[00:10] <rick_h_> widox doh! I see the query problem. Stupid stupid
[02:09] <hatch> huwshimi: hey how goes the battle?
[02:12] <huwshimi> hatch: Just doing some sandbox qa at the moment
[02:14] <hatch> huwshimi:  cool - you should go through the guide on setting up juju for ec2 and doc any issue you have
[02:17] <huwshimi> yep will do
[05:55] <urulama> morning all
[07:11] <rogpeppe1> urulama: morning!
[07:11] <urulama> rogpeppe1: hey
[07:14] <urulama> rogpeppe1: how are you? got rid of that cold?
[11:21] <rick_h_> morning party people
[13:35] <rick_h_> kadams54: around?
[13:35] <kadams54> rick_h_: yup
[13:35] <rick_h_> can you join me in the standup room?
[13:37] <kadams54> Sure
[13:38] <kadams54> daily or friday?
[13:38] <rick_h_> I'm in friday 
[13:38] <rick_h_> I can jump over if you did the other one
[13:50] <frankban> guihelp: I need two reviews + 1 QA for https://codereview.appspot.com/136570043 (GUI charm). anyone available?
[13:50] <rick_h_> frankban: looking
[13:51] <frankban> thanks
[14:08] <jcsackett> rick_h_: i have discovered a problem with updating bundles for machine placement. there doesn't appear to be away to tell a unit it's hosted on a machine with another unit of itself.
[14:08] <jcsackett> all the "to" syntaxes are in terms of put this unit on a machine with this other service.
[14:09] <rick_h_> jcsackett: ok, well then let's ignore that one for now and we'll have that updated in the updates with deployer/charmstore/gui export
[14:10] <jcsackett> ok, so when we have colocated units of the same service, just export as we do now (i.e. ignore placement?)
[14:10] <rick_h_> I think it's rare enough that there's value in the current work if it doesn't work 100% for folks that put two of the same thing colocated. I'm not sure how that works really. 
[14:10] <rick_h_> yea, ignore placement and put them on their own 
[14:10] <jcsackett> rock on. i can do that. :p
[14:10] <rick_h_> you can't really run two mysql or whatever on the same machine anyway
[14:10] <rick_h_> so I think the issue is small at first glance
[14:11] <jcsackett> rick_h_: that's an excellent point.
[14:11] <jcsackett> might become an issue with lxc placement, but that's still a weird one.
[14:12] <rick_h_> right and that's not available for most providers yet so we have time to update our formats/story by then
[14:13] <rick_h_> frankban: code looks good, starting trusty qa/test run/review. 
[14:13] <rick_h_> frankban: so if the other reviewer could do precise we can split it up
[14:13] <frankban> rick_h_: thanks! guihelp ^^^
[14:20] <rick_h_> kadams54: comments added, I think we're missing some functionality. Let me know if you've got questions 
[14:21] <kadams54> rick_h_: wowza, you were on that. I realized I missed some commits and am pushing them now.
[14:21] <rick_h_> kadams54: cool thanks
[14:22] <kadams54> rick_h_: I'll also have QA instructions on it shortly :-)
[14:22] <rick_h_> kadams54: rr
[14:22] <rick_h_> rgr
[14:23] <rick_h_> bah
[14:24] <rick_h_> hatch: heh, YUI Mail -> reactjs
[14:24] <rick_h_> makes sense
[14:27] <hatch> rick_h_: hah yeah....I've written Y! off
[14:27] <hatch> the way they treated teh community I'm even scared to use purecss
[14:28] <rick_h_> hatch: :/
[14:29] <hatch> "a developer scorned"
[14:34] <hatch> issue is that pure I think is the best out ther haha
[14:34] <rick_h_> yea, same here
[14:50] <kadams54> guihelp: https://github.com/juju/juju-gui/pull/546 is ready for QA and review.
[14:50] <hatch> I can look kyle
[14:51] <jrwren> standup in 9? is it really this early in the day yet?
[14:52] <kadams54> thanks
[14:52] <hatch> kadams54: did you address rick_h_'s message?
[14:52] <kadams54> hatch: yes
[14:52] <hatch> jrwren: haha what time is it there?
[14:52] <jrwren> hatch: 8 till 11am.
[14:53] <hatch> wha?
[14:53] <rick_h_> jujugui call in 8
[14:54] <jrwren> it feels later than that. busy morning, that is all I'm saying.
[14:54] <hatch> ohhhhh 
[14:56] <urulama> jrwren: one of those days :) i finished lunch not long ago (and it's 5PM here :D)
[14:57] <jrwren> :)
[14:57] <rick_h_> frankban: trusty is good, jujugui need a second live env qa/test run of the charm on precise please
[14:57] <rick_h_> jcsackett: you were stabbing at ec2 recently, up for background'ing a test run for a bit?
[15:00] <jcsackett> rick_h_: i'm going to need my ec2 env to throw generated bundle files at in the not too distant future--how long are we talking?
[15:01] <rick_h_> ant__: ^ 
[15:54] <urulama> hatch, frankban: FYI, parallels now support utopic as well
[15:55] <hatch> urulama: like....actually supports it? Or just claims they do?
[15:56] <urulama> hatch: there was an update today and now it works, with HW acceleration and all, just as trusty
[15:56] <hatch> wow they might finally be getting their shit together
[15:58] <urulama> hatch: naaa, it'll break with next beta, don't worry :)
[16:00] <hatch> lol
[16:07] <frankban> urulama: yesterday for quickstart QA i have been working in ssh from my working vm to other three vmware machines (precise, trusty and utopic) + an ssh to the host for osx QA. It's been crazy but worked really well ;-)
[16:14] <urulama> frankban: as long as you have 16gb in that machine, it works great
[16:14] <hatch> haha
[16:14] <urulama> frankban: i also like such workflow, and when done, just throw away the VM and make a next clone of a "clean" VM
[16:15] <frankban> urulama: exactly
[16:16] <urulama> frankban: maas without maas :D
[16:19] <urulama> jujugui: enjoy, have fun. bye
[16:40] <hatch> kadams54: so any luck in the css stuff? I see you updated the tests
[16:40] <kadams54> hatch: yeah, should have that landed shortly.
[16:43] <hatch> Makyo: can we close #535?
[16:43] <hatch> jujugui Makyo needs another review on https://github.com/juju/juju-gui/pull/545
[16:44] <Makyo> hatch, yes
[16:44] <kadams54> hatch, Makyo: yeah, I'm working on that
[16:45] <hatch> Makyo: closed
[16:58] <kadams54> Makyo: done with https://github.com/juju/juju-gui/pull/545
[17:03] <kadams54> hatch: where is the more menu used at? Want to make sure my stuff doesn't cause regressions…
[17:03] <kadams54> Oh wait, n/m, that's for the new MV menus, right?
[17:03] <hatch> yeh
[17:12] <kadams54> hatch: OK, CSS re-swizzled. Take a look and let me know what you think.
[17:19] <hatch> will do
[17:26] <hatch> kadams54: +1
[17:28] <kadams54> hatch: 'cept for the legit test failure on build :-(
[17:33] <hatch> yeah well....there is that
[17:42] <hatch> jcsackett: Makyo looks like there will be another contributor to the ghost charm https://github.com/hatched/ghost-charm/issues/25
[17:43] <Makyo> \o/
[17:55] <jcsackett> w00t!
[18:23] <kadams54> hatch, Makyo: the bug I'm working on is an uncaught exception that happens when an undeployed service is destroyed. It's happening because there are data-bound elements in the (just destroyed) inspector that are trying to update.
[18:24] <kadams54> hatch, Makyo: so… how to go about unbinding everything before destroying the service and inspector?
[18:24] <hatch> unbind()
[18:24] <hatch> :)
[18:24] <Makyo> what's the context, hatch?
[18:24] <Makyo> this?
[18:25] <Makyo> Or this.bindingEngine?
[18:26] <hatch> it's in the viewlet managert
[18:26] <hatch> destroying the inspector should unbind
[18:27] <hatch> https://github.com/juju/juju-gui/blob/develop/app/views/viewlet-manager.js#L438
[18:27] <hatch> Makyo: kadams54
[18:29] <kadams54> hatch: yeah, I can confirm that's being called before I hit the error, and yet…
[18:29] <hatch> sounds like you have a real bug on your hands ;)
[18:34] <kadams54> Could there be multiple bindindEngines?
[18:34] <hatch> kadams54:  well first I'd start by tracking the instantiations of the binding engines then make sure they are properly destroyed
[18:34] <hatch> if you get any left over...well then you know
[18:35] <hatch> beyond that....
[18:35] <hatch> probably follow the traceback in the error to see where the event is being fired from and see if maybe something isn't being detached wher eit hsould be
[18:37] <hatch> kadams54: it could be as simple as moving that unbind to the top of the destructor so that it unbinds before it destroys the views
[18:38] <hatch> and the event detachers for that matter
[18:38] <kadams54> Alright, did hatch just win the internet?
[18:39] <kadams54> Womp, womp. No, sorry.
[18:39] <hatch> well hey I can't do everything!!
[18:39] <hatch> If I always won there would be no internet for everyone else
[18:39] <hatch> I just lose on purpose....
[18:41] <kadams54> lol
[19:33] <hatch> kadams54: soooooo any luck?
[19:59] <kadams54> hatch: not yet.
[19:59] <hatch> are you understanding the databinding system?
[20:00] <kadams54> Though I went for a run to take a break from the problem, so we'll see.
[20:00] <kadams54> I think I have a very wierd understanding of it :-)
[20:01] <kadams54> I've been looking at it from the inside out, stepping through in a debugger
[20:12] <jcsackett> jujugui: do we have a good way of checking if a machine is uncommitted? is it still just check for "new" in the name?
[20:15] <kadams54> jcsackett: we really should have a isCommitted function on the model, even if it just checks for "new" in the name.
[20:15] <jcsackett> kadams54: agreed.
[20:15] <jcsackett> however, for now, "new" it is. :p
[20:15] <kadams54> :-)
[20:18] <Makyo> jujugui could use some thoughts on https://github.com/juju/juju-gui/pull/547 - namely, trying to fix https://bugs.launchpad.net/juju-gui/+bug/1360183 without jeopardizing the integrity of _dirtyFields
[20:18] <mup> Bug #1360183: When configuring a service the orange asterisk is visible after you have pressed save <juju-gui:Triaged> <https://launchpad.net/bugs/1360183>
[20:19] <hatch> jcsackett: kadams54 I'm adding that right now..."commitStatus"
[20:19] <kadams54> Yay!
[20:20] <hatch> loooking Makyo
[20:20] <Makyo> It works, but feels evil.
[20:21] <hatch> Makyo: you say mv....but the *'s are in the inspector...right?
[20:23] <hatch> Makyo: so dirtyfields is set when the values in the input no longer match the values in the config attr right?
[20:26] <Makyo> hatch, yes, but the bug is that we don't want to show them alongside the uncommitted icon.  That update doesn't happen until the callback; lazySetConfig only modifies _dirtyFields
[20:27] <hatch> so yeah...first I do not like this approach it;s kind of like a bandaid and I'm guessing you agree heh
[20:27] <Makyo> Yeah, I do.
[20:27] <Makyo> I also don't want to introduce yet another set of config things to keep track of.
[20:28] <hatch> What should happen (totally spitballing) is on save dirtyfields should be synced with the diff between environmentConfig and config
[20:28] <hatch> wait...will that work...
[20:28] <hatch> no
[20:28] <Makyo> Wouldn't that leave asterisks on all the uncommitted fields?
[20:28] <hatch> hmm
[20:28] <Makyo> Yeah
[20:29] <hatch> quick call?
[20:29] <hatch> standup
[20:29] <Makyo> Sure, one sec.
[20:30] <hatch> er friday standup i guess
[20:30] <hatch> i gots an idea
[20:53] <kadams54> hatch, Makyo: let me know when you're done chatting. I've got more databinding questions :-)
[20:54] <hatch> we were done long time ago
[20:54] <hatch> we are efficient
[20:54] <hatch> :P
[20:54] <kadams54> Hah
[20:54] <kadams54> OK, so the unbind function iterates through this._models and this._viewlets
[20:55] <kadams54> But _updateDOM iterates through this._bindings
[20:55] <kadams54> Why doesn't unbind do anything with this._bindings?
[20:56] <hatch> tbh i have no idea. it's been a long time. lemme take a look
[21:00] <hatch> kadams54: if I remember correctly bindings is just a collection of objects and doens't actually contain any listeners
[21:01] <hatch> am I way off base? What are you seeing as bindings?
[21:01] <kadams54> Well at some point, after everything's been unbound, _updateDOM gets called.
[21:01] <kadams54> At that point, it iterates through the bindings, hitting on the one for the service-overview
[21:02] <kadams54> And tries to render a new value for unit_count
[21:02] <hatch> ok and did you put a debugger in there to trace back to why it's being called after being unbound?
[21:02] <kadams54> At which point it fails
[21:02] <kadams54> The stacktrace doesn't really help
[21:02] <kadams54> http://pastie.org/private/72jfurp6pnc3ohdnkhpg
[21:02] <hatch> even with async?
[21:03] <hatch> right...i mean put a debugger in there
[21:03] <hatch> then do that before it throws the error
[21:03] <hatch> to see why updateDOM is being called 
[21:03] <hatch> because it shoudnt be
[21:03] <hatch> there are only a couple places which call it
[21:04] <hatch> kadams54: I'm thinking it's the debounce stuff on line 845
[21:04] <hatch> just another guess
[21:04] <hatch> maybe I'll win the internet with this one
[21:04] <hatch> lol
[21:04] <kadams54> Yeah, it is…
[21:05] <hatch> and holy crap can I fly with this keyboard when my fingers go where they are supposed to
[21:05] <hatch> (going where they are supposed to is the tough part however)
[21:05] <rick_h_> hatch: lol, takes a couple of weeks to really take off
[21:05] <kadams54> If you look at the very bottom of the stacktrace, line 6011 in yui.js is the implementation for later
[21:05] <hatch> OOoo
[21:05] <hatch> :)
[21:05]  * hatch can taste the internets.......
[21:06] <kadams54> Does it taste like chicken?
[21:06] <hatch> nope.....bad coffee
[21:10] <rick_h_> hatch: what ended up being ayrton's issue getting his ghost setup working?
[21:10] <hatch> rick_h_: which time? lol
[21:10] <hatch> the first time was his hardware
[21:11] <hatch> the second time was he related to mysql with some failed units and then after resolving they didn't re-run the relation script
[21:12] <hatch> so he removed the relation and re added it and then he had a truly horizontally scaled blog :)
[21:12] <rick_h_> ah coolio
[21:12] <rick_h_> yea, just saw his blog post and realized it was him in irc that day
[21:13] <hatch> oh he wrote a blog post?
[21:14] <hatch> ahh I see it now
[21:14] <rick_h_> yea, hit the planet
[21:14] <rick_h_> <3 4g at the campground since it's a little rainy 
[21:20] <hatch> haha yeah it's nice having tech in the country soemtimes
[21:25] <kadams54> guihelp: https://github.com/juju/juju-gui/pull/548 is ready for review and QA - should be an easy one.
[21:25] <Makyo> hatch, a little more elegant? https://github.com/juju/juju-gui/pull/547/files
[21:26] <hatch> kadams54: lol nice
[21:26] <hatch> Makyo: much, don't ya think?
[21:27] <Makyo> Yeah.  Think I should move forward?
[21:27] <Makyo> Simpler to test, too
[21:30] <hatch> Makyo: definmitely
[21:30] <hatch> definitely even
[21:50] <Makyo> Just needed a test modification
[21:50] <Makyo> hatch, https://github.com/juju/juju-gui/pull/547
[21:50] <hatch> cool, lookinf
[21:57] <Makyo> jujugui calling it an early day since I'm fading, but will check in over the weekend on mine/others PRs
[21:58] <hatch> cyaaaaaa