=== schwuk_away is now known as schwuk === schwuk is now known as schwuk_away === schwuk_away is now known as schwuk [12:26] teknico, bcsaller could you please make sure that you have your London tickets approved, purchased, and entered on https://docs.google.com/a/canonical.com/spreadsheet/ccc?key=0AtC9etoysSQldFRfYU0zZjE5WThuNUVWRXlEVmQwOVE#gid=8 by EoD today? (Same goes for me) [12:47] hi benji, i landed my branch that borrowed from yours and then made some changes. you may see mild conflicts. [12:47] bac: ok, thanks for the heads-up [12:55] gary_poster: will do. got a sec for a quick call re: vacation? [12:57] teknico, sure, come on by guichat [13:11] gary_poster: travel request submitted, waiting for approval and booking code [13:13] teknico, approved [13:14] gary_poster: got the code, thank you [13:14] ehi jujugui! tired of all that javascript? there's a nice and long python diff waiting for you at https://codereview.appspot.com/12022044 but hurry, only two seats left! ;-) [13:23] jujugui, we have one critical task that I'd like someone to focus on: try to figure out why redeploys of the jujugui charm to jujucharms do not successfully inform browsers to clear their caches, and fix it. this is a "drop what you are doing" sort of task unless you are near tying something up, that should be wrapped up tomorrow so we can deploy Wed. [13:24] jujugui, meanwhile, in addition to Nicola's review request (https://codereview.appspot.com/12022044) we also have Kapil's (https://codereview.appspot.com/11966044/) and Huw's (https://codereview.appspot.com/11999043/) [13:25] gary_poster: I'm merging a branch now and could take this up. I'm also annoyed at the cache issue, so there's anger I can harness ;) [13:25] jujugui, finally, another urgent task is that IE 10 unit tests are failing, . [13:25] benji, cool thanks :-) [13:25] I'll put your head in the card benji [13:25] on [13:25] k [13:26] teknico: i'll do your review. have an orange stand up in a few minutes though [13:26] bac, I don't think you do. combined standups start today [13:27] bac: do we? Are we trying the joint standups this week? [13:27] gary_poster: oh, right. cool! [13:27] * bac fixes calendar [13:27] bac: thank you [13:29] hazmat: looking at your proposal, unclear on how to review it :-) (also, I cannot find its card, is there one?) [13:30] teknico, no card, please create in maintenance [13:30] teknico, you should try finding a charm with a bool [13:30] in the congig, make an environment with it, export it [13:31] old version should have issue described by related bug, new should not. extra credit if you try running it through the deployer [13:33] gary_poster: yeah, but even before QAing it, I was wondering how to understand the change by looking at it :-) [13:34] it'd be nice to be able to look at a diff on non-minimized code [13:36] teknico, first line of the compressed file should tell you where to find the diff source, and related yaml parser issue. if it is insufficiently clear, that's something for the review. after you've tried to figure it out, if you are still unclear let me know and I'll try to clarify. I don't want to spoil your blank-slate impressions though: we should be aiming for clarity for someone in your shoes [13:37] gary_poster: oh, right, thanks, I'll try to put together the change [13:37] cool teknico [13:39] gary_poster: when/where is the standup? I don't see it in my google calendar [13:42] adeuring, thank you for reminding me. I just invited orange squad to the meetings. Friday is a different time than the others. we always meet in http://tinyurl.com/guichat . Look good? [13:43] morning all [13:43] morning hatch [13:43] gary_poster: sure [13:43] cool adeuring [13:50] gary_poster: in a few minutes I can start on huw's review [13:51] thanks hatch [13:52] trying to disconnect/reconnect for better network; back in a sec, hopefully [13:52] nope, no real improvement [13:52] :-/ [13:53] it didn't show you dc/rc :) [13:54] it got to 10C last night....summer is over :( [13:56] * hatch straps his snowboard on and sits on the stairs waiting [13:57] heh [13:58] teknico, doing your second review [13:59] gary_poster: thank you [14:01] teknico: your branch looked good. doing a QA deploy now. [14:01] rick_h: saw your q in #yui, any progress? [14:03] hatch: yea, some progress. I'm buried into the ns-routing code atm [14:03] wheee! lucky monday! [14:03] oh alright - keep in mind that the majority of that code is the yui router code with some stuff injected into it [14:03] hatch: yea, but I'm into how we parse/split/combine urls atm [14:03] hatch: so all our own doing :) [14:04] ahh - [14:04] have the bug fixed now but getting #bws-configuration#bws-configuration in the url :) [14:04] hatch: but thanks for checking in [14:05] np - I may be able to lend a hand with the routing stuff if you run into any 'wtf' areas :) [14:05] hatch: just hard to trace. still have double dispatch along with a couple of passes through the code for the orig url vs incoming and such. #needmoarcoffee [14:06] my coffee is watered down for some reason....not impressed - might have to take a walk to a coffee shop [14:10] gary_poster: as far as QA'ing Huw's branch goes - is this supposed to be an incremental branch? I have found a few 'bugs' but they aren't critical [14:11] hatch, not sure it is supposed to be incremental, but we can treat it as such. +1 on getting this landed and then either shooting Huw/list an email with qa issues, or if they are trivial putting something together quickly yourself [14:20] gary_poster: we should have a "sprite all the things" card so that we don't forget in the future :) because all of these new icons are being added un-sprited [14:21] hatch, oh suck. :-/ ok, will do. I assume you'll mention that to Huw as well, so that in future branches he does these with sprites [14:22] done fwiw, in maintenance: high [14:23] sinzui: could you have another at my MP from friday? [14:23] yup will do [14:24] adeuring, I am just giving your my r=me [14:24] sinzui: coll, thanks! [14:25] woooooooooooo! it works! [14:25] adeuring, you can approve the review to merge it. [14:26] sinzui: ack [14:26] orangesquad. I think we need to build a new staging env [14:27] rick_h, yay!!! :-) [14:27] teknico, LveryGTM with comments [14:29] gary_poster: thanks, replying [14:31] adeuring, I would like you to work on https://bugs.launchpad.net/charmworld/+bug/1199780 [14:31] <_mup_> Bug #1199780: search requests errors when a charm has no last_change [14:31] sinzui: I started already ;) [14:31] :) [14:32] adeuring, I am moving the card to the juju-gui board [14:32] sinzui: ok [14:32] gary_poster: reviewed/replied to huw's branch - I can fix the code issues and take a stab at the qa ones to get it landed if you like [14:33] I could also investigate the IE10 failures [14:34] hatch, big +1 on landing Huw's branch with fixes, thank you. IE10: you ok with that, or would you rather someone else look at it? You've certainly paid your dues there :-) [14:35] haha well there looks like there is a CI failure and an intermitant IE failure [14:35] I don't know how long the fixes in huw's branch will take (shouldn't be too long though) so if someone else can pick it up first that'll be fine :) [14:35] I just started the IE10 stuff, but haven't gotten much farther than turning on the VM [14:36] hatch cool, thanks. Makyo, awesome, thanks. [14:41] Mocha desperately needs a stop button. [14:42] heh [14:42] yes [14:45] * rick_h runs make test-debug and ducks [14:48] rick_h: so in your new branch the #'s will be gone? [14:48] from the tabview [14:48] hatch: no [14:49] hatch: they'll be there still since we need them for the charm browser side [14:49] ahh, hmm [14:49] hatch: but the big lesson here is that the tabview doesn't add the #, the Y.App does [14:49] which took me a LONG time to figure out and quit blaming the wrong code [14:49] ohh from the 's? [14:49] hatch: riiiiight [14:49] which is why all of this is a routing issue, not a tabview/browser subapp issue [14:50] see a's are bad [14:50] :P [14:50] oh hush [14:51] lol [14:51] sorry coudln't resist [14:52] rick_h: I'm guessing that has something to do with the pjax stuff [14:52] hatch: well, it's the Y.App wanting to update the url without a reload [14:52] hatch: ah, yea...that might be part of it as well I guess. using hash for content/page id [14:53] yeah - pjax listens to all a's and then updates the url directly without clicking [14:53] without redirecting i mean [14:54] right [14:54] so i'd say first place to start would be monkey patching pjax out of there [14:54] I remember we left it in because it didn't appear to cause us issues [14:54] we were wrong :) [14:54] heh, no. The fix is actually not bad. I've got it all working just needs tests. [14:54] oh cool [14:54] https://code.launchpad.net/~rharding/juju-gui/routing-issues/+merge/177401 for diff [14:55] that and hash comes after QS in a url [14:55] doing it the other way is un-good [14:55] while this is a good fix its really just bandaiding the real problem [14:56] Hmmm, I'm not 100% sold ont hat [14:56] err that [14:56] because that hash shouldn't be added to the url at all [14:56] we don't use pjax [14:56] yes it should! [14:57] we want to generate a unique url so you can link directly to the readme of a charm [14:57] it's part of the UX design. [14:58] ohh, hmm that should probably be /precise/ceph-13/readme then [14:58] at least imho [14:58] that we can debate perhaps, but it's not a different resource to view the readme [14:58] only a small section of content changes [14:58] which is what anchor tags are for [14:59] that's arguable because webapp not website [14:59] I agree it's arguable [15:00] just don't agree with your side :) [15:00] lol [15:00] teknico: i tried to deploy the charm with your changes and got an install error. investigating still. [15:00] this is the url from the inspector though :8888/#bws-readme which is no good [15:01] hatch: agree, and it's an easy fix. [15:01] we just have to catch the click event when deployed via the inspector and halt [15:01] hatch: but only when done through the inspector, so we need an init flag for it [15:01] bac: uhm, I run "make ftest" before proposing without errors... [15:01] sure - so why wouldn't we remove the pjax and then enable it, instead of disable it [15:02] it's a side effect that we just happen to be using in a single place in the app [15:02] hatch: either we disable it and add it for browser or leave it and disable it for inspector. [15:02] either way we have to work around this stuff. [15:02] I'm saying we manually add the hash when we need to [15:02] look how long it took you to find the source and you wrote the code haha [15:03] and none of that is the router code's fault? That is improperly parsed QS and hashes? [15:03] and I didn't write the router code :P [15:04] well when I 5why it I get to pjax [15:04] because the hash is added via pjax. Then we improperly hanled hashes in our routing [15:04] bac, gary_poster: reproposed with changes that address your remarks [15:05] if you remove pjax, we're still handling them wrong in routing. [15:05] that's not how 5why's works [15:05] lol [15:05] I thought 5why was a typo :/ [15:05] I'm not saying we handle them properly (although I don't know what the issue in router is) [15:05] * rick_h hits the web for wtf 5why is [15:05] oh hahaha [15:06] 5why's is a technique for determining the root cause of a problem [15:06] it also exposes issues along the way [15:07] yea, I see. I'll write up my 5why's later after I get these tests in [15:07] on a different angle....should we pass the viewlets container to the render method? [15:08] https://codereview.appspot.com/11999043/diff/1/app/views/viewlets/charm-details.js [15:08] see that as to why I think so... [15:09] hatch: so isn't that the slot name? Or is it not? [15:09] the viewletManager should handle this I'd think when you show/hide a slot [15:09] and not the viewlet itself [15:09] though the way things work 'render' !== 'show' [15:09] :/ [15:11] yeah hmm ok this method should be slightly refactored then [15:11] hatch: +1 [15:11] render should handle generating the content [15:11] should only* [15:11] hatch: yea, I'd look to the viewletManager for controlling the slot visibility imo if possible [15:11] as part of the showSlot() stuff [15:11] yep and it does [15:11] hideSlot [15:12] oh, then why is this necessary? [15:12] * rick_h isn't caught up. [15:12] not sure [15:12] haha [15:12] but agrees with refactor of it [15:12] * hatch looks further [15:13] rick_h: about bug 1199780: How is charm['code_source'] dictionary used? Can we simply drop charm['code_source']['revision'] and charm['code_source']['last_log'] if it is not avaliable, should we set it to None/null? (Note that this can also happen with the data for related charms) [15:13] <_mup_> Bug #1199780: search requests errors when a charm has no last_change [15:14] adeuring: sec, looking to be sure [15:15] adeuring: so we'd have to refactor to add sanity checks it exists. We don't have a 'if model.get('code_source') as it seemed you'd have to have source to have a branch/pushed/etc. [15:16] rick_h: right, but we can have charms whose LP brnach is deleted. And then we have this situation [15:16] rick_h: if you're going to leave pjax in can you put a comment in the tabview code which explains where the hashes are coming from [15:16] I figure that's the logical place someone might look in the future [15:16] adeuring: right, so we'd have to update the front end to handle it being option. Right now it doesn't. [15:16] hatch: sure [15:17] rick_h: OK... What about js using the text "not available"? [15:17] adeuring: it's expecting a nested object. So no matter what it's changed to we need to adjust [15:17] adeuring: if we don't have it then I'd say we leave the value out. [15:17] jujugui: IE10-able folks, need some reviews/test runs for a fix https://codereview.appspot.com/12033043/ [15:17] then we can look if we can provide a decent default that prevents too much work [15:17] rick_h: OK [15:20] Makyo: i can review. can run tests with some pain if really necessary. [15:20] Makyo: I can run the tests [15:20] .....and are you serious? this was the cause....ugh IE [15:20] bac, how about hatch runs the tests, and you sanity check? :) [15:21] sure [15:21] Makyo: is the flags fix because window.flags was undefined at that part? [15:21] hatch, yeah, for running that suite in isolation. [15:22] for folks working on charmworld.. with elastic search.. came across this library this weekend, made working with elasticsearch much more pleasant and pythonic. http://elasticutils.readthedocs.org/en/latest/ [15:23] builds on top of the py connector library we're already using. [15:24] Makyo: I wonder if we should have a follow-up branch which adds window.flags to every suite and then cleans it up after every suite [15:24] hatch, I put that in the MP. [15:25] oh.... /me learns to read [15:25] Makyo: nm my comment about the same. [15:25] * bac learns to read too [15:26] Makyo: sorry, windows 8 keeps crashing [15:26] I just did an update so,.....ya know [15:26] hatch, I feel your pain, yeah [15:30] hey Windows....nice of you to boot up [15:31] I used to be afraid to do Linux updates, now I'm afraid to do Windows ones, my how the times have changed [15:32] make test-server [15:32] blarg [15:36] jcsackett, I am looking at the charmworld board. I think card 3.1 is done and card 3.2 is coding and already represented on the juju gui board. [15:37] sinzui: you're right; i've moved 3.1 to the done-done. [15:37] sinzui: do you want us updating this board as well? [15:42] jcsackett, thank. you. I am dismantling the board. moving work to jujugui when it isn't already there [15:43] Makyo: lgtm'd [15:44] hatch, thanks [15:45] sinzui: dig. [15:48] how long after I feature/unfeature should it show up on the site? [15:49] jcastro: pretty quick. We might hit a cache on the production site but the 'change' is instant as far as the data is concerned [15:49] jcastro, I think the squid cache is the factor in play [15:50] ok [15:53] jujugui, orangesquad, call in 8. please update kanban now for a quick overview call. [15:53] 7, that is :-P [15:53] woah....new board [15:53] gary's been workin [15:58] heh [15:59] jujugui orangesquad call in 2 [15:59] in guichat [16:00] jujugui I can haz critical review please? https://codereview.appspot.com/12036043 [16:01] Makyo, you here? [16:01] rick_h: critical as in negative? :P [16:01] critical as it bugs are marked critical [16:02] yup I'll take one [16:05] teknico: I heard that there is no Alfredo sauce in Italy? And that you don't put chicken in pasta.....true? [16:07] jcsackett: can you link the branch? I tried to click the bug in the card to get to the review but failed [16:07] jcsackett: where is your mp for that card? [16:07] rick_h: yeah, i haven't been good about linking to bugs. [16:07] :) [16:08] rick_h: one sec, i'll go find it. [16:08] jcsackett: ty [16:08] rick_h: https://codereview.appspot.com/12031043/ [16:10] hatch: what?!? and, what?!? [16:11] a friend of a friend has a friend from Italy and they claim that there is no 'Alfredo' sauce and that you don't put chicken in your pasta [16:11] 10 minutes on a call with that many people. nice! [16:11] we pro! [16:11] bcsaller: rick_h making hangout [16:13] jcsackett: done, thx [16:24] that's so odd - the camera was reversed during that call compared to the guichat ones [16:33] hatch: I know of no "Alfredo" sauce and of no recipe for putting chicken in pasta [16:34] crazystuffs [16:34] here - Italian pasta has Alfredo sauce and chicken lol [16:35] sinzui: gary_poster any preference on next task priority? Pick a bug any bug? [16:35] hatch: I'm not a food tradition purist, and I've been eating no pasta for years now, so I'm not offended [16:40] rick_h: issue with the QA [16:40] when I follow https://bugs.launchpad.net/juju-gui/+bug/1205468 [16:40] <_mup_> Bug #1205468: fullscreen charmbrowser tabs are broken

[16:40] and then hit the back button [16:41] it stays in the same page [16:41] because it puts a history entry for the hash [16:41] ah crap...double dispatch fml [16:41] oh wait, back just removes the hash? [16:41] yeah [16:41] trying on serviceInspector which has no double dispatch [16:42] same [16:42] right, that's kind of 'working as intended' because it does dispatch when you click a tab [16:42] pjax must also add a browser entry [16:42] * rick_h is thinking [16:42] so when it's removed it should dispatch then and go back to summary [16:43] hatch: so I'd propose this is a low priority follow up bug to worry about the back button history atm. It fixes 4 bugs in the system currently. [16:44] one critical and one high [16:44] I'd tend to agree [16:44] hatch: because fixnig that isn't easy as the hash is part of the app 'state' and the state comes from the App dispatching. The fix is to not 'dispatch' but then we dont' get our app state change atm [16:45] so it'll take some thought/testing [16:46] and jcsackett :P [16:46] wait,what? [16:46] jcsackett: your comment in your LGTM [16:47] aaaah. [16:47] i was suddenly concerned you were throwing work at me, and i couldn't figure out how. e.g. "it'll take some thought, testing, and jcsackett". [16:47] lol, +1! [16:48] this is not me endorsing the plan that work should be thrown at me. :-P [16:50] rick_h: found a real bug [16:50] shall I reply with repo steps on the commit? [16:50] stop doing that [16:50] hatch: yes please [16:51] just finding the quickest repro [16:52] ok, well food fetching while you do that and will look back [16:54] ok replied [16:57] hatch: thanks, easy fix. Appreciate it [17:00] excellent [17:05] * hatch wishes for a parent css selector [17:06] bcsaller: so...any ideas for anything better than this._slots[viewlet.slot].container.ancestor('.slot') [17:07] i'd add a slot class to the slot [17:07] or could make it an attribute [17:07] which would probably be nicer [17:07] but container.ancestor() still sucks [17:07] yeah, thats not great [17:08] I could go container.get('parentElement') [17:08] but that isn't very robust [17:08] chat real quick? [17:09] sure [17:09] in guichat [17:21] hey rick_h, you still looking for something to do? I've got ideas if so [17:21] sorry was lunching [17:22] gary_poster: I will be. Working on a last bug/issue in getting the other branch landed [17:26] rick_h: so in this branch of huw's when I close the left panel after viewing some of the tabs it leaves the #'s in the url [17:26] rick_h, cool. Looking for choices. First one is this. Trivial but important: https://bugs.launchpad.net/juju-gui/+bug/1202636 [17:26] <_mup_> Bug #1202636: Charm Details Page Under Providers Change Openstack to HP Cloud [17:27] rick_h: interestingly enough, when I realod the page, that hash gets stripped lol [17:27] hatch: yea, it's not what I thought it was. [17:27] hatch: when I dupe it works when I click the juju, but then fails when I change viewmode. There's some hidden bit somewhere I'm missing [17:28] re the QA bug? [17:28] hatch: yea [17:29] hmm I can't really be of any help there [17:29] hatch: yea, all good. [17:29] thought that's what you were talking about [17:29] I haven't spent much time in your homebrew routing code ;) [17:29] oh no sorry [17:29] there's an existing bug for chrome that clears the url out when manually hitting a url :( [17:29] that's not part of this fix [17:29] ohhhhhh [17:29] ok that must be waht I'm hitting now [17:30] no prob doesn't effect this branch negatively [17:30] the status bar plays tricks on my eyes [17:30] the red/green bars look different heights [17:30] even though I know they can't be :) [17:30] hatch https://launchpad.net/bugs/1199392 [17:30] <_mup_> Bug #1199392: can't view fullscreen charm details from url

[17:31] we have too many router technologies acting in one place [17:31] :) [17:33] sinzui, I'd like to triage https://bugs.launchpad.net/charmworld/+bug/1206158 as high. We're having multiple questions about this on sales calls, so this would be a (presumably easy?) thing to add for good sales win. Is that ok with you? Am I badly underestimating the effort involved? I hoped it would be a day or two max for one engineer. [17:33] <_mup_> Bug #1206158: On manage.jujucharms.com show download breakdown by last 30 days, last week [17:33] gary_poster, okay [17:33] sinzui, what do you think about effort? [17:34] gary_poster, a single branch that adds the model properties and and template presentation [17:37] sinzui, so day or two max? [17:37] yep [17:38] taking off for lunch, call cell is ya need me [17:45] gary_poster: so in reading that bug does that maen that we're going to start testing on 'openstack' somewhere? Or just not going to have that visible? [17:45] rick_h, we're not going to have it visible. apparently, according to mramm and others, openstack installations are too different [17:45] so in the futrue [17:46] we may have results for HPCloud [17:46] and some other OpenStack provider [17:46] k [17:46] and we will not be at all surprised when they have different results [18:03] hatch, when you are back, no rush, but is this inspector card/task no longer valid? "Add mutation observer to viewlet to rebind viewlets when updated out of band" [18:11] gary_poster: That one was from me, its still an option but its an optimization not a requirement [18:11] hatch: also when you're back, branch updated for a second run through QA. [18:12] bcsaller, ok cool, thanks. clean uptask, then? [18:12] hatch: think I've caught it all there. One other QA point found during this was to sidebar, open a charm, click a tab, then minimize the browser, and bring it back up. [18:12] yeah, pretty much [18:13] and now my head hurts...ugh routing/state FML [18:20] is anyone else having problems building (NPM downloads failing) [18:22] benji, I had to make clean; didn't look closely [18:27] rick_h, playing around with your branch. lots of fixes, cool. this one is not supposed the query string related bugs, right? for instance, http://localhost:8888/fullscreen/search/precise/ceph-13/?series=precise&text=ceph&type=approved#bws-readme does not load properly. I already saw you say that back arrow bugs were for later, I think, which I agree with. [18:27] "this one is not supposed the query string related bugs, right" -> this branch is not supposed to fix the query string related bugs, right? [18:29] rick_h, those are the only bugs I can find. Way better. thank you! [18:29] gary_poster: looking [18:29] gary_poster: right, it doesn't fix the chrome removing the query string bug [18:30] gary_poster: I did not run into that cause in my list. [18:30] oh hmm, that happened to me on FF as well. Thought it was a chrome-only bug. [18:30] gary_poster: but yea, that's still out there on our board. [18:30] rick_h, cool, thanks! [18:31] gary_poster: but it should hit 4 bugs currently in the tracker. [18:31] rick_h, I saw, sweet :-) maybe even more [18:32] the back button I think we could chat on. I keep going back/forth on how that 'should' work since we're treating tabs as unique urls, vs 'back button, ignore tabs' [18:36] I see [18:37] rick_h: qs is removed in both ff and chrome,but in ff the correct view is shown [18:37] so it's partially an FF bug. [18:37] jcsackett: ah, in nightly it went back to search just like chrome [18:37] jcsackett: in FF nightly that is. [18:37] rick_h: interesting. [18:37] rick_h: i'm not sure what that suggests,but it certainly suggests something. :P [18:37] jcsackett: but yea, that's a routing/double dispatch bug I think. [18:38] jcsackett: yea, suggests there's work to be done to fix that bug...for someone... [18:38] if it's still around when i get done with charmstore-apocalypse i'll be happy to look. [18:41] back [18:42] rick_h: ok will re-qa [18:42] bcsaller, hey. did you see my September/London flight request? please get that done today (which means, you propose via form, I approve, you go to travel provider, they get you flight, and you enter details on spreadsheet) [18:43] gary_poster: yeah, I saw it, I have it on my calendar to do it today as well. [18:44] cool bcsaller thanks [18:49] hazmat what I was trying to say in last reply to review is that I'm happy for you to give this to us to land if you want. we appreciate very much the work you've already done [18:51] gary_poster: CI is broken....can't deploy charm [18:51] hatch, may be because canonistack is broken. :-/ sinzui did you say that you can't get a juju env on canonistack any more? [18:53]