[00:33] morning huwshimi === rogpeppe is now known as rogpeppe1 === rogpeppe1 is now known as rogpeppe [12:27] frankban: do you still need a review? If so can you shoot me the link and I'll look. [12:28] rick_h_: thank you! https://codereview.appspot.com/84630043 [12:28] thanks for the link [12:52] * frankban lunches [13:23] what's the status of icons and local charms? [13:28] hazmat: on the board for this week hopefully [13:28] hazmat: got delayed due to juju 1.18 changes messing with quickstart a bit [13:29] rick_h_, i need to discuss that a bit more.. we're using local charms for a demo.. afaicr.. gui was still fetching icons from the store based on what the store told it, but apis where put in place to fetch from state server? [13:30] er. demo this week [13:31] er. apis on the state server where put in place to fetch/return files from a charm.. [13:31] hazmat: yep, that's the work we've got to do to hook to those [13:32] hazmat: free if you want to chat now [13:33] hazmat: https://plus.google.com/hangouts/_/canonical.com/weekly-gui [13:34] or https://plus.google.com/hangouts/_/canonical.com/weekly-gui?authuser=1 for the second account business if you need [13:35] morning all [13:35] morning hazmat [13:35] err hatch [13:35] damn tab complete [13:35] haha - I'm so close to finishing this darn url parser [13:35] man we have a lot of urls [13:36] hatch: very cool [13:37] http://www.regexr.com/ it's like regex pal but "modern" :) [13:38] rick_h_, grabbing coffee, 1m [13:38] hazmat: rgr [13:45] rick_h_ I can't seem to boot ubuntu with an external monitor plugged in - have you tried that with yours yet? [13:46] nope [13:46] my only monitor is my 4k on the desktop [13:46] so no spares around [13:48] ahh ok - it just kernel panics so I'll have to do some digging [13:48] one of these days I'll get it working hah [13:48] usually I wait a few months before trying out a new version, this is very out of character :) [14:11] jujugui lf two reviews on the url request to state parser https://github.com/juju/juju-gui/pull/225 plz and thanks (no qa) [14:11] hatch: will look in a bit [14:11] it's *snif* beautiful [14:11] frankban: are you looking for reviews? kadams54 or bac might be able to help? [14:11] hatch: taking a look [14:11] thanks [14:12] rick_h_, guihelp: yes, I need one review for https://codereview.appspot.com/84520047 (quickstart). thanks! [14:12] frankban: heads up, after quickstart we've got a important thing to try to get local charm icons for this week. Will bring it up on call today [14:13] frankban: i'll look [14:13] rick_h_: ack [14:13] I can look [14:13] bac: thank you! [14:13] oh [14:13] too slow :) [14:14] what does token mean? [14:15] hazmat: it's the widget used in the browser [14:15] rick_h_ kadams54 I'm purposefully leaving out reviewer comments because I wrote it to try and be as easy to read as possible - let me know if I failed in that :) [14:15] for search completion suggestions, sidebar, etc [14:15] rick_h_, ic.. not to be confused with auth token etc [14:15] hazmat: correct [14:15] rick_h_, thanks [14:16] rick_h_, iotw a viewlet by another name/impl ? [14:16] hazmat: sure [14:16] hazmat: it's pre the viewlet work and the charm tokens were done as YUI widgets. so it was a non 'widget' word [14:17] gadget is open ;-) [14:17] haha [14:17] go go gui gadget! [14:17] "gadget all the things!" [14:17] jujugui: during the weekend I had some fun writing a little juju plugin which allows deploying charms hosted on github. Maybe it can be useful for someone: https://github.com/frankban/juju-git-deploy [14:17] frankban, hatch i saw that one.. [14:17] frankban: sweet [14:18] nice [14:18] frankban, very nice ++ [14:18] was trolling pypi looking for juju this morning and came across it [14:18] s/trolling/trawling [14:18] frankban awesome! [14:18] hazmat oh c'mon you were trolling it :P [14:19] hazmat: yeah, uploaded this morning, I wanted to hack something with python3 [14:19] * kadams54 tries to imagine what trolling pypi would look like [14:19] hatch: thanks, I used your ghost charm for examples [14:19] kadams54 someone uploaded a bunch of null packages to npm as barack_obama ..... that's trolling...I think? lol [14:19] lol [14:19] Yeah, that qualifies [14:20] frankban I hope it was good enough quality :) [14:20] hatch: it deploys... good enough ;-) [14:20] lol [14:20] If I could just figure out haproxy I could get it back in for review [14:21] I can't seem to get haproxy to proxy anything ( even wordpress ) so I'm guessing I'm doing something wrong [14:47] hatch: that's a lot of code. What's going to be replaced? [14:48] loadRequest… [14:48] kadams54 yeah loadRequest and anything loadRequest calls [14:48] chunks of routeView [14:48] lots of browser.js stuff [14:48] that too [14:48] yeah [14:48] Filter? [14:48] filter doesn't technically exist right now [14:49] it's old code there [14:49] ? [14:49] we don't filter anything anymore [14:49] this.filter.update(query) <- line 261 in loadRequest [14:49] k, will look into it. Thought we still needed it [14:50] ohh I thought that we didn't use this because we only used search now [14:50] jujugui call in 10, kanban please [14:50] hatch: it's used in search still [14:50] you can't select by series and whatnot any longer [14:50] So is state.js taking on functionality that had previously existed in browser.js? [14:50] kadams54: yes [14:51] kadams54: well, the plan is it'll exist in state.js and statecontroller (not there yet) soon [14:51] All of the new functionality seems like a good for state.js, so I approve :-) [14:51] yay! [14:51] rick_h_ kadams54 I was pretty sure that the only line in use in filter was `this.set('text', data.text);` [14:52] Still not sure we need a state manager… I started down that route but decided it was overkill [14:52] bac: thanks for the review. +0 == , correct? ;-) [14:52] after the call we can chat and you can educate me on this filter class [14:52] maybe I missed the functionality somewhere [14:52] kadams54 entirely possible....the functionality is still required, just weather it's required in a capacity that requires a whole new class :) [14:54] whether even [14:55] ;-) [14:55] good afternoon luca__ [14:56] good morning hatch [14:56] we need to invent some cross-timezone greating [14:56] greeting [14:56] bleh I can't type [14:56] Good day to you, fine [sir/madam]. [14:56] old-english has to have something that would sound awesome [14:57] Good post-pre noon [14:57] haha that's not very old-englishy [14:57] kadams54 is so far winning [14:57] A merry Greenwich morning! [14:57] rofl [14:57] old english isn’t my forte [14:58] what? Isn't everyone in England fluent in old english? [14:58] just like all Canadians live in igloos [14:58] It's just English, only all olde timey, right? [14:58] hmmm, don’t think so [14:58] english is a melting pot of other languages that's why it's such a mess lol [14:59] My wife was a Chaucer nut, so I had some exposure Middle English. [14:59] though its the best language melting point [14:59] clearly - it's winning, that must mean it's the best [14:59] right? right? [14:59] ... [14:59] I think Mandarin is winning [15:00] jujugui call in 1 [15:00] luca__ but outside of their borders too? [15:01] hatch: think so, and then it’s spanish [15:01] kadams54: standup please [15:01] hmm, it sounds to me like I'm grossly misinformed [15:01] which woudln't be the first time [15:01] tech diificulties, one moment [15:03] *sigh* [15:03] "It's taking too long to connect you to this video call. Try again in a few minutes." [15:11] I'll review/qa huws branch [15:11] hatch: thanks [15:13] ohh rick_h_ did you want to educate me on the filter stuff? [15:13] hatch: let's plan that for later in the day? I want to go through the branch before my call at noon [15:14] yeah sure np [15:14] hatch: and I've got to review myself. I remember kadams54 and I talking about it and there was something there [15:14] but hazy memory atm [15:18] ok sure - take a peek at it when you get a chance, I'm pretty sure it's not being used anymore with the exception of that one line [15:21] * frankban bbiab [15:27] hatch: comments added, have one big note I'd like to chat on [15:27] ok looking [15:34] hatch: setup a link when you're ready. Would like to chat pre-call in 30. That's a 2hr call so will be afk for a bit [15:34] yep almost done replying [15:37] rick_h_ replied, generating hangout [15:37] rick_h_ https://plus.google.com/hangouts/_/76cpjkvbtggmqe4fcb1apod8cs?authuser=1&hl=en [15:50] * hatch puts thinking cap on [15:52] hatch: sorry to throw wrenches, duck harder lol [15:52] haha - no this is good, we definitely need the state to be coordinated by UI 'section' [15:52] hatch: but <3 the well thought out series of urls/parsing/etc. Exactly what we need [15:53] I am thinking of a way to parse it and allow for expansion/modification easier [15:53] know that was a pita to go through but it'll save so much time later [15:53] great [15:53] should actually have jcastro look at it and go "here's your new urls" :P [15:54] hatch, show me! [15:55] https://github.com/juju/juju-gui/pull/225/files#diff-2a3133ac229f3cbcee63940b2d270c50R300 [15:55] 100 lines of supported urls lol [15:57] will chrome stream .ogg files by default? I'm trying to test before it's up but the file 'could not be found' [15:58] hatch, I don't think icecast is up yet [15:58] any idea if chrome will stream .ogg? [15:58] It will [15:58] ok cool [15:59] There it goes. [15:59] I just didn't want to be scrambling to find a plugin heh [15:59] oh there we go [15:59] elevator music [15:59] cool [15:59] I feel like I'm in a uk sitcom [15:59] Haha [15:59] oh billy go grab my trunk from the boot [16:00] haha, that's all I got [16:00] You're weird. :) [16:00] truth! [16:01] This coffee shop is playing country - it's a weird mix. [16:02] haha [16:04] jujugui remember this call is internal and keep any notes/discussion in internal irc [16:05] * rick_h_ puts away paranoid boss hat [16:05] :) [16:05] rick_h_ I think I figured out a good way to do this state thing - I'll throw it together after this call [16:05] so don't mention the facebook buyout offer? [16:05] hatch: cool [16:06] lol - that was such a poorly executed april fools joke [16:17] ok, 2hr call ended [16:17] well cancelled [16:17] so yay me [16:54] rick_h_ before implementing I wrote a proposition in the PR, check it out when you can [16:55] hatch: will load it up [16:58] * Makyo runs back home real quick. Back on there. [16:59] hatch: replied [16:59] kadams54: appreciate your feedback based on your state understanding as well please [16:59] thanks - I'm trying to think/read/listen at the same time heh [16:59] brain isn't built for three tasks [17:00] hatch: understand, same here [17:01] ok call done, now back to single tasking [17:02] rick_h_ I'm going to keep replying in the PR so that others can get in on the convo/stay up to date on it...ok? [17:02] it's kind of an important architecture part of the app [17:03] hatch: I suppose, but let's not go an hour typing what we can talk in 10 [17:03] haha for sure [17:03] hatch: what I'm looking for is the example realistic use case that requires the dual state purpose [17:04] sure [17:04] typing also helps to think things through before hitting send [17:04] :_ [17:04] :) [17:04] rgr [17:12] rick_h_ ok replied, after you read through that if you want to chat on it fire me a link [17:13] reading [17:13] I think I prefer typing then discussing....helps flesh out ideas before trying to stumble through them in a call first :) [17:14] hatch: https://plus.google.com/hangouts/_/7acpjlm8eiil44q1btqpkg163o?hl=en [17:14] kadams54: ^ if you'd like to join [17:14] or anyone else [17:50] jujugui if anyone has any questions from that call today let me know and happy to see if I can answer or get answers. [18:26] jujugui: i'm going to be starting late due to dr appt tomorrow. [18:27] bac: rgr, thanks for the heads up === BradCrittenden is now known as bac [19:15] Makyo have you had any luck getting the tests to pass in vagrant? [19:15] I haven't... [19:15] * hatch crys [19:16] hatch, not even in test-server? [19:17] Makyo yeah test-server works, but it's over 4x slower :) [19:17] "ain't no body got time for that" [19:20] It's not THAT bad, yeesh. [19:20] At least, not for me. [19:20] Also, xchat's broken for me, sorry if I don't see replies right away. [19:23] oh yeah it's about 20s in test-debug, and in the browser it's ~100s [19:26] so I guess that's 5x heh [19:40] Hasty, hasty :) [20:15] * rick_h_ runs off to get boy from day care. Have a good evening all [20:16] lata [20:39] hmm [20:39] rick_h_ I ran into a slight glitch with this new object, ping whenever you get a chance [20:41] we actually have three sections, sidebar (sectionA), machineView(sectionC), and charm details(sectionB) [20:41] ^ kadams54 [20:41] unless the charm details is in section C [20:42] I'd create the new section...but I'm not sure if there are any new designs which have the charm details go into the machine view area [20:42] Sorry, context switching is hard… we're talking about sections (substrings) of the URL, right? Not sections (panes) within the UI? [20:43] Also, I haven't seen any designs that incorporate pulling details up in machine view… FWIW. [20:43] sections in the UI [20:43] ok then in that case we 'need' three sections [20:44] * hatch puts foots down [20:44] I guess when rick_h_ returns he can tell me to undo it all lol [20:44] From what I've heard, details would be in the sidebar? [20:45] when you look at all of the information in the charm details pane, don't you think that would be hard to make legible in the sidebar? [20:45] especially considering bundle examples [20:45] well really it's pretty trivial to change in my code, just a pita to change 100 url results in the test [20:46] Ah, yeah… some of my conversations two weeks ago with Rick are starting to come back [20:46] We talked about three sections: sidebar, content, and canvas [20:46] Well, maybe canvas was default content [20:46] So then it would just be two [20:46] then the charm details would go over the canvas, like the machine view? [20:47] So details would display in the content section [20:47] Content section would display these: canvas, machineView, details [20:47] ok our current UI requires three states though [20:47] Sidebar would display these: search/browse, inspector, ...? [20:47] I doubt he would have had that UI change also done by vegas [20:48] Not sure what you mean by three states [20:48] sidebar: search/browse/inspector [20:48] details: charm details [20:48] content: canvas/machineview/networkview [20:48] I'm not following why details can't be just another type of content [20:51] the state needs to be organized by UI blocks so that we can parse this state to generate urls [20:52] so each block that can be rendered at the same time needs it's own 'section' in the state object [20:52] so since search, charm details, machine view can all be rendered at once, that's three sections [20:52] yup that's it....that was the defining argument - thanks mr sounding board :) [21:06] hatch: ping [21:07] pong [21:07] * hatch points to the scrollback [21:07] hatch: yes, there are designs to make all details go into the machine view area by making it full size [21:07] ok but that's not coming before..... [21:07] and there's designs trying to find a way to make a mini-details in the sidebar [21:07] whatever the conf coming up is called [21:07] hatch: well, if it's easier to do that than deal with it as a 3rd component then yes [21:08] which I think it might be eaiser to just make the thing full width [21:08] from my end here it's really easy to add another section [21:08] the harder part is writing all these darn tests [21:08] heh [21:08] as in, it took me about 15s to add another section :) [21:08] maybe less [21:09] right ,but then you have to test that as a 3rd combination of sections [21:09] vs just showing the view or not showing the view [21:09] which is a dispatch problem really and not for this branch [21:10] yep [21:10] it makes this branch simpler, and dispatch has to destroy machine view and show details view [21:10] no I'd render it overtop [21:10] I'm not seeing what the big work is, it's some CSS to adjust the width, which doesn't have to happen right away [21:10] that's an optimization when we get there [21:10] right now they're two views, we show one or the other [21:10] well the toggle buttons need to go to a 'blank' state, then when you click on it needs to hide the details [21:11] no, because details have a close button [21:11] you can have 'machine' clicked due to the url or your activity [21:11] and when you close it, the url updates, and it renders the machine view [21:11] and what happens if the user clicks the machine/canvas buttons? [21:11] it will need to close the details [21:11] the url updates, dispatch determines there's nothing to do [21:12] (if it's full screen) [21:12] because you've not closed details [21:12] it's part of the priority/ "should show XXX" work in dispatch [21:12] that feels like a bug to me [21:12] at least for machine view 1.0 [21:12] sure, then we can work on the bug, but adding a 3rd state right now is asking for trouble we don't need [21:12] call? [21:12] sure [21:12] link [21:12] on it [21:13] https://plus.google.com/hangouts/_/7acpiv3m6of4d4nfejfu4or3ns?hl=en [21:19] hey rick_h_, you're back. what was the ES book you found to be good? [21:20] bac: looking [21:21] bac: I think it was http://www.amazon.com/gp/product/B00BAOC2KQ/ref=oh_d__o08_details_o08__i00?ie=UTF8&psc=1 [21:25] rick_h_: ping me if you discover it was something else. thanks [21:25] bbiab [21:32] bac: that was the one [21:33] rick_h_: cool. now i have light reading for the evening [21:37] bac are you going to be our ES expert? [22:30] so the console version of mocha doesn't output the assertion failure message to the console.... [22:35] mocha is fail [22:37] lol [22:39] I wonder if jasmine would fix the mocha bugs [22:40] probably will fix the mocha ones and introduce others [22:40] lol [23:01] Morning [23:13] I wish our tests would all run :( [23:26] morning huwshimi [23:28] huwshimi so the trash can is a different size and colour? Can you just pick one and use it for both then? kind of odd to have two trash cans that are so similar but different [23:29] hatch: How about we make all our fonts the same colour and size as well? [23:30] hatch: Ooh, ooh, we could do the same with our charm icons too! [23:30] we don't need to load in multiple fonts to change the size [23:30] we also don't need multiple trash cans to change the size either :P [23:31] hatch: Well, it's a sprite, not a whole lot of extra bytes [23:31] hatch: You could bring it up with spencer if you feel strongly about it. [23:32] hatch: I'm not even sure where we use delete.png [23:32] relation delete dialogue [23:33] Ah [23:33] I just remember when we had you go through the sprites and clear out all the old unused junk [23:33] that's going to happen again unless we keep on top of the assets [23:35] hatch: Sure, but it's a bit crazy to say that things shouldn't be different sizes and shapes [23:36] erm sizes and colours [23:36] no - but you can scale images down [23:36] so a single asset becomes multiple [23:37] hatch: Not with our spriting system you can't. [23:37] hatch: Also recolouring rasters is not fun. [23:37] We're at least not set up to do that either [23:38] OH KAY then [23:38] just trying to look out for the future! [23:39] hatch: Additionally these are rasters so you can't do any pixel hinting on resize. [23:40] we can't do a css scale transform on the element? [23:41] hatch: Not with any pixel hinting [23:41] what's pixel hinting? [23:42] css blend modes are coming soon [23:42] that's gona be cool [23:43] hatch: Well, if you scale an image you'll end up with blurry edges as they won't line up with the pixel edges, that's why images are redesigned, or at least recreated for different sizes. [23:43] yeah I suppose, I figured they were so small it wouldn't matter [23:44] hatch: The smaller the image, the more this matters. [23:44] hatch: Or at least, the more you will notice it. [23:45] hmm interesting [23:45] well then they should be the same asset then lol [23:45] if they're different size then they're png's. We should be able to size somewhat [23:45] if they're different colors we need to address with UX [23:45] that seems like an issue and we've got a lot of trash cans coming in the machine view [23:45] so better to get it straight [23:45] so much garbage [23:45] :P [23:46] however, it's fine to land until we get answers, but we should get an email out at the very least and bring it up on comingsoon [23:46] hatch: Well, the point is that they need to be different assets as you need to recreate them for each size so you can get the pixels in the right places. [23:47] huwshimi yeah I am not saying your arguments aren't valid [23:47] I'm saying that we should probably use the same one in both places [23:47] like...change the design [23:48] hatch: But that's what I'm saying, that argument is like saying all the fonts should be the same size. [23:48] well no, it's more like saying all the fonts should be the same font [23:48] :) [23:48] hatch: did you review/qa the rest of the branch? [23:48] trash can aside? [23:48] e.g. is it ok to land and bring up with UX [23:48] rick_h_ trying to do it while cooking supper [23:48] hatch: rgr, thanks. I can look in a bit as well after the boy goes to bed [23:48] hatch: You want to change the design for technical reasons, but it is more than logical that the icons might be a different size to fit with different UI elements. [23:49] but let's agree we need to communicate with UX and not debate on it here. We can't move it forward without involving them and their wireframes [23:49] and from my pov, the change of color/etc is more interesting to debate than the size. [23:49] rick_h_: The colour may be because of the background it is on. [23:50] it may well be because of outdated colours too. [23:50] huwshimi are you firing events for some reason? Or just to allow for testing? [23:50] deleteToken, selectToken for example [23:51] hatch: I assume we'll want to interact with the token somehow... [23:51] ok - I wasn't sure if there was an actual reason or if it was just put there as a sort of placeholder [23:51] I'm not sure how we would interact with it tbh [23:52] events do seem to be a good way though [23:52] hatch: I actually don't know what will happen with those events yet... [23:54] code looks good with some tidbits, just qa'ing now [23:55] rick_h_: Should I move onto the container token next? [23:55] huwshimi: yes please [23:55] huwshimi: and you should be able to wire up the click on the machine token to fill out the containers. You might need to talk with kadams54 about that at some point [23:56] rick_h_: ok sure [23:56] huwshimi: and I want to make sure you see https://www.dropbox.com/sh/lvdydgiu7jeuuso/7MlRX5-IPu#lh:null-05.4-container.png [23:56] that folder has a list of container-y wireframes [23:56] huwshimi: I'll be around if you want to chat. Working on taxes right now. (Yay.) [23:56] including the container token stuff [23:56] kadams54: yay you [23:56] I'm so glad to get that over with, always hangs over my head [23:57] huwshimi: check out 8 of 22 and 7 of 22 [23:57]