[00:26] <rick_h_> lol
[00:40] <gary_poster> heh
[00:40] <gary_poster> I wondered about that myself
[00:42] <huwshimi> hatch: Also, thanks for fixing that :)
[00:51] <rick_h_> gary_poster: did we get any word on inspector qa?
[00:52] <rick_h_> poke your head up at night and get ping'd :P
[00:52] <gary_poster> rick_h_, no.  hoping no news is good news.
[00:52] <gary_poster> :-)
[00:52] <rick_h_> cool, ok well night. 
[00:52] <gary_poster> night :-)
[11:09] <luca__> anyone here who's working on the inspector?
[11:23] <rick_h_>  luca__ yep, what's up?
[11:24] <luca__> rick_h_: hey
[11:24] <luca__> rick_h_: I need help with a error message, I'm not sure what's possible on the dev side of things
[11:24] <luca__> rick_h_: I'm just writing an email about it
[11:24] <rick_h_> luca__: sure thing
[11:24] <rick_h_> k
[11:25] <luca__> rick_h_: I just sent it, maybe you could check it out
[11:25] <rick_h_> luca__: will do
[11:26] <luca__> rick_h_: also, it's always been planned that if you edit your config and then press the close button that it should prompt you with a "You have unsaved changes, do you want to save your progress? Discard/save"
[11:27] <rick_h_> luca__: k, file a bug on that one and we can get that updated
[11:27] <luca__> rick_h_: ok
[11:27] <rick_h_> luca__: please let us know where/how we're to present that to the user 
[11:27] <luca__> rick_h_: yeah, I had it at the top but I think it looks weird
[11:27] <rick_h_> the issue is location/space. They could be scrolled around and such. I'm assuming we'd do something like the popup slide thing like on destroy service?
[11:27] <luca__> rick_h_: yeah
[11:28] <luca__> rick_h_: at the top
[11:28] <luca__> rick_h_: but I think it will look weird so Spencer is mocking it up
[11:28] <rick_h_> ok, yea please put that into the bug then to help with that. 
[11:29] <rick_h_> luca__: and we have to remember that it needs to work for both the ghost/non-ghost inspector
[11:29] <rick_h_> pre-post deploy
[11:29] <luca__> rick_h_: yup
[11:30] <rick_h_> luca__: for your email, what if we were to do something like red-border the input field and replace the description text under the field with a red error message?
[11:30] <luca__> rick_h_: oh, that could work
[11:31] <luca__> rick_h_: can we do that on the fly or would we need to have the user press save first to get the error message?
[11:31] <rick_h_> I think we can do it on the fly
[11:31] <rick_h_> but I've not looked at how the input validation is going on atm
[11:32] <luca__> ok, afaik we are only checking the string and integer thing, I haven't been told about any other error types
[11:32] <rick_h_> luca__: sure, we could also look at highlighting the (string) part of the name section as well if that helps with that issue
[11:33] <luca__> rick_h_: interesting :)
[11:33] <rick_h_> but I think if we were to add an error message "wordpress port must be an integer" where the description is it might be clear/not require space movements too much
[11:34] <luca__> yeah, I'll get spencer to mock something up\
[11:34] <luca__> thanks for the idea, it's better than the ones I've been throwing around :)
[11:34] <rick_h_> luca__: cool, yea file a bug on both of those and we'll try to get those updated asap 
[11:35] <rick_h_> luca__: other than that things looking ok for release? 
[11:38] <luca__> rick_h_: I think styling wise yes, but I've found a few UX bugs, some I think are UX stuff that hasn't been implemented at all and that the inspector is ok to launch without and the others are just minor tweaks I think.
[11:38] <luca__> rick_h_: I have them listed in a email which I was going to send once I had finished the wireframes for the save dialogues.
[11:38] <rick_h_> luca__: ok, well cool overall booo if we missed some stuff. Thanks for the run through
[11:38] <luca__> rick_h_: there isn't anything big
[12:15] <gary_poster> hey luca__ .  I was going to see if I could get the "build relation" menu to always show and hide with the inspector this morning, because you and I had talked about that and it seemed to me like a real potential annoyance, especially to a new user.  Not sure if it is a quick job or not but will try.  Do you agree with that UX prioritization or do you have something higher on your list?
[12:35] <rick_h_> gary_poster: does hatch's branch from yesterday meet the requirements for the inpsector card on "service name selection should give live feedback"?
[12:36] <gary_poster> rick_h_, yes, I think so.  It doesn't have the spinner but that's because the check is effectively instantaneous.  Do you have concerns?
[12:37] <rick_h_> gary_poster: no, just looking at the cards left and it seemed like that one could go away
[12:37] <gary_poster> rick_h_, it is on my board.  I will refresh
[12:38]  * rick_h_ didn't follow that...but ok :)
[12:38] <gary_poster> oh!
[12:38] <gary_poster> we had two for the same thing
[12:38] <gary_poster> you deleted
[12:38] <gary_poster> thank you :-)
[12:38] <rick_h_> right, I was trying to determine if they were different in some way beyond the text on the card
[12:38] <gary_poster> yeah, cool
[12:38] <gary_poster> sorry for being confused
[12:38] <rick_h_> mornings... hah
[12:38] <gary_poster> :-)
[12:39] <gary_poster> so the only ones to take have hatch's face on 'em
[12:39] <gary_poster> and only one of those is not blocked
[12:39] <rick_h_> yea, and a couple are blocked
[12:39] <rick_h_> I was giong to run through and do the charm model rename real quick and maybe luca__ will have the updates they want after that
[12:40] <rick_h_> if you read the back scoll, they've got a couple of outstanding things they're working on an email about, but not sent yet
[12:40] <gary_poster> ok.  that one kinda scares me right before release but it is probably fine.  s/BrowserCharm/Charm/g seems safe on the face of it
[12:40] <gary_poster> yeah I saw
[12:41] <rick_h_> yea, the idea was to run one full release before we change the names back to catch anything
[12:41] <rick_h_> and so we're about ready for a big release, been a couple of them now actually
[12:41] <gary_poster> oic
[12:41] <gary_poster> we made a release with everything except for that change you mean, right?
[12:41] <gary_poster> everything on that refactor path, I mean
[12:41] <rick_h_> right, we've left that change back in case anything poppuled up on the old Charm name
[12:41] <gary_poster> ok cool
[12:41] <gary_poster> yeah, that's a good one
[12:42] <rick_h_> popped...wtf is poppuled
[12:42] <gary_poster> :-)
[12:56] <luca__> gary_poster: That's fine to work on, it's an annoyance hehe, I'm just about to send some inspector bugs via email to you, to get it all wrapped up./
[13:08] <gary_poster> cool thanks luca__ .  That bug may turn out to be bigger than I I have time for--it may be that reimplementing the new relation UX will be a better use of time once we get there, because the inspector and the current menu implemention are too disconnected with one another right now.  investigating a bit more still
[13:09] <luca__> gary_poster: no worries, the new UX for building a relation is much better anyway :P
[13:09] <gary_poster> luca__, agreed.  we don't have wireframes for that yet, do we?  I know you showed me but I don't recall where
[13:10] <luca__> gary_poster: not yet, Mark S didn't sign the design off :(
[13:10] <gary_poster> ah :-(
[13:11] <luca__> gary_poster: he signed the UX off but not the design, so it's most probably not worth implementing
[13:11] <gary_poster> right
[13:38] <hatch> morning
[13:40] <gary_poster> morning hatch
[13:41] <hatch> how does juju deal with load balancers?
[13:41] <rick_h_> how so?
[13:41] <hatch> I don't think I've ever heard anyone mention anything about them
[13:41] <rick_h_> for juju itself? there are charms for things like haproxy
[13:41] <gary_poster> hatch, you  deploy a service that is a load balancer and hook it up
[13:41] <gary_poster> yeah
[13:41] <hatch> would that be a charm level feature?
[13:41] <gary_poster> no
[13:41] <gary_poster> well
[13:41] <gary_poster> not really
[13:41] <rick_h_> hatch: well you'd run the haproxy charm, which has interfaces for load balancing things like a web service
[13:42] <gary_poster> a charm has to connect to the right interface.  Like Rick says
[13:42] <hatch> ahh ok gotcha
[13:43] <hatch> I was just thinking because we add all these units but we never connect it via anything to split up the requests :)
[13:44] <gary_poster> luca__, thank you very much for the detailed feedback!  Some of them we've talked about before as being technically difficult/impossible now.  I'll reply with details back, and we can start working on the rest.  Do I understand correctly that you don't regard any of these as show stoppers though--we could in theory launch as is and you would be OK with it, with the knowledge that we would be improving subsequently? 
[13:44] <gary_poster>  If that's not the case, I'd like to work with you to categorize and prioritize
[13:44] <rick_h_> yea, so charmworld in production is meant to have some squid proxying and such to do that
[13:44] <gary_poster> hatch also some charms balance themselves or have cleverness built in
[13:44] <gary_poster> if a load balancer is sufficient, then great
[13:44] <gary_poster> but, for instance, databases have to be a lot smarter
[13:44] <gary_poster> and the charms support that
[13:45] <gary_poster> I think our mysql charm is pretty good, and we have our postgres expert working on the postgres charm so I am guessing that it rocks but have no firsthand knowledge of that
[13:45] <rick_h_> jujugui I need two reviews and a QA of this mechanical branch to bring jcsackett's model awesomeness work to an end please. https://codereview.appspot.com/13273047/
[13:45] <hatch> ahh ok ok
[13:46] <rick_h_> https://jujucharms.com/precise/haproxy/#bws-readme has some notes on using it for instance. 
[13:47] <hatch> I went kiting last night and tangled up my kite so I had a lot of time to think about juju while floating to shore :P
[13:47] <rick_h_> lol
[13:47] <gary_poster> :-)
[13:47] <rick_h_> isn't it chilly these days? I've starte to see boats around here getting covered/shrink-wrapped
[13:47] <rick_h_> 35F this morning
[13:47] <gary_poster> rick_h_, I'll do code review.  then going to go back to trying to reply to email
[13:47] <luca__> gary_poster: the only ones that I think are high priority is the Quick action notification bugs, apart from the resolved unit bug.
[13:48] <luca__> gary_poster: I think the terminology is the most important thing
[13:48] <hatch> rick_h_: yeah the mornings/nights are cold but the days are still warm +20C. The bodies of water are pretty big so it takes them a while to cool off
[13:49] <luca__> gary_poster: everything else is ok to do post release
[13:49] <gary_poster> ack luca__, thank you.  hatch, IIRC there was a technical issue on this quick action notification item.  could you comment, please?
[13:49] <gary_poster> >       o There should be multiple error tabs, each type of error should
[13:49] <gary_poster> >         be pulled out into it's own Quick action notification, for
[13:49] <gary_poster> >         example, 10 Start errors, 10 relation errors, etc.
[13:51] <hatch> split them up in the unit overview?
[13:52] <bac> rick_h_: do you need another reviewer?
[13:52] <gary_poster> hatch, yes, main inspector
[13:52] <rick_h_> bac: yes please
[13:52] <bac> okeydokey
[13:52] <rick_h_> bac: thanks
[13:52] <gary_poster> bac could you be qa dude?
[13:53] <bac> sure
[13:53] <gary_poster> bac for rick's branch I mean.  thank you
[13:53] <hatch> gary_poster: yes there was, just pulling up the code to see what it was
[13:53] <gary_poster> thanks hatch
[13:54] <gary_poster> rick_h_, trivial comment, but you could also adjust file names in this branch if you wanted.  e.g. test_browser_charm_details.js -> test_charm_details.js.  <shrug>
[13:54] <rick_h_> gary_poster: sure thing
[13:55] <gary_poster> rick_h_, though in that case the file name might be intended to be test browser: charm details.  there is also a test_browser_search_view.js, for instance
[13:56] <rick_h_> gary_poster: true, I think some of it is "this is browser code" and some of it was just naming collisions when there were two models
[13:56] <gary_poster> so rick_h_ whatever you think
[13:56] <jcsackett> abentley: so the diff for all the work is 665 lines, or you can look at each thematic bit if you want to look at the pipes. have a preference? i'll create an MP for the entirety of it if you would prefer.
[13:57] <abentley> jcsackett: I'll look at the whole thing.
[13:57] <hatch> gary_poster: right now the go backend tells us when a unit is in error but doesn't tell us why without parsing the string which is very fragile, they are working on providing us an error object so that we have a proper dataset to act on cc) luca__
[13:57] <gary_poster> hatch thanks.  right so luca__ we are working on that one in Juju Core but Juju does not allow us to do it yet in a safe way
[13:59] <jcsackett> abentley: dig.
[14:02] <gary_poster> rick_h_, code LGTM.  jcsackett, we almost have your great BrowserCharm/Charm unification work past the very last step thanks to rick_h_ doing the last s/BrowserCharm/Charm/g.  Yay!
[14:03] <jcsackett> gary_poster, rick_h_: huzzah!
[14:03] <gary_poster> :-)
[14:03] <rick_h_> thanks for the review gary_poster 
[14:03] <gary_poster> np, thanks rick_h_ 
[14:03] <rick_h_> gary_poster: I'm not seeing this email from luca. Is there a next thing to start getting done to unblock release? or start fixing for post release?
[14:04] <rick_h_> gary_poster: and was this email sent to something non-peeps or is my email fubar
[14:04] <gary_poster> rick_h_, sent to peeps, titled "Inspector bugs"
[14:04] <gary_poster> can forward?
[14:04] <gary_poster> I can, if you want, I mean
[14:04] <rick_h_> gary_poster: thanks, will check the gmail spam folder I guess. I got his input validation one this morning 
[14:05] <rick_h_> oh look, bunch of messages in this spam folder. You spammer luca__ ! :P
[14:05] <rick_h_> gary_poster: got it, thanks
[14:06] <gary_poster> rick_h_, heh.  I'm writing a reply to the email, but the only blockers are these small ones, assuming luca__ is OK with the technical limitation Jeff mentioned above that we are trying to work around for later.
[14:06] <gary_poster> >       o "XX Running" should read "XX running units"
[14:06] <gary_poster> >       o "XX Pending" should read "XX pending units"
[14:06] <gary_poster> >       o "XX Needs reboot" should read "XX machine need to be restarted"
[14:06] <gary_poster> >       o "XX Security upgrade" should read "XX security upgrades available"
[14:06] <rick_h_> gary_poster: ok, so is preference to grab something from bundles or a small item from this list. 
[14:06] <gary_poster> rick_h_, branch fixing those plus any other small things from his email that you can throw in quickly would be great
[14:06] <rick_h_> gary_poster: rgr
[14:07] <gary_poster> thank you
[14:07] <luca__> gary_poster: that's fine with me, it'll get in there some day :)
[14:07] <gary_poster> :-) cool luca__ thanks.  I'm writing a reply with status of all of those you raised so we can refer back to them later.
[14:08] <gary_poster> hatch, luca__ , I thought we had UX direction to no longer have the "hide help text" button.  Do either of you remember anything in that vein?
[14:08] <gary_poster> >       o The "hide help text" button in the configuration section is missing.
[14:09] <hatch> yes that was removed
[14:09] <hatch> as per ux ;)
[14:10] <gary_poster> luca__, so yeah, we thought you explicitly told us that you had changed your mind, and not to do that
[14:10] <luca__> gary_poster: oh, right, never mind that one then. I didn't have that catalogued in wireframes :(
[14:10] <gary_poster> cool thanks luca__ 
[14:20] <gary_poster> luca__, another one.  You said this:
[14:20] <gary_poster> >       o I haven't seen any resolved unit notifications pop up, I imagine
[14:20] <gary_poster> >         that is because it doesn't work in the simulator.
[14:21] <gary_poster> I'm not sure what you mean by resolved unit notifications and don't see them in the wireframes.  What am I missing?
[14:22] <gary_poster> luca__, last one:
[14:22] <gary_poster> >       o When you press "Save" to save any changes the input fields flash
[14:22] <gary_poster> >         green, this shouldn't happen. It should show the green checkmark
[14:22] <gary_poster> >         for a short period of time.
[14:23] <gary_poster> I had mentioned to you at the sprint that a fading background checkmark didn't work well with CSS animations, and I wanted to do the green field flash, and you said that was OK.  Maybe I didn't explain clearly?
[14:26] <bac> waiting to hear if my airbnb reservation will be accepted by the hosts.  it's like waiting to be picked for the kickball team.
[14:26] <rick_h_> bac: lol, did you send them a gift basket? I hear that helps smooth the selection process
[14:26] <rick_h_> bac: I'll be curious to hear how that goes sometime. I've never had the guts to mess with it.
[14:27] <bac> no, but i did get a recommend from tom berger.  that's got to be worth something.
[14:27] <bac> rick_h_: we've used it a lot.  never had this review process before.
[14:27] <rick_h_> bac: interesting
[14:28] <bac> i want to put our guest suite on there to rent.  but i've been overruled.
[14:29] <jcsackett> abentley: new MP up with info about the whole mess. https://code.launchpad.net/~jcsackett/charmworld/better-stats-window/+merge/186044
[14:39] <abentley> jcsackett: looking.
[14:41] <gary_poster> rick_h_, LGTM with one comment
[14:42] <rick_h_> gary_poster: hah, I hadn't put it up yet. Just finished my proposal comments :P
[14:42] <gary_poster> rick_h_, oh, sorry :-)
[14:42] <rick_h_> gary_poster: and my proposal comments covered your ? :P
[14:42] <gary_poster> rick_h_, yes, thinking about it
[14:43] <rick_h_> gary_poster: yea, I originally updated it to keep in sync with the rest of the boolean/checkboxes. I didn't want to worry about the html/JS issues with having this one be different. 
[14:44] <rick_h_> I could try to scope the .hidden-checkbox to be inside of a config div or something, but they're generated the same so that they share styles
[14:44] <gary_poster> rick_h_, I dunno, I'd prefer changing the handler to look for the textvalue node and only make the change if it exists, and then remove the node from the HTML.  I'll regard it as bikeshedding and leave it to you though.
[15:00] <bac> jujugui: is it a known issue that 'view all notifications' messes up everything?  i don't see a bug for it.
[15:00] <bac> i can describe more precisely if needed.
[15:00] <hatch> bac: I thought it just showed under the browser/inspector
[15:01] <rick_h_> bac: it's been brought up. I mentioned it at the sprint and there's talk of reworking but you're right. I don't remember seeing a bug
[15:01] <gary_poster> pretty sure there is one, but maybe I just imagined one
[15:01] <gary_poster> looking
[15:01] <bac> hatch: that plus it brings up the bottom of the canvas leaving a black void that doesn't go away
[15:01]  * gary_poster needs to appoint someone  bug tracker wrangler
[15:01] <bac> pretty much makes everything unusable
[15:01] <rick_h_> gary_poster: heh, started to poke at it in slack time yesterday. It's a mess
[15:02] <gary_poster> bac, can click on juju logo to escape
[15:02] <bac> gary_poster: oh, then never mind.  :)
[15:02] <gary_poster> thanks rick_h_ .  volunteers welcome. :-)
[15:02] <rick_h_> gary_poster: hah, I've got commitment problems :P but can try to help
[15:02] <gary_poster> heh ok thanks
[15:04] <gary_poster> bac, hatch, rick_h_ https://bugs.launchpad.net/juju-gui/+bug/1218011
[15:04] <_mup_> Bug #1218011: full notification list is difficult to read and to use <juju-gui:Triaged by lucapaulina> <https://launchpad.net/bugs/1218011>
[15:05] <rick_h_> gary_poster: ah, thanks
[15:06] <rick_h_> ooh, it's official http://insights.ubuntu.com/news/press-releases/ubuntu-on-windows-azure-gets-the-juju-magic/
[15:07] <rick_h_> thanks for the review bac
[15:19] <hatch> gary_poster: so there is a small issue with this new approach for the service name
[15:19] <gary_poster> hatch, uh oh :-)
[15:19] <hatch> actually no sorry it's pre-existing
[15:19] <hatch> but anyways...
[15:19] <hatch> I change it's name, it updates the canvas
[15:19] <hatch> I click 'save' to close the inspector
[15:20] <hatch> I open the inspector, it shows the original value, but the canvas shows the custom name
[15:20] <gary_poster> hatch, what does the model say?  the original value?
[15:21] <gary_poster> oh
[15:21] <gary_poster> wait you said save
[15:21] <gary_poster> that
[15:21] <gary_poster> hatch that is an issue with save sucking
[15:21] <gary_poster> forgot about that :-(
[15:21] <hatch> lol
[15:21] <hatch> rename save to 'suck' ?
[15:21] <hatch> haha
[15:21] <gary_poster> heh
[15:21] <gary_poster> hatch, in general save discards all values atm right?
[15:22] <hatch> right - I thought I saw a branch go through which fixed that
[15:22] <hatch> but the code doesn't reflect that
[15:22] <gary_poster> dunno, I didn't see that
[15:22] <gary_poster> hatch, ok, sounds like another showstopper
[15:22] <hatch> well we could rename 'Save' to 'Close'
[15:22] <gary_poster> heh
[15:23] <hatch> because that's what it does now
[15:23] <gary_poster> close should reopen with same values I'd argue
[15:23] <hatch> 'Cancel'
[15:24] <hatch> :D
[15:24] <gary_poster> :-P
[15:25] <gary_poster> hatch I suggest you finish up your branch, I'll finish up mine, and then we can discuss options for this other issue, after whoever finishes first.  sound good?
[15:25] <hatch> sure thing
[15:25] <gary_poster> thanks
[15:27] <bac> rick_h_: ...and i got declined!  wow.
[15:28] <rick_h_> bac: :( sorry, I called them and warned them about your party animal side
[15:28] <gary_poster> huh!
[15:28] <gary_poster> weird
[15:28] <gary_poster> heard you were a crazy American
[15:28] <bac> gary_poster: it was for sf, the weekend before the sprint
[15:29] <gary_poster> bac, oh!  even weirder
[15:29] <bac> teh germans still get their shot
[15:29] <gary_poster> heh
[15:30] <dimitern> rogpeppe, https://codereview.appspot.com/13720045/ I have a review for you :)
[15:40] <hatch> we are so cool even the core guys are hangin out here now
[15:40] <hatch> look at us go
[15:42] <hatch> I think the HACKING doc is missing something...rapi is telling me that zookeeper isn't installed but I definitely just installed it
[15:43] <hatch> anyone have any ideas?
[15:43] <rick_h_> hatch: so it's a python thing. the virtualenv either needs a --site-pacakges flag or you need to isntall zookeeper into it
[15:43] <rick_h_> hatch: I did the later
[15:44] <hatch> ENGLISH SON!!!
[15:44] <gary_poster> lol ????
[15:44] <rick_h_> http://paste.mitechie.com/show/1025/ where pip is the virtualenv's pip 
[15:44]  * rick_h_ goes to look where it is these days
[15:45] <gary_poster> I think that's an expression lost from translation from the Canadian
[15:45] <hatch> haha clearly
[15:45] <luca__> gary_poster: Resolved unit quick action notification is a list of units that have been resolved.
[15:45] <rick_h_> hatch: so run `source virtualenv/bin/activate`
[15:45] <rick_h_> hatch: and then run those three pip commands in the paste and get back to me
[15:46] <gary_poster> luca__, ah ok.  how long do they stay there?  /me goes to look at wireframes again
[15:46] <luca__> gary_poster: it's in the list of quick actions in the wireframe
[15:47] <luca__> gary_poster: I think you did explain yourself clearly but it doesn't look good after seeing it…there is no way we can have a green checkmark? even if it doesn't fade and instead just appears and disappears.
[15:47] <hatch> rick_h_: did you forget some of that path?
[15:47] <rick_h_> hatch: it's assuming you're in your working tree
[15:47] <hatch> well I'm in the rapi dir but there is no virtualenv
[15:47] <gary_poster> luca__, yeah we can have a checkmark appear/disappear without fade
[15:48] <rick_h_> hatch: yea, go back up a level
[15:48] <luca__> gary_poster: there isn't much more UX for it than just a bar. They go to live there until dismissed for a second time. It to allow you to find units that you have marked as resolved, instead of them being lost.
[15:48] <rick_h_> hatch: that's from the root of the gui dir
[15:48] <luca__> gary_poster: ok, that would be fine
[15:50] <gary_poster> luca__, checkmark: ok cool, I'll get it done.  resolved unit notification: so you want a a select all checkbox and a dismiss button?  or a dismiss all button?  I wonder if that needs to be sketched out a bit more
[15:50] <Makyo> jujugui call in 10
[15:50] <gary_poster> thanks Makyo 
[15:50] <luca__> gary_poster: I'll get the UX done
[15:52] <hatch> rick_h_: ok ran all those commands, they executed successfully but still no luck
[15:52] <hatch> ImportError: No module named zookeeper
[15:52] <gary_poster> cool thanks luca__ .  hatch, any concerns with above discussion of resolved unit notification?  luca__, note that this will be specific to your browser.  If you reload or go to another browser or whatever, data will be gone.  hatch, I envision us doing this with local annotations.  luca__, what happens if a resolved unit gets a new error?  it stays in resolved list and also is in error list?
[15:53] <rick_h_> hatch: let's hang on after the call and we can get it going
[15:54] <luca__> gary_poster: hatch do you think there is value in this quick action? The only reason I added it was because it seems quite dangerous to resolve something that still potentially has an error. I'm not tied to it.
[15:55] <hatch> I think there are too many edge cases to add this feature before release
[15:55] <luca__> hatch: oh yeah, of course, this isn't a high priority thing.
[15:55] <luca__> hatch: but is there value in the feature?
[15:55] <luca__> hatch: it's like a tracking thing
[15:56] <bcsaller> Resolve should indicate things are working properly, if that is true you don't need them, if its false they should generate new errors on usage
[15:56] <hatch> so there is a unit in error, you resolve the error, then you click resolve - that's it
[15:57] <hatch> if there is another error, it'll go back into error state
[15:57] <hatch> I don't think they stay in any 'limbo' mode
[15:57] <gary_poster> luca__, that's certainly a fair observation (the danger).  hatch we won't add this before release, no worries.  we're talking longer term.  luca__, I think the observation is good but I suggest that we should wait for user feedback in this regard before designing/building.
[15:57] <gary_poster> hatch bcsaller, don't quite agree
[15:57] <luca__> bcsaller: gary_poster hatch if that is the case then I guess there is no need for it :)
[15:57] <gary_poster> "retry" I agree
[15:57] <gary_poster> but "resolve" is going to say "juju, this is fine, trust me"
[15:58] <gary_poster> and any further errors will only be operational outside of the purview of juju
[15:58] <bcsaller> retry being explicitly rerun hooks, which will either succeed or fail right away
[15:58] <gary_poster> right
[15:58] <bcsaller> but if you actually resolved the issue with resolve you might not need that 
[15:58] <hatch> right - but how would we know when to remove the resolved unit from the 'tracking' section to 'running'?
[15:58] <gary_poster> but "resolved" is a marker not an action
[15:58] <bcsaller> not with resolve, and then 'told juju it was resolved' 
[15:59] <gary_poster> hatch it would be in both AIUI
[15:59] <gary_poster> jujugui call in 1
[15:59] <gary_poster> in link from calendar
[15:59] <bcsaller> it didn't used to be, it just says take it out of the error state and being running hooks again (but without forcing the last hook to run)
[15:59] <gary_poster> sure, but if last hook was start...
[16:12] <bac> gary_poster: from our conversation i wrote a rewrite rule ~bac/charms/precise/wordpress -> ~bac/precise/wordpress.  curtis says it is unneeded.
[16:13] <gary_poster> Makyo, reading your guidocs branch, cool, thank you!  Trivial suggestion so far: remove "search" and "-[REVNO]" from links.  For example, href="https://jujucharms.com/fullscreen/search/precise/juju-gui-76/" could become href="https://jujucharms.com/fullscreen/precise/juju-gui/"
[16:13] <Makyo> gary_poster, ah, cool, will do.
[16:13] <gary_poster> bac why?  I thought you already had a "remove charms" rewrite anyway
[16:14] <gary_poster> Makyo, also trivial: THis is
[16:14] <gary_poster> +            good for a monitoring type scenario.</p>
[16:15] <gary_poster> Makyo, you may want to mention that readonly does not protect the API behind a higher security: it is a visual change only.
[16:15] <gary_poster> That sucks, but it is the truth
[16:15] <bac> gary_poster: the first one is only for '^/charms' and a separate for '~user/charms' (done correctly)
[16:16] <gary_poster> bac, ah ok, that makes sense to me.  why does curtis not like it?
[16:16] <bac> gary_poster: 'user urls don't have "charms" in them'.  but this rule is to account for mapping old urls, no?
[16:16] <gary_poster> Makyo otherwise LGTM, thanks again!
[16:18] <gary_poster> bac, that's the intent, but maybe I was wrong and they never looked like that?  Certainly if I go to https://manage.jujucharms.com/~gary/oneiric/buildbot-slave that works and https://manage.jujucharms.com/~gary/charms/oneiric/buildbot-slave
[16:18] <gary_poster> that doesn't
[16:19] <gary_poster> bac, so IOW maybe I was wrong.  If Curtis says we never had old urls like https://manage.jujucharms.com/~gary/charms/oneiric/buildbot-slave then I'm fine with going with his suggestion
[16:20] <bac> gary_poster: that is curtis' assertion
[16:20] <gary_poster> (or https://jujucharms.com/~gary/charms/oneiric/buildbot-slave)
[16:20] <gary_poster> he would know better than I
[16:20] <gary_poster> sorry for any wild goose chasing I may have sent you on, bac
[16:21] <bac> np.  just don't want it any more cluttered than necessary
[16:21] <hatch> rick_h_: I just kicked off CI again for ya
[16:21] <rick_h_> hatch: ty sir
[16:21]  * rick_h_ corsses fingers
[16:22] <hatch> gary_poster: bcsaller did we want to chat about this 'tracking' block?
[16:22] <gary_poster> hatch, oh right
[16:22] <gary_poster> hatch, bcsaller, no, I mean yes. ;-) https://plus.google.com/hangouts/_/6be4e6d76bd266bd36ec5c5626c7d5a61d0ea88c ?
[16:29] <gary_poster> all solved. I'll write it up. :-)
[16:30] <luca__> gary_poster: rick_h_ hatch bcsaller Makyo I need a name for the mockups that is believable instead of "Environment on demonstration," have you got any other ideas?
[16:30] <gary_poster> Environment on openstack, luca__ 
[16:30] <gary_poster> luca__, more realistically would be
[16:31] <luca__> gary_poster: but do we have any examples of what people call their production environments?
[16:31] <gary_poster> Environment on HP Cloud or EC2 or LXC, but I don't have the exact strings
[16:31] <hatch> I'd call mine 'The Death Star' then name all my machines after starwars characters
[16:31] <gary_poster> lol
[16:31] <hatch> because I'm a nerd like that
[16:32] <hatch> haha
[16:32] <luca__> hatch: lol
[16:32] <gary_poster> you can't name your machines in The Cloud!
[16:32] <luca__> rookie mistake
[16:32] <hatch> or - wookie mistake....
[16:32] <luca__> haha
[16:33] <Makyo> Nerd cred revoked.  It's spelled Wookiee. :|
[16:33] <gary_poster> lol.  luca__, I'm actually not sure what that string is anymore.  I *think* that it is the same as the environment names we have in environments.yaml.  my names are things like "ec2" or "ec2gud" (gud is the name of my desktop) or "canonistack"
[16:33] <gary_poster> lol
[16:34] <hatch> lol!!
[16:34] <luca__> nerd off!
[16:34] <luca__> or is it, geek off?
[16:34] <Makyo> I've spent all my life preparing for this. :T
[16:34] <luca__> haha
[16:34] <hatch> haha
[16:35] <luca__> ('''\(';…;')/''')
[16:36] <gary_poster> heh
[16:38] <bcsaller>  or sorry I missed that call, I was having coffee
[16:38] <bcsaller> s/or/oh,
[16:39] <gary_poster> luca__, replied to your problem email with comments.  worth a scan when you have time.
[16:40] <luca__> gary_poster: I'll take a look now
[16:40] <gary_poster> thanks
[16:41] <hatch> luca__: that's pretty scary!
[16:42] <luca__> :P
[16:46] <hatch> so far I"m digging this standing desk thing - but adjustable is key. definitely can't do it all day :)
[16:46] <rick_h_> hatch: yea, we'll convince you to get a geekdesk yet :P
[16:47] <luca__> hatch: I don't think your meant to stand all day…isn't it really bad for blood pressure?
[16:48] <hatch> rick_h_: nope going to get one of these https://myupdesk.com/upwrite
[16:48] <gary_poster> yes
[16:48] <hatch> everything we do is bad for us
[16:48] <rick_h_> hatch: wow
[16:49] <gary_poster> yes, it is bad for blood pressure, hatch and luca__ , fwiw.  you should either change position every 20/30 minutes (sitting standing) or walk
[16:49] <hatch> rick_h_: you like?
[16:49] <rick_h_> hatch: you're not worried about getting marker stuff all over everything?
[16:49] <rick_h_> hatch: meh, to each their own
[16:50] <hatch> nope - I write so many temproary notes that it'll be awesome to do it right on the desk then wipe it off
[16:50] <hatch> also a motivation to keep my desk clea n
[16:50] <hatch> haha
[16:50] <hatch> us shipping is $129, into canada...probably $1000
[16:50] <hatch> lol
[16:51] <hatch> that's a lot to spend on a desk though so we'll see what happens
[16:52] <rick_h_> yea, my geekdesk is about half that
[16:52] <rick_h_> but no whiteboard top, though I guess you could whiteboard paint the top of the desk 
[16:52] <hatch> I'm guessing you didn't buy it from geekdesk then
[16:52] <gary_poster> bah, type=number doesn't really work yet cross-browser :-/
[16:52] <hatch> because their prices are not much cheaper
[16:53] <rick_h_> hatch: http://www.geekdesk.com/default.asp?contentID=634
[16:53] <hatch> gary_poster: and it falls appart on mobile
[16:53] <rick_h_> gary_poster: yea :(
[16:53] <gary_poster> hatch, :-(
[16:53] <hatch> rick_h_: yeah would be pretty hard for me to balance my keybaord and monitor on those bars :P
[16:53] <rick_h_> hatch: yea, I bought a solid core door from the local hardware store for $50 and screwed it on
[16:53] <hatch> haha that's not a bad idea
[16:54] <rick_h_> hatch: added $50 to the cost and I could cut the door to the size of the desk top I wanted
[16:54] <rick_h_> hatch: ended up with a larger top than the one that came with it and if I destroy it, I don't feel back throwing it away and getting another one
[16:54] <hatch> yeah and the shipping was probably way less
[16:54] <rick_h_> hatch: and as I said, costs about half as much
[16:54] <rick_h_> hatch: yep, saved some $70 on shipping sand top
[16:54] <rick_h_> hatch: overall about $300 cheaper than with the top/shipped
[16:55] <rick_h_> hatch: so as I said, I got mine for about half that :P
[16:55] <hatch> haha yep
[16:55] <hatch> the whiteboard top is so cool though!
[16:55] <rick_h_> hatch: yea, I've got whiteboards. I don't know about notes on the desktop. That's the "to each their own" part
[17:00] <gary_poster> I can imagine the "ugh I have dried whiteboard marker gunk all over my desk...my hands...and my keyboard" to be a downer
[17:06] <hatch> haha maybe
[17:06] <hatch> rick_h_: looks like CI might be a real error
[17:08] <rick_h_> hatch: yea looking. It's a different error this time
[17:15] <hatch> gary_poster: when clicking 'deploy' it can take a few seconds to appear in the canvas at which point it appears that it's locked up....the old inspector used to grey out the 'deploy' button - maybe we should do that and add a spinner or something
[17:16] <gary_poster> hatch +1 on making it gray out now, and sending an email to luca about the spinner.  You up for that, or you want me to make a separate card??
[17:16] <gary_poster> jujugui 1 review request: https://codereview.appspot.com/13583044/
[17:16] <Makyo> gary_poster, on it
[17:17] <gary_poster> Thanks Makyo!
[17:19] <hatch> gary_poster: I'd rather another branch
[17:19] <hatch> want to keep these things small and focused if possible :)
[17:19] <hatch> I can still do it
[17:19] <hatch> but in a followup
[17:37] <rick_h_> hatch: got a sec to go over the CI issue?
[17:38] <hatch> lets do it
[17:39] <gary_poster> hatch +1 on separate
[17:47] <rick_h_> jujugui anyone have chrome 28 installed right now?
[17:48] <Makyo> rick_h_, 29 :/
[17:48] <bcsaller> rick_h_: yeah
[17:48] <gary_poster> I have 29,sorry
[17:48] <rick_h_> bcsaller: can you jump in https://plus.google.com/hangouts/_/c14546ec56b44c0d1b00303658863100c83b3038?hl=en to help test something?
[17:49] <benji> jujugui: I have a smallish branch up for review (my branch was too big so I broke it up): https://codereview.appspot.com/13627048
[17:50] <gary_poster> benji I will look
[17:50] <benji> thanks
[17:52] <gary_poster> benji LGTM
[17:52] <benji> cool
[17:57] <gary_poster> hatch your branch ready for review or are you commenting?
[17:58] <hatch> ready to go https://codereview.appspot.com/13457052/ I think it's pretty self explanatory
[18:00] <gary_poster> cool, on it
[18:01] <hatch> thanks
[18:10] <gary_poster> hatch items to tweak or discuss.  stepping away, back in a few
[18:22] <rick_h_> hatch: I got to watch the current in progress test run. It's a race condition there. Submitting a one liner to see if that will help keep it from ever happening.
[18:22] <rick_h_> will put getting selenium and maybe sst going locally on the todo list
[18:30] <hatch> rick_h_: saw your fix - curious if you could have picked a better element to click, like the footer or something - isn't the environment-switcher supposed to 'do' something :D
[18:30] <rick_h_> hatch: well, not in dev mode which we test in. Footer is going away soon
[18:30] <rick_h_> hatch: if this runs perfectly open to follow ups, but seemed a good test-target
[18:31] <hatch> yeah I guess if that link doesn't do anything
[18:31]  * hatch grabbing lunch
[18:58] <gary_poster> hey hatch, when you get back let's talk through the branch
[18:58] <benji> I have to go AFK but I have an lbox propose chugging away that will hopefully make its way to rietveld soon.  A review of it while I am away would be very effecient of you.
[18:58] <gary_poster> :-)
[18:59] <gary_poster> k 
[19:33] <hatch> gary_poster: back
[19:34] <gary_poster> hey hatch.  was hoping to be farther along when you came back.  ok lemme get your branch here then will call
[19:40] <gary_poster> hatch hey 
[19:40] <gary_poster> https://plus.google.com/hangouts/_/3fdaf445e6d2062f47cbba4c6278e6310da88876
[20:09] <gary_poster> hatch, problem is that you can't define multiple background colors :-/
[20:09] <hatch> ohhhh
[20:09] <hatch> well there we go
[20:10] <hatch> ok for the solution we will use little canvas elements which play nyan cat in the background when it's accepted
[20:12] <hatch> gary_poster: we could use the transparent imput background trick
[20:12] <hatch> the old way of putting things 'in' an input
[20:12] <gary_poster> hatch, like the checkbox?
[20:13] <hatch> you absolutely position the input over top of another element and set the inputs background to transparent
[20:13] <hatch> then anything you do to the background element is visible through the input
[20:13] <gary_poster> ah right
[20:15] <hatch> then you could put the checkmark in the back element and fade the background of the input
[20:15] <hatch> hacky and will only work in the constraints
[20:15] <hatch> but doable :)
[20:17] <gary_poster> ack :-) thanks.  not sure if I want to implement, but maybe I'm being a CSS snob :-P
[20:17] <gary_poster> probably more like lazy
[20:17] <hatch> I'd like it if the border of the input pulsed green
[20:18] <hatch> that would be my favourite interaction
[20:18] <gary_poster> hm, that would be doable
[20:18] <gary_poster> could propose to luca
[20:18] <gary_poster> my current fallback is to rip out animations entirely
[20:18] <gary_poster> webkit and chrome working towards images being able to transition actually
[20:22] <hatch> it's something we can work towards
[20:22] <hatch> we can release pretty fast once the inspector lands
[20:22]  * hatch hopes
[20:28] <bac> gary_poster: webops merged the rewrite branch and then replied that the RT had been completed.  i see no difference in behavior.  :(
[20:28] <gary_poster> bac :-(
[20:29] <gary_poster> bac, all I know is to report the problem on #is or #webops and see who you can get to help
[20:29] <gary_poster> I wonder if they need to manually kick apache or something
[20:29] <bac> gary_poster: yep, that's where i'm headed
[20:30] <gary_poster> cool, sorry 
[20:30] <gary_poster> and thanks
[20:33] <bac> gary_poster, luca__: where might i find the import/export assets shown in https://docs.google.com/a/canonical.com/file/d/0B7XG_QBXNwY1Tm9PVWhMZzB4alk/edit ?
[20:33] <bac> longs for juju-gui/graphics or some such
[20:34] <gary_poster> bac, this approximates what you long for
[20:34] <gary_poster> https://drive.google.com/a/canonical.com/?tab=co#folders/0B7XG_QBXNwY1Qi13X3RqR1dfaFE
[20:34] <gary_poster> though I don't think it has what you need
[20:34] <luca__> bac: hey Brad, what are you implementing that you need them for?
[20:35] <bac> luca__: was beginning to look at the export card
[20:35] <hatch> Makyo: bcsaller I have a d3 q if either of you are around
[20:35] <Makyo> hatch, shoot
[20:36] <hatch>     node.select('.name').html(function(d) {
[20:36] <hatch>                       return d.displayName;
[20:36] <hatch>                     });
[20:36]  * bcsaller reads along
[20:36] <hatch> 'this' is the proper element
[20:36] <hatch> so why does this not set the elements innerHTML
[20:36] <bac> oh, thanks gary_poster.  oscon is one folder i didn't explore
[20:37] <hatch> node is the service elements title
[20:37] <hatch> it doesn't throw any errors, it just doesn't do anything
[20:37] <bac> luca__ (i know you're way past eod) is there a problem with that proceeding?
[20:37] <Makyo> Can you breakpoint in there, hatch?
[20:37] <hatch> yep and d.displayName is (ceph)
[20:37] <bcsaller> yeah, looks like that should work
[20:38] <Makyo> What is .name?
[20:38] <hatch> the title of the service element on the canvas
[20:38] <bcsaller> the first match
[20:38] <Makyo> .text(), not .html()
[20:38] <bcsaller> which if node is correct will be what you expect
[20:39] <benji> gary_poster: thanks for the good review.  I replied to your comments (https://codereview.appspot.com/13368056/).
[20:39] <benji> oh, and I'm back now
[20:39] <bcsaller> Makyo: ahh, right, SVG doesn't support innerHTML
[20:40] <Makyo> Yeah; ran into this with testing a while back.
[20:40] <hatch> thanks Makyo :)
[20:40] <hatch> the API is confusing
[20:40] <hatch> it seems like it should work
[20:40] <Makyo> SVG != HTML, according to w3c, so the internal APIs are different.
[20:40] <hatch> I didn't even think of trying text though
[20:41] <luca__> bac: Nope, just wondering how they were going to be implemented. I haven't got the assets to hand but I can get Jamie or Spencer to create them first thing tomorrow and send them over.
[20:41] <gary_poster> bac, use a placeholder now, I suggest.  Also know that the location of the placeholder may change
[20:41] <gary_poster> (it *will* change long term, but may change sooner)
[20:42] <bac> cool, so i can just slap down a smiley face anywhere and run with it?
[20:43] <luca__> sure, it'll make the GUI more personable hehe
[20:43] <luca__> night all
[20:43] <hatch> DO IT!
[20:43] <hatch> he's gone!
[20:43] <hatch> lol
[20:43] <gary_poster> lol
[20:44] <gary_poster> bac, yes, assuming it is behind feature flag :-)
[20:44] <hatch>  /:flags:/happytime
[20:45] <gary_poster> heh
[20:45] <gary_poster> benji LGTM * 2 with small comment
[20:45]  * benji look
[20:45] <benji> s
[20:46] <benji> gary_poster: what does "It would have been nice to also see
[20:46] <benji> the new tests" mean?
[20:47] <gary_poster> benji, it means that I'm distracted and making less sense than usual, and that it would have been nice to have you highlight which tests were copied and which tests were fresh from your work
[20:47] <gary_poster> in review comments
[20:48] <benji> gary_poster: ah!  yep that makes sense.  I intentionally structured the code changes so that v2 to v3 changes would be apparent, but the copy/paste method of doing the tests precluded that.  Next time I do something like that I'll construct a diff and pastebin it somewhere so that those changes can be seen too.
[20:48] <gary_poster> cool, thanks benji.  anyway, all that said, looks like great work.  thanks.
[20:53] <hatch> I think I'll have a variable in my random id generator called the 'gary-coefficient'
[20:53] <hatch> :P
[20:53] <hatch> lol
[20:53] <hatch> i kid
[20:54] <gary_poster> yah yah yah
[20:58] <gary_poster> small CSS review request: https://codereview.appspot.com/13747043
[20:58] <gary_poster> needs IE 10 qa
[20:58] <gary_poster> jujugui ^^^
[20:58] <benji> gary_poster: looking
[20:58] <gary_poster> thanks benji
[20:59] <gary_poster> benji, ".animation-fill-mode(forwards, forwards);" is legacy from earlier attempt; reverting
[20:59] <benji> ok
[20:59] <benji> gary_poster: do you need QA?
[20:59] <gary_poster> forgot to review my review before announcing :-P
[20:59] <gary_poster> benji yes, on IE10
[21:00] <benji> gary_poster: ooooh...
[21:00] <benji> I'm IE-challenged right now.
[21:00] <gary_poster> benji, I can request elsewhere, but you and I should get back on the IE10 train :-)
[21:00] <Makyo> I'll do it.
[21:00] <gary_poster> thank you Makyo
[21:01]  * benji reluctantly buys a ticket.  The ticket is printed on used napkins from a BBQ restaurant.
[21:01] <Makyo> Need a bzr plugin to branch a repo from a reitveld link.  I am the laziest.
[21:01] <gary_poster> lol
[21:11] <gary_poster> hey hatch, I'm heading out.  you want me to check back in an hour or so for follow-up review?
[21:11] <Makyo> gary_poster, it works.
[21:11] <Makyo> LGTM on IE.
[21:11] <gary_poster> cool, thanks Makyo!
[21:12] <bac> gary_poster: rewrite is rewriting
[21:12] <gary_poster> bac, but?
[21:12] <bac> no butt
[21:12] <bac> it works now
[21:12] <gary_poster> bac, awesome!  just needed apache to be kicked?
[21:12] <bac> dunno what he did
[21:12] <hatch> gary_poster: sure if you have time, if not it can wait until morning
[21:13] <gary_poster> cool hatch, thx
[21:13] <hatch> have a good one
[21:13] <gary_poster> thank you bac!  follow up to jcastro--will improve his day
[21:13] <gary_poster> thanks ttyl!
[21:14] <bac> is jcastro having a bad day or is this just good news to make any day better?
[21:14] <bac> let's see if we can conjure him
[21:14] <hatch> double double toil and trouble
[21:15] <bac> jcastro got me hooked on quenepas and they ended up staining all of my clothes.
[21:15] <bac> quenepas [21:20]  * bac dog walks
[21:56] <hatch> benji: looks like your branch may have broken CI
[23:04] <huwshimi> Morning
[23:07] <hatch> morning huwshimi
[23:28] <gary_poster> hey huwshimi .  hacth, looking at your branch fwiw
[23:37] <gary_poster> hatch LGTM with small suggestion.  Thanks!
[23:40] <hatch> gary_poster: looks good thanks I'll get to that tonight and get it landed
[23:42] <gary_poster> thanks hatch!  I was hoping to have a plan for the ghost inspector with you today.  I have half a plan that makes me sad in a hacky sort of way ;-) and I have an alternate plan that involves wondering if we can release without fixing the bug and I have an alternate alternate plan which involves databinding the ghost inspector to the model.  We'll talk tomorrow, and hopwfully get this out the door, assuming qa goes ok
[23:43] <hatch> haha ok, sounds good to me