[00:02] hatch, hi. if unit destroyed but service is there, go to service view. if service destroyed, go to environment view in all cases. notification would be nice-to-have addition, maybe. [00:03] alright - yeah I figure that a notification would be less jarring if you're looking at it and all of a sudden 'boom'....where did it go [00:03] yeah [00:04] fyi - I just triggered a new run on the CI as I coudln't reproduce the failure locally [00:07] ugh, not that [00:07] 2 failure(s) running 80 tests. Retrying. [00:07] 2 failure(s) running 74 tests. Retrying. [00:07] 2 failure(s) running 72 tests. Retrying. [00:07] 2 failure(s) running 72 tests. Retrying. [00:08] not working tonight, hatch. might check in later. [00:08] yeah no problem :) [00:08] have a good night [00:10] ugh now 179 tests are failing in chrome :/ [00:27] I'm guessing someone introduced another bad cleanup :-/ [00:27] I thought you weren't working tonight ;) [12:26] bac, hi. does the config issue you were mentioning last night possibly mean necessary changes in juju core? [12:27] gary_poster: no, i can deal with it on the gui sie [12:27] side [12:27] it is just a matter of the juju-core developer (me) not communicating with the juju-gui implementer (me) [12:28] bac, lol ok cool [12:28] thanks [12:28] sending Config expecting Settings [12:28] ah gotcha [12:30] benji, I ran out of time yesterday evening in the add button conversation. I'm concerned that the visual approach is going to be untenable for full screen and problematic for side screen unless it is *truly* side screen [12:31] benji, do you know if that's resolved, or is rick going to take that to UX? I think it needs to go to Jovan [12:32] gary_poster: so for now it ditches the details part of the UX and I've got on my list of items for our standup to bring it up with jovan [12:33] gary_poster: or actually, because my branch hasn't landed it can't redirect anyway. So I recall saying it could be left alone and I can follow up [12:35] gary_poster: I've got a fix for the UX from Huw. I'm pulling it down and giong to get it into lbox and qa in a sec [12:35] great thanks rick_h_ [12:35] * gary_poster remembers to turn the volume back up from zero :-P [12:38] rick_h_, I'd like to be part of that jovan call please--the interaction with charmbrowser and rest of UX is something I want to make sure is clean, since the two have been separate for so long [12:39] gary_poster: k, I figured to bring it up on our stand up 9:30am est but can see if something seperate works better? [12:39] rick_h_, I am happy to show up then if I have a link [12:40] gary_poster: sure thing [12:40] cool thanks [12:40] heh, sorry no good reason to /msg that. Anyway [12:45] rick_h_: is staging.jujucharms.com being down a known thing? [12:45] gary_poster: have a sec to peek at https://codereview.appspot.com/8853043/ then? qa's ok here [12:46] benji: yes, it's gone down the last two nights. I want to leave it to allow abentley and sinzui to debug [12:46] rick_h_, looking thanks [12:46] benji: they should be up/at it in the next 30min or so [12:46] rick_h_: is there a way to develop the sidebar without it? [12:47] benji: run a local charmworld instance and change the url in the config [12:47] benji: you might try using the manage.jujucharms.com 'production' site [12:47] benji: mthaddon is working on it. Up at the moment, but only has around 50 charms [12:47] ok [12:47] http://manage.jujucharms.com/ [12:48] thanks [12:59] rick_h_, huw's branch breaks resizing windows while charmbrowser is open [12:59] * gary_poster checks uistage [12:59] gary_poster: looking [13:01] gary_poster: yea, see that it overlaps the bottom nav. Looks like it only does it on window shrink, but on grow it works again. It's on uistage and this branch so don't think it causes it. [13:02] rick_h_, ack agree. I'll give it a LGTM, thanks to Huw and you. [13:02] gary_poster: thanks for looking it over [13:31] jovan2: standup? [13:32] the two failures coming from CI are real failures unfortunately I can't reproduce them locally - I get a number of failures locally in FF [13:44] annnnnd now I get no failures lol [13:44] FF takes 334s to run through the tests [13:47] if I have FF minimized while the tests are running I get 6 failures [13:47] if it's focued I get 0 [13:59] oh hey hatch [13:59] I just sent an analysis to the list as well [14:00] I've seen these two errors before but they were very rare. They happen all the time now on CI. [14:00] I can dupe them, with effort, as I said. [14:00] oh, rick_h_, are you going to clarify with benji about the behavior we heard from jovan? [14:01] gary_poster: sure thing, sec [14:01] I can always use a good clarification. [14:06] thanks rick_h_ . [14:09] gary_poster: ahh ok I see the email - do the tests also run 3x slower in FF than Chrome for you? [14:09] jovan2, are you available for a quick call? [14:10] hatch, 19 seconds chrome vs 33 seconds FF [14:16] 38s chrome vs 334 ff ....ok so 9x slower [14:16] :) [14:39] gary_poster: so the updating of service settings doesn't seem to work in pyjuju or juju-core. this is expected? [14:39] bac, uh? [14:40] bac, I would guess that's a no. [14:40] benji: howdy [14:41] gary_poster: give it a try on uistage if you have a moment [14:41] ok [14:41] benji: so after talking with UX they say: "In the fullscreen mode, you click add, the browser goes away entirely" [14:41] benji: "If sidebar mode, you click a charm, the charm detawils closes and the sidebar stays open [14:41] ok [14:41] " [14:41] k [14:41] * benji will include the string "detawils" in the code. [14:42] benji: however, all the url changes stuff is in the branch I've got now I need to get tests finished for so I'm still thinking 'leave it' and we'll add the url updates once I can get this through [14:42] sounds good [14:43] bac, you click update and nothing ever happens? [14:44] gary_poster: yes [14:44] unclear to tell whether there is activity or if nothing happens. haven't looked at the server side yet. [14:44] bac, did you investigate? I'm happy for you to divide this up into "fix bug" and "integrate go" if that makes sense [14:45] gary_poster: i'll make a card to investigate. it is outside the scope of my current task, which i'm about to wrap [14:46] bac, cool. on improv, the message is sent to improv and a reply is returned [14:46] gtk [15:02] ugh, CI errors are painful to watch [15:02] agreed [15:16] guihelp: is there an YUI event that is fired when a node is removed using node.remove()? [15:18] frankban: negative [15:18] hatch: as I supposed [15:37] a guy I know is building a ~2000sq/ft house here and his quotes to do the plumbing are between $60-70,000!!! [15:38] I think trades are a little out of hand here lol [15:44] eek [15:49] yowzer [15:49] i'll bet it is all of the insulation for the pipes [15:49] lol [15:55] bcsaller: is there a way I can get the url from a specific namespace from an nsRouter instance? [15:55] hatch: we could add a method, but nsRouter.parse(Y.Location()) or similar should do it [15:56] ok that's kind of what I was thinking [15:56] is that for testing? [15:56] thinking normally we shouldn't need that [15:56] my issue is that when you destroy a unit you need to return to the service view, but there is no data anymore so I need to parse the url for the service name [15:57] sorry...when a unit you're viewing is destroyed [15:58] gary_poster: you missed your ten minute warning! [15:59] bac, sorry :-) [15:59] sounds like a one-off fix though, this is the kind of thing we should have a pattern for, a kind of 404 handler in the client, at min could take you back to '/' [15:59] jujugui call now, please get kanban up to date asap [16:01] benji starting [16:06] jovan2, ping? [16:06] gary_poster hi, on my way... [16:07] teknico, let me see what I can do about network. [16:07] jovan2, you missed, but would like to talk [16:07] Makyo, maybe you can try the low bandwidth hangout settings? [16:07] gary_poster you're so efficient with stand ups ! [16:08] teknico, Can try, though that turns off video for screen sharing. If you're fine with just voice, that's okay. [16:08] jovan2, getting better :-) guichat now? maybe quick question [16:08] gar_poster: sure [16:08] ty [16:08] Makyo, oh right. ok, let's try: https://plus.google.com/hangouts/_/02bb45411739e441fe107c9f66e2a8cc36ba4ba7 [16:08] jovan2: gary_poster I got the icon for the browser, but not the gui charm. Thanks jovan2. [16:08] gary_poster: just FYI we have cut the "final" release (1.10.0) [16:08] ack rick_h_ cool [16:08] gary_poster: just waiting for someone to turn it into a package [16:09] gary_poster: i hope it's got the stuff in it that you guys need [16:09] rogpeppe3, awesome! congrats. Yes, I thinnk we are golden. Everything seems to be working well. I'll make a release of GUI once you all make a release and we can announce [16:10] gary_poster: wonderful === rogpeppe3 is now known as rogpeppe [16:10] gary_poster: now for a beer! [16:12] rogpeppe, lol, yes I figure we deserve a break. I'm not sure what you plan to do for the next couple of weeks while we prepare for oakland. I'd like to do polish and revisit some topics, which would include being able to set a service's units directly, rather than add/remove; and investigating what kind of debug log story we can get over the API. Maybe we can do both, maybe neither, but now would be a nice time t [16:12] o investigate those sorts things, I think. [16:12] Anyway, rogpeppe topics for next week perhaps [16:12] gary_poster: sounds good [16:12] gary_poster: yeah [16:12] cool thanks [16:12] gary_poster: or even tomorrow if you want [16:13] rogpeppe, cool :-) [16:13] gary_poster: except my broadband is down currently and i'm connecting through my phone [16:13] jovan2, I'm all lonely in guichat [16:13] gary_poster: which is actually surprisingly ok [16:13] gary_poster: but costs [16:13] rogpeppe, lol yeah I bet [16:13] when is your broadband supposed to be back, rogpeppe ? [16:14] gary_poster: yesterday :-) [16:14] heh [16:14] gary_poster: seriously, i have no idea [16:14] gary_poster: the phone co knows [16:14] gary_poster: sorry, be there i [16:14] gary_poster: and i know no more than that [16:14] rogpeppe, I hear you, similar over here for these sorts of things [16:22] teknico, rick_h_, Makyo: I just talked to jovan2 about the placement of the environment and new charms and so on. His preference is that we ake *no change* to the current behavior until we have more information from users about their expectations. So, my take is that we move that card to trash and pick a new one, with apologies to teknico and Makyo. [16:22] s/we ake/we make/ [16:23] gary_poster, oh well, ok :-) [16:23] rick_h_, ack? [16:24] rick_h_, gary_poster, Juju GUI charm should be ready for tomorrow - Friday [16:24] thanks jovan2 [16:25] hmm darn my approach won't work for the redirects :/ [16:30] gary_poster: defintiely, sorry about the wasted time [16:30] cool rick_h_. I think everything else you gave us clearly needs to be done. [16:31] http://uistage.jujucharms.com:8080/bws/sidebar does fix gettnig at the env now though so yay [16:35] gary_poster: what happens when you log out from uistage? the login screen no longer shows here, could anyone confirm? [16:36] frankban, :-( confirmed [16:36] I'll make another card [16:36] gary_poster: also in trunk, it doesn't show up if you run make prod [16:37] that's critical regression [16:38] yes :-/ [16:39] gary_poster: it doesn't look like there is any way I can reliably check if the view is waiting for service data, or if it has been deleted :/ [16:42] hatch, hm. guichat? [16:42] yup [17:34] bcsaller: did you watch the keynote yesterday? [17:37] bac: I skimmed it, did I miss something? [17:37] bcsaller: no, i forgot to watch. was wondering if it was good. [17:38] bac: I think it was less compelling that previous ODS demos, but thats just me [17:38] bcsaller: +1 was more business-y kind of presentation it felt like [17:40] yeah, the tech part was small. there was only a small bit with the GUI, the tech part was more focused on landscape. The part with the GUI looked rather chaotic to me, it was a cluttered layout. [17:43] agree, not ideal presentation of it [17:53] he is a good presenter [17:54] maybe we need an 'organize gui' button which will align everything nicely :) [17:54] annnnnd Ben can write it [17:55] :P [17:56] hatch: we did depend more on automatic layout in the past, it makes less sense to invest in that since we have persistent shared positions now [17:56] I can see it now - one of the techs will stack all of the icons ontop of eachother just to irritate his coworkers [17:56] haha [17:57] frankban still needs a review. i need two. [17:58] hatch: at one point we were going to write automatic layout with relationship routing that curved lines around elements on the page [17:58] Ooooo fancy [18:00] hmm I need lunch [18:02] rick_h_: how do you search the charm store by charm URL (e.g., cs:precise/apache2-6)? [18:03] benji: not following, You want to search the charmworld manage.jujucharms.com? [18:04] benji: or do you mean the search UX that's getting worked on for the browser? [18:04] rick_h_: for example http://jujucharms.com/search/json?search_text=provides:http returns all charms that provide HTTP, I want to search for a charm with a particular URL [18:04] I've tried things like http://jujucharms.com/search/json?search_text=url:cs:precise/kusabax-0 but haven't found a form that works. [18:05] benji: hmm, yea not sure we have that. [18:05] benji: you can load a charm via the id of the charm http://staging.jujucharms.com/api/0/charm/precise/couchdb-9 [18:05] but that's not the whole url [18:05] that may work, thanks [18:06] benji: what are you doing? The charm data/model should be there for you already. The Charmworld0 store has the charm method that takes the id to load from the new api as well [18:07] ok, I'll take a look there [18:08] benji: but the CharmView has a ATTR charm that's a BrowserCharm that the plan was to do send to the environemnt with a new Charm(view.get('charm').getAttrs()) [18:08] so it implements turning the new api data into the format for the old Charm model [18:09] that sounds good === deryck is now known as deryck[lunch] [18:31] gary_poster: chat? [18:31] bac, I is there [18:31] okey doke [18:33] ooo boy I like these changes [18:33] now to lunch [18:33] ding if ya need [19:13] has anyone ever seen this error before.... Uncaught Error: InvalidStateError: DOM Exception 11 [19:14] it's throwing it on ws,send(data) in reconnecting-websocket.js [19:14] hatch: I've seen that. It was trying to append to itself or something [19:14] well, it was something liket hat [19:15] hatch: http://stackoverflow.com/questions/3488698/invalid-state-err-dom-exception-11-webkit comes up in a quick search. traceback? [19:15] hatch: but when I got something like that I was trying to append dom to something that could not happen [19:16] alright [19:16] the traceback is pretty much useless but maybe I'll have to step through it [19:16] :( [19:29] right now it doesn't appear to be having any issues [19:45] gary_poster: the terrible deed is done. el raton has gone to a better place. [19:45] bac, heh, good work === deryck[lunch] is now known as deryck [20:00] does the python/go backends have a way to receive a unit from the environemtn? [20:00] I can't find anything [20:31] he he heeee [20:36] guihelp: why does the first exception test work but the second (clearer) one not? http://paste.ubuntu.com/5719840/ Other tests using assert.throws do work. [20:36] the second test fails with: Error: the string "Exactly one of config and data must be provided" was thrown, throw an Error :) [20:37] test failures should not have smiley faces [20:37] no idea, I've never used assert.throw [20:38] when you say doesn't work do you mean it never fails? [20:39] hatch: no, i mean the test fails, as shown in the second message above, the one with the smiley face [20:40] how are you throwing the error? [20:40] hatch: the code is throwing the error. the assert is not recognizing it for some reason. [20:40] test_sandbox.js uses assert.throws() [20:41] throw '...' [20:41] in the nodejs doc examples assert.throws() always throws a new Error() [20:42] so it could be that it's not expecting only a string? [20:42] bac I think you need to pass a callable [20:42] I am not sure - I can't remember the last time I intentionally wanted to throw an error in js [20:43] looks like you need throw new Error('string') and its passing a string directly? [20:43] bac, you are passing the called result [20:43] gary_poster: ahhhhh [20:43] * hatch confused [20:43] so I was right? [20:43] lol [20:44] you can throw a string [20:44] but I'm 80% sure that assert.throws wants a callable [20:44] so IOW [20:45] assert.throws(function() {DO STUFF THAT THROWS AN ERROR}, 'Error object [20:45] ') [20:45] ahh ok [20:45] right, can't catch it otherwise [20:45] gary_poster: is test_sandbox passing a callable? [20:45] hatch, you want to receive a unit from an environment? [20:46] * gary_poster looks [20:46] gary_poster: well I did a workaround but [20:46] am running into another issue [20:46] bac, yes [20:47] bac, bind is a JS thing. so... [20:47] my issue now is that if you're viewing a unit and someone deletes the service....then it's totally baffled [20:48] bac, "client.open.bind(client)" == "function() {return client.open.apply(client, arguments);}" [20:48] hatch, well, beat some sense into it! :-) [20:48] gotcha [20:48] thanks for the eyeballs [20:48] welcome [20:49] gary_poster: got a couple minutes to walk through this code? [20:49] hatch want to guichat? sure [20:49] sure [22:22] rick_h_: Any idea what keeps bringing the staging server down? Anything I can do to fix it in my timezone? [22:26] it's scared of your huge bugs [22:30] :) [22:39] huwshimi: is it down again? It seems zookeeper went boom. To fix it they've had to rebuild it [22:39] huwshimi: what I'd do is change the charmworld url in the config to manage.jujucharms.com [22:39] rick_h_: No, it's up now, it's just that it has gone down two days in a row :) [22:40] rick_h_: Ah, that's good to know, at least I can get stuff to show up then :) [22:40] huwshimi: rgr [22:40] rick_h_: Thanks :) [22:40] huwshimi: yea, I'm not sure what they figured out yet [22:41] huwshimi: cool, landed your branches, tweaked the loop stuff in the charm.js [22:42] rick_h_: Ah I see. [22:42] rick_h_: Thanks for landing everything [22:43] huwshimi: also added a card about the sidebar. It's adverse to window sizes < 1200 [22:44] huwshimi: and jovan warned they'd have some font changes coming. They think the font size is a bit small [22:44] rick_h_: Yeah, I made a regression when I was fixing the sidebar overlay div issue [22:44] rick_h_: OK, fun, fun. [22:45] huwshimi: cool, yea figured now that it 'works' can work on making it nice. So just added a card and can go from there. Thanks for the updates! [22:45] rick_h_: No problems, I'm fixing that bug now [22:45] awesome [22:46] rick_h_: Any update on the search or homepage? [22:48] huwshimi: no, jcsackett is rebasing off my branch. I didn't get all my tests done so still working on my branch. It's holding everything up. [22:48] huwshimi: won't have it tonight. So it'll have to be a next week thing. [22:48] rick_h_: Yep, sure. I'll style from your branch today. [22:49] huwshimi: cool, sorry to hold up. [22:49] rick_h_: It's all good, it's such a rush :) [22:51] huwshimi: if you need anything on my branch let me know. It's a bit complicated but should work out to do /bws/fullscreen with the editorialview [22:51] huwshimi: and abentley landed the code to mark 'featured' but don't think it's on production :( [22:51] rick_h_: OK, I'll take a look after this. [22:52] rick_h_: Is there anything else pressing for me to take a look at today? [22:53] (that you know of) [22:53] huwshimi: just take a peek through the cards for anything that jumps out. Otherwise nothing I know of. [22:54] rick_h_: No problems. Thanks [22:54] huwshimi: the fullscreen landing is really next with the category icons and such [22:54] great....success [22:54] hatch: woot [22:55] the code is ugly as sin but the proof of concept works [22:55] so now that means I can convert to promises [22:55] (this section) [22:55] and we have removed reliance entirely on one dispatch call [22:56] one small step for devs, one giant leap for dispatch destruction [22:56] huwshimi: there's also some work to update the sidebar editorial to match the design. Under the three charm containers. [22:57] rick_h_: Oh right, what needs to be done? [22:58] huwshimi: I'd have to check the UX assets. Sorry, on my tablet so don't have easy access. [22:58] huwshimi: I thought there was some stuff under those first charm containers, but can't recall what was there. [22:58] huwshimi: then there's also still the toggle to show/hide the sidebar. [22:59] rick_h_: Oh well we don't have any of the charm icons being set yet... [22:59] huwshimi: sorry, just trying to think off the top of my head since you'll be coming back monday during my sunday. [22:59] huwshimi: so but we do have the default svg now in the assets directory. [22:59] huwshimi: and we've got the category icons in there [22:59] huwshimi: so we'll default to the charm store one unless they have a category, then we'll use that [23:00] rick_h_: Ah I see, OK, I'll do that. [23:01] huwshimi: cool, hope that helps while jcsackett and I work on getting things caught up. [23:01] huwshimi: anyway, I'm off. Have a good day! [23:01] rick_h_: Yep, thanks. I have a list of little things to get to as well. Thanks. [23:01] rick_h_: Have a nice night [23:10] anyone need any reviews before I EOD it?