[12:09] hi everyone [12:10] frankban, could you please beg/borrow/steal some juju-core reviews for your branches? :-) [12:10] benji, you up for reviewing bcsaller's branch for #1185982? [12:11] https://codereview.appspot.com/9906044/ [12:11] gary_poster: sure [12:11] thanks benji [12:17] hi alejandraobregon. Do you know when you will get a chance to have some of Charline's time for reviewing the user data questions? I'd like to get that moving, but I do want to have her input and advice if we can. If you want me to contact her directly I'd be happy to, but I also appreciate you talking with her about it. [12:50] gary_poster: I'll do :-) https://codereview.appspot.com/9810044 would like to be reviewed by someone in the golden horde [12:51] thanks frankban :-) bac, you around for frankban's review request ^^ , by chance? [12:51] gary_poster: sure [12:52] thanks bac [12:52] thanks bac. frankban, fwiw looks good to me, but I'd want to look at it more closely for an official review. [12:53] frankban: why do you consistently use the word 'machiner'? is that just a typo for 'machine' or am i missing something? [12:53] machinery? [12:54] bac: it's how the machine worker is called in juju-core, I have no problem in using "machine" if it makes it more understandable [12:54] machiner (plural machiners) [12:54] One who operates a machine. [12:55] gary_poster: hi! nice weekend? Have a meeting booked in with Charline tomorrow to discuss user feedback :) [12:55] if that is the correct word for this context then please use it. but i don't want to see something clumsy and juju propogated [12:55] frankban: ^^ [12:56] alejandraobregon, yes, nice weekend, thanks! dad was in for a visit. hope yours was nice too. Charline meeting tomorrow: fantastic, thanks! [12:56] bac: ack, 2 out of 2 reviewers got confused by machiner, enough to change it IMHO [12:56] bac, you saw frankban's reply about juju core's nomenclature? [12:56] ok [12:57] gary_poster: you mean in this discussion or elsewhere? i didn't read responses to other reviews yet. [12:57] bac, simply above: """it's how the machine worker is called in juju-core, I have no problem in using "machine" if it makes it more understandable""" [12:58] gary_poster: yes, i read that [12:58] I worry a bit about moving away from their names, even though I agree they don't make sense [12:58] well, that's too strong [12:59] but maybe it is reasonable to define their terms somewhere and then use them [12:59] for instance, I think William has to translate in his head when he talks to us [12:59] to handle our use of the "delta" term, for instance [13:00] the more we have a shared understanding and vocabulary the better [13:00] end of thought. [13:00] fine with changing it to "machine worker" for this branch [13:00] gary_poster: +1 [13:01] cool. so another approach would be to start a "definitions" doc file. Not advocating, brainstorming. [13:05] gary_poster, teknico: when you have time, I'd like us to have a quick call re juju-test [13:06] gary_poster, frankban: i just looked at the juju-core code and they have a Machiner class. it is a real conceptual thing, not just a passing odd word. so i think we should use it. i like the idea of defining it, though. [13:06] frankban, cool. I should have time within the next hour or so. [13:06] bac, great, thanks for looking at that [13:06] gary_poster: cool, thanks [13:11] frankban, gary_poster: I'm ready [13:11] thanks [13:12] gary_poster, bac: on the other end, I don't like the idea of the charm having to deal with internal juju-core implementation concepts, like machiner or agent.conf. In this context, they were added as a temporary workaround. We still need those for backward compatibility but we will eventually get rid of those. [13:13] s/end/hand/ [13:15] gary_poster, teknico: ready on the chat when you are. [13:15] ok thanks. still trying to finish something else up. will ping/ [13:16] frankban, re machiner name: not sure. I trust whatever decision you and bac agree on. [13:20] good morning all [13:21] morning hatch [13:22] gary_poster: model/view databinding in yui functional prototype https://gist.github.com/hatched/285d3aadb0dbdd94b2db [13:22] :) [13:22] hatch, rock :-) [13:22] thanks [13:27] hatch, looks simple and very good. I see you filled out the google doc too, thank you. Maybe send demo and doc around to gui team + rick and jc? [13:30] Review available, 50% off: https://codereview.appspot.com/9953045 [13:30] orangesquad: going to be a moment late, fighting g+. [13:31] sinzui: ping for standup [13:33] benji: i was just about to do that one but couldn't find the url. on it now. [13:33] benji: coincidentally, i have one ready for eyeballs too! [13:34] bac: thanks. I'll take a look at yours as well. [13:35] benji: sorry, i couldn't get my bzr config set properly for lbox so it is an old school LP merge proposal [13:36] * benji feels a rush of nostalgia. [13:37] gary_poster: sure thing I'll put up a jsbin in a minute [13:44] cool thanks [14:01] frankban, teknico call in 14 minutes, at quarter past hour? [14:02] gary_poster: antonio and I have a call with mark this morning @ the same time as the site/GUI call; if you guys need anything from us just let us know [14:02] gary_poster: sounds good [14:02] frankban, gary_poster, works for me [14:02] thanks frankban & teknico [14:03] jcastro, will do, thanks. I talked with Antonio about some topics, esp. the quality tab, that I'd like your ack on. Maybe we can catch up later today [14:03] bac: I need to talk with you re: tarmac, got some time? [14:05] teknico: let me finish this review [14:05] won't take long [14:05] bac: sure, thanks [14:07] rick_h_: I finally remembered why I pinged you the other day :) [14:07] hatch: party on [14:16] teknico: guichat? [14:17] bac: I am in ongoing call there with frankban and (soon) gary_poster [14:17] teknico: oops, i see it is busy [14:17] teknico: ok, ping me when you want to talk [14:17] bac: I will [14:23] email sent [14:23] or not.... [14:27] power went out [14:28] jujugui: seeking one more review for https://code.launchpad.net/~bac/charms/precise/juju-gui/unified-ppa/+merge/167039 [14:29] glad I had a UPS heh [14:29] orangesquad, charmworld r420 is deployed. we are waiting for ingest to do a full run to verify non-charmers promulgated is fix released. [14:30] sinzui: rgr, awesome [14:33] bac: coast is clear [14:48] marcoceppi: hi, got a minute for a quick chat? [14:49] teknico: certainly [14:49] marcoceppi: thanks, tinyurl.com/guichat ? [14:49] uhm, http://tinyurl.com/guichat [14:49] teknico: works for me [14:55] rick_h_ or jcsackett: Please have a look at this: https://code.launchpad.net/~abentley/charmworld/api-2-docs/+merge/167058 [14:55] abentley: looking [15:06] man the power is still out [15:07] abentley: so I'd mention to sinzui you are headnig to api2 [15:08] there can be only 1 [15:09] rick_h_: I think you just did :-). sinzui, docs are here. https://code.launchpad.net/~abentley/charmworld/api-2-docs/+merge/167058 [15:09] the spike is now a feature? [15:10] sinzui: No, not implemented yet. I wrote the docs to reflect our conversation. [15:17] gary_poster: I summarized my chats with b.a.c and marco.ceppi in the card description; what's a better place for that info, if any? it's my card in Maintenance - Active - Branch start [15:24] bah, got disconnected [15:46] grr, after tweaking my code through review I now get test failures when run in phantomjs (but not Chromium) [15:56] abentley: can you peek at https://code.launchpad.net/~rharding/charmworld/add-category-api/+merge/167078 when you get a sec please? Short/sweet [15:56] rick_h_: sure. [15:57] thank you teknico1 [15:57] sinzui: For your work, when the author has not specified any categories, do you handle it as None, [] or no-such-attribute? [15:58] guichat in 3? [15:58] 2 [15:59] abentley, I intended to return []. You can disagree now and tell me to return None. [15:59] yay power just came back on [15:59] sinzui: I like []. rick_h_, how do you feel about changing categories to be [] when not supplied? [16:00] abentley: fine by me. I went with the change that failed the least number of tests :) [16:00] abentley: I'll change to [] and update the rest of the tests [16:00] abentley, I have chosen to return None for no icon because an empty string is bad. the views/model will need to omit the attr from the json when icon is None [16:01] ^ rick_h_ [16:01] sinzui: there is no icon in the api any more [16:01] sinzui: ^ [16:01] sinzui: it's a direct check for the icon.svg in files [16:01] fab. I can remove it from my model. [16:02] sinzui: I guess the problem is we haven't actually deleted API0. jcsackett, were you able to figure out whether it's safe to delete API0? [16:02] jujugui Call now? [16:02] hey gary_poster, you coming? benji? [16:02] coming [16:02] abentley: it should be safe currently. [16:02] abentley: yes it is safe. thought i told you on friday. my apologies. [16:02] abentley: uistage has been on api1 for a while [16:03] jujugui I am coming [16:03] jcsackett: Thanks. It's entirely possible you did tell me and I forgot. [16:04] abentley: well, either way. kill it. :-) [16:04] sinzui: I will kill API0 and the icon as my next task. [16:05] understood [16:09] gary_poster, browser froze. Going well. [16:10] rick_h_: Did the API0 tests pass for your branch? I would have expected them to break due to the introduction of categories. [16:10] abentley: yes, all tests passed in the original branch. Getting some failures now in a couple of places. [16:11] abentley: So they passed because if there was no categories or it was [] I didn't return it. [16:12] rick_h_: Ah, right. [16:19] orangesquad: Could you please review https://code.launchpad.net/~abentley/charmworld/fix-orphan-chunks/+merge/167084 ? [16:20] * sinzui looks [16:22] rick_h_: r=me. Thanks! [16:23] abentley: cool thanks, was just waiting for the diff to update [16:30] sinzui: gary_poster anyone: can you set constraints for a bundle? === hatch_ is now known as hatch [16:31] bac: what does "c-n-p" mean (re: your review of benji's branch)? [16:31] rick_h_, do you want to be at tech discussion? [16:31] gary_poster: loading now [16:31] cut-AnD-paste [16:31] ooh, right [16:31] sorry [16:32] bcsaller: can you set constraints for a bundle? or edit settings? Also, what settings would be available when your deploying a bundle? [16:33] luca__: I'm in a meeting now but can reply after [16:33] bcsaller: np, thanks :) [17:13] ok I'm going to finish my breakfast now :) [17:14] 11:13...more like lufast [17:14] since it's more lunch than breakfast [17:14] and luca is gone now :( [17:17] bcsaller: so is it decided that we are sticking with d3 for the env then? [17:17] I remember the discussion from oakland about alternatives [17:18] but now with the request for more with d3 experience [17:19] when the power went out the second time I learnt that my phone does not have enough processing power to run a hangout with 10 people :) [17:20] hatch: we don't have a sufficient alternative at this point. There was some talk of rendering with WebGL but IE10 support is only via a plugin so its a no-go [17:21] ahh and the canvas tools available were subpar? [17:21] additionally we have UX that zooms, canvas is raster, svg is vector and scales [17:22] ahh right right [17:22] webgl would be so damn cool [17:22] its different, we'd give up the CSS skinning, the DOM event model (for something similar but different) [17:22] does IE10 auto update to 11 when it comes out? 11 will support GL :) [17:23] hatch: quit talking crazy [17:23] I don't know the plan there [17:23] rick_h_: oh c'mon - webgl, occulus rift, nintendo powerglove, treadmill [17:25] sinzui, do you have a few minutes to help me write the email to Mark about the quality metric for charms? Or is there someone else who would be able to give me the context you think I might need? [17:25] rick_h_: care to take the first review on https://codereview.appspot.com/9965044 ? [17:26] rick_h_, was their discussion about including the charm tabs in the URL? I have a use case for it right now. :-) [17:27] there [17:27] gary_poster: bcsaller should we have a chat about where to go from here? [17:28] hatch: yeah, when ever you have time we can hope on g+ again [17:28] hatch, bcsaller was thinking something similar. [17:28] I'm on guichat now [17:28] could talk now? [17:28] okee [17:29] gary_poster: yes, it's an open bug to add that feature in [17:29] gary_poster: just lost some priority with recent api fixes/etc. [17:29] ok cool thanks rick_h_ [17:30] gary_poster: #1175149 if you care enough [17:30] <_mup_> Bug #1175149: URLs cannot take you to a particular tab in charm details [17:30] jcsackett: looking [17:30] jcsackett: actually, will look in a sec. Today's flown by and need to fetch lunchables. [17:44] rick_h_: dig. [17:46] gary_poster, I think this summarises my understanding of the issue: https://docs.google.com/a/canonical.com/document/d/1I25jG9rcZY6K6x250LhY-8bfEy0KgDXZYrk9UgagFCo/edit [17:47] thank you very much sinzui. Incorporating [17:50] gary_poster: my branch is landed and I am hungry; after lunch I'll check with you about... whatever it was that you wanted me to do next ;) [17:50] benji, ok thanks :-) [18:05] sinzui: The 02 migration (removing periods from keys) has tests that now fail. Should I remove the tests? Remove the whole migration? [18:06] oh. [18:07] abentley, Yes is the immediate answer. Maybe we want to note this circumstance in the docs. We can remove migrations when we believe every supported site is migrated...and that is just mange.jujucharms.com [18:09] sinzui: Cool. For some reason I thought we'd have to support jujucharms.com, but of course we've already gone past it. [18:34] sinzui: We don't currently have an elasticsearch migration tool. Maybe we should add one, maybe we should migrate mongo and elasticsearch using the same tool, maybe we should just reindex mongo. Thoughts? [18:35] jcsackett: LGTM inbound with a couple of things to look at please. Small enough changes I think, but let's chat if I'm off base if that's cool [18:36] abentley, I favour reindex mongo...I think we had agreed that in the past [18:37] sinzui: Oh, I don't remember that, but it works for me. [18:38] sinzui: abentley didn't we talk about adding a function to the current migration code so that you could create a new migration file who programattically did a reindex in the upgrade function call? [18:38] abentley, I suppose I was jumping to conclusions in the past. [18:38] abentley: sinzui thus kind of re-using our "do it once when we need to" code already written without manual intervention on deployments? [18:39] rick_h_: I don't remember that, but it makes sense. [18:40] abentley: sorry, maybe I had the chat with sinzui along those lines. It's been a bit. [18:41] abentley: but yea, ideally I'd think we'd add a migration helper functino to do a re-index and just create a migration that touches ES using the helper. [18:41] abentley, I assumed sync-index would be the way to update elastic search after ever mongodb change. [18:42] The things we currently do are: in the charm, index mongo into ES as soon as both are available, and on change to the ES mappings, reindex ES into ES. [18:42] ! [18:42] abentley, that reminds me... [18:44] abentley, rick_h_ ...I think I want to change that rule to know if the services have current data. I wait to wait for a new ingest when I brought up the app with the current db and search [18:44] s/wait to wait/had to wait/ [18:46] sinzui: So that was a case where we could have done a migration, and maybe should have. Or else requested webops to do a manual ingest as well. [18:48] abentley, every time we remove and add db+search relations, we do a full ingest. That means adding a unit will try to do a full ingest (we ignore its attempt). If I bring up the db from a dump file, I only want sync-index to run [18:50] rick_h_: replied, happy with most changes. working on those now. [18:50] jcsackett: cool looking [18:51] sinzui: When there is already a unit and we add a unit, I believe we don't do that. It's only when there are no units and we add one. [18:51] gary_poster: I'm back [18:51] hey benji thanks. checking guichat... [18:51] gary_poster: it's not free [18:52] benji, k, making room [18:52] sinzui: I guess I'm not clear what the new rule would look like. Never auto-ingest, because we assume there was a dump? Have a new relation variable to indicate that there was a dump? [18:53] abentley, if config-changed could check both db and search have an agreeing set of data, then go right to the expose step [18:54] abentley, If just db, then sync-index and expose [18:56] jcsackett: thanks, I'm pulling down the branch to play with the delegate. Hold off landing until I check that out please. [18:56] sinzui: I think we can make sync-index fast enough that we don't care whether it runs, by using bulk_index. [18:56] abentley, I am not proposing you do this config-change work. I was pondering it when we had the agent state non-sense last week. [18:56] abentley, oh, lovely. [18:57] hah, forget fixing issues, just bypass with new tools ftw! [18:59] jcsackett: hangout when you get a sec please? [19:15] jujugui: my isp just announced they are on their way to service our tower so i may drop off in a bit [19:16] ack [19:16] gary_poster: i'm looking for a card to pick up. you have any suggestions? [19:16] looking [19:18] bac I suggest the high from Story A, which I can describe, or one of the three maintenance high cards [19:18] the first two are d3-related [19:18] the first two maintenance I mean [19:19] gary_poster: so you mean the analytics story? [19:19] yes bac. [19:20] i'm happy to look at that. i'd like to see the names 'crazy egg' rejected. [19:20] lol [19:20] bac, ok want to come by guichat and we'll talk about it? [19:20] sure [19:21] gary_poster: i'm there [19:21] redirect loop between docs.google.com and accounts.google.com; whee! [19:21] i think [19:21] oh, huh [19:23] jcsackett: email back with notes. I can't replicate the currentTarget not working correctly [19:37] "lunch": biab [19:59] hatch: I'm back when you're ready [20:00] alright one minute [20:00] just grabbing something to drink [20:03] no rush [20:03] alrighty all ready [20:04] just finished the Ghost in The Wires book that I was reading in oakland [20:04] what a great book [20:05] I have 40 books in my amazon wishlist and about 5 unread on my bookshelf - I'm not sure why I even bother lol [20:06] hatch: in the chat room now [20:07] gary_poster: are you free to join chat? [20:08] orangesquad: I might disappear. I am experiencing a heavy down pour and the water is now over the door sill. [20:08] sinzui: Sorry to hear. [20:09] I have two weeks left in this house. \o/ this could be my last flood [20:17] bcsaller, guichat? [20:46] bac: did you have something to chat about? [21:08] a WYSIWYG editor isn't worth its salt unless it has orphan control [21:08] what's orphan control? [21:09] it is a way to say "this chunk of text should not be separated by a page break from this other bit of text"; for example, a header should always be on the same page as the first paragraph of the section it heads [21:09] gary_poster: It needs more work, but the vitals doc is in a good shape for you to take a look at it [21:09] oh...I don't even think Word does that [21:10] jujugui: if anyone's free, can i get one more review on https://codereview.appspot.com/9965044/ [21:11] yep, in (the last version of Word I used in anger, about a decade ago) you could select a chunk of text and choose the "keep together" option [21:12] jcsackett: sorry, I need to run [21:12] * benji runs. [21:13] jcsackett: sure [21:19] benji awesome thank you looking [21:37] TIL my NAS doesn't turn back on when the power goes out [21:37] * hatch just spent 15 minutes trying to debug the connection [21:37] lol [21:38] * hatch fail [21:48] how many people here use lastpass? [21:49] * gary_poster raises hand [21:50] any issues you have found? [21:50] benji added comments, thank you! [21:50] hatch, mm, I wish it could accept both yubikey and google authenticator simultaneously [21:50] hatch, otherwise +1 [21:56] man lastpass for android is bloody ugly [22:17] hatch: i use it; i don't *love* the mobile app, but it works, and the service is good. [22:21] I set it up on a site and now it auto logs me in [22:21] * hatch is a little creeped out [22:21] lol [23:02] does the lastpass log in fail the first time intentional? [23:30] hey huwshimi [23:30] gary_poster: Hello! [23:30] gary_poster: I'm assuming green highlights are us? [23:30] :-) [23:31] hatch, no, we are everything that is not orange squad or Ecosystems [23:31] alrighty [23:32] hatch, so inspector, (bundles, containerization,) gray masthead, analytics [23:32] gotcha [23:32] OSCON sounds like something from Spiderman [23:32] OSCORP [23:33] huwshimi, I don't know if you are able to work on the gray masthead today, but I just noticed a few minutes ago that there appears to be two similar, related bugs: https://bugs.launchpad.net/bugs/1172735 and https://bugs.launchpad.net/bugs/1185006 [23:33] <_mup_> Bug #1172735: Implement grey masthead with logout, alerts and get juju [23:33] cool I wish I knew about this sooner this conf looks packed [23:34] the branch would assume that the right-hand search panel could just disappear in a cheap hack for now [23:34] gary_poster: The masthead icons refer to Charm Browser masthead (and I've implemented them in a branch that I've been working on). [23:35] huwshimi, ah! [23:35] gary_poster: For the main masthead is the plan to implement that under the browser_enabled flag? [23:36] gary_poster: If so, any hints on how to use that flag in index.html which is not a handlebars template? [23:36] huwshimi, the plan is that, once #1175019 and #1186299 are fixed, we make the charm panel the default and rip out the flag [23:36] <_mup_> Bug #1175019: staging has issue with black bar at top of fullscreen charm details [23:36] <_mup_> Bug #1186299: The charm browser quality tab does not have data [23:36] huwshimi, and the hope is that this will be extremely soon [23:36] so you could simply change the main masthead directly in your branch huwshimi [23:37] and if we somehow have issues with those two bugs [23:37] then we will implement the flags our selves somehow (/me waves hands and hopes it doesn't happen) [23:38] while you atre on vacation [23:38] wdyt, huwshimi? sound ok? [23:40] huwshimi, also could you please subscribe to juju-gui-peeps, on bottom left of https://launchpad.net/~juju-gui-peeps [23:41] gary_poster: That sounds ok, but there are three parts to replacing the header. 1. Style the header 2. Replace the Alerts Bootstrap dropdown with a YUI widget and 3. Remove the Charms search and panel button [23:41] gary_poster: So I'm not sure if that can all really be done without breakage in one branch [23:42] huwshimi, do we really need to to #2 right now? [23:42] agreed we need 1 & 3 now [23:42] huwshimi, but if you deliver incremental branches that is AOK. Are you saying you'd like others to deliver 2 and/or 3? [23:43] huwshimi, I am assuming you are subscribed to https://lists.ubuntu.com/mailman/listinfo/juju-gui , but if not please do [23:44] gary_poster: I'm not sure if we can really do #1 without #2 without a lot of fighting with Bootstrap and having to use all their specific HTML. It something that we can do, it just makes things a lot uglier. [23:44] huwshimi, oh I see :-/ [23:44] huwshimi, ok, so the order needs to be #2, #3, # 1? [23:44] I was waiting until Friday but huwshimi we could look into http://purecss.io/ [23:45] * gary_poster steps away [23:45] not sure if it helps much [23:45] but it's yui's answer to bootstrap [23:45] gary_poster: The cleanest scenario for me would be #2 and #3 (in any order) and then I can come along and do #1. [23:47] huwshimi: did you have a module in mind for the alerts dropdown? [23:48] hatch: I imagine it would just be an Overlay [23:48] gary_poster: did we already know that uistage is broken? [23:49] huwshimi: I wonder if we could do it all with css with a styled radio button [23:50] OR [23:50] we it could be as simple as a class toggle on the list [23:50] and have the result list and button in a wrapper [23:50] so they align [23:52] hatch: I'm not sure what we really get with doing some kind of CSS implementation. We'd still need to build a YUI widget to handle building the menu items etc. and we may as well use something that is built in rather than making something ourself, right? [23:53] yeah............. [23:54] overlay just feels so heavy for such a simple thing [23:55] blah it's past EOD and I'm just talkin garbage, feel free to ignore me :) [23:55] hatch: Well, it's exactly and overlay :) [23:56] *an [23:56] well it IS but at its purest form it's really just a list of errors [23:56] so as long as it can be turned on/off and aligned to the button that's all that's really necessary [23:59] hatch: But that's pretty much exactly what an Overlay is designed for, it's just that someone's already made it for us :) [23:59] well in reality an overlay is so much more code [23:59] it's certainly a lot less work to use one though [23:59] :) [23:59] my hacking project tonight will be to try and make what I'm blabbing about