[00:00] bcsaller: back [00:01] hatch: let me push what I have but I know its your EOD [00:01] cool - I'm just about to make some supper then I can take a longer look at it [00:07] hatch: lp:~bcsaller/juju-gui/service_inspector behind :flags:/serviceInspector/ [00:07] * bcsaller takes a break [00:56] bcsaller: I think I have c9 setup but need someone to try it out with.... [01:02] hatch: got a link on what I need to do? [01:02] well you need to sign up [01:02] which is very easy [01:02] then I need your username to grant you permissions [01:03] hatch: its bcsaller [01:03] ok now to find out how to add you [01:03] :) [01:03] https://c9.io/hatched/juju-gui [01:03] there is the url [01:04] ok granted you read/write access [01:05] open the COPYING file [01:06] seems to be working though I don't seem to have an indicator of what you have open and are editing [01:07] I Just typed under Preamble in COPYING [01:07] hmm [01:09] I just highlighted the to paragraph of the COPYING file....can you see that? [01:09] according to the vids this should 'just work' [01:09] hatch: doesn't seem to be showing edits [01:10] well that's lame [01:10] does the chat work on the right? [01:18] going to get back to debugging the rendering [01:19] alrighty [01:19] :) [01:19] you know we could easily parse data-bind="name" out of rendered DOM, but we'd lose the easy programatic config [01:19] well, not lose, be at odds with [01:19] yeah I had that in one of my refactors [01:19] but I was torn [01:19] and reverted :) [01:19] yeah [01:20] * hatch goes to delete the c9 account [01:22] thx for the help [12:15] bac, looks like we have some stats. now the question is. how do we read them :-) [12:16] mysql-bac has average value of 14.72, mysql-x has average of 7.56, service (?) has average of 7.39 [12:16] bac, looks like we have some stats. now the question is. how do we read them :-) mysql-bac has average value of 14.72, mysql-x has average of 7.56, service (?) has average of 7.39 [12:17] oh interesting. [12:17] gary_poster: sadly i'm not a good scientist as for this round i didn't keep records. i recall setting one to 15 units, so that would correspond with mysql-bac [12:18] ok [12:18] gary_poster: and there was the simulator perturbing the data [12:19] rt [12:40] rogpeppe, hi. I'd like to get your replies to the "Juju GUI API addresses handling" email thread (William said you had some plans in that regard) and to the "Juju core tasks for GUI project" thread (William and I both would like your thoughts on a reasonable way forward with the restricted mode that I mentioned and you prototyped. Do you have a chance of replying to those this afternoon? If not, np. Those two ar [12:40] e not immediately pressing, though I want us to agree on our way forward. I have a "ping Roger" to do item that I can fire off again next week sometime to try and get your thoughts :-) [12:41] gary_poster: yes, i should be able to do that [12:41] thanks rogpeppe :-) [12:44] fwereade__, thank you very much for your replies to us about the gui needs. The set unit work would would ideally be implemented first of our juju core bits, but did I understand correctly that, even were we to decide to try to implement it ourselves, we need to wait for the same changes (a major version upgrade mechanism and/or moving the DB entirely behind API) that core would be waiting on? that is, it is not m [12:44] erely a matter of bandwidth but also technological infrastucture? [12:44] teknico: ping (or are you gone today?) [12:45] gary_poster: ^^ ?? [12:45] gary_poster, I'd like to chat to someone about that [12:45] gary_poster, I had a suggestion that might not involve a schema change, and might make for a useful model [12:45] involving SetMinimumUnits? [12:45] gary_poster, yeah [12:46] bac I think he is here today. out wednesday this week only [12:46] ok [12:46] no rush [12:46] gary_poster, can't chat today though, I'm afraid, just driving by [12:46] fwereade__, ok, maybe I can try to get a meeting with you, me, and frankban sometime next week? I'll propose on calendar? [12:47] gary_poster, perfect, early monday is fine by me, I'm really keen to unblock you [12:47] fantastic, thanks fwereade__ . ttyl [12:47] gary_poster, but whenever works :) [12:47] gary_poster, cheers [12:51] frankban, do you mind helping out with analysis of the concurrent set unit discussion Monday at 1430 UTC? It might lead to you leading that implementation eventually, so be warned. :-) [12:53] gary_poster: I am definitely interested :-) [12:53] cool thanks frankban :-) [13:07] jcsackett: morning, I'm still working on getting tests into the branch, but if you get a second can you QA it and make sure it does solve the black bar problem you saw? https://code.launchpad.net/~rharding/juju-gui/details-double-dispatch/+merge/168053 [13:08] jcsackett: I don't see it, but I want to make sure I'm doing the exact steps you did to find the issue. [13:09] rick_h__: looking. [13:10] rick_h__: it doesn't. [13:10] bac: I'm back from lunch, what's up? [13:11] rick_h__: going to this url http://jujugui.local:8888/fullscreen/search/precise/juju-gui-61/:flags:/browser_enabled/?series=precise&text=gui&type=approved still causes black bar horror. [13:11] teknico: not much, i was just wondering about the spelling for the config setting to disable GA. i'm writing some docs [13:12] teknico: and my thoughts about exposing it in the charm were truly cracktastic. please ignore. [13:12] rick_h__: oddly though, if i move the ":flags:" bit to be the first bit (which i think is correct) no black bar. [13:12] bac: the idea was to use "use-analytics: true|false" [13:13] rick_h__: the latter url was the one i was actually using yesterday. so partial success at least. the flags thing may not be an issue, since i think flags are supposed to be the first component of a url anyway? [13:13] teknico: ok. that works [13:16] jcsackett: k, looking [13:21] hey sinzui, I went to http://tinyurl.com/orange-standup but hangouts said the party was over. what's the new daily url? [13:23] gary_poster: https://plus.google.com/hangouts/_/cf491220dfb7736cee2876baf29110eeb52f5997?authuser=1 [13:24] thank you rick_h__ [13:25] I think that old URL is from deryck's days time with orange. [13:25] * sinzui deletes the url === BradCrittenden is now known as bac [13:48] mornin [13:50] gary_poster: who is that hairy man kissing karen on facebook? [13:50] bac, I dunno, but he's in some big trouble. [13:50] gary_poster: so, zooming in and out should be on the basis of a service, eh? [13:50] gary_poster: h.a. [13:51] :-) [13:53] benji, something like that, yeah. Maybe scrolling should be on the basis of the center of the viewport instead? [13:53] benji, goal is to handle case of services way out from to pleft origin [13:53] benji, current zoom is always from top left [13:54] which doesn't work well in some cases [13:54] gary_poster: it has always felt wierd to me that zooming wasn't centered on the mouse cursor [13:54] benji, I suggest having a pre-imp with Makyo when he is available, if you want to tackle it [13:55] gary_poster: ok (the card said you or Ben, and he out timezone-manuvered you) [13:55] [I need to learn to spell one day.] [13:55] benji, heh, you talked with me to get the goal of the card (make env work better in that case, which is what happens with silly people on uistage [13:56] hey hatch, do you recall where/what causes the second dispatch? [13:57] deltas [13:57] every time a delta comes in it will dispatch [13:57] Makyo: yo [13:58] so if i just load '/', there's always a second dispatch. where in the code might i find the event handler (or whatever) that is doing the delta dispatching? [14:00] grep for dispatch [14:00] its in a method called on_delta_something_something [14:00] it's similar to the "something_something_dark_side" method [14:01] * jcsackett laughs [14:01] thanks. [14:01] :D np [14:05] alejandraobregon: ping, any chance your around and available? [14:07] hey rick, soo sorry, been flat out! [14:07] rick_h__: will ping you when am out of my meetings [14:07] alejandraobregon: np, I'd be great if we can just schedule something? [14:07] alejandraobregon: ok, yea just want to let sinzui know what time table I'm looking at on the branch. [14:07] alejandraobregon: have fun with your meetings :) [14:08] benji, yo. [14:09] Makyo: hi, can we have a pre-imp call about the "zooming in and out should be on the basis of a service" card [14:10] benji, sure, guichat's free. [14:10] sounds good [14:18] Makyo: ping me when you're done? [14:18] teknico, ping. :) [14:19] gary_poster: it seems that zooming happens relative to the mouse cursor already so Makyo and I are not able to determine what the card is about (there is a bug that means that if the cursor is over a service no zooming happens, I could work on that) [14:19] Makyo: guichat? [14:19] teknico, yep [14:20] gary_poster: could you please review my juju-test branch? https://codereview.appspot.com/9874050 [14:20] benji frankban on call will ping [14:20] thanks [14:20] gary_poster: ok thanks [14:32] benji, Makyo, what happens if you use the zoom controls on the bottom of the page? [14:32] good question; trying now [14:33] rick_h__: is it valid for a url to have both '#' and '?' [14:33] and if so, is ordering of those two important/specified? [14:33] hoping you can save me a trip to URI specification docs. :-P [14:34] jcsackett: yes [14:34] jcsackett: no, I think usually the ? is first ,the # is second. JS makes location.hash available for the # part [14:34] rick_h__: cool. thanks. [14:35] jcsackett: sorry, other way aroud on the order [14:35] gary_poster, that's set to zoom around (0, 0), zooming around the center of the current view would be easier, but would rely on code from my branch. [14:35] #someanchor?query=true [14:35] rick_h__: yeah, was just writing a sample url the wrong way and thought, "this doesn't seem right". :-P [14:37] Makyo, cool, thought it was 0,0. +1 on zooming around center of current view. Is that something you should slip into your current branch, ir that benji should work on as a follow on to your branch, or...? [14:37] s/ir/or/ [14:38] gary_poster, I can do that in my branch, once I get the scale stuff working properly. There's still the issue of zoom events not bubbling [14:38] Makyo, ok thanks. benji ^^ s'ok with you? [14:39] yep [14:39] thanks benji. Makyo, benji, I'll combine text of benji's card with Makyo's and then delete card. :-P [14:39] Makyo: it shure looks like it zooms around the center when using the zoom slider [14:39] k [14:39] Cheers. [14:39] Makyo: deleting all cookies wasn't a great idea :-) [14:40] teknico, oops :) [14:40] teknico, I found you can delete them from the console in the resources tab. [14:41] benji, on uistage, zoom out. drag canvas to left so services are just to the left of where you can see them. zoom in. [14:41] hatch: around? [14:41] you bet [14:41] just reading your branch [14:41] hatch: and you got those emails last night? [14:42] from last night :) [14:42] yup [14:42] ok cool [14:43] frankban, fwiw, I don't have a lot of time. I will review code but will not have time for qa till later today or Monday [14:43] let me know how much you want me to do [14:43] gary_poster: (ui stage isn't working for me, but I used a local server) I'm not sure what that is supposed to prove. If it zooms to the center of the visible part of the canvas (the current behavior and a good one, IMO) then if you follow those steps you will not be able to see any services, but that's not unexpected [14:44] benji, want me to show you on guichat? :-) [14:44] gary_poster: sure [14:44] benji oops it is busy [14:44] 1 sec [14:45] heh, yeah I just noticed [14:45] bcsaller: yeah I see the issue here [14:46] we can get together after the calls [14:46] hatch: sounds good [14:46] we definitely need a flag to stop people from opening multiple of the same inspector [14:48] hatch: we can do that with the tracking map in the env view, but I think we might want a better fix than just that [14:54] gary_poster: code review would be great, but please don't worry if you don't have time. [14:54] cool [14:54] jujugui call in 6 [14:55] kanban now [14:55] * benji opens up the branch start check list. [14:59] jujgui call in 1 [14:59] jujugui [15:00] jcsackett: ok, try to break please lp:~rharding/juju-gui/details-double-dispatch :) [15:02] rick_h__: works! [15:03] jcsackett: party, now how the $#@#@ to test it all... [15:03] hah, all existing tests pass...just lp-submit now in a hurry :P [15:28] rick_h__: hi! [15:28] rick_h__: so sorry, crazy day! [15:28] hang out now? [15:29] alejandraobregon: give me 3min (hopefully) in the middle of submitting another branch before I can switch over to dmeo [15:29] rick_h__: no problem [15:30] rick_h__: do you have a live thing like uisearch but with trunk handy? [15:31] jcastro: http://uistage.jujucharms.com/:flags:/browser_enabled/ ? [15:32] alejandraobregon: invite coming in a sec [15:35] alejandraobregon: https://plus.google.com/hangouts/_/129fd072762c185e15013cb63d854c7e87c68cd4?hl=en [15:38] rick_h__: sorry couldn't understand what you were saying... [15:43] rick_h__: i have to leave shortly to pick up my 1YO from nursery... [15:43] rick_h__: can we cover it in ten mins? [15:47] alejandraobregon: yep, moved to my home connection sorry [15:47] alejandraobregon: starting up a quick hangout, should take just a couple of min [15:47] rick_h__: ah, there you are! [15:47] cool! [16:02] sinzui: gary_poster so we have the ok to land the quality tab changes for now with agreement to re-evaluate/discuss it with luca upon his return. [16:02] rick_h__, +1 [16:02] sinzui: gary_poster and if there's a meeting I'd be happy to sit in and present what we did and why I think it makes sense. [16:03] yay, rick_h__ . thank you to you and alejandraobregon . rick_h__ would it be relatively quick to write up an email to juju-gui with the reasoning? then we can see if we can luca agrees with basic idea without a meeting? [16:04] gary_poster: I can, is everyone on that list then? [16:04] rick_h__, yeah pretty sure. cc luca and ale just to be sure [16:04] gary_poster: do you have a direct list then? /me does't see ~juju-gui with a list [16:05] oh! yeah [16:05] pretty sure you are on it but could be wrong [16:05] rick_h__, Juju GUI Developer List [16:05] * rick_h__ checks list membership [16:05] rick_h__, it is an Ubuntu-style list, not LP [16:06] gary_poster: ah, gotcha. [16:07] teknico: in guichat [16:13] jcsackett: sinzui hatch bcsaller or anyone else that's intersted, can I get a couple reviews to defeat the great tabs-induced black bar of doom once and for all? https://codereview.appspot.com/10086045 referring to bug #1175019 [16:13] <_mup_> Bug #1175019: staging has issue with black bar at top of fullscreen charm details

[16:13] jsPlumb looks pretty cool [16:14] rick_h__: looking. [16:15] rick_h__: looking as well [16:15] thanks guys [16:25] * rick_h__ fetches remote food for lunch today [16:33] trunk is broken with make prod [16:41] oh wait [16:42] ignore me :) [16:42] hatch, Done and done. [16:42] :P [16:42] I kid, I kid :) [16:42] bcsaller: shall we chat? [16:43] hatch: joining [16:51] gary_poster: thanks for the review! [17:19] gary_poster: aargh, i realise i totally haven't replied to your request and i've reached eod. i will try to do it monday morning - please hit me over the head if i don't! [17:21] gary_poster: do we have our call in 9? [17:22] rogpeppe, :-) np thanks. have a good weekend [17:22] hatch yes [17:25] great I didn't forget this time :) [17:36] abentley, rick_h__, the charmworld deploy script does not get the current charmworld charm, and since it is tainting the charm, the versions will never be synced. I am working to get the correct version of the charm installed now [17:41] sinzui: interesting. So we need a long term solution for this then as we update the charm? [17:42] we do [17:43] I am currentl;y discussing branch tags to compare version [17:44] gary_poster: email away to the list now that uistage is updated. [17:44] yay thanks rick_h__ [17:44] gary_poster: also heads up that the black bar issue is closing up. Should hopefully unblock some. [17:46] hatch: currently we are not limited to a single panel, we should think about how we want to handle that. I'm tempted to add the YUI drag handler to the panels on creation so we can just move them around for testing [17:48] yay rick_h__ ! ripping out charmbrowser flag today or next week? [17:48] gary_poster: not sure, I don't know what the current checklist is for that. [17:49] jcsackett: ping, how goes it? Want to chat now that I've got this stuff landed? [17:49] rick_h__, I think the checklist has been checked. black bar, bad typing experience of -, and quality tab [17:49] " bad typing experience of -" ? [17:50] rick_h__: we can chat. i found the problem just before grabbing food. [17:50] jcsackett: k [17:50] rick_h__: but i would love to chat about best solution. :-P [17:50] jcsackett: why don't you invite away [17:54] bcsaller: yeah I see no issue in that, you should be able to just add it to the view-container template [17:56] rick_h__, I think your dash bug is fixed in production: https://manage.jujucharms.com/api/1/charms?categories=app-servers&series=precise&text=&type=approved [17:59] bcsaller: mind if I grab some lunch and then we can get back at it? [17:59] hatch: go for it [17:59] cool - i'll be here so just ping if ya need [18:03] jujugui Is uistage's improv down again? [18:03] Makyo, not for me [18:04] gary_poster, Huh, will try again. Getting "websocket is closed before connection is establised" [18:04] only spelled right. [18:04] looks okay on the server [18:04] Makyo: w/o the :8080 wfm, the objects are off the canvas by default [18:04] :-) [18:04] bcsaller, oh, derp. Forgot we removed the :8080 [18:04] 8080 wfm too [18:05] gary_poster, I get the error only on :8080. [18:05] wfm. just did a hard reload [18:06] hatch: need to grab you for a quick chat when you're back from lunch. [18:06] sinzui: yay! [18:06] sinzui: so maybe you can get with gary_poster on the un-feature flagging of the browser. [18:06] sinzui: sounds like we're darn close and might be able to monday or tues. [18:07] +1 [18:07] yes [18:07] sinzui: I'll add a card to the board to remind me to bring it up monday then. Should be a short 3 or 4 line change [18:08] sinzui: so jcsackett is working through our routing/url dropping QS and #xxx issues. I'm giong to move on to api2 then for now and wait for his work to fall through if that's ok [18:09] +1 rick_h__ [18:38] hatch: give it a try when you get back, you can have multiple open bound draggable panels now [18:50] will do [18:53] bcsaller: awesomer [18:54] hatch: what do you think about adding tabs for the viewlets next and then a second one? [18:55] sure - just for prototype though right? UX will probably have something specific in mind [18:59] yes [19:10] ""Regenerated Templates"" haha friggen finally! [19:10] I don't think I have EVER seen that before - that must have been broken forever [19:19] bcsaller: w00t tabs work :) [19:19] that was easy.... [19:19] lol [19:19] mesa thinks all that planning made the coding a lot easier :) [19:22] hatch: skip on needing to chat; can you look at https://codereview.appspot.com/10119043 [19:22] jcsackett: we needed to chat? [19:22] oops sorry :) I missed the ping [19:22] i pinged you earlier about talking when you got back from lunch. it's all good. :-) [19:22] i' [19:23] i'm not sure we actually needed to. [19:23] reviewing [19:27] jcsackett: guichat? [19:27] sure [19:30] hatch: you have a branch link I can merge? [19:32] bcsaller: is this good enough? can bzr 'apply' diffs? https://gist.github.com/hatched/4bedc5bd0e9dceb96f32 [19:33] hatch: bzr push lp:~hatch/juju-gui/service_inspector and then I can merge it [19:33] cool one sec [19:35] bcsaller: done [19:36] jcsackett: lgtm'd with comment as per the hangout [19:36] thanks! [19:36] hatch: cool, thanks [19:38] bcsaller: I used data-attrs for the tabs as you can see - figured i'd follow the same convention as with the mappings [19:40] hatch: yes, I think that makes sense [19:41] I improved the styling, changed it to inline-block [19:42] yeah I just hacked some inline css in there haha [19:43] so next step is to add some inputs to see if we can get the updating to work [19:43] hatch: should work out of the box [19:43] but I can change the template now [19:44] sorry I meant with the units side [19:45] hatch: chat? [19:45] yup [19:46] orangesquad: Could you please review https://code.launchpad.net/~abentley/charmworld/hide-same-series/+merge/168167 ? [19:46] * sinzui looks [19:59] bcsaller: can you look at https://codereview.appspot.com/10119043/ ? [20:05] orangesquad: could you please review https://code.launchpad.net/~abentley/charmworld/migrations-tests/+merge/168170 ? [20:05] abentley: sure. [20:05] abentley, jcsackett sorry, it was r=me [20:06] sinzui: No, that's a second branch that jcsackett agreed to review. [20:06] oh, sorry. nm [20:07] abentley: r=me. [20:07] jcsackett: Thanks. [20:20] folks saw ubuntu.com today right :-) [20:26] OOoooo look at that [20:26] hah! cool :-) [20:27] gary_poster: is there a reason why we don't have units as part of a service model? [20:27] just wondering if there was a big architectural reason I'm missing [20:30] hatch, probably because that's how they are modeled in juju. Probably also because it is a bit relational-db-like the way it is. Maybe because there's no guarantee that we will hear about a service before we hear about one of its units? Maybe there is a guarantee like that.... Anyway, it was a decision from before my time, but I don't have an argument against it. Do you? :-) [20:30] I do! :) [20:31] but I think I have a workaround [20:31] heh [20:31] ok [20:31] just an issue we are running into with the bindings [20:31] feel free to send an email or make a card [20:31] yeah will wait to run my idea past ben when he gets back [20:32] *I will [20:33] cool [20:34] jcsackett: it seems like you could just put that code on nsRouter.split and it should work, no? I assume you tried that and hit an issue? [20:35] hatch: back, btw [20:35] that was quick [20:35] guichat? [20:35] there [20:43] * gary_poster running [20:43] have a great weekend [20:43] you too cya! [20:47] bcsaller: it's more that .split is implicated in more places and i'm cautious about what i modify without any understanding of this code. [21:01] ok, benji your sphinx checks have bitten me trying to propose my branch. hope you're happy. :) [21:01] heh === BradCrittenden is now known as bac____ [21:14] bcsaller: the events object in the view container should probably be user defined [21:14] or....do we want people to subclass view-container instead of init directly? [21:15] events and var viewletContainer = container.one('.viewlet-container'); should be user defined IMHO so that it can match the templates [21:15] and view-container can then be a general use wrapper [21:15] hatch: are events even used/needed here? [21:15] well it's the view container which houses the navigation [21:17] so we should really be subclassing the view-container [21:17] which is probably overkill :) [21:20] hatch: agreed, at this time I dont' see the need either. The controller should configure that layer I think [21:22] ok so I should make the events object user configurable? [21:22] and the viewlet-container element [21:23] hatch: I'd think the container element should be ok to be hard coded. Provides consistancy and you can always drop in your own node inside of it. [21:24] hatch: nvm, I was reading that as the container itself, but you're talking about the node inside the container already [21:24] :) [21:34] hatch: it works and its pretty cool, pushing. [21:34] awesome - I'm just finishing up my changes to view-container and it's tests [21:36] all pushed [21:36] you can watch the inspectors of multiple panels update the unit viewlets with the simulator running [21:37] excellent - lp:~hatch/juju-gui/service_inspector was updated [21:38] so you can pull those changes [21:41] I think my simulator is broken [21:41] oh nm there it goes [21:41] right on [21:41] :) [21:42] this is pretty damn cool [21:42] yeah [21:43] like I want other people to see this already :) [21:43] haha [21:43] * hatch ships a bunch of redbul to UX [21:44] ha [21:44] excellent I'm glad that idea worked out [21:45] yeah, I thought you meant creating them on the fly at first and I was opposed, but I keep them up to date in the delta stream [21:45] extended process_delta to be able to handle this case [21:46] I'd get rid of the global list now :) [21:46] I wasn't sure if that was being used anywhere [21:46] only to select the subset by service [21:47] we need a close button on these inspectors [21:48] yup [21:48] I can do that [21:48] did you merge in my changes? [21:49] yes [21:49] ok I'll blow this branch out and pull down a new one [21:49] and then add that [21:49] you can just merge mine back in and keep going [21:50] ahh ok that'll be faster [21:51] yeah, just pushed again [21:54] cool - this is going to take a few minutes [21:54] then I'll write the tests for it [22:06] almost done [22:07] aggregated status is computed and doesn't fire a change event. might make sense to aggregate as status_error: {Number}, etc so the binding will work with it [22:07] directly on the service as attrs I mean [22:08]