] <rick_h__> bac: never met
[11:54] <rogpeppe1> urulama: i hadn't heard of CADF before
[11:55] <urulama> rogpeppe1: ah, its just some standard for auditing in clouds
[11:55] <rogpeppe1> urulama: have you got a good overview link by any chance?
[11:55]  * rogpeppe1 is wary of standards
[12:08] <frankban> urulama, rogpeppe1, guihelp: I need two reviews for https://github.com/juju/charmstore/pull/78 . anyone? thanks
[12:09] <rogpeppe1> frankban: on it
[12:09] <frankban> thanks
[12:17] <urulama> frankban: will be on it after a short lunch
[12:17] <frankban> urulama: thank you
[13:38] <hatch> morning all
[13:39] <rick_h__> morning hatch 
[13:39] <hatch> kadams54:  so we also had a monstrous storm last night :) took the power out :)
[13:40] <kadams54> yippee
[13:45] <hatch> http://www.quora.com/How-good-is-YUI-as-an-open-source-JavaScript-library-compared-to-current-alternatives-2014?srid=dOo&share=1
[13:45] <rick_h__> " But now that JavaScript itself will soon have a built in module system and loader, the way forward is to embrace ES6 modules.
[13:45] <rick_h__> "
[13:45] <rick_h__> "will soon" bothers me
[13:46] <rick_h__> :( on the lack of noting the niceities of a deep event system, common api across code, etc. 
[13:46] <rick_h__> but yea, it's running its course
[13:47] <hatch> heh there is no way we'll be able to use es6 modules/loader any time soon
[13:47] <hatch> maybe 1yr
[13:47] <rick_h__> maybe
[13:48] <hatch> My next personal project whatever that may be will not use YUI so I can try and find a good collection of modules to provide a similar functionality 
[13:48] <hatch> so far I can't find anything that has the event stack :/
[13:48] <hatch> those darn events.....
[13:49] <hatch> maybe I could write something
[13:49] <rick_h__> yea, I bet you still end up writing a lot more code
[13:49] <rick_h__> it's kind of sad that there's not more of a desire to modernize YUI vs deprecate
[13:49] <hatch> Y! is too busy buying mobile companies....lol
[13:59] <luca> rick_h__: just rebooting
[14:43]  * rick_h__ heads back to the house
[14:44] <hatch> kadams54:  did you end up doing the review of my branch yet?
[14:44] <kadams54> hatch: in progress. I'm also having problems with changeState in my branch.
[14:45] <kadams54> So I'll finish off QAing yours and then pester you with state questions ;-)
[14:47] <hatch> yeah sure np
[14:47] <hatch> changeState is really easy to understand and debug once you wrap your head around how it works
[14:47] <luca> hatch: jcastro kadams54 btw is the uncommitted progress icon just the yellow triangle? The yellow triangle technically means pending...
[14:50] <hatch> luca:  right now there is no uncommitted progress icon
[14:51] <hatch> there is the pending icon but that's after juju has ACK'd it and it's really pending
[14:51] <hatch> but there is this odd place between uncommitted and the ACK
[14:51] <hatch> which I have seen be up to 30s 
[14:51] <hatch> where it's not pending or not uncommitted 
[14:56] <rick_h__> jujugui call in 5
[14:56] <rick_h__> or 4
[14:56] <rick_h__> kanban away please
[14:58] <urulama> rick_h__: i get "this party is over" :(
[14:59] <rick_h__> jujgui https://plus.google.com/hangouts/_/gw35roelf4huu6vmyvyl3z6rgea?authuser=1 
[14:59] <luca> hatch: I see, thanks
[14:59] <rick_h__> see if this works
[15:00] <hatch> I'm in the normal standup room
[15:00] <hatch> with ant and matt
[15:00] <rick_h__> ok, it's working now
[15:00] <rick_h__> just doens't like me
[15:00] <rick_h__> jujugui standup time!
[15:00] <urulama> rick_h__: https://plus.google.com/hangouts/_/canonical.com/daily-standup?hceid=cmljay5oYXJkaW5nQGNhbm9uaWNhbC5jb20.j0rk5d371ph8331ijtf48t2uj0
[15:00] <urulama> rick_h__: try this one
[15:01] <rick_h__> bac: ^
[15:01] <bac> oops
[15:01] <bac> joining
[15:03] <bac> having trouble
[15:10] <hatch> rick_h__:  are all of the cards in the ready to code of equal priority? It seems like some of these may be able to be pushed to post 1.0 
[15:10] <bac> jujugui: sorry, i cannot get into the hangout
[15:11] <hatch> it's over
[15:11] <hatch> :)
[15:17] <bac> hatch: miss anything good
[15:17] <bac> ?
[15:18] <hatch> rick_h__:  did some break dancing to lady had a little lam
[15:18] <hatch> lamb
[15:20] <kadams54> guihelp: where is views.Templates originally defined/setup?
[15:20] <rick_h__> hatch: we can debate over them
[15:20] <rick_h__> hatch: but they all seemed 1.0 stuff when I looked monday
[15:20] <rick_h__> kadams54: it's built from the /lib/templates or something, /me pulls up a browser to look
[15:21] <rick_h__> kadams54: sorry, the things in app/templates is built into lib/templates.js as compiled handlebars
[15:22] <rick_h__> kadams54: what are you trying to do?
[15:24] <urulama> jujugui: bye, see you later
=== urulama is now known as uru-afk
[15:59]  * rick_h__ goes to get lunch
[16:27] <hatch> jujugui do we have any assets or mockups for ""handle changes from the delta stream and alert with conflicts in ecs"" ?
[16:27] <kadams54> not that I've seen
[16:27] <rick_h__> hatch: no
[16:28] <hatch> I can work on the backend of it np, but I'll send something to luca about it I guess
[16:29] <rick_h__> hatch: yea, I think that card is one that breaks into smaller. we need to plan out where/how to keep the 3 results (current, ghost, delta) and how to resolve that
[16:29] <hatch> yep ok on it
[16:44] <rick_h__> hatch: thanks for the great qa notes on huw's branch!
[16:44] <hatch> :) np
[16:48] <Makyo> jujugui it appears we can't clear config changes, because we don't retain the old ones, just a list of what has changed.  Any thoughts?
[16:48] <rick_h__> Makyo: chat with hatch as I think we need it for that as well
[16:48] <hatch> hmm
[16:48] <rick_h__> Makyo: hatch quick chat?
[16:48] <hatch> yeah.....standup?
[16:49]  * rick_h__ will try again wheee
[16:49] <hatch> if it says it has ended just click refresh a couple times
[16:49] <rick_h__> went to FF
[17:02] <frankban> rogpeppe1: so, should we return nil, nil in a meta endpoint if, for instance, it makes sense only for charms and a bundle is passed?
[17:03] <frankban> rogpeppe1: IIUC, if the meta result is nill, then the result is not included in meta when include is used
[17:04] <rogpeppe1> frankban: yes, that's right.
[17:04]  * rogpeppe1 needs to stop now. wedding anniversary requires prompt stopping :)
[17:04] <rogpeppe1> g'night all
[17:05] <frankban> rogpeppe1: good night
[17:06] <rick_h__> happy anniv rogpeppe1 !
=== uru-afk is now known as urulama
[18:23] <hatch> jcsackett:  thnks for the review, I replied
[19:20] <hatch> kadams54:  review?
[19:21] <kadams54> hatch: yeah, still wrestling with a local env
[19:21] <hatch> hmm, what issues are you having? I've probably had it before haha
[19:21] <kadams54> hatch: the behavior was funky when I ran through, so I wanted to verify and make sure it wasn't just my setup.
[19:22] <hatch> oh ok, what was happening?
[19:23] <kadams54> hatch: the machine would go through OK, but the bare metal container would be left as uncommitted state. I had to force a re-render on the container token before it properly displayed the state.
[19:23] <hatch> oh ok that's possibly a real bug
[19:23] <hatch> actually probably is
[19:23] <kadams54> hatch: So I ran through the QA instructions, deployed, confirmed, waited, then saw the machine change from uncommitted to deployed. But the container stayed as uncommitted.
[19:24] <hatch> the uncommitted state rendering is kind of implemented funky
[19:24] <hatch> the bare metal cannot be uncommitted so that's definitely a bug
[19:24] <kadams54> As far as my local env woes… I think it's the same thing jcsackett's seen. I've run the juju set command, it seems to run successfully, then I refresh in the web browser and nothing is running anymore.
[19:25] <kadams54> Annnnnddd now it works fine. Of course.
[19:25] <hatch> are you juju setting the source?
[19:25] <jcsackett> kadams54: on precise you have to use sha to set the source. 
[19:25] <jcsackett> On trusty you can use branch. 
[19:25] <hatch> it takes time to pull down and build the source
[19:25] <kadams54> Hmm, I'm on precise, but it just worked with the branch…
[19:25] <hatch> and what jcsackett just said
[19:26] <hatch> ( which I also learnt just recently)
[19:26] <kadams54> hatch: yeah, I gave it 10 minutes :-)
[19:26] <jcsackett> hatch, I'm not sure I'm +1 with your reply; you say it's overkill to put update id on the machine list, but that's the only place you use the method. 
[19:26] <jcsackett> Seems like overkill to patch, to me. 
[19:27] <hatch> I just figure because it's a limitation of the library and not of our code it should go to the lib
[19:27] <hatch> kadams54:  can you write that bug in the PR and I'll look into it after lunch
[19:27] <jcsackett> I agree it's unnecessary to create our own extension, but can't you just add update I'd for the machine list? We already have one as an object, right?
[19:27] <kadams54> hatch: Yeah, now that I've finally managed to confirm it :-)
[19:28] <kadams54> hatch: FYI, I added the changeState code pretty much as it is in scale-up.js, for navigating to machine view… but it seems to do nothing. That is, if you're in service view when you click on the link, you'll still be in service view.
[19:29] <hatch> jcsackett:  no LML has an internal ID map which is only updated on add and remove so you cannot update the id of a LML object and still have it functional 
[19:29] <jcsackett> Aaaah. 
[19:29] <jcsackett> That's the missing part of info. 
[19:29] <hatch> kadams54: does the service view bubble the change up to the browser.js?
[19:29] <hatch> jcsackett:  oh sorry :) 
[19:29] <hatch> I'll add that to the PR
[19:30] <kadams54> Ah, good question. I'll make sure it does.
[19:31] <hatch> kadams54:  typically you will include the steps to reproduce the QA failure
[19:31] <hatch> even though I know how....just for good measure
[19:31] <jcsackett> ok. hatch, knowing that, I'm good. Can you add a test for the patch's added method, though?
[19:31] <hatch> jcsackett:  I can
[19:31] <hatch> plz add to the PR
[19:33] <kadams54> hatch: done.
[19:33] <jcsackett> Will do. Thanks, hatch. 
[19:33] <hatch> thanks - I'll get on that right after lunch
[19:33] <hatch> I'm trying to figure out a good strategy to port the gui to Dart without rewriting :D
[20:55]  * rick_h__ runs away night all!
[21:18] <hatch> cya rick_h__
[21:18] <hatch> oh that was like 25mins ago
[21:18] <hatch> heh
[21:37] <lazyPower> heyyyyyy hatch
[21:37] <hatch> wassup boi!
[21:38] <lazyPower> Wanna be a peach and proof something for me? (rough video cut, but almost done)
[21:39] <hatch> yeah sure 
[21:39] <hatch> I'm just finishing up a review
[21:39] <lazyPower> https://www.youtube.com/watch?v=CEfFy6tODrQ knew I could count count on ya! ty ty
[21:39] <lazyPower> oooo you're doing a charm review?
[21:42] <hatch> haha no, js review
[21:42] <hatch> lazyPower:  I'll start looking at it in about 5
[21:43] <lazyPower> no rush :) and ooohhh yeah. i almost got excited there for a second.
[21:48] <hatch> lazyPower:  your quiet...
[21:48] <hatch> can you preamp the audio?
[21:48] <lazyPower> that has been preamped
[21:49] <lazyPower> Whats your youtube volume at perchance?
[21:49] <lazyPower> i preamped it and ran it through a compressor
[21:49] <hatch> 100% heh
[21:49] <lazyPower> wow
[21:49] <hatch> and my laptop is also at 100%
[21:49] <lazyPower> really?
[21:49] <lazyPower> oi
[21:49] <hatch> :) maybe it's just my computer
[21:49] <hatch> I'm running in a vm
[21:49] <lazyPower> let me solicit additional feedback - you're on mac right?
[21:50] <hatch> I'm on a mac, in Ubuntu
[21:50] <hatch> lol
[21:50] <hatch> maybe my soundcard default volume is set to low :)
[21:50] <hatch> anyways, continuing
[21:52] <lazyPower> hatch: ty for the feedback though - if i need to up the amps on the audio i'll try to do that on teh final copy
[21:52] <lazyPower> s/amps/decibels/
[22:02] <hatch> lazyPower:  notes pm'd
[22:04] <lazyPower> Thanks man! Much appreciated.
[22:09] <hatch> anytime :)
[22:10] <hatch> When it's done let me know so I can spread it among my network too
[22:10] <hatch> anything to get the juju love out there
[22:39] <jcsackett> hatch: how does one turn the simulator off in gui?
[22:39] <jcsackett> it's massively mucking me up.
[22:39] <hatch> ctrl s
[22:39] <jcsackett> thank you.
[22:39] <hatch> "One does not simply....turn the simulator off"
[22:39]  * jcsackett sort of wants the simulator off by default
[22:39] <hatch> jcsackett:  you can also set it to off in the config-debug.js file
[22:39] <jcsackett> yes, but then i have to fix that before i commit.
[22:39] <hatch> I've been asking for that but always get voted down
[22:40] <jcsackett> i might bring it up on friday (again)
[22:40] <hatch> come to the dark siiiide
[22:40]  * hatch hands you a pin
[22:40] <jcsackett> it is *so* annoying when you're dealing with machine creation/unit creation etc.
[22:40] <hatch> yop
[22:40] <hatch> I'll +1 your vote to banish
[22:41] <hatch> now and convince others behind closed doors like a good politician and make it look like a true democracy when we vote
[22:41] <hatch> lol
[22:41] <hatch> rofl
[22:41] <hatch> oh I crack myself up
[22:53]  * jcsackett laughs
[23:06] <huwshimi> Morning
[23:08] <hatch> morning huwshimi
[23:09] <hatch> lots of comments on your branch but good to land once those are taken care of (they are all tidy up things)
[23:20] <rick_h__> morning hju
[23:20] <rick_h__> mornin huwshimi 
[23:20] <rick_h__> tab completion works better when you type correctly :/
[23:21] <huwshimi> haha
[23:24] <hatch> :D
[23:41] <huwshimi> rick_h__: If you happen to have the time, I'd love another review of the more menu: https://github.com/juju/juju-gui/pull/498
[23:43] <rick_h__> huwshimi: will look in a sec here. 
[23:44] <hatch> huwshimi:  thanks for making those fixes
[23:44] <hatch> be sure to rebase before landing :)
[23:47] <huwshimi> hatch: Thanks for the review, yeah, lots of changes in there :)
[23:47] <hatch> heh mine is the same, I'm going to have to have fun rebasing that one