rick_hhatch: no, it was what I was going to start but IE bugs come first00:06
hatchfishing done?00:07
rick_hhatch: yes...why yes it does00:10
hatchcatch anything?00:10
rick_hhatch: no, it was not a good trip 00:10
rick_hlost my pole/reel into the lake... /me isn't proud00:11
hatchhaha darn I hate it when that happens00:17
rick_hnew rod/reel, just over a week old. Not feeling the love of today00:18
hatchrick_h: last year 2 of my buddies went out fishing in a canoe with all brand new gear - an hour later they returned soaked....they tipped the canoe and lost everything00:54
hatchcounting cell phones and glasses it was probably a $3000 tip00:55
hatchyup that was an expensive weekend!00:57
MakyoGood news, everyone.01:22
rick_hMakyo: I like good news01:22
MakyoThe viewBox attr on SVGs fixes both FF and IE1001:23
rick_hMakyo: sweet!01:23
Makyorick_h, Going to correct the default icons first, then we can figure out a story for charm/icon.svg01:23
bcsallerMakyo: I think the api server can patch them on ingest 01:36
rick_hbcsaller: yea, it completely can. Just needs to be added01:37
rick_hbcsaller: and should get it added to the demo/template provided in the docs perhaps01:37
Makyobcsaller, yeah, that's what I was thinking.  Gary was concerned about that as a solution over requesting charm authors to do it themselves.01:38
bcsallercharm authors no more known what they SVG editors save as does than... well... I bet they don't know01:39
MakyoHah, fair.01:39
MakyoWhere's the template, rick_h ? The attribute would change based on the final width/height values, curious how those are specified.01:39
rick_hwe'll never have it 'fixed' if we rely 100% on charm authors01:39
rick_hMakyo: sec, in the juju docs. Looking. 01:40
rick_hMakyo: https://juju.ubuntu.com/docs/authors-charm-icon.html01:40
MakyoI like that the title is "None" :P01:40
rick_hbut the template is at https://juju.ubuntu.com/docs/media/icon.svg I bet jcastro or marcoceppi can update that if we need. 01:41
Makyorick_h, Alright.  From the looks of it, authors may be editing the width/height in Inkscape, as well.  I may prowl around through there and see if there's an option in Inkscape itself for viewBox.01:42
rick_hhatch: I submitted without the extra tests but poked at the env and db. Would like to chat about how those 'work' tomorrow if you've got time01:42
rick_hMakyo: cool01:42
rick_hthanks for looking into it. 01:42
rick_hhatch: e.g. getting a charm from the db doesn't have the config values in it that I could tell. And there wasn't an idea of fetching the charm from the env I could see either01:43
Makyorick_h, Okay, found it.  Hopefully it's this easy!01:46
hatchrick_h: yeah remind me when I get in and I'll fill ya in03:20
hatchunless you're around now03:20
hatchMakyo: thanks :D03:20
rick_hhatch: around now but past my bed time. I'll hit you up tomorrow03:27
marcoceppirick_h: what did I do?07:11
=== schwuk_away is now known as schwuk
rick_hheh, well it's closer I guess. Friend was trying to convince me FF was faster than chrome now. Except test suite 37s in chrome and 56s in FF12:15
rick_hit used to be twice as much though, so getting better12:15
bacabentley, benji, sinzui: in a fit of OCD i reorganized the task list on our google doc from yesterday.  i also made cards for my tasks in kanban.  might be nice to add the ones you've been assigned too.  i'll make cards for the unassigned ones.13:19
benjithanks; oh, yeah, I need to add a card for mine too13:19
sinzuithank you bac13:19
rick_hlol, love fits of OCD13:20
bacor CDO as jeff bailey calls it.  everything should be alphabetized.13:20
rick_hyay! I have defeated IE! Where is my victory parade?13:23
bacsinzui: if you understand #10 on the list could you re-write it to make sense?  it is currently gibberish to me.13:24
jcsackettjujugui: can someone else with ie10 ready confirm that bug 1112783 appears to be fixed? i have tried to reproduce and it works for me. if it works for someone else i'll mark it resolved.13:26
_mup_Bug #1112783: 'destroy service' never returns from confirmation dialog <ie10> <juju-gui:Triaged> <https://launchpad.net/bugs/1112783>13:26
rick_hjcsackett: sec, will in just a min13:27
jcsackettrick_h: thanks.13:27
abentleybac: ack13:30
rick_hjujugui IE QA/review please https://codereview.appspot.com/1287104613:30
benjirick_h: looking13:30
sinzuibac, I cannot edit it13:31
rick_hjcsackett: cannot reproduce13:31
rick_hjcsackett: it's from Feb so not surprised we've fixed something between now and then13:31
jcsackettrick_h: huzzah!13:31
jcsackettrick_h: move the card to releasable, or just delete it?13:32
bacsinzui: your gmail id had edit permission.  now your canonical does too13:33
rick_hjcsackett: I'd just delete. 13:33
rick_hjcsackett: and mark the bug as invalid. If it's releaseable then we'd do fix-released I'd think13:33
jcsackettrick_h: i would say it's fix-released; it wasn't invalid, we just fixed it at some point. but i'll delete the card.13:34
rick_hjcsackett: k13:34
rick_hjcsackett: then keep the card and get bi-weekly bug credit :)13:34
jcsackettMakyo: i see you assigned to the card for bug 1207035, but assigned in LP to 1207036; are you working on both?13:36
_mup_Bug #1207035: Some charm icons are improperly sized in IE10 <charmbrowser> <ie10> <charmworld:Triaged> <juju-gui:Triaged by makyo> <https://launchpad.net/bugs/1207035>13:36
jcsackettoops, mup didn't pick up the second one. bug 1207036.13:36
_mup_Bug #1207036: Service SVG icon in inspector is stretched vertically <ie10> <juju-gui:Triaged by makyo> <https://launchpad.net/bugs/1207036>13:36
rick_hjcsackett: he thinks he's got a fix for both with one change13:38
rick_hjcsackett: so those should be good coming soon13:38
rick_hthough might need a charmworld branch as well. 13:38
abentleybac: Could you please give me edit rights on Charmworld URLs and Pages notes?13:39
bacabentley: your gmail identity has them.  i think it automatically granted to all identities in the original chat.13:40
bacabentley: you want me to add your canonical?13:40
abentleybac: Yes, please.13:41
bacabentley: try now13:42
abentleybac: Looks good.13:43
bacabentley: cool.  just got smart and opened to all canonical13:43
rick_hjcsackett: can you qa https://codereview.appspot.com/12871046 for me in your fresh IE setup please? 13:44
jcsackettrick_h: sure.13:48
jcsackettrick_h: have you seen this happen on IE10 before? happened after i hit the search bar in fullscreen. http://d.pr/i/PpyH13:53
rick_hjcsackett: looking13:54
rick_hjcsackett: oh...interesting...not seen that one13:54
jcsackettrick_h: i'm checking if that happens in trunk. can't imagine it's related to your branch.13:54
rick_hjcsackett: so it does it every time?13:54
rick_hah, I can dupe it. In fullscreen, missed tghat part13:55
jcsackettrick_h: it does it in trunk, so it's not your branch.13:55
rick_hjcsackett: yea, I bet it's the autocomplete widget which is on comingsoon, but not jujucharms. Will need to file that as a new one13:56
jcsackettalso, in IE10 if i just load jujugui and click the "browse" button i'm not going to fullscreen.13:56
jcsackettrick_h: can you reproduce that?13:56
* jcsackett hopes its something weird locally.13:56
rick_hjcsackett: no, it changes modes for me13:56
jcsackettmorning, hatch.13:57
rick_hjcsackett: ok, will file the AC bug and add a card for that13:57
jcsackettrick_h: ok.13:57
jcsacketti'll update your MP with my QA ok.13:58
jcsackettrick_h: got any notions of why AC pushes down the rest of the screen? i'm free to pursue the bug, but would like pointers if you've got 'em.14:00
rick_hjcsackett: thanks for the search link then14:01
rick_hjcsackett: errr, the notion that it's used in search as well14:01
rick_hjcsackett: so basically going to be something in the css of the autocomplete that's not being position absolute I guess. 14:01
rick_hjcsackett: I'd just start going through the .aclist* classes and nodes and seeing if they can be tweaked14:02
jcsackettrick_h: ok.14:02
jcsackettrick_h: i'm actually less free than i thought--hatch emailed me a bug last night i need to look into as it appears to be browsercharm fallout.14:03
jcsackettprobably simple though, so i should be able to get to AC soon.14:03
rick_hjcsackett: rgr14:03
jcsacketthatch: i saw your email and the bug, but i can't reproduce.14:06
hazmatsinzui, incidentally that rack charm store issue is caused by a bug in the charm itself14:07
sinzuihazmat, great.14:07
sinzuihazmat, details14:07
sinzuior update the bug14:07
hazmatsinzui, i replyd via email, its going to take a few for lp to publish14:08
hatchjcsackett: hmm I just did http://comingsoon.jujucharms.com/:flags:/serviceInspector/ dd mediawiki, click 'destroy service' click 'confirm', see error in console14:08
jcsacketthatch: i missed destroy was available before deploying.14:13
jcsacketthatch: btw, it's "Jon" not "John". :-P14:18
hatchyeah I got nothin14:19
hatchoops :)14:19
abentleyjcsackett: I always thought jc was short for Jesus Christ :-P14:21
jcsackettonly in my most meglomaniacal moments.14:22
rick_hI have slain the mighty IE! All bow before my...oh wait...my bad14:22
* jcsackett laughs14:22
hatchor that his last name was Penney14:22
jcsackettyes, because none of these are jokes i have heard a thousand times. :-P14:23
hatchor Chasez14:23
hatchhave you heard THAT one?14:23
jcsacketti have not. well played.14:23
jcsackettonly half credit though, as it's just a variation on a theme. :-P14:23
hatchlol true true14:24
jcsacketthatch: fix up for that bug https://codereview.appspot.com/1279504614:26
jcsackettcan you review?14:26
hatchon it14:26
jcsackettjujugui: i suppose even though it's one line, i need one other reviewer.:-P  ^14:27
hatchthere are two now14:27
rick_hjcsackett: already done, but figured you'd trivial it14:27
jcsackettrick_h: i usually eschew trivial when the fix is for actual breakage.14:28
jcsackettand thanks for the review.14:28
hatchrick_h: when did you want to go over the db stuff14:30
rick_hhatch: after the standup call ok?14:30
hatchyep no problem14:30
sinzuiabentley, jc is short for our lord and savour John Cleese.14:32
jcsackettok, *that* one is new.14:33
hatchjcsackett: oh crud it's a little late now but we should have added a test for that bug14:39
* hatch hasn't had his coffee yet14:40
abentleysinzui: chat?14:43
hatchjcsackett: I can add that test14:43
jcsacketthatch: ok, i'm pursuing a css issue with AC in IE10. if you've not gotten around to the test before i'm done with that, i can pick it up then.14:44
hatchI'll start on it now, I need a break from this constraints stuff14:44
sinzuiabentley, in 15? I am in meetings14:45
abentleysinzui: sure.14:45
abentleysinzui: How's it going?15:10
sinzuiabentley, all clear. we can talk15:11
abentleysinzui: hangout at https://plus.google.com/hangouts/_/666d51fb65e7f2ce7442c104ad381044fefac89a?hl=en-GB15:12
marcoceppiHi juju-gui people. Question about the canvas in gui. (well, two)15:15
MakyoShoot, marcoceppi 15:16
marcoceppiOne, is it worth filing a bug to disable right-click context menu from the browser on the gui? We had 21 users who were operating under the assumption that in order to bring up the menus they needed to right click (they also though that the long click was right click only) and were getting frustrated by having the browser menu pop up when using right click15:16
marcoceppiafter showing them it was all Mouse 1 they got it, so I'm just doing due diligence as I said I'd bring it up with you guys15:17
Makyomarcoceppi, jujugui, I say file it, and then we can triage it during calls.  From my point of view, it makes enough sense, at least on the canvas.15:17
hatchyeah that's a tough one from past experience15:18
hatchif you're going to disable it, it needs to have a custom one15:18
hatchelse I've found people get equally frustrated hah15:18
* hatch included15:18
rick_hwhy would people right-click on the web? /me never would have even thought of thatr15:18
benjiif that's what the users expect, we could also actually make right click bring up the menu15:18
hatchbenji: right15:18
marcoceppithe second question is more of a bug, but I wanted a sanity check on it. Currently I can't highlight text when drilling down in to the unit view (ie, click on service, click on unit, try to select IP address to copy), is this a known issue? Expected behavior? If not I'll also file a bug15:18
rick_hMakyo: hmm, maybe we can deploy flag something like that? /me doesn't want to lose 'right-click - inspect element' even in the canvas15:19
marcoceppirick_h: maybe enable a "developer" mode which doesn't override right click?15:19
marcoceppiAny event I'll create a bug for you guys to track it15:19
hatchmarcoceppi: I think the selection was disabled, checking15:20
hazmatthe inspector mode will change that15:20
Makyorick_h, agreed.15:20
hatchmarcoceppi: just to confirm this is in the new inspector?15:20
marcoceppihatch: I've just been deploying cs:precise/juju-gui15:20
marcoceppiSo whatever that constitues15:20
hazmatthe old unit/service view it sounds like15:20
hatchbut do you add /:flags:/serviceInspector to the url?15:21
Makyohatch, marcoceppi - no, drilling down is old behavior.15:21
marcoceppiIf there's a new inspector coming I'll just let them know that new cool shit is coming and to just dealwithit.gif15:21
Makyohatch, yeah, stuff under feature flags wouldn't be used in these contexts, I don't think15:22
hatchmarcoceppi: fyi on comingsoon.jujucharms.com I can't repro your select issue15:22
hatchbut yes big changes coming down the pipe soon :)15:22
* marcoceppi peeks15:22
marcoceppiI like the slider15:22
marcoceppihatch: ah, yeah selecting text works. Thanks!15:23
marcoceppiWe had about 25 people using the GUI for the first time, all really overwhelmingly good feedback outside of the above questions!15:23
hatchoh that's so awesome to hear :)15:24
rick_hmarcoceppi: good to hear15:24
marcoceppiI saw them using it more than they were the cli ;)15:24
MakyoExcellent :)15:24
hatchhaha perfect ;)15:24
jcsackettrick_h: i think i have a fix up for the AC thing. care to review? https://codereview.appspot.com/12760045/15:44
rick_hjcsackett: sure thing, doing the lbox dance with mine right now. Will pull/check it out when it's done15:44
rick_hand swap you reviews :)15:44
rick_hjcsackett: oh, can you see if that fixes the home bug I filed as well then? Seems it might be related. #121269515:45
_mup_Bug #1212695: home link moves during search in ie10 <ie10> <juju-gui:Triaged> <https://launchpad.net/bugs/1212695>15:45
rick_hjcsackett: maybe not, nvn15:46
jcsackettrick_h: i'll look.15:46
jcsackettbut yeah.15:46
jcsackettrick_h: nvm on my MP. 15:49
jcsackettin looking at yours i've realized mine introduces a new horror.15:49
jcsackett(your bug, not your MP)15:50
jcsackettgoddamn IE requiring absolute positioning...15:50
hatchjujugui guichat in 10 kanban now15:50
rick_hjcsackett: ugh, yea I was afraid of that. 15:51
rick_hjujugui couple of reviews please, not IE specific. Happens in all browsers. https://codereview.appspot.com/12997043 but the bug was noted in IE15:51
jcsackettrick_h: looking.15:51
Makyorick_h, on it.15:51
rick_hthanks guys15:54
hatchjujugui looking for 2 very quick reviews on https://codereview.appspot.com/12998043/ as well15:55
Makyohatch, on it15:58
bcsallerthink its already got two :)15:58
hatchjujugui guichat in 215:58
Makyobcsaller, hatch Well...bonus!15:59
bcsallereveryone likes more tests15:59
hatchjujugui guichat on now16:00
jcsackettok. i'm going to get lunch. IE10 has temporarily dispirited me.16:28
rick_hjcsackett: hah, don't let it beat you down!16:29
hatchare constraints supposed to be 'dumb' right now? Or are we supposed to be pulling the options from juju?16:42
rick_hno idea16:46
hatchI love lines that are 80 characters long16:59
hatchsuch - a -rebel16:59
rick_hdown with him!16:59
hatchugh I can't wait until we switch to sass and have sourcemaps for our css17:01
loveaQuick question about charm deployments - Juju client run on node that has internet access and on isolated internal LAN, Juju boostrap node on isolated LAN/no internet access, Juju-GUI node on isolated LAN/no internet access. Q1: Does Juju client pull charms from the charm store and push to Juju bootstrap node? Q2: Is Juju-GUI node useless without an internet connection to pull pickable charms from?17:07
rick_hlovea: the final point is true. The Gui needs access to https://manage.jujucharms.com to show the list of available charms to pick from. 17:08
rick_hicons and data are pulled from that data endpoint17:08
lovearick_h: ok, thought so17:09
rick_hhatch: Makyo do you guys know how a 'deploy' goes down? Does the gui send the charm over to the bootstrap node? Or do we just tell it to fetch it?17:12
Makyorick_h, we just tell it to fetch it.17:12
hatchjust fetch it17:12
rick_hMakyo: cool, thanks. What I assumed. 17:12
rick_hlovea: so does that answer the other part then?17:12
loveaForgive my ignorance but would the theory be that, given an internet enabled Juju-GUI node, I could wire up my charms to meet my specific deployment needs as a "model", export the "model", use the Juju client to actually implement the "model" in my isolated environment.17:13
Makyorick_h, lovea GUI will work as a monitoring tool behind the firewall, would still show the environment, but charm-store would be empty.17:13
Makyolovea, yes.  You can use the GUI and then export (or even just write down) your environment to be deployed elsewhere.17:14
Makyolovea, There is a section on "Deploying behind a firewall" at https://jujucharms.com/fullscreen/search/precise/juju-gui-74/?series=precise&text=juju-gui&type=approved#bws-readme17:15
loveaor are we saying that Juju bootstrap node also needs to be internet enabled because it gets fetch instructions from Juju client?17:15
loveaMakyo: ok I'll read up on this. Many thanks17:15
Makyolovea, the bootstrap node would only need internet if deploying charms from the charmstore (eg: "cs:precise/mysql").  Using the CLI, one can deploy local charms, however.17:15
lovearick_h, Makyo: Thans for te clarifications17:16
benjiabentley: the bundle API tweak branch is up for review if you have a moment (after lunch): https://code.launchpad.net/~benji/charmworld/update-bundle-api/+merge/18038417:24
abentleybenji: Okay.17:24
hatchrick_h: what was the trick you used to get templates auto generating again?17:31
rick_hhatch: there was a typo in the dir name in the config of places to look?17:31
hatchahh right17:31
rick_hmanually I'd do rm build-shared/juju-ui/templates.js  && make build-shared/juju-ui/templates.js17:32
hatchyeah that's it, I'll try that17:32
abentleybenji: I have an important call in 15 and I need to do some prep.  Is it cool if I delay your review by an hour or so?17:43
abentleybenji: I have some thoughts, but haven't gotten everything done.17:43
benjiabentley: sure17:44
hatchbcsaller: this collective event stuff has gota be fixed lol17:46
jcsackettdoes IE have any sort of dev tools?17:48
rick_hjcsackett: yea, in the options is "developer tools"17:48
hatchthey are 'sort of dev tools'17:48
rick_h(right side gear)17:48
hatchF12 opens it17:48
rick_hyea, I noticed 'debugger;' didn't do me any good :(17:48
rick_hand the element selector only half works17:49
hatchnope you have to click 'start debugging'17:49
hatchthe element selector works only if you hit 'refresh' in the devtools first17:49
jcsackettrick_h: i do not have a developer tools option in my gear menu. f12 worked though.17:49
hatchjcsackett: they actually market it as "The F12 Developer Tools" :D17:50
rick_hjcsackett: oh, ok then17:50
* jcsackett continues to loathe IE.17:50
rick_hcarry on17:51
hatchthey say the IE11 ones are actually real devtools17:51
* hatch is hoping17:51
hatchI used to have to support IE7+17:51
* jcsackett is hoping to never need IE11 tools, b/c he's hoping to never to IE work again.17:51
hatchAND the 'simulator' in IE10 doesn't even come close to simulating IE8,9 properly17:51
hatchsorry :)17:51
jcsackettrick_h: got a bit to chat? my current approach is so error prone i think i need to step back and maybe hear other ideas.17:52
rick_hthe 9 one did a good job with 817:52
rick_hjcsackett: sure thing17:52
jcsackettrick_h: guichat is open.17:52
hatchrick_h: also lemme know when you want to chat about the db stuff17:53
MakyoHm.  Good news, bad news.17:57
MakyoGood news: ingest now adds viewBox to icons.  It works in FF and IE.17:57
MakyoBad news: it only MOSTLY works in IE.17:58
MakyoThis wouldn't be easy >:/17:58
hatchwhat's mostly? :)17:58
MakyoIt fixes the icons in the service blocks, but not in the browser (which might be because they're not styled with a size - investigating), and not in the inspector (for potentially the same reason)17:59
hatchARG!!!! I can't propose my branch until I upgrade jshint so I can tell it about globals......18:00
hatchMakyo: ahh ok, well good luck :)18:00
MakyoWhat globals?18:00
hatchthis._unitSpinner = new Spinner({18:00
MakyoOh, and Spinner isn't provided by YUI or anything?  I mean, that's how we get d3.18:01
hatchI should be able to do /* global Spinner */18:01
hatchI think d3 is defined by the namespace methods18:02
hatchso it's not actually a global18:02
hatchwe dont jslint the html so that's why this hasn't come up before18:02
MakyoNah, not that I can see.  It's just used immediately in app/topology/topology.js18:03
MakyoOh well, though.18:03
MakyoWe've used the spinner before in JS, though.  It hasn't come up yet?18:04
hatchhmm you're right18:04
hatchnope the spinner is only ever used in index.html18:04
hatchand in my new code18:04
hatchohh wait18:05
hatchI was grepping crappy18:05
Makyoand app/widgets/overlay-indicator.js18:05
hatchguess I'll have to make this spinner in the index.html as well18:05
hatchI am curious as to how d3 is being defined there18:06
hatchin topology.js18:06
hatchas in why the linter isn't going wako18:06
MakyoWhat is Y.spinner.getSpinner()?18:07
hatchI guess I could refactor that to be a generator18:08
MakyoIt's already there, that's what I'm saying.18:08
hatchbut even there it's not blowing up...18:08
MakyoI'm sure you'll figure it out.  We're all counting on you.,18:08
hatch /rage18:09
MakyoSo why won't Y.spinner.getSpinner() work in your case?18:09
hatchit's mergingness of it's options will not work for some of the options i need18:10
MakyoCan't you provide an options object?18:10
hatchso I can fix that....but I'm not sure if that's the right place18:10
MakyoI'm not sure what you mean by mergingness, so...18:10
hatchit merges it's defaults with the config that I need18:11
hatchso I'll either need to specify more options resetting them to other defautls18:11
hatchor patch the file18:11
rick_hjcsackett: http://paste.mitechie.com/show/1002/18:12
MakyoPatching sounds notsogood.  Is there an issue with specifying all the options you need?18:12
hatchwell nothing beyond I have to go find out what the defaults are18:13
hatchwhen the issue at hand is really the tools18:14
MakyoThat's fair, and perhaps working with upstream is required, but working with tools is kinda what we do.18:15
rick_hhatch: reuse the existing spinner. What's wrong with it?18:15
hatchit doesn't fit in an input box and isn't blue :)18:15
bcsallerhatch: can't you just use an animated gif for the little in <input> field spinner you need?18:15
rick_hwell, wtf is 'blue' a thing? Spinner is standardized.18:15
rick_hand we can adjust the size in the widget if we need I believe18:16
rick_hah, nvm. We use the same animated gif, not js spinner. We had a bug for that at some point18:16
hatchbcsaller: only if you have one :)18:16
bcsallerhatch: http://www.ajaxload.info/ ;)18:17
rick_hhatch: why are we spinning a single input box?18:17
hatchindicator to let the user know units are being spun up18:17
rick_hlink to the UX mockup?18:17
hatchoh it's on the dropbox one18:18
hatchI don't have the link on this machine18:18
bcsallerin google drive if you search inspector and pull the v11 hit though its there18:18
rick_hyea, loading/looking18:18
hatchit's ok I have updated the linter already18:19
hatchthere is only 2145 errors!18:19
rick_hthe little blue refresh-looking thing?18:19
hatchoh that was only 33% of the files heh18:20
hatchI think it needs some config options18:20
hatchrick_h: yes18:20
rick_hhatch: well where's the assets for it then :P18:20
rick_hwhy are you complaining about Spinner when you can't use that anyway18:20
* rick_h is confused18:20
MakyoThis sounds like a less than ideal path, and also counter to what we've done in the past of two-step implement-and-then-style branches.18:21
hatchso drop the spinner?18:22
rick_hyea, until you have the assets to implement it imo18:22
MakyoOr implement A Spinner, and style it later.18:22
rick_hyea, I guess I'd do it textual with a "replace with gif when available"18:23
hatchit is implemented heh18:23
MakyoEh, nevermind, I'm out.  Gotta eat, then propose the charmworld stuff.18:24
hatchbcsaller: I'm going to grab the asset from there I guess :)18:27
hatchinteresting the old jshint accepted our json files, the new one does not18:30
hatchlooks like it doesn't accept the old config file18:31
hatchyeah it'll take a bit to reconstruct this config18:34
hatchbcsaller: it appears that the databinding system doesn't like to have multiple fields bound to the same value19:10
hatchohh wait nm19:12
hatchwait nope it's definitely broken19:13
hatchit only ever calls the update method once regardless of how many fields are bound19:14
hatchthe 'target' gets overridden by the last element with the binding set to it19:16
Makyojujugui charmworld branch for viewbox on icons (default and ingested) https://codereview.appspot.com/1300104419:16
benjiMakyo: I assume you're looking for a review.  I'll take a look.19:17
hatchlooks good to me but I'm not really qualified :)19:17
bacMakyo: i'll look19:17
Makyobenji, bac Yes, thanks.19:17
bacMakyo: have you run 'make lint'?19:20
Makyobac, not yet, had assumed .lbox.check did that, will do it now.19:21
bacMakyo: hey you could add that to .lbox.check if you wanted.  but, the charmworlders are not generally using lbox.19:22
Makyobac, Oh, okay, maybe I'll do that.19:23
bacMakyo: actuallly, charmworld is using tarmac to land, triggered by the merge proposal on LP.  so using lbox is actually problematic.19:27
Makyobac, Only once the MP is marked approved, though, right?19:28
bacMakyo: i'll go vote approved on LP.  you'll then need to set a commit message and mark the whole thing Approved when you're ready to land.19:28
Makyobac, so propose shouldn't affect it, just don't submit?19:28
Makyobac, okay.  Only done this once before :/19:29
bacMakyo: votes need to be recorded over there too.  btw, charmworld is only requiring one review19:29
MakyoOh, okay.19:29
bacMakyo: for now, it may make sense to change .lbox.check to be only run /bin/false19:30
Makyobenji, single/double quotes: the processing instruction is generated by ElementTree with single quotes, but attributes are generated with double quotes.  I could switch to single quotes and escape the internal ones, but I figured this was cleaner.19:32
benjiMakyo: that makes sense; maybe it's worth commenting19:33
Makyobenji, Alright.  As for the logs issue, I was under the impression that we should be adding the attribute on ingest, since we can as long as we have width/height (it's just 0 0 <width> <height> every time).  Should I log it anyway, or die, or..?19:34
benjiMakyo: ah, so we just need the attributes, even if they are 0?  If so, a warning (or info, or nothing) seems fine.  I would add a comment about why we're doing what we're doing though.19:36
Makyobenji, I can check for 0 as well, since that would be a malformed SVG.  Logging and commenting sound fine too.19:37
benjiwell, you can use lbox for inline code reviews if you want, but landing has to be done through LP19:39
abentleysinzui: I just landed versioned charm-ids.19:57
abentleysinzui: Or rather, deployed them to staging.19:58
hatchbcsaller: I'm going to work around this now but I'd like your input whenever you have some spare time https://bugs.launchpad.net/juju-gui/+bug/121281320:17
_mup_Bug #1212813: databinding only allows a single field to be bound to an attribute <juju-gui:New> <https://launchpad.net/bugs/1212813>20:17
sinzuiabentley, benji: do either of you have time to review my unsexy branch that adds the concept of API3 https://code.launchpad.net/~sinzui/charmworld/add-api-3/+merge/18042620:43
abentleysinzui: Sure.20:44
abentleysinzui: For the test_health changes, it looks like thge test cases are almost identical.  Have you considered using two TestCases and just overriding the check_api call?20:54
sinzuiabentley, I can do that.20:54
abentleysinzui: The indentation at 173-174 looks odd to me.  I'd expect it to line up with the opening paren, or else for it to be indented once.20:55
abentleysinzui: check_api3 isn't used yet, right?20:56
sinzuiabentley, nasty pep8 check wanted the if continuation separated from the lines executed20:56
sinzuiabentley, check_api3 is used by the /heartbeat view20:57
sinzuiIts check though uses the same rules as API220:57
sinzuialmost the same rules20:57
sinzuiall the checks have a common control structure20:58
abentleysinzui: How does check_api3 get referenced by /heartbeat?20:58
abentleysinzui: nm, found it.20:58
sinzuiline 24420:59
abentleysinzui: r=me.21:00
sinzuiabentley, thanks. I'll refactor the tests before landing21:00
MakyoGood news everybody.21:16
bcsallerhatch: having looked at it I don't see why that would be true (it used to be when we used a map for _bindings, but now its an array)21:16
MakyoI was actually insane when I said things didn't work on Chrome.  Thinks work fine :)21:17
hatchMakyo: yay for insanity!21:17
Makyo"things are actually alright" is the best kind of insanity.21:17
hatchI think that's insanity after the medication21:17
MakyoHah! I reverted my config change before I checked the icons last time.21:17
MakyoIt looks like the branch approved earlier fixes most icon issues.21:18
MakyoMinus the one in the inspector, but that appears to be styled wrong.21:18
Makyojujugui need one quick review for trivial: https://codereview.appspot.com/12828045/21:36
hatchMakyo: lgtm21:37
Makyohatch, thanks.21:37
* hatch is furious - sent my wifes HTC One in for repairs because the screen is separating and they just called saying that the repair depo says it's physical damage and that it'll be $25021:40
hatchbut of course I can't talk to anyone about this bs until tomorrow21:42
hatchgood thing I took pictures of the device before it was shipped21:43
* Makyo finally dogwalks21:59
hatchbcsaller: hmm not sure - I've been fighting getting the tests to pass again so I haven't looked into it much but I know that the 'update' method was only called with a single value (the last one) so we'll just have to look at that sometime in the future22:10
bcsallerhatch: I'll try adding a test for it to what I'm doing now. Does your branch already include the change to sum up the changeKeys between updateDOM intervals, because this would look like that as well I think22:12
hatchbcsaller: here is my branch https://code.launchpad.net/~hatch/juju-gui/scale-up-ui22:14
hatcha....plethora of tests fail now hah22:14
hatchif plethora was 722:14
hatchwell...7 is when it stops22:14
hatchthere is probably a lot more22:14
hatcharg I broke the databinding22:34
hatchbcsaller: have a minute to take a look at my databinding code?22:44
hatchI'm sure I'm missing something obvious22:44
bcsallerhatch: what part do you want me to look at?22:44
hatchfor some reason this causes the conflict code to apply the conflict style to the wrong element22:44
hatchwhich element....I'm not sure....possibly the mirror one?22:45
hatchbcsaller: if I comment out 194-204 it works as expected22:45
bcsallerhatch: I think you put the key collapse code in the wrong place though, I saw this living in modelChangeHandler (dealing only with keys) and cleared in updateDOM. By dealing only with keys at that layer we can select all bindings that match those keys in deltaFromChange22:46
bcsallerdoes that make sense?22:46
bcsalleryou try and remove dup bindings, but this is the binding, not the change, we just want all the bindings matching a key at this point22:48
bcsaller(or a set of keys)22:48
bcsallerif you'd like I can take a stab at it22:48
hatchhmm maybe I've been staring at this too long because I was under the impression that I was aggregating the deltas using this technique22:49
bcsallerdeltasFromChange is only the delta in the keyspace of the model, not the change events22:50
bcsalleryou are close though22:50
hatchok I am totally missunderstanding what's happening here then - maybe if you wouldn't mind fixing it so I can see what I'm missing22:51
* hatch slides a pint of beer across the table22:52
hatchhow about now?22:52
bcsallerhatch: happy to, I'll write it now and you can merge when I have it ready22:52
bcsallervirtual beer... 22:52
hatchnext sprint - real beer22:52
hatchhey huwshimi how are you doing?23:04
huwshimihatch: Good thanks, yourself?23:04
hatchcould be better :)23:06
huwshimihatch: Are you sick?23:08
hatchnah, just stupid things happening today23:08
hatchHTC repair depo not wanting to fix my wifes One set me off23:08
hatchhuwshimi: so hows that IE stuff coming?23:11
bcsallerhatch: so I *think* this should fix it, but I didn't generate a full test for the multiple key case as getting the timing right on that would delay you actually testing it. lp:~bcsaller/juju-gui/databinding-safeupdate 23:12
hatchbcsaller: ok looking thanks23:13
bcsallerhatch: if that doesn't help I'll keep at it with you23:13
hatchbcsaller: ok I gota merge this in manually so it'll take a second23:15
bcsallerhatch: if you didn't change databindings.js for anything else take that version23:15
hatchyeah I added the prev value stuff23:16
huwshimihatch: Our animating input fields are being a pain. We need to set the inputs in the inspector to a flexible width as the size of the scrollbars varies between browsers.23:16
bcsallerrouter issue, might lose wifi here23:18
hatchok merged...trying23:19
hatchbcsaller: awesome, fixed23:21
hatchthanks a ton23:21
hatchnow to figure out the difference23:22
hatchI think I saw it though23:22
bcsalleryeah, I was going to ask23:22
hatchahhhh now I see now I see23:22
bcsallerthis just accumulates keys until updateDOM is called23:22
hatchyou only deal with the keys not the datasets23:22
bcsallerkeeps it simple23:23
hatchI'm not entirely clear why my technique broke it though23:23
hatchclearly it was wrong...but I don't know why....23:23
bcsalleryou thought delta in deltaFromChange means ws delta changes or something23:23
bcsallerwe are selecting the binding objects for model change keys here23:23
bcsallerso a set of keys becomes a set of bindings23:23
bcsallernot change events23:24
hatchohh that's why it was pointing to the incorrect element23:24
hatchok got it now23:24
bcsallerthe delta is vs the keyspace of the model in this case. I'm open to renaming it so this doesn't happen to another dev23:24
bcsallerbindingsFromModelChange or  somesuch23:25
hatchyeah it's not  bad idea....now I"m going to tackle the remaining failures :)23:26
hatchthanks again!23:26
hatchohh my prevValue code doesn't support pojo's23:27
hatchhuwshimi: did you need a hand or need to bounce any ideas off anyone for that?23:30
huwshimihatch: It's ok I just need to modify the scrolling code which is all fairly hardcoded to be more flexible23:31
hatchahh ok, I think there was a scrolling change landed today23:34
hatchso you might want to merge trunk23:34
huwshimihatch: Ah right, thanks23:36

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!