[00:57] <gary_poster> huwshimi, hi.  to see what I mean, go to http://comingsoon.jujucharms.com/sidebar/search/bundle/~benji/wiki/5/wiki/:flags:/charmworldv3/?text=benji#bws-bundle
[00:57] <gary_poster> then click on the charm token
[00:57] <gary_poster> This appears:
[00:57] <gary_poster> mysql
[00:57] <gary_poster> Charm Name:cs:precise/mysql-27
[00:57] <gary_poster> Unit-Count:3
[00:57] <huwshimi> gary_poster: Oh, I figured it out, it's the little overlay of info :)
[00:57] <gary_poster> huwshimi, that's it :-)
[00:57] <gary_poster> thanks huwshimi .  everything going ok?
[00:58] <huwshimi> gary_poster: Yeah, just about to finish up here with the recommended stuff and then do that
[00:59] <gary_poster> awesome thanks huwshimi.
[01:00] <gary_poster> hatch btw changed comingsoon to point to staging
[01:00] <gary_poster> running away again
[01:07] <hatch> thanks :)
[12:37] <gary_poster> hey hazmat, on SF CDO sprint topics, do you want to be required for both "Colocation support for Bundles" (which I think clearly needs you) and "Bundles v2: Config and modify before deployment" (which is GUI specific) or just the first one?
[12:45] <gary_poster> hey rick_h_ , you looking at huwshimi's follow-on recommended charms branch or shall I?  I know you had (have) a lot on your plate with the charmworld problems on top of your other work.
[12:47] <rick_h_> gary_poster: I've not looked at it yet. If you've got the bandwidth that'd be cool
[12:47] <gary_poster> rick_h_, great.  on it.
[12:47] <rick_h_> thanks
[12:52] <rick_h_> bac: welcome back!
[12:52] <rick_h_> bac: you mean fitbit stepping machine
[12:53] <gary_poster> hey bac!  hope you had a great time
[12:54] <antdillon> gary_poster, Morning, did you want a quick catch up?
[12:54] <gary_poster> hey antdillon! welcome back to you too! :-)
[12:55] <antdillon> gary_poster, Thanks :) let me know when your free for a chat if you want one
[12:56] <gary_poster> antdillon, yeah...would it work to have a call in 30 minutes?  Or I can schedule something any time after that for the next couple of hours
[12:56] <antdillon> gary_poster, Anytime is good with me. Which ever works best for you
[12:57] <gary_poster> antdillon, great.  just set something up in calendar for half hour from now
[12:57] <gary_poster> thank you
[12:59] <antdillon> gary_poster, Great thanks
[13:10] <gary_poster> rick_h_, all of a sudden http://comingsoon.jujucharms.com/precise/juju-gui-77/ no longer shows our icon (it is connected to staging).  could this be related to one of your recent charmworld branches?
[13:12] <bac> thanks gary_poster, rick_h_.  yes good to be back and somewhat staitionary for a bit.
[13:12] <gary_poster> :-) ool
[13:12] <gary_poster> c
[13:16] <rick_h_> gary_poster: probably, looking
[13:17] <gary_poster> rick_h_, thank you.  also I'd like to consider tackling https://bugs.launchpad.net/juju-gui/+bug/1235012 this morning, if it can be a quick fix: maybe can have a pre-imp with you in 45 min or so?
[13:17] <_mup_> Bug #1235012: Fullscreen charm details load tabs above content <charmbrowser> <juju-gui:Triaged> <https://launchpad.net/bugs/1235012>
[13:17] <rick_h_> gary_poster: sure thing
[13:17] <gary_poster> thank you
[13:18] <rick_h_> gary_poster: so I think we've found something of our bug with the new site. 
[13:18] <rick_h_> gary_poster: notice that your charm on stagins is missing all code, the description
[13:18] <rick_h_> gary_poster: I think ti's not showing the icon because the files aren't there to show
[13:18] <gary_poster> rick_h_, ah! :-(
[13:19] <rick_h_> gary_poster: yea, very :( 
[13:19] <rick_h_> gary_poster: because other recommended charms have their icons atm and have their files and such
[13:19] <rick_h_> strange that it's just the one charm. /me checks other 'new' charms
[13:20] <gary_poster> rick_h_, at least we can diagnose this one locally.  are you ok with looking at this?  I suspect bac is doing email triage but we could ask him to investigate with you or alone?
[13:20] <rick_h_> gary_poster: yea, I think I have staging access somewhere. Looking now. 
[13:20] <gary_poster> cool
[13:20] <bac> gary_poster, rick_h_, yeah i'm just wading through emails.  glad to pair on something if useful.
[13:20] <gary_poster> cool thanks bac
[13:21] <bac> rick_h_: and i do have staging accessa and know how to use it!
[13:21] <rick_h_> thanks bac 
[13:30] <rick_h_> jujugui orangesquad heads up I'm commenting out all cron jobs on staging.jujucharms.com so that I can debug the current state of things. Please don't land any changes to the codebase for a bit as well. 
[13:30] <benji> k
[13:30] <bac> rt
[13:31] <abentley> rick_h_: ack
[13:31] <gary_poster> x
[13:31] <gary_poster> (the unknown)
[13:32] <gary_poster> antdillon, I am in  https://plus.google.com/hangouts/_/calendar/Z2FyeS5wb3N0ZXJAY2Fub25pY2FsLmNvbQ.mph7p8n965vj59586qkphghkckbut no rush
[13:32] <antdillon> gary_poster, Loading
[13:41] <hazmat> gary_poster, hmm.. i'm fine with just doing the first
[13:42] <gary_poster> ack thanks hazmat will update
[13:55] <gary_poster> benji, I suspect you have seen or will see soon the email from Mark B.  When you have an idea of the answer, or when you think it is reasonable for me to ping Kent about it, please let me know.
[13:57] <benji> gary_poster: Kent just came online, I'm working on an issue he had late yesterday afternoon.  If we can get it fixed soon we could schedule then.
[13:57] <gary_poster> cool thanks very much benji.  please let me know when you know more
[13:57] <benji> gary_poster: will do
[14:02] <frankban> hey gary_poster: the charmers' charm is updated to revision 77, and I tested the upgrade steps. we are now able to 1) test it in a firewalled environment and 2) make a new jujucharms.com release
[14:06] <hazmat> what's ice cream?
[14:07] <gary_poster> frankban, awesome! :-)
[14:07] <hatch> hazmat: frozen cream
[14:08] <gary_poster> hazmat, a requested ODS demo
[14:08] <hazmat> hatch, wouldn't that make it hard to use on a screen, it melts and obscures everything ? ;-)
[14:08] <hatch> haha
[14:09] <gary_poster> frankban, I can ask kentb to try  it out in a firewalled environment.  Or are you doing so?
[14:10] <gary_poster> rick_h_, hey.  when you have time for a pre-imp let me know
[14:11] <rick_h_> gary_poster: pre-imp on the z-index?
[14:12] <gary_poster> rick_h_, pre-imp on that bug I mentioned.  I don't think that is z-index.  but lemme reboot after this upgrade...
[14:12] <rick_h_> gary_poster: oh, sorry yea. Sure thing. Any time.
[14:12] <gary_poster> thanks
[14:12] <rick_h_> gary_poster: rgr
[14:12] <frankban> gary_poster: please do it thanks!
[14:22] <gary_poster> rick_h_, https://plus.google.com/hangouts/_/10fe9c48d75b9ebbe273219cb03f4e53869ce4f7 when you have a chance.  no rush.
[14:28] <rick_h_> benji: or bac either of you have time to do a quick "does this make you think of anything" chat?
[14:28] <benji> rick_h_: I will in about 10 minutes
[14:28] <rick_h_> benji: thanks
[14:29] <bac> rick_h_: sure
[14:30] <rick_h_> bac: maybe wait the 10 and then we can all see if this issue rings any bells to anyone
[14:30] <rick_h_> ?
[14:32] <hatch> hey gary_poster did you ever reply to my q as to where your sort differs from a lexographical one?
[14:32] <bac> rick_h_: sure.  i wish hangouts could have theme music.  http://www.youtube.com/watch?v=kPYese-Hl8M
[14:32] <gary_poster> hatch I did in the code, sorry.   will reply to the MP
[14:32] <hatch> oh woops I missed that
[14:32] <hatch> I was able to get the same results with just using .sort() so was curious
[14:33] <rick_h_> bac: lol
[14:35] <gary_poster> hatch, that may indicate that I should write a better test.  I just made a rv reply with the comment though
[14:36] <hatch> cool thanks
[14:37] <gary_poster> hatch, I had a real world example that showed more but I ripped it out at the last instant because, even though it passed in IE 10, FF and Chrome it failed in phantom.  I am 99% sure it is because of differences in localeCompare
[14:37] <gary_poster> (a real world test)
[14:37] <hatch> ahh - entirely possible
[14:38] <gary_poster> the sort was still reasonable
[14:38] <gary_poster> just not the same
[14:40] <hatch> apple has to be the only company I know of that has a press conference within 10 days and still doesn't tell anyone about it
[14:43] <bac> hatch: and it'll be SRO
[14:43] <hatch> sro?
[14:43] <bac> packed.  standing room only
[14:44] <hatch> oh - right probably
[14:44] <hatch> it would be funny if no one showed up because they had made other plans
[14:44] <hatch> 'sorry went on vacation'
[14:48] <gary_poster> hatch did my explanation make sense?
[14:48] <hatch> it did yes thanks
[14:49] <gary_poster> cool
[14:51] <gary_poster> luca__, should we cancel the web team mtg until further notice?  I think we all have the best intentions and all are swamped :-P  We might be able to look at the SASS change after SFO but tbh all of the other UX items are higher priority
[14:51] <luca__> gary_poster: yeah, I agree hehe
[14:51] <gary_poster> :-/ :-)
[14:52] <hatch> those orange corners look out of place :)
[14:52] <gary_poster> hatch as in bug or as in they surprise you?
[14:52] <hatch> like they feel too far away from the content - there is so much whitespace between the content and the triangle
[14:53] <hatch> so it kind of feels like it was just 'put there'
[14:53] <gary_poster> ack hatch.  not sure I agree, but tbh not an engineering bug is my current interest :-P :-)
[14:53] <gary_poster> classifying as "not an engineering bug" I mean
[14:54] <hatch> nope not that
[14:57] <gary_poster> hatch but please raise with luca if you have concern
[14:57] <gary_poster> not denying your opinion, just determining whether I have to prioritize what you are saying :-)
[15:02] <hatch> gary_poster: I probably wouldn't ;) But from my QA day yesterday https://bugs.launchpad.net/juju-gui/+bug/1237457 you might want to schedule :)
[15:02] <_mup_> Bug #1237457: Hitting enter quickly in autocomplete causing result list to appear over real results <juju-gui:New> <https://launchpad.net/bugs/1237457>
[15:02] <gary_poster> wow.  if you go to a fullscreen view of a charm directly, the page renders sidebar, fullscreen, sidebar, (broken) fullscreen
[15:03] <gary_poster> ack thanks hatch
[15:03] <rick_h_> gary_poster: lol, so it's a bigger bug than thought. 
[15:03] <hatch> can we get rid of fullscreen yet? :D
[15:04] <gary_poster> :-) not yet
[15:04] <rick_h_> gary_poster: it's supposed to draw the fullscreen details, the sidebar is rendered but empty and hidden (I think) and originally we had a fullscreen editorial behind the fullscreen details but I thuoght we got rid of it
[15:04] <gary_poster> gotcha
[15:05] <rick_h_> gary_poster: the key is that you hit all three routes and their callables per request, and then double dispatch hits the second pass through. Only one should 'run' though and the other two should bail out. 
[15:06] <gary_poster> rick_h_, so should charm render be called four times?
[15:06] <gary_poster> (i.e., where is the bailing out supposed to happen)
[15:07] <hatch> gary_poster: now that we have moved to the new inspector we could entertain assigning some dev days to removing the double dispatch completely
[15:07] <rick_h_> gary_poster: routeDefault, routeDirectCharmId (which you should be hitting), and routeView 
[15:07] <gary_poster> hatch, after SFO we are going to have a "prep for 1.0/papercuts" time period
[15:07] <rick_h_> gary_poster: those determine which main call to hit fullscreen, sidebar, minimized
[15:07] <hatch> cool cool
[15:08] <rick_h_> gary_poster: and then each of those methods does the real logic to call the various _renderXXX required
[15:08] <gary_poster> hatch, I am proposing a sprint session for us to prioritize what papercuts to tackle
[15:08] <rick_h_> gary_poster: those completely might get hit twice, but should not be 4 times :/
[15:08] <rick_h_> hatch: what's the YUI combo file builder tool again?
[15:08] <rick_h_> hatch: searching YUI builder gets me all kinds of shifter crap
[15:08] <hatch> combinator
[15:09] <hatch> nope thats wrong
[15:09] <hatch> configurator
[15:09] <hatch> http://yuilibrary.com/yui/configurator/
[15:09] <hatch> the combinator is something else
[15:09] <rick_h_> bac: still got time? I'm ready to confess or something before i lose my mind
[15:09] <bac> sure
[15:09] <rick_h_> invite on the way
[15:09] <gary_poster> rick_h_, ack.  fwiw, the this._details.destroy(); approach in _cleanOldViews did not address
[15:10] <gary_poster> even as a band aid
[15:10] <rick_h_> gary_poster: :/ ok. If I can figure out this charmworld issue I"d love to pair on that and help. I know that browser state stuff is a pita and I know the buried bodies more than anyone
[15:10] <gary_poster> ack rick_h_ thx
[15:11] <rick_h_> bac: https://plus.google.com/hangouts/_/39287b62294faad7d485563b5753a945d51623a1?hl=en and benji if you get time
[15:12] <gary_poster> benji, re: keyboard shortcuts, I am after a quick win pre SFO.  are they truly so poorly thought out that we can't show them?
[15:13] <hatch> gary_poster: should we put our timeoff  on the juju ui calendar?
[15:14] <gary_poster> benji, I was hoping for a Huw styling branch, after which we say, "look, keyboard shortcuts!"  I think what we have is better than nothing, and we are unlikely to prioritize the keyboard bits in the near future unless someone finds the time to adopt them as a hobby/slack (ha ha? :-( ) project
[15:14] <gary_poster> hatch, not necessary for me.  if you think it would be helpful for team, propose for weekly call?
[15:15] <hatch> sure - I'm just doing it because I usually forget what others are doing :)
[15:15] <gary_poster> hatch, fwiw do you know about https://canonicaladmin.com/canonical/intramaster.nsf/2Week?openform
[15:15] <gary_poster> it's what I use
[15:16] <gary_poster> not ideal but sufficient
[15:19] <hatch> I didn't but cool
[15:19] <hatch> now I do :)
[15:20] <gary_poster> hatch :-) available from left hand side "Reports" -> "Holiday/Leave Calendar"
[15:32] <gary_poster> rick_h_, this solves the immediate problem.  As the comment says, it's reasonable, but also a band-aid for a bigger issue.  I'll try to investigate the bigger issue for a bit now.  http://paste.ubuntu.com/6214235/
[15:35] <gary_poster> I really think we should be using the db as a charm cache for the browse, or at the very least give browse its own charm cache.  ATM I don't see a reason to separate them.
[15:36] <gary_poster> but tha's also a separate issue
[15:43] <rick_h_> gary_poster: yea, so that's the thing. the view should worry about the state/double dispatch bug. I've tried to capture any mess of that into just browser.js so that all the crazy is in once place
[15:44] <rick_h_> gary_poster: the fact that ignoring a call works is a sign there's a missing state check. It's trying to re-render for some reason in browser.js when it should be saying "This is already done, skip"
[15:44] <rick_h_> gary_poster: as for the charm cache, we can move it where we want. Originally the models weren't comaptible and the browser/inspector share the same cache now. 
[15:44] <gary_poster> cache: right, cool
[15:44] <rick_h_> gary_poster: I don't recall where Makyo set that up, did it move from browser.js up to app.js? 
[15:45] <rick_h_> gary_poster: but it's why the store methods take a cache arg is so that one cache could be setup/used/shared
[15:45] <benji> how does one specify that a charm can have multiple relations of a given type?
[15:45] <rick_h_> benji: I think the charm just has to handle it in the hook/config. 
[15:46] <gary_poster> benji, it's a bit of a mess, in core and gui.  there's a bug.  but anyway, practically speaking:
[15:46] <rick_h_> benji: I think that's the big discussion that went round with the haproxy/etc the other day
[15:46] <gary_poster> - GUI disallows requires to make more than one connection
[15:46] <gary_poster> - GUI allows provides to make more than one connection
[15:46] <gary_poster> so
[15:46] <gary_poster> ice cream should provide ice cream
[15:46] <gary_poster> and toppings should require ice cream
[15:47] <gary_poster> and I think it will do what you want
[15:47] <gary_poster> benji, clear as mud?
[15:47] <benji> gary_poster: that's (esentially) what I tried, it only allows one relation
[15:47] <benji> (I had one relation type for "topping" and one for "sauce", which of course lets there be two relations to an ice cream)
[15:48] <gary_poster> benji that is the opposite of what I said
[15:48] <benji> oh! reverse the requires/provdes
[15:48] <gary_poster> see what I mean?
[15:48] <gary_poster> yes
[15:48] <benji> I'll try that now.
[15:48] <gary_poster> cool
[15:50] <Makyo> jujugui call in 10
[15:58] <gary_poster> jujugui call in 2
[16:00] <hatch> rick_h_: last night my HTC One got the Android 4.3 update
[16:01] <hatch> the odd ui lags that were sometimes there appear to be all gone now
[16:02] <rick_h_> cool, no 4.3 here :(
[16:02] <hatch> well mine has Sense 5 so it's not 'true' 4.3 but the core improvements all seem to be there
[16:28] <Makyo> jujugui That was it, was trying to use a real store.  Will add that to the card and propose.
[16:28] <gary_poster> yay :-)
[16:30] <Makyo> Async was just hiding things, since the only error was a timeout.
[16:30] <hatch> the mock test suite I've been thinking about would also be a good way to avoid these issues
[16:33] <bac> rick_h_: i did the ingest locally and then in mongo see this: http://paste.ubuntu.com/6214424/
[16:34] <bac> rick_h_: it looks complete.
[16:34] <bac> rick_h_: won't answer mystery but may try doing enqueue/ingest again on staging to see what it produces
[16:36] <rick_h_> bac: yep, thanks. Verifies what I did locally as well
[16:36] <bac> rick_h_: why don't i take this card and you go on to the revision problem?
[16:37] <rick_h_> bac: k, if you've got caught up enough to look that'd be great. 
[16:37] <bac> rick_h_: sure
[16:38] <rick_h_> bac: did you awnt me to walk you through how I was debugging on the staging server?
[16:38] <bac> rick_h_: maybe later.  i'm about to have lunch and i just want to poke around a bit first on staging
[16:39] <rick_h_> bac: rgr
[16:39] <rick_h_> geeze, it is almost 1pm. where did today go?
[16:58] <rick_h_> gary_poster: http://paste.mitechie.com/show/1035/ is the bug fix. Will work on adding test cases in there. 
[16:58] <hatch> do we have a bundle which has any non-recommended charms?
[16:58] <rick_h_> gary_poster: basically the routeDirectCharmId should have skipped maye. 
[16:59] <rick_h_> hatch: I don't think so atm. 
[16:59] <hatch> alrighty
[16:59] <hatch> TDD it is
[17:00] <benji> gary_poster: ice cream is go
[17:04] <rick_h_> gary_poster: actually, this version is 'more' correct. That route should just ignore the call and pass it along. http://paste.mitechie.com/show/1036/
[17:07] <benji> gary_poster: if we are going to have a call with Mark B. we have less than 3 hours from now to do it ("kentb:#canonical> benji: ok. I need to pack these up by 3pm Central time, though.  Gotta get them to the loading dock for a Thursday a.m. pickup")
[17:08] <gary_poster> benji, ack
[17:08] <gary_poster> trying
[17:08] <gary_poster> will lunch later :-)
[17:09] <gary_poster> rick_h_, cool.  I'd like to have both if you don't mind: mine still handles less-likely scenarios of views changing quickly
[17:10] <rick_h_> gary_poster: but the only way views change quickly is through the browser which destorys before it creates 
[17:11] <rick_h_> gary_poster: I guess disagree but if you'd like them can add them. It doesn't 'hurt' just crosses the bounds of isolating dispatch rules into just browser.js
[17:11] <gary_poster> rick_h_, ack.  may do yours only, or if you want to land go for it.  will be returning to lunch soon I hope
[17:12] <rick_h_> gary_poster: rgr, lunch is in the oven so working on tests over lunch and will submit shortly. 
[17:12] <gary_poster> thanks rick_h_ !
[17:17]  * hatch needs a bigger callstack, it's been exceeded
[17:17] <gary_poster> benji, how soon UTC-wise would a demo work for you?  I was going to set up a meeting with Mark but his schedule is booked, so I'll just offer a time
[17:17] <gary_poster> I mean a range of times
[17:17] <gary_poster> benji, tomorrow morning I mean, sorry
[17:19] <benji> gary_poster: and time tomorrow from 8 Eastern on will work for me 
[17:19] <gary_poster> ok thanks benji
[17:25] <Makyo> jujugui 2 reviews and QA, please! https://codereview.appspot.com/14438061
[17:26] <gary_poster> Makyo, I'm doing a review
[17:31] <Makyo> jujugui Would like real upgrade charm QA, but you can fake it by adding a service and then running app.db.services.item(0).set('charmChanged', true)
[17:34] <rick_h_> jujugui tiny branch/qa please https://codereview.appspot.com/14441054/ see bug referenced for any missing details. #1235012
[17:34] <_mup_> Bug #1235012: Fullscreen charm details load tabs above content <charmbrowser> <juju-gui:Triaged> <https://launchpad.net/bugs/1235012>
[17:36] <jcastro> heya rick_h_
[17:37] <hazmat> gary_poster, any chance gui export can drop empty constraint values its currently dumping via shift-D
[17:37] <rick_h_> jcastro: howdy
[17:37] <gary_poster> hazmat, yes, I believe part of bcsaller's current work; if not we want it too and will happen hopefully within next couple of weeks
[17:38] <jcastro> rick_h_, nm, gary answered my question
[17:38] <jcastro> I have some ideas for bundles but I need export to work
[17:38] <jcastro> but I can work around
[17:39] <hazmat> gary_poster, cool this is re bug 1233234.. i can work around in deployer, but we've already got versions for saucy in.
[17:39] <_mup_> Bug #1233234: don't export empty / non existent constraints <juju-gui:Triaged> <https://launchpad.net/bugs/1233234>
[17:48] <hatch> bcsaller: you're a proponant of the Object.keys().forEach technique ... can you tell me why this one is breaking out of the closure? http://jsbin.com/AtoYElI/1/edit
[17:49] <hatch> I should know this but I'm drawing a blank
[17:49]  * bcsaller looks
[17:50] <bcsaller> hatch: you mean that name is being resolved from somewhere else?
[17:50] <hatch> yeah name isn't available where the debugger statement is
[17:52] <bcsaller> so if you console log name before the Object keys it works, I think its getting collected as unused by something there
[17:52] <bcsaller> bad jit or the like
[17:55] <Makyo> rick_h_, will review yours, if you still need it.  Swap reviews if you have time?
[17:55] <rick_h_> Makyo: sure thing
[17:57] <hatch> bcsaller: thanks - that seems to be the consensus elsewhere too
[18:01] <gary_poster> Makyo, trivial css to fix before release: http://comingsoon.jujucharms.com/sidebar/search/precise/django-rainbow-1/?text=django-rainbow shows text wrapping ignoring whitespace for word breaks in the description.  I bet this is to try and keep long strings from hosing the presentation, but hopefully there is a happy medium
[18:02] <gary_poster> hey jcastro, I got a complaint that a user could not connect mysql to django-rainbow in the GUI.  that's because the charm apparently doesn't work that way.  May I send the user your way?
[18:03] <Makyo> gary_poster, alright, will take a look.
[18:03] <gary_poster> thank you Makyo 
[18:04] <jcastro> gary_poster, marco would be better probably
[18:04] <gary_poster> jcastro, ok thanks
[18:06] <rick_h_> Makyo: are there suggested real qa instructinos then? (the non cheating db kind?)
[18:07] <rick_h_> gary_poster: or are you doing QA on Makyo's branch?
[18:07] <gary_poster> rick_h_, no on calls
[18:07] <rick_h_> rgr
[18:07] <Makyo> rick_h_, Set up  an environment with this branch of the gui (either through changing config or deploying and setting juju-gui-source), then juju deploy cs:precise/mysql-15 and upgrading; should get the message that the service was upgraded.
[18:07] <rick_h_> Makyo: rgr
[18:08] <rick_h_> Makyo: does it not work/do anything in sandbox then?
[18:08] <Makyo> rick_h_, sandbox will always grab latest charms.
[18:08] <rick_h_> gotcha
[18:22] <rick_h_> Makyo: got a second to hangout?
[18:22] <Makyo> rick_h_, sur
[18:23] <Makyo> e
[18:24] <rick_h_> Makyo: https://plus.google.com/hangouts/_/719c922cf26bc270b30d1bd557d17218caf34af2?hl=en
[18:36] <gary_poster> rick_h_, do bundles have working changelogs?  if so, how do we get them?
[18:36] <rick_h_> gary_poster: /me goes to look
[18:36] <gary_poster> thanks rick_h_ 
[18:37] <rick_h_> gary_poster: they're not in the current API passed out, I'd imagine we coudl generate them as new metadata like we did charm_metadata and files as a feature
[18:37] <gary_poster> ok cool rick_h_ thanks.  will add card
[18:37] <rick_h_> gary_poster: rgr
[19:02] <bac> rick_h_: i think i've found a bug in ingester that only occurs when a new versio of the charm existed.  trying to replicate locally but not sure i can enqueue a previous version.
[19:03] <rick_h_> bac: ah, benji added some logic to go back and find old versions? wonder if it's something out of that. 
[19:03] <benji> rick_h_: not yet (that work was paused for ice cream)
[19:03] <rick_h_> benji: ah, ok. So we don't have a way to go back and enqueue the previous version atm then?
[19:03] <benji> (what did land was code that checked out the actual version of the branch which was enqueued, instead of the most recent version)
[19:04] <benji> rick_h_: nope; my intent is that at ingest time we will query the charm store for the last N revisions and ingest them if they are not in the DB
[19:06] <gary_poster> hey marcoceppi_ .  I had user feedback that the GUI didn't allow him to connect django-rainbow to mysql.  Of course, it doesn't allow that because the charm doesn't allow that.  Could I connect the user to you to help them figure out how to move forward?  This thread is 10 days old, so it may lead anywhere, but it's worth trying
[19:06] <gary_poster> it may not lead anywhere
[19:07] <marcoceppi_> gary_poster: definitely
[19:07] <gary_poster> awesome thanks Makyo 
[19:07] <gary_poster> marcoceppi_, thanks I mean
[19:07]  * Makyo curtseys, anyhow.
[19:07] <rick_h_> bac: ok, want to go through it? Can we test out the issue in tests/mock data perhaps? 
[19:08] <rick_h_> now that the juju-core site is acting ok I'm not sure it's a huge deal. It appears that things are starting to work
[19:08] <rick_h_> might have jsut been several unreleated things breaking at once that seemed to be related to each other sending me scambling
[19:08] <rick_h_> scrambling that is
[19:08] <bac> rick_h_: sure, you want to chat?
[19:08] <gary_poster> Makyo, :-) actually something for you.  in your qa, could you try deploying a service from juju core and then seeing if dragging works properly in the gui?  I had an issue with that when I was hacking on deployer bits to get them deployable: after deploying the bundle, the charm drag doesn't work properly
[19:08] <rick_h_> bac: sure
[19:09] <rick_h_> bac: https://plus.google.com/hangouts/_/N5mZjt8pongxvvIFSRzcyA?hl=en
[19:10] <Makyo> gary_poster, sure
[19:13] <benji> gary_poster: the ice creams are going in the box
[19:13] <gary_poster> cool benji.  with relations?
[19:13] <benji> the condor flies at midnight
[19:14] <gary_poster> :-)
[19:14] <benji> gary_poster: yep
[19:14] <gary_poster> awesome benji!  thanks a lot
[19:15] <gary_poster> benji, back to ingestion, at least after catching your breath, then?  We have some bundle story 1 tasks left, but you were kind of in the middle of the ingest.
[19:15] <benji> gary_poster: yeah after I make a quick branch to fix a couple of small bugs I found while doing the ice cream stuff
[19:16] <gary_poster> cool thanks benji
[19:32] <gary_poster> rick_h_, do you mind if I forward your two observations about the upgrade charm UX to luca, cc-ing peeps?
[19:33] <rick_h_> gary_poster: sure thing, please do
[19:33] <gary_poster> thanks rick_h_ 
[19:35] <gary_poster> hey jcastro, for bug 1237605, we do that now.  I've seen it work in ec2.  it doesn't work quite as I want it to, but it works, as opposed to your screenshot.  what environment is this on?
[19:35] <_mup_> Bug #1237605: Public URL in the inspector should be clickable <juju-gui:New> <https://launchpad.net/bugs/1237605>
[19:35] <bac> rick_h_: are you turning the cron jobs back on?
[19:35] <gary_poster> oh
[19:35] <gary_poster> duh
[19:35] <jcastro> gary_poster, ec2. :)
[19:35] <gary_poster> it is ec2
[19:36] <jcastro> I deployed what's in the charm store as of 20 minutes ago
[19:36] <gary_poster> jcastro, ack.  weird.  I'll investigate; was working for me.  thanks for bug
[19:36] <rick_h_> bac: no, not right now. I think the plan is 1: review change, 2: land change, 3: watch for change to hit jenkins/staging update 4: see if cron job is auto turned back on by charm on upgrade and then either uncomment or check that the data is *fixed*
[19:38] <bac> rick_h_: sounds good
[19:41] <bac> benji: would you have a moment for a quick charmworld review?
[19:41] <benji> bac: sure
[19:41] <bac> https://code.launchpad.net/~bac/charmworld/update-branch/+merge/190229
[19:42] <benji> bac: ooh, that was probably my fault
[19:43] <bac> benji: well...i wasn't going to bring that up.  :)
[19:43] <rick_h_> gary_poster: so per mthaddon they managed to get the package into IS repos. The deploy is looking good: http://manage.jujucore.jujucharms.com/charms/precise/liferay has qa data and such in it
[19:43] <benji> :)
[19:43] <bac> benji: but 'bzr blame' spares no feelings
[19:43] <benji> bac: Looks good. It would be nice to add a test that would have caught this bug (i.e., really call the function with a dict).
[19:43] <rick_h_> gary_poster: so they're looking to continue the migration/etc tomorrow now. 
[19:43] <rick_h_> gary_poster: fyi and all that before I run away for today. 
[19:44] <gary_poster> rick_h_, thanks :-)  any quick update on the juju-gui probs on staging?
[19:44] <rick_h_> gary_poster: bac is proposing a branch now that should fix it. Once it lands it should re-ingest and hopefully 'self-heal' 
[19:44] <rick_h_> gary_poster: I've got the other bug as my next todo and will have that up tomorrow sometime if all goes well
[19:45] <gary_poster> awesome.  thanks rick_h_ !  thanks bac!  rick_h_ hopefully tomorrow you can emerge from emergency-land :-)
[19:45] <gary_poster> have a nice evening
[19:45] <bac> benji: test_update_branch_passes_on_subprocess_errors_if_not_retrying does pass a dict and it works
[19:45] <bac> benji: the test i correct now shows that it requires a dict.  i *think* we're covered.
[19:45] <benji> bac: then how did the bug survive?  
[19:46] <bac> benji: ah, because the call site of update_branch was wrong
[19:47] <benji> bac: I guess there is an agument to be made that there is still a missing test then, one that would have caught this bug
[19:47] <bac> benji: yes, looking the function 'fetch_branch' is completely untested.  i'll do that now
[19:48] <benji> cool
[19:49] <Makyo> OT: anyone know how to file a claim with BCBSMA?
[19:49] <benji> gary_poster: oh, I took this screen shot to show you the finished product and forgot to actually show you: http://i.imgur.com/aURfaEl.png
[19:50] <hatch> benji: :D I love it!
[19:51] <bac> benji: that's a lot of units of ice cream
[19:51] <gary_poster> benji, awesome :-)
[19:51] <gary_poster> benji, I will share with Mark B :-)
[19:51] <benji> bac: I have a big appetite
[19:58] <gary_poster> hey benji, on your taskbar on the top right, what is the blue icon of an S being squished by robot hands stand for?
[20:00] <benji> gary_poster: that is synapse (sp?) which is the launcher I use (shhh, don't tell anyone I don't use that keyboard-unfriendly unity launcher)
[20:00] <gary_poster> benji :-) ok secret is safe with me
[20:05] <arosales> gary_poster, hello I may have missed this and if so I apologize, but did the charm auth config option land?
[20:06] <gary_poster> arosales, yes, sent you an email last week.  It was in the dev charm then.  As of today, actually, it should be in the released charm.
[20:06] <arosales> gary_poster, ah great to hear
[20:07] <hatch> rick_h_: in your experience with widget children is it expected that the attributes aren't set until after render? Or is there just something bonkers with the Token ?
[20:07] <arosales> gary_poster, sorry for missing the mail last week (filters failing me)
[20:07] <gary_poster> arosales, np :-)
[20:15] <abentley> orangesquad, benji, bac: I've updated Jenkins so that permission to land is based on membership in juju-jitsu.  Both Yellow and Orange are members, so this shouldn't be a problem, but let me know if it is.
[20:15] <jcsackett> abentley: ack, thanks. :-)
[20:15] <bac> abentley: great
[20:31] <gary_poster> Makyo, first charm I deploy on comingsoon always has 0 units :-(
[20:31] <Makyo> gary_poster, oh, hmm.  Been working in local, let me switch to sandbox.
[20:32] <rick_h_> hatch: no idea. widgets are Y.Object so I'd expect the normal initialzier(cfg) setting ATTR as any other Y.Object
[20:33] <hatch> rick_h_: intersting - because setters in Token are not called if it's not rendered
[20:34] <hatch> so must be some wakoness with the Child thing
[20:41] <gary_poster> Makyo, I bet I know the problem.  trying something.
[20:43] <benji> I'll note something I heard from Kent (something I have heard from a few people): the outline of a big charm that is displayed over an empty canvas looks like the app is broken; people interpret it as a drop-shadow of an image that didn't load
[20:44] <gary_poster> benji, please either send to luca and cc peeps, or I will :-) though I think that's going away in his plans already as part of the onboarding.
[20:45] <benji> gary_poster: I'll send a message
[20:45] <gary_poster> thanks benji
[20:52] <gary_poster> Makyo, bzr+ssh://bazaar.launchpad.net/~gary/juju-gui/fixGoSandboxWatcher/ has fixes which makes it work, but also two test failures, and the change to _prepareDelta in sandbox.js is not right.  At least one of the tests is wrong.
[20:52] <Makyo> gary_poster, alright, will merge that with the style branch and see what I can do with it.
[20:53] <gary_poster> Makyo, looking more but have to go soon
[20:53] <gary_poster> cool
[20:53] <gary_poster> thank you
[20:59] <rick_h_> hatch: http://yuilibrary.com/yui/docs/api/files/widget-parent_js_Widget-Parent.js.html#l453 it just calls new XXX on the child and returns it
[20:59] <rick_h_> hatch: nothing strange in the parent there
[20:59] <gary_poster> Makyo, fixed one test and pushed; looking at other
[20:59] <rick_h_> hatch: setters aren't called on any Y.Object init I thought :/
[20:59] <hatch> rick_h_: it would probably be in the childs init cycle not the parent
[20:59] <hatch> rick_h_: Y.Base.create() you mean?
[21:00] <hatch> and yes, attr setup is done before initializer is called
[21:00] <rick_h_> hatch: the child is just an instance of a widget
[21:00] <hatch> so Y.WIdgetChild is empty?
[21:00] <rick_h_> http://yuilibrary.com/yui/docs/api/files/widget-child_js_Widget-Child.js.html#l9
[21:01] <hatch> interesting, it's just a constructor
[21:01] <rick_h_> it mixes some stuff in but doesn't touch init/construction really. Just adds a few attrs and helper methods
[21:02] <rick_h_> hatch: but I thought setter wasn't called in normal Y.Object init. I know I"ve got a comment somewhere about 'better way to do this' because of that issue
[21:02] <hatch> why do you keep saying Y.Object init?
[21:02] <hatch> am I missing something
[21:02] <hatch> ?
[21:02] <rick_h_> hatch: see the store/charmworld.js and the initializer there 
[21:03] <rick_h_> hatch: it seems you're complaining that a setter isn't called when a WidgetParent creates instances of WidgetChildren
[21:03] <rick_h_> hatch: because your setter isn't called
[21:03] <hatch> right...but what is a Y.Object?
[21:03] <rick_h_> hatch: but that's never going to be true for YUI Objects as I ran into with store/charmworld.js
[21:03] <hatch> where are you creating instances using Y.Object?
[21:03] <rick_h_> Y.Object == my shorthand for a YUI Object with ATTRS/etc
[21:03] <rick_h_> hatch: sorry, bad notation for that. 
[21:03] <hatch> ok so Y.Base.create() :)
[21:03] <gary_poster> Makyo, other test shows that either my change to normalize to "annotations" is correct, or it should be "annotation".  Thing to do there is verify with Go environment
[21:03] <hatch> ok got it
[21:04] <hatch> ok now I have to write up a quick bin
[21:04] <hatch> I can't believe I've had this wrong the whole time haha
[21:04] <rick_h_> hatch: so anyway, I think your issue is nothing to do with widgets at all much less parent/children, but YUI classes
[21:04] <gary_poster> I don't know which but would not be surprised if "annotation" were correct (and so you'd need to revert those parts of my diff and change the sanbox to generate "annotation" not "annotations"
[21:04] <gary_poster> )
[21:04] <gary_poster> Makyo have to run now
[21:04] <gary_poster> ttyl and thanks
[21:04] <gary_poster> bye all
[21:05] <Makyo> gary_poster, thanks, cheers.
[21:05] <hatch> rick_h_: I'm just writing a bin right now to see if I've been wrong this whole time
[21:07] <hatch> http://jsbin.com/AwaMUWI/1/
[21:07] <hatch> ^ rick_h_
[21:07] <rick_h_> hatch: looking
[21:07] <hatch> I knew I wasn't going nuts
[21:08] <hatch> I'll create one with a widget now
[21:08] <hatch> just to see
[21:08] <rick_h_> hatch: well I'd love to see why the charmworld.js one didn't work then :/
[21:09] <hatch> http://jsbin.com/AwaMUWI/2/edit
[21:09] <hatch> rick_h_: same with the widget.....now I'm confused
[21:09] <rick_h_> hatch: sec testing
[21:10] <rick_h_> hatch: yea, have to do family stuff so not sure. There's something there. 
[21:11] <hatch> rick_h_: ok cool - well we should look into this to find out what's going on sometime in the future :)
[21:11] <rick_h_> hatch: agreed
[21:13] <Makyo> Going to walk dogs, back to this in a bit.
[21:23] <bac> hey benji, hate to nag but your r409 of charmworld introduced a lot lint.  weird things you wouldn't expect like wrong python indentation levels.
[21:27] <bac> benji: though i think some may be things reasonable people could  disagree with the linter
[21:31] <benji> bac: we really need to get the linter into the landing process; either start using lbox or have "make propose" and "make land" targets 
[21:32] <bac> benji: or the lander?
[21:32] <benji> bac: I would say "and the lander"
[21:32] <benji> the tighter the feedback loop the better
[21:34]  * benji goes away.
[21:40] <hatch> bcsaller: do you have some time today to run through the issue I'm having with promises?
[21:40] <hatch> see if your technique, or a modification of it, can solve these issues
[21:55] <bcsaller> hatch: sure
[21:55] <bcsaller> hatch: want to talk now?
[21:55] <hatch> sure
[21:56] <hatch> calling
[22:03] <huwshimi> Morning
[22:03] <hatch> morning huwshimi
[22:06] <hatch> huwshimi: so were you able to get a direct flight for the next sprint?
[22:07] <huwshimi> hatch: I always have a flight within Australia (no international flights from my local airport) and then limited flights direct to SFO, so ended up flying via LAX again.
[22:07] <huwshimi> hatch: So, three flights :)
[22:08] <hatch> haha, that middle one is still going to be a long one :D
[22:09] <huwshimi> hatch: Yep, about 14 hours
[22:10] <hatch> lol
[22:10] <huwshimi> hatch: I think the same flight is 15 hours on the way back
[22:10] <hatch> clearly they need faster planes
[22:11] <hatch> fly to space....wait.....fly back down
[22:11] <hatch> clearly the most efficient
[22:12] <huwshimi> :)
[22:29] <hatch> jujugui looking for a review on https://codereview.appspot.com/14591044/ plz
[23:02] <huwshimi> hatch: Just wanted to check in with you and see if there's any more bundle work that needs my attention yet?
[23:02] <hatch> well my branch lands the icons but I think that layout stuff was already done
[23:02] <hatch> I'm working on the source tab now which I think should be able to keep the same styles as the charm one
[23:02] <hatch> so....
[23:03] <hatch> thats all I know of
[23:03] <hatch> maybe you could do more markup/styling of the bundle data overlay thing
[23:03] <hatch> (the white box when you click a charm in the bundle topology)
[23:03] <hatch> it needs a icon in there
[23:03] <hatch> an
[23:04] <hatch> but that's pretty trivial I think?
[23:04] <hatch> I'm not really sure how far down that rabbit hole you want to go :)
[23:05] <huwshimi> hatch: OK, I'll take a look at that icon...
[23:06] <hatch> you can see it in the two bundle mockups
[23:41] <hatch> jujugui now I'm looking for a review on https://codereview.appspot.com/14591044/ and https://codereview.appspot.com/14595045/ thanks
[23:46] <gary_poster> hatch trying to squeeze them in quickly...
[23:49] <hatch> gary_poster: haha, just gettiner-dun
[23:50] <gary_poster> awesome
[23:51] <hatch> gary_poster: for the AM, which card/bug would you like me on next?
[23:51] <gary_poster> hatch bundle visualization?
[23:51] <gary_poster> caption box stuff we were talking about
[23:52] <hatch> cool, assigned to myself and put it in starting
[23:52] <gary_poster> cool thank you
[23:54] <gary_poster> hatch, for charm icons, "conditional LGTM on filtering out the non-approved charms, per below."  trying qa now
[23:55] <hatch> oh woops
[23:56] <hatch> I should also add the ellipsis
[23:57] <gary_poster> hatch, yes qa good if you add the ellipsis.  ellipsis should only be added if charms have been omitted (because we didn't show a non-approved charm, or because we had more charms than we could show (10 is max?)
[23:58] <hatch> right, yep