[00:06] huwshimi: ok done [00:06] ;) [00:06] hatch uistage isn't broken. zoom out. silly kids [00:07] huwshimi: http://jsbin.com/itaqem/2/edit click the box :) [00:07] gary_poster: lol Uncaught TypeError: Cannot read property 'browser_enabled' of undefined is what I see [00:09] hatch, wfm. hard reload? [00:10] hatch: Nice! Now you just need to make it close when you click outside of it, make it work with the keyboard and build a YUI widget to handle the creation etc ;) [00:10] http://uistage.jujucharms.com:8080/ and http://uistage.jujucharms.com:8080/:flags:/browser_enabled/ [00:10] huwshimi: umm....yeah...about that.... :P [00:10] hehe [00:10] gary_poster: ohh here is what happened Application Cache Error event: Failed to parse manifest http://uistage.jujucharms.com:8080/juju-ui/assets/manifest.appcache [00:11] we didn't properly set it to clear the appcache before we updated it [00:11] so the index.html is still cached on my system [00:11] hatch, ah. so yeah, clear with chrome://appcache-internals/ and prob fine [00:12] this brings up something curious to me....what is the proper way to remove appcache on a live site [00:12] hmm [00:14] ok gota run, will be back later to try and make 'click outside' to work with css :P [00:15] ^ was a joke :) [00:15] :-) [00:18] :) === rogpeppe2 is now known as rogpeppe [12:28] hey bac, do the recent jenkins failures look like they have anything to do with images? [12:29] doesn't look like it to me :-/ [12:29] looking [12:30] looks like random stupidity and a reminder that we should switch to a more production-centric system [12:35] gary_poster: yeah, it looks like a bunch of hooha to me. CI should not require so much finger crossing and chicken sacrificing. [12:35] yeah :-( [12:36] jujugui: so yesterday afternoon my ISP dudes couldn't find parking so they just left. now they are circling the block again. connectivity may (or may not) be flaky soon. [12:37] I guess the only things stopping us moving to ec2 are (1) corporate account, or me taking one for the team; and (2) opening up the firewall from the test center to ec2. [12:37] The first one is the one that slows me down the most. :-/ [12:37] gary_poster: heck, i'll volunteer my AMEX. think of the points! [12:38] no, really [12:38] :-) [12:38] but then anyone who can log into the machine can get your ec2 creds [12:38] bac: that's hilarious [12:40] oh [12:40] so, carlos is here to replace my radio. [12:40] bac, wow.. thats silly. [12:40] er? [12:40] bac, isp dudes leaving, cause they can't find parking.. [12:40] oh, yeah, well... [12:41] afternoon cerveza sounds more apt [12:41] i would've gladly given them a cerveza [13:01] gary_poster: I'm in guichat [13:02] benji, darn, ok [13:02] coming [13:16] * bac o/ [13:53] morning [13:55] looks like canonistack errors on jenkins [14:02] hi abentley, do you have a moment to answer some bzr lightweight checkout setup questions? [14:02] bac: Sure. [14:04] orangesquad: should we move the green categories card into the coding lane? I think it is now bound on rick's card as the last thing that needs to land for it to be done, yeah? [14:04] or does it just sit and go to done when we're done with other cards? [14:04] jcsackett: seems strange to have a green card in coding without a face/etc on it [14:05] rick_h_: don't disagree, but having it in "Next" when all activity on it is going seems weird too. [14:11] rick_h_: : in the case of a review with lbox -req (i.e. a prereq branch) does lbox submit properly merge the branch, or does it just apply the patch set? [14:11] jcsackett: oh hmm, I think since it's a req the other one needs to land first? It just builds the right diff/MP [14:12] jcsackett: I don't think you can submit both in one swoop. [14:12] jcsackett: but I could be wrong, not tried it tbh [14:16] rick_h_: yeah, i don't really think now's the time to experiment, since it'll break the build if it doesn't work. :-P [14:28] benji, gustavo suggested that we use full ip address, not merely port. Could you take a look at the only non-resolved comment on https://docs.google.com/a/canonical.com/document/d/14W14YAPYdivL9nF7Wvrb6kHX0xrsxV0fAEMfXNt7L3Q/edit# and see if you can see where to incorporate the idea, or the question (since I don't really understand where he is coming from) in the new doc? [14:30] gary_poster: I can't claim to understand the comment containing the explanation. [14:31] benji, you and me both then. Maybe add a comment in the new doc highlighting the fact that we still need to resolve this with Gustavo? [14:35] orangesquad. I was/am leaving the feature cards in the next lane because they remind us that the task/defect/improvement cards should relate to the features. I move the features to acceptance when I think we have delivered all the work. This is a bit risky since acceptance can find issues that will go into the next lane [14:36] bac: y u no put python-charmhelpers in juju-gui-charmers ppa? charm deploy sad [14:36] I don't think the story1 and story2 makes sense if we continue to let the feature cards define the theme of work [14:37] BradCrittenden: ^^ [14:39] gary_poster: so I'm reading through this doc from luca__ and I'm wondering how I give feedback, is there some way to comment on these points? [14:40] hatch: which parts do you want to comment on? [14:40] teknico1: sorry, the charm-tools in stable is from juju/pkgs and does not include the python-charmhelpers. i am currently copying it from juju-gui-charmers/devel === BradCrittenden is now known as bac [14:40] luca__: so far Route A and Route B - I feel that we should have some kind of 'setting up' indicator to indicate that it's being set up instead of just hanging around as a ghost for 15 minutes [14:41] hatch: you can add that to the comments section at the bottom of the post [14:41] hatch: if you click into it [14:41] ahh ok [14:41] * hatch votes to use google docs next time :P [14:41] teknico1: unfortunately it'll take at least an hour [14:42] luca__: also """Is there ever a case where you just want to add charms without reading the details?""" yes I do with npm all the time [14:42] hatch: https://sites.google.com/a/canonical.com/juju-meeting-notes/scenarios-1/buildinganenvironment [14:42] bac: the charm tries to install the "python-charmhelpers" package, it's not going to be happy if it does not find it in the ppa [14:42] hatch: great :) [14:42] bac: or am I missing something? [14:43] luca__: :) I coudln't comment on that section so I'll leave that to you [14:43] teknico1: i understand. that's why i just told you my actions to rectify the problem. [14:43] hatch: I've just changed the user needs to a separate post so it can be commented on [14:43] oh ok awesome [14:43] I've never heard of sites.google.com before [14:44] bac: then it's me the one who's not understanding :-) what do you mean by "the charm-tools [...] does not include the python-charmhelpers"? [14:44] teknico1: by "i am currently copying it" i mean on launchpad i have requested the package be copied from the devel PPA to the stable PPA [14:45] bac: but I don't see it in the devel PPA either... [14:45] that is, here: https://launchpad.net/~juju-gui-charmers/+archive/devel [14:45] teknico1: it's weird. you have to click on 'details' and then look at charm tools. it'll show you the packages withing that, uh, package [14:45] the charm-tools source package builds three different installation packages. python-charmhelpers is one of them [14:46] last week the build recipe was changed to *not* build python-charmhelpers for ppa:juju/pkgs [14:46] bac: oh my gosh, I see it now, thank you. way to go hiding stuff :-) [14:46] teknico1: that was quite a mystery to unravel last week [14:47] teknico1: anyway, i screwed up i getting the charm-tools from the wrong spot for stable [14:47] bac: I wonder why that happened... (not building python-charmhelpers) [14:47] teknico1: marco changed his debian/control file to explicitly delete it. unsure why. [14:48] teknico1: if you need to deploy a charm right now you can set repository-location=ppa:juju-gui-charmers/devel [14:48] otherwise the copy from one ppa to stable will complete in an hour or so [14:48] bac: ok, thanks, at least I know what the problem is :-) [14:49] * bac regrets having lost his PPA administration super powers [14:49] * bac wonders if that is the cause of jenkins failures [14:49] it would seem likely [14:49] (bac: did you ask around? someone may have found them ;-) ) [14:53] bac, maybe follow up with marco about python-charmhelpers? does this mean that he broke gui deployments? [14:55] gary_poster: i talked to him last week but decided to use my own build recipe for our devel and stable ppa. while his recipe is no longer building p-ch it was not deleted from the juju/pkgs repo, though it is not visible via the UI. [14:55] ok thanks bac [14:56] bac, maybe clarify with him that deleting it Would Be Bad? Or is it not possible to delete? [14:56] i erred by testing against our new devel ppa and not stable [14:56] If it is possible to delete, maybe that would be an answer to our "what do you do if you have a package in devel that doesn't work out" question from yesterday? [14:56] gary_poster: it is not possible to delete unless he were to delete the entire charm-tools source package from that ppa, which he has no reason to do [14:56] ok [14:56] thanks [14:57] gary_poster: what is the risk you see now that our charm does not rely on juju/pkgs anymore? [14:58] bac, old charms only. but until you push the new charm with dependencies on the new ppa to ~juju-gui-charmers (did you do that?), the main charm is the old charm [15:00] gary_poster: that was done: https://code.launchpad.net/~juju-gui-charmers/charms/precise/juju-gui/trunk [15:00] cool bac, thanks. [15:00] bac, then I guess I'm not too worried. Paranoid, but not worried. :-P [15:00] but i think it was premature and not following our guidelines [15:00] because it was released prior to full testing? [15:01] it should gone to the old spot but due to my broken setup i pushed to lp:charms/juju-gui. (so my new blog post is wrong) [15:01] ah ok [15:01] yes, the official charm is currently broken until the PPA publishing for stable is complete [15:02] :-/ k [15:02] soory [15:02] you're on it [15:02] you are fixing it [15:03] I suppose we should have another "what can we do so this doesn't happen again" conversation, in which we all say "use tarmac" [15:03] bac: while you're at it, please also fix my head and my wall ;-) [15:04] want to ask matsubara where we are on that but have not seen him [15:05] hey luca__ , could you copy over the things you want us to comment on to google doc instead? [15:05] I'd like to use the google doc line by line thing [15:05] gary_poster: sure, I'll do that once I get out of this meeting [15:05] thank you! [15:07] jujugui: to avoid the charm problem i just had, everyone should verify lbox is merging to lp:~juju-gui/charms/precise/juju-gui/trunk not lp:charms/juju-gui [15:07] bac, I suggest we make a .lbox file in charm , like the one in gui [15:08] instead of [15:08] this blog post has bzr setup instructions if using lightweight checkouts: http://jujugui.wordpress.com/2013/06/04/bazaar-lightweight-checkouts-and-charm-development/ [15:08] propose -cr -for lp:juju-gui [15:08] we would simply have [15:08] propose -cr -for lp:~juju-gui/charms/precise/juju-gui/trun [15:08] k [15:08] and then anyone checking out the branch would have the proper lbox behavior immediately [15:08] AIUI at least [15:14] gary_poster: ok, i'll add that now. w [15:14] thanks bac [15:16] gary_poster: luca and i just had the meeting with charline [15:17] alejandraobregon, cool! [15:17] call? [15:17] wrote up comments and suggestions beneath your current questions [15:17] https://docs.google.com/a/canonical.com/document/d/1bSR1nfSsZs4XQoTv1IZu-so5FM-_R4JsvRsW3IAZyag/edit [15:17] bcsaller: are you in yet? [15:17] looking [15:21] hatch: I've added google docs to the posts [15:21] gary_poster: ^ [15:21] :) You didn't have to [15:21] but thanks! [15:21] :) [15:22] I'll move my comments there [15:22] hatch: I am now [15:22] I am writing the viewlet too if that's alright - I can't really do much without it [15:23] alejandraobregon, really fantastic, big thanks to you, Luca, and Charline. I'll make a few small revisions, and then send it to the lawyer. [15:23] * gary_poster needs to step away [15:23] back in a few [15:23] hatch: sounds good [15:23] gary_poster: no wories [15:23] worries [15:27] bcsaller: https://gist.github.com/hatched/2463d15dd51d12e04258 got a sec for a guichat? [15:40] gist updated [15:45] congrats on landing core branch, frankban :-) [15:46] hatch: I can hop in chat now [15:46] gary_poster: \o/ [15:46] :-) [15:47] bcsaller: ok there [15:51] luca__: ping, any chance we can get these category icons in 160x160 for the full charm details icon view please? https://drive.google.com/a/canonical.com/#folders/0B1IM--9A1RkTMndkVG8zeHFlVmM [15:52] rick_h_: I have no idea, let me ask the guy who made them :P [15:52] luca__: thanks, appreciate it [15:52] luca__: I had three sizes and ran with it and now see that we changed the sizes so I've got 2/3 I need to complete the category work [15:55] rick_h_: He's going to make the new sizes, hopefully won't take too long. I'll add them to the drive and email you when you they are up. [15:55] luca__: appreciate it [15:56] jujugui kanban now call in 4 [15:59] jujugui call in 1 [16:00] orangesquad: Could you please review https://code.launchpad.net/~abentley/charmworld/remove-api-0/+merge/167324 ? [16:00] * sinzui does [16:04] rick_h_: I've just uploaded the 160x160px icons to that folder: https://drive.google.com/a/canonical.com/#folders/0B1IM--9A1RkTMndkVG8zeHFlVmM [16:05] abentley, is test_reindexed_no_client_charms testing that an error is not raised? There is no asserts. [16:06] sinzui: Yes. [16:06] sinzui: Should I add asserts or a comment? [16:07] A comment would be helpful [16:09] sinzui: I've added a similar comment to test_create_replacement_misssing [16:10] thank you abentley. r=me. [16:10] sinzui: Thanks. [16:17] bcsaller: reasonable to assume that we could have something like https://gist.github.com/hatched/3876ff1fc8ca1bff5bb0 for the view-container? [16:17] jcastro: ping [16:19] julianwa: pong! [16:21] jcastro: hi, I had a problem to login juju-gui [16:21] jcastro: juju gui is deployed by charms in charm store, for python juju stable version [16:22] * gary_poster listens [16:22] but failed to login with admin-secret key [16:22] 2013-06-04 11:56:22,399: juju.rapi.ws@INFO: Invoking method > with {u'password': u'FDU2K9N7bsJZkfh6s6:829z85ytjayxqYKLPU:Tvmtbwcku9FsSyjuVWzPxmAEBUaBjA9E', u'user': u'admin'} [16:22] 2013-06-04 11:56:24,092: juju.rapi.ws@INFO: Returning results {'failure': 'Traceback (most recent call last):\nFailure: zookeeper.NoAuthException: not authenticated\n', 'request_id': 1, u'user': u'admin', 'err': True, 'op': u'login', u'password': u'FDU2K9N7bsJZkfh6s6:829z85ytjayxqYKLPU:Tvmtbwcku9FsSyjuVWzPxmAEBUaBjA9E', 'log': [('error', 'Invalid credentials')]} [16:23] hatch: wouldn't it be a loop around viewlets where you put the container class use for switching around them automatically? [16:24] jcastro gary_poster, any ideas? [16:24] julianwa, do you have multiple environments in your ~/.juju/environments.yaml file? We haven't encountered any problems like that ourselves, so my first guess woudl be that you used the wrong password for the active environment [16:25] back in a sec [16:25] gary_poster: thanks, I will check :) [16:28] jujugui: would like one review of the code backport just for sanity check: https://codereview.appspot.com/9975046/ === BradCrittenden is now known as bac [16:28] bcsaller: updated gist https://gist.github.com/hatched/3876ff1fc8ca1bff5bb0 [16:29] the specifics we can work out after design - but the concept should be similar [16:35] jujugui: python-charmhelpers is now properly in stable PPA: https://launchpad.net/~juju-gui-charmers/+archive/stable/+build/4641114 [16:35] thank you bac [16:35] bac: thank you [16:37] gary_poster, teknico1: can either of you do the quick review i requested? i'd like to get this landed to finish cleaning up the mess. [16:38] I will if teknico1 can't. Otherwise trying to finish something up. teknico1, are you already at EoD? [16:40] gary_poster: I'll squeeze it in, it seems simple enough [16:40] thanks teknico1 [16:41] bac: it looks exactly the same as yesterday's branch, isn't it? [16:42] :) [16:42] i did say sanity check [16:42] that's the nature of a backport [16:42] bac: done [16:42] thanks === deryck is now known as deryck[lunch] [16:52] jcsackett: sinzui can I get some review <3 please? https://codereview.appspot.com/9975047 note that it's blocked on kanban until we get the 160px category images but that doesn't effect the code at all. [16:52] hatch: ^^ as well if you have a chance [16:55] alright [16:56] hatch: ignore all the svg files :P [16:56] #DiffsAreSmallerThanTheyAppearInMirror [16:57] lol [17:04] gary_poster, bcsaller - would like to start "zoom to a service if any on first delta". It sounds reasonable, but I'm game for a pre-imp call if either of you want. [17:04] Makyo, don't need one from me :-) [17:04] Makyo: not unless you want one [17:04] bac, you done with your critical card in Maintenance? [17:05] gary_poster, bcsaller - Alright. Will start for now. [17:05] sinzui: ping, got a sec? [17:07] rick_h_: review done [17:08] hatch: cool, I'm changing it up right now :) [17:09] gary_poster: moved [17:09] hatch: chat? [17:09] thanks bac [17:09] sure [17:21] rick_h_, I do. Sorry for the delay. I had a late lunch to argue with my ISP [17:24] sinzui: invite inbound [17:28] I am having an issue generating a new template file [17:28] no matter what I do it doesnt' add the new .handlebars file into the rollup [17:28] I tried make clean and then make again [17:28] but no luck [17:29] hatch: so when you run that make command I give you do you get any error output? [17:30] hatch: if there's a syntax error in the template you'll get a one line in the make comment output that tells you 'something went wrong' and is partially helpful [17:30] i just rm'd the whole dir's and then remade [17:30] lol [17:30] well, copy the output next time you make build-shared/juju-ui/templates.js [17:30] I'll bet you have a mismatched tag or missing # or something [17:31] * rick_h_ did that more than a few times the last couple of days [17:31] rick_h_: yup it was a syntax error [17:31] :/ [17:31] argz [17:32] hatch: party on [17:36] ok, gary_poster, I'm out of time (plus some). Do you want to review the draft together or just make changes and annonce? [17:36] benji, I need to actually run and do something else. Let's have lunch and then reconvene? I'm not too worried about it since I kind of announced it already via the calendar announcement. [17:37] does that sound ok to you? [17:37] k [17:37] cool, thank you [17:46] rick_h_: looking now. === deryck[lunch] is now known as deryck [17:50] rick_h_: nm, just saw curtis's comments hit. [17:59] jcsackett: yea thanks [17:59] going to grab some lunch [18:04] sinzui: It turns out my migration was bogus (it only updated a single document, and my tests only tested a single document). Should I do a new migration or bugfix the existing one? [18:04] abentley, I think a new one. [18:05] abentley, Wouldn't a bug fix be more risky for a production deployment? [18:06] sinzui: We'd have to manually futz with staging to force migration 5 to run again. [18:06] well, that is what I hoped a new migration would avoid. [18:07] abentley, how is a bug fix different? [18:08] sinzui: I was describing a bugfix. A bug fix would update migration 5. If we introduce migration 6, we still have to have migration 5, though I can change it to a no-op. [18:09] I like a new migration. making migration 5 a no-op is fine since we reserve the right to choose which migrations are no longer supported [18:25] sinzui: https://code.launchpad.net/~abentley/charmworld/really-remove-icon/+merge/167364 [18:25] * sinzui looks [18:28] ah, like the RE g option when doing substitutions. r=me abentley [18:28] sinzui: Thanks. [18:33] sinzui: I would like it if we could remove migrations that are no longer supported, but the current code requires an unbroken numerical sequence. [18:34] abentley, I think we can live with dummies that just document the historical intent. [18:40] jcsackett: ping, did you want the background on the bar issue stuff at all or are you ok? [18:45] bcsaller: around? [18:45] hatch: yes [18:45] have a moment for a screenshare? [18:46] sinzui: tarmac appears to have gone on strike, and I can't find matsubara. [18:48] abentley, I see === _mup___ is now known as _mup_ [18:51] abentley, the only option I can think of at this time is to test and merge ourselves, then call upgrade charm [18:52] sinzui: I'm not rushed with this one. [18:57] * benji_ cleans hundreds of messages about PyPI projects out of his inbox. [18:57] * benji_ wonders when he grew a tail. === benji_ is now known as benji [18:58] * benji feels much better now. === _mup___ is now known as _mup_ [19:40] new yui bugfix release http://www.yuiblog.com/blog/2013/06/04/yui-3-10-2-released [19:40] we should look into upgrading to 3.10 at some point :) [19:42] jcastro: Great idea on the app infrastructure :) [19:44] gary_poster: i've learned more about google analytics and it doesn't look like it is going to work for us. they expect a single tracking code to be tied to a specific web site, which is not our use case. [19:44] bac, oh :-( [19:44] we want to track the use of an app that'll appear in lots of places [19:45] y [19:45] gary_poster: for instance, they won't even begin gathering data until their robot finds the tracking code at the specficied url ... which cannot include a custom port, btw [19:45] so i cannot even test it on uistage [19:46] bac, any other options you found for our use case? [19:46] not yet [19:48] crazy egg has the same setup [19:48] ok :-/ [19:49] I see for-fee ones that would seem to work [19:52] or maybe we have to set up our own app for this? in which case this is a stretch goal [19:52] that would be a shame [19:54] bah, benji, neither of us pinged the other and I've been doing other things. looking at your doc [19:55] k [19:59] gary_poster: you use sublime text 3 right? Are there any reasons to upgrade? [19:59] hatch, no, for me currently only anti-reasons. hopefully better later [20:00] alrighty :) [20:01] I use this as my icon http://1024jp.deviantart.com/art/Sublime-Text-Icon-373837347 [20:01] thought you might like [20:17] :-) cool [20:19] gary_poster: it *may* be possible to use ga if we get it bootstrapped with a fixed location [20:19] bac, how would that work? [20:20] gary_poster: it looks like a code that their robots have verified at a fixed location can be used from other sites if you turn off the url filtering [20:20] oh [20:20] well, we have jujucharms.com.. [20:21] yeah, that's true. we also have uistage.jujucharms.com if we were on port 80 [20:21] we have that bac [20:21] http://uistage.jujucharms.com/ [20:22] oh, we do don't we? [20:22] i've been going to 8080 out of habit and bookmarks then [20:22] me too [20:22] benji: nm [20:51] I'm about to introduce a Y3K bug into the GUI. I've made a note to myself to fix it in 2999. [20:53] you better! [20:56] benji, +1 on vitals proposal [20:56] made some changes [20:56] cool [20:56] benji, please review my last two comments/changes? [20:56] * benji looks [21:33] BradCrittenden, are you optimistic or pessimistic about GA working out for us at this point? [21:42] gary_poster: hey [21:44] typeof null === "object" just bit me ....ouch it stings [21:56] alejandraobregon, hey [22:12] TDD works pretty damn well once you get the hang of it [22:33] has anyone ever had any issues with setStyle on elements in mocha? [22:34] for some reason it's throwing on a setStyle claming the method doesn't exist when it clearly does (from stepping through and manually calling it) [22:37] hmm it thinks nothing exists... [22:41] blah [22:50] figured it out - it was reporting the wrong line :/ [22:51] using a source map and pretty print at the same time does break the line numbers [22:52] now to figure out why THIS line doesn't work [22:52] heh [22:57] well I'm baffled [22:59] well I really wish I could solve this so I could commit this code and be done [23:04] I know the feeling [23:07] well I'm not getting anywhere it's pushed up lp [23:08] have a good night