=== alejandraobregon is now known as alejandra-lunch [12:25] teknico, frankban: reviews done [12:25] bac, thanks! [12:34] teknico: the last time gary asked me to be more specific about parameters we discovered the API was not spec'ed properly and it was an opportunity to fix it. so good things come out of the digging to write those docs, even if they are a pain sometimes. [12:35] bac, I totally agree with that too :-) [12:38] ok, lunch time! [13:24] teknico, land with changes. thank you [13:25] bac, could you review benji's branch too? It would be awesome to get that taken care of [13:26] gary_poster: ok [13:26] ty [13:26] gary_poster and bac: Ben has already reviewed it. [13:27] benji, oh! kanban board did not have tag. thanks [13:27] gary_poster: I'll add the tag. [13:27] cool, thanks [13:28] That's another good job for a script. [13:32] benji: i don't see been on the RV as a reviewer. what's going on? [13:34] bac: is this the one you are looking at: https://codereview.appspot.com/7129049/ ? [13:34] benji: nope. why would i look there and how would i know to? [13:35] bac: I bet you aren't; apparently when you re-propose it generated a new review and the old review is still linked to the board [13:35] That is something that would be good to fix. [13:35] ugh [13:35] Otherwise we risk people reviewing non-up-to-date branches. [13:37] benji: yeah, so a bug with two linked merge proposals is a problem [13:37] i'll make a slack card [13:42] benji: and i don't even see a way to mark a rietveld as abandoned or obsolete [13:42] hmm [13:43] bac: under "Edit Issue" there is a "Closed" checkbox [13:43] bac: thanks for the review [13:44] benji: oh, on https://codereview.appspot.com/7069068 i have "Can't edit" instead b/c i'm not the owner [13:45] that might make editing it via script harder; unless we can have an uber user [13:46] re: docs, as the Chinese are purported to say, be careful what you ask for, you might get it... [13:53] goodspud, Stefanie Danevoy wrote that she booked Dijkstra for Monday for us instead of Blue Fin for the daily standup. Since (barring snow issues) you and Alejandra are going to be with us in Austin I'm going to reply that we don't need it. Make sense? [13:54] gary_poster, certainly does [13:54] cool thanks goodspud [13:55] We are being booted out of our Juju room [14:02] bcsaller, good morning, I need to pick your brains some more, when you're around [14:02] gary_poster: did you get any snow? [14:03] bac, dusting. schools delayed two hours [14:09] teknico: I need about 10 minutes, but happy to talk then [14:09] bcsaller, great, thanks [14:10] gary_poster, we might have to delay our meeting a few minutes, is that ok? [14:10] teknico, NEVER! I mean, sure. [14:10] :-) [14:10] :-) [14:22] hi benji, are you reviewing #1099909 ? [14:22] <_mup_> Bug #1099909: We are trying to talk on the web socket before we have connected < https://launchpad.net/bugs/1099909 > [14:23] frankban: I reviewed it a few days ago, if there is something new I can review that too [14:24] benji: ah... so that's why your name is in the card. yes please, https://codereview.appspot.com/7129055 [14:24] thanks [14:29] frankban: The branch looks good. I had a couple of small comments. [14:31] thanks benji [14:32] gary_poster, if bcsaller can talk to me in half an hour or so, maybe it's best if we go ahead and talk now, if you still can [14:32] teknico, now is fine for me, or can wait a bit. I'm in the hangout now. [14:32] gary_poster, ok, getting there [14:33] cool [14:36] teknico: when ever you get done with that hangout I'm free now [14:36] bcsaller, I will, thanks [15:09] I just noticed a story we should probably address: When looking at a service a user may want to see the available relations and which services could fulfill those relations. [15:16] bcsaller, so your definitely not seeing annotations in the initial delta stream? [15:17] hazmat: are you? I wasn't, I get the change events just fine from the stream but didn't see them in the initial pump [15:19] bcsaller, well it annotates current, which is stored as previous, which is used for initial delta to new client [15:19] * hazmat writes a test [15:20] bcsaller, this is live / position stuff is live on uistage also? [15:20] hazmat: the annotations get assigned x/y but not the whole story [15:21] hazmat: I think it sets up last state based on current and then diffs it vs current and sends nothing on the initial pump [15:23] bcsaller, confirmed re not seeing them in the browser [15:23] bcsaller, the delta manager is global across clients [15:24] so previous w/ annotations should have it.. anyways obviously doesn't.. investigating.. eta is like +2hrs, treating as priority so we can show for next week [15:26] hazmat: when that changes the code I have in my branch should handle the rest, and thanks [15:29] bac bcsaller benji frankban hazmat teknico call in 1 [15:30] bcsaller, you should be able to put your branch into review even without yet? [15:30] bcsaller, sorry, it took a while :-) maybe we can talk after the weekly call? [15:31] hazmat: I took things a little further in the branch than I should have and am cutting a version with only the part of the feature gary expects to show, so soon but not yet [15:45] gary_poster: unfortunately I have a conflict for today's retrospective [15:45] arosales, ok, feel to change time [15:45] or talk onday [15:45] M [15:46] gary_poster: sorry I was referring to the standup in a few minutes [15:47] oh ok cool arosales [16:17] bcsaller, having a hard time reproducing this behavior in unit tests.. [16:17] hazmat: because you're seeing the annotation data or just hard to implement? [16:18] bcsaller, because i'm seeing the annotation data [16:18] hmm [16:18] wierd. still playing with it [16:19] deploying a service, setting annotations, connecting a client and it gets the annotations.. might need to step up to a higher level w/ browser and improv to get a better sense of it [16:20] testing guihelp testing :-) [16:21] teknico, works :-) [16:21] heh [16:21] not for me though? [16:22] teknico, most clients are smart about that.. guihelp by me doesn't highlight for me [16:22] teknico, but perhaps it works for you? [16:22] yes, it does if you write it :-) [16:23] cool [16:31] bcsaller, ah ic the problem [16:31] hazmat: :) [16:32] the annotations on previous where getting overwritten, because the annotations are coming in as a stream, in the dstream from the browser you can even see them getting removed. [16:32] not sure if that made sense.. but working on it [16:35] so is it "guihelp" or "gui-help"? [16:35] benji, former [16:35] k [16:35] * benji edits his IRC config. [16:58] bcsaller, fix is @ https://codereview.appspot.com/6936063 its a trivial (one-liner) w/ test .. going to proceed [16:59] hazmat: thanks [17:05] bcsaller, merged [17:05] * hazmat wanders off to arrange exit visa [17:05] hazmat: thanks, I'll test it now, don't expect any issues [17:29] bcsaller, there's still one problem with the third parameter of toggleServiceModule [17:29] the one we renamed from "context" to "module" [17:29] and now we're doing: module.get('component').get('container').one('#service-menu'); [17:29] that line raises TypeError: Object # has no method 'get' [17:31] verify each component in that call chain making sure its what you expect. I didn't think container was a DOMNode but it looks like it might be in the test (rather than a Y.Node) [17:31] Random crazy idea: introspection panel & charm panel could use http://angularjs.org/ . Maybe worth investigating. [17:32] gary_poster: thats a different framework [17:33] Looks like it has community ties to jquery [17:33] but is standalone [17:33] bcsaller, yes. usable in isolation, for only parts of an app. [17:36] bcsaller, the container in the test is created by Y.Node.create [17:39] bcsaller, the value of the "module" parameter of the call (formerly "context") is an SVG "g" element [17:40] maybe we cannot yet get rid of the "view" parameter [17:41] teknico: hmm, in that case its still view->module, but context is... I'd have to read through to find that, I'm guessing its containing the rendered service [17:43] bcsaller, it doesn't look like the case, it only contains other SVG elements [17:43] unless the inspector is not showing everything [17:44] teknico: it might be higher up the tree, we should see if the usage is required, it sounds like it shouldn't be needed to make this process work, I'll peek at 'active_context' [17:47] teknico: it looks like its passed around in old code, but not actually used (in the relations module) [17:48] bcsaller, I think I lost you [17:48] bcsaller, for the sake of this branch, do you mind if I put the "view" parameter back in [17:48] and leave the signature refactoring to you? [17:49] my EOW is fast approaching, and I'd like to land this before then :-) [17:52] teknico: sorry, yeah, its fine, I can clean the rest up next week [17:52] bcsaller, ok, thanks [17:55] gary_poster, benji: either of you have a moment to review https://code.launchpad.net/~bac/lp2kanban/pick-newest-mp/+merge/143925 [18:05] does anyone else see the kanban board lose its brain where one card shows links from another? for instance selecting the RV link from one card takes you to the link from the one above it? a full page reload cures it. [18:18] so current trunk i'm not seeing the login at all.. [18:19] bac: I'm lunching at the moment but I will check if you still need a review when I'm done. [18:20] hazmat: how did you run it? [18:20] bac, make debug [18:20] hazmat: debug - no login required, prod - login [18:20] bac, debug should still send th elogin [18:20] else with login-required its broken [18:20] hazmat: yes, sends but doesn't prompt [18:20] bac, its not sending [18:20] this relates to my comments on the branch [18:21] hazmat: ok, that's different. i just meant the credentials are in config-debug.js but not config-prod.js [18:21] simply shortcircuiting if not connected is broken for the login caller [18:21] as you were. :) [18:21] because it results in not actually logging in at all, even though the caller did env.login [18:21] the caller needs to be updated to do it on env connected [18:30] else we're left with the current behavior which unfortunately is just broken [18:33] hazmat, trunk of juju-gui running make debug and make devel I consistently see a login on rapi logs [18:33] 2013-01-18 13:32:30,605 juju.rapi.ws:INFO Process message {"op":"login","user":"admin","password":"admin","request_id":2} [18:34] like that ^^ [18:34] hazmat, is that not what you are talking about? [18:35] gary_poster, yes. but i'm not seeing the login request [18:36] i'm trying to test the login-required before i land it to ensure no breakage. [18:36] hazmat, not sure what to tell you. I'll try a make clean and then do it again. I'm pretty sure the code path should login as well. Note that we wait for connection before we login now...that's not the behavior you commented on. we changed that on the basis of your comment. [18:38] gary_poster, i'm not seeing the wait behavior here.. https://codereview.appspot.com/7129055/ [18:38] just a moving the shortcircuit from env to app [18:38] on check creds [18:38] gary_poster, the redispatch will auto login ? [18:39] hazmat redispatch will go through routes again, yes, and our first route is check_user_credentials [18:39] We have to be connected in order to try and login [18:39] hmm [18:39] so the connected flag is not being set when run against your branch I guess [18:40] which should happen in env.js...looking [18:41] hazmat, are you no longer sending ready as part of your initial response? [18:41] looking at line 88 of env.js [18:41] and preceding comment [18:41] gary_poster, on the ws server sends.. {"state": "login-required", "version": 0, "extensions": [], "ready": true, "default_series": "precise", "provider_type": "dummy"} [18:41] clients not sending anything [18:42] hazmat can try to debug for 18 min. where's your rapi branch? [18:42] gary_poster, lp:~hazmat/juju/rapi-login [18:42] k [18:43] gary_poster, thanks. if not let me know and i'll continue debug [18:43] ok hazmat [18:49] something is setting connected to false again... [18:49] no... [18:49] but connected is not true [18:50] I bet we need this.env.after [18:50] try [18:50] ing [18:50] yeah that does the trick [18:50] gary_poster, or env.once('connected') [18:51] we don't want it once, I think...I think we need it every time...thinking [18:51] well we need to reset the db every time [18:51] the redispatch is ok to do [18:57] arosales: here are the screen shots you requested; let me know if you have ideas for any additional ones I can make: http://imgur.com/a/HqkQR [18:57] * arosales looking [18:57] benji, thos are for jcastro on #juju. not arosales, unless arosales also asked for them [18:58] gary_poster: ah! [18:58] :-) [18:58] arosales: feel free to enjoy them anyway ;) [18:58] benji they look cool to me, thanks :-) [19:00] benji: will do [19:00] gary_poster, it's too late, sorry, I give up, not landing my branch, feel free to adopt it, your or anyone else [19:01] s/your/you/ [19:01] teknico, what's wrong? [19:01] I mean with the branch [19:01] it's well past your EoD, I picked up on that :-) [19:01] benji and bcsaller need to agree whether the firsts parameter of toggleControlPanel is a service or a box [19:01] last merge from trunk subverted all the assumptions :-) [19:02] s/firsts/first/ [19:02] heh [19:03] the changes are all there, it only needs that little agreement :-) [19:03] ok thanks teknico. benji, can you run with that branch? [19:04] tests are failing, you'll see very quickly what the problem is :-) [19:05] have a nice weekend everyone! [19:11] thanks benji [19:12] gary_poster: sure, I assume it is "double click on service leaves service menu open" we are discussing [19:12] yes benji [19:12] apparently everything was workling till merge [19:19] hazmat: where is your branch/MP/Rietveld for the jitsu speed up work? [19:20] bac, its a juju-core branch [19:20] hazmat: ok [19:20] bac, i think it still needs a little work.. but its at https://code.launchpad.net/~hazmat/juju/test-sans-stat/+merge/143598 [19:21] ie. the testing env var needs to propogate to the provisioning agent, else it only applies to the bootstrap node [19:22] benji: are you picking up teknico's branch? [19:22] otherwise i can [19:22] bac: yep [19:22] * benji realizes he needs to add his head to the card. [19:24] gary_poster: we missed our call yesterday. any time/desire to make it up this afternoon? [19:24] bac yes will ping in a few [19:24] k [19:37] bac in 53 minutes? [19:37] ok [19:41] benji those screen shots are great I think [19:41] gary_poster: cool, I think [19:41] ;P [19:41] benji, :-P [19:41] heh [19:42] I wonder if I can script my IRC client so it adds "... I think" to the end of everything I type. [19:50] I am confused. I think I migth have been bitten by the cards-point-to-the-wrong-review bug. [20:09] benji: did a kanban page reload help? [20:09] or was it the old MP? [20:10] bac: no, but I don't think that is actually what is going on; there appears to be only one MP [20:17] guihelp: have any of you guys used lbox's -adopt switch? As far as I can tell Nicola's branch is good to go, I just need to land it [20:17] no [20:18] I'll just make my own branch and go straight to lbox submit with a comment about what is going on. [20:20] ...right after I figure out why the tests are failing. [20:26] benji, yes.. i have [20:26] benji, it basically sets bzr commit --author [20:27] benji, i just extract someone's identity from the change log to use for it,.. so it looks the same as there others [20:27] thanks hazmat [21:09] benji: are you going to be landing nicola's branch today? [21:10] bac: I think/hope so. [21:54] benji do you need reviews or are you just going for it? I see you linked the branch [21:54] gary_poster: yeah, it was a one-line fix; I am just going to land it [21:55] cool benji [21:55] gary_poster: i lost the tab with your review and i don't see a card. send me the link again, please [21:56] bac https://codereview.appspot.com/7141059 [21:57] gary_poster: done [21:57] thanks bac [22:01] hazmat gui trunk has necessary fixes. verified that you can do a bad login and then a good one [22:01] (so IOW the disconnect after bad login was fine) [22:22] gary_poster, thanks [22:22] gary_poster, cool.. hopefully that fixes staging as well [22:22] welcome [22:23] hazmat staging works for me [22:23] yup works for staging [22:23] cool [22:24] have a great weekend all. bye [22:24] bye