[10:00] gmb, hi, do you perhaps know how to get the return value of a method with named_post? I get "undefined" for addBugSubscriptionFilter, and I need the newly created filter [10:02] danilos: Hmm, that's odd. Maybe the return type needs to be declared somehow. It should just work. Hang on, let me take a look. [10:02] gmb, thanks [10:04] gmb, I guess it returns stuff in the 'on.success' handler instead [10:04] gmb, so, my error in thinking it might be serializing stuff for me [10:04] danilos: Ah, yes! Sorry, I was thinking about when calling the webservice in Python :) [10:06] danilos: It might be the second or third parameter passed to the success handler. [10:06] Or I may be thinking of what happens when you call .patch(). [10:06] (This should be more straightforward) [10:08] danilos: Oddly, I can't find a good example of us doing a named_post and getting something back from it. [10:08] At least not in the bugs JS. [10:09] gmb, yeah, neither could I and I even tried grepping through entire lib/lp :) [10:09] Haha. [10:09] Boldly going where no dev has gone before... [10:09] danilos: So, it might be that if you have something like: [10:09] gmb, we mostly do it in on.success it seems, it's just non-obvious to someone just diving into this JS stuff [10:10] success_handler(some_param, some_other_param) {...} [10:10] gmb, I was just thinking named_post might be hiding that from me [10:10] yeah, it gets the returned stuff as the first parameter afaict [10:10] Ah, cool. [10:10] But yeah, completely non-obvious. [13:07] gary_poster: Morning. Your branch can now land; my DB patch is now on db-devel after some testfix tomfoolery this morning. [13:07] yay, gmb! that must mean we are out of release...thingy too [13:07] So I'll land my branch [13:08] oh, and what about the api branch? [13:08] gary_poster: It's working through PQM now. [13:08] awesome! [13:08] I'll make my UI MP after that [13:08] In fact, it's now on db-deve. [13:08] *l [13:08] woo hoo [13:08] OK, I'll get started on all that in a mo [13:08] how goes the bug 1 hang thing? [13:08] <_mup_> Bug #1: Microsoft has a majority market share my question was, why is that code being called at all? [13:10] gary_poster: I haven't had chance to dig around much on it today; I've been OCR this morning. However, the review queue is for now quiescent, so I hope to have an answer to that question in the next hour or two. [13:13] ack gmb. you started that before my suggestion to only start things that you think you can finish in 2 hours, but maybe you can divide this up into diagnosis and such, so that if you diagnose this, you can record the results somewhere and move to something else if necessary. [13:13] RIght. [13:14] gary_poster: I started it because it was on the board and annoying me :). I need to perhaps think twice about scratching my own itches when they're complicated itches. [13:14] :-) [13:26] bac benji danilos gmb: just wanted to let you know that HR has all of my reviews, so if I'm called away for baby duty, all that is taken care of. bac, Sarah said she would enter yours for me from the text doc I gave her because of the technical issues we discussed. [13:26] bac benji danilos gmb: please make sure kanban board is up to date within 2 minutes [13:26] we will have our call in 4 [13:29] * gary_poster looking at board [13:31] gary_poster: mumble or skype? [13:31] mumble still doesn't work for me :( [13:31] skype please gmb, all [13:31] danilo you are not logged in :-P [13:31] gary_poster, but I am! :) [13:32] heh ok skype lies a lot [13:32] haven't actually logged out since yesterday, skype is sometimes slow [13:32] Godsdammit [13:32] gary_poster: Please call again. [13:41] gmb, my ui branch is 814 line diff. I can split it up into 100+ line python and 600+ line JS if you like, or just throw it at you [13:41] what do you prefer? [13:42] gary_poster: Just give me the whole thing; simpler than splitting it up and after 2000 line diffs a couple of weeks back 814 lines is a cakewalk. [13:42] :-) ok thanks gmb [13:44] gary_poster, also, epiphany has started to crash quite often for me, which hasn't helped with my debugging efforts, enough that I switched to firefox and chromium [13:45] ack :-/ [13:49] gary_poster: So far as I can tell, the reason that parent messages are looked up is "because that's the default behaviour." However, I can find no test to reference *why* it's the default behaviour, and indeed the default on IBug is for include_parents=False, where as on the model it's True. I've changed the model to match the interface and I'm running it through EC2 to see what flies off. [13:50] huh [13:50] cool [13:50] This might be one of those it-made-sense-in-2005 things. [13:50] I don't know. [13:53] gary_poster: did the storm blow y'all away last night? [13:54] btw, how does one get a user URL for a team object? filter.subscriber_url is calculated in the view directly :/ [13:54] bac, yeah it was crazy! not too many new branches on the ground [13:54] gary_poster: not actually bad here. the weather radio kept going off talking about tornadoes next door in alamance county [13:55] huh [13:55] the sound was the craziest thing here [13:55] danilos I did it that way because I didn't know :-/ [13:55] (you mean Python view class I think, yeah, and you want it done in the JS) [13:57] brb [13:58] bac, benji, gmb, if you have any wisdom for danilos, please share :-) [13:58] Ffff... [13:59] it seems simplest to parse team.link (which is an API link) but that's _very_ broken; round-trip to server sucks as well, though I don't even know what would I get that way [13:59] I don't have anything off the top of my head. [14:00] I can't think of anything that's not "parse team.link" [14:00] yeah, pulling stuff out of API links and reconstituing a new link sounds bad [14:01] try deryck as last resort, then go for it :-/ [14:01] If you know at page render time that you'll need the link, you can stuff it into LP.cache for use on the client side [14:19] yeah, that's what we do for the other stuff, but this is for the dynamic add story, so no such luck [14:20] although... [14:20] danilos, what about changing the LP.cache collection of teams to include links? [14:20] as constructed by code in lib/lp/bugs/browser/structuralsubscription.py [14:21] the team thing is a constrained set [14:21] and we are already iterating for names [14:21] so you could include the link [14:21] gary_poster, yeah, probably the best option [14:21] cool [14:22] benji, is there a LP-preferred way to run the Crockford JS linter? [14:23] gary_poster: not that I know of; I've been using jslint.com [14:23] ack, thanks benji [14:29] gary_poster: doing QA i've discovered a problem i'm looking into. the overlay is being rendered in the page on the bugs root: https://bugs.qastaging.launchpad.net/firefox [14:30] bac, argh, good catch [14:40] gmb, damn, I have to run, and didn't finish MP prep (got caught in lint) [14:40] will be back in a few hours [14:40] bye [14:41] bac: I believe we want the new subscription feature flag rule in production, right? [14:42] benji: don't know [14:42] i don't see what it would hurt as long it is team:yellow [16:00] hi gmb, in structural-subscription.js we're using Y.fail...but it errors "Y.fail is not a function" [16:01] Hah. [16:01] i assume we need to add something to our 'requires' but i don't know what [16:01] it is part of the test package. is that even appropriate to use in production code? [16:03] bac: I thought it was part of "base" [16:03] gmb: ok. we're not using 'base'. i'll add it and see [16:06] gmb: nope, base doesn't help. where should we look to find a definitive answer in the docs? [16:07] bac: Urr. I don't know. I just tend to dig around in http://developer.yahoo.com/yui/3/ until I find what I need to know. [16:08] bac: Although just looking, and the only refereces I can find to Y.fail() seem to involve tests... [16:08] filebug_dupefinder.js uses it [16:08] benji did you add the use of Y.fail() to structural-subscription.js? [16:09] bac: yep [16:09] benji: so it doesn't seem to work [16:09] it was suggested in a review [16:09] darn [16:13] benji: who was the reviewer? [16:13] hmm, let me look [16:24] hmm, there has to be a good way to find that MP [16:27] bac: found it! https://code.edge.launchpad.net/~yellow/launchpad/accordion-client-1/+merge/54715 [16:27] hah! it was gmb! [16:27] how very circular [16:28] :) [16:28] so, i cannot figure out where Y.fail comes from. [16:29] it seems to be for tests, at least the docs imply that [16:29] we have it in a number of places in production code [16:31] gmb: has something odd happened with filebug_dupefinder.js? it uses Y.fail so i was trying to exercise it on lp.dev and get [16:31] NotFound: Object: , name: u'filebug-dupefinder.js' [16:32] bac: Not AFAIK. Have you tried cleaning and rebuilding? [16:32] yes, many many times [16:32] very odd [16:33] from my reading of the YUI docs, it doesn't seem like .fail is for tests but a shortcut way to raise an assertion error [16:37] gary_poster, I am out now, but the branch is in a state where everything works (I've fixed a few bugs not related to this branch, but not all: eg. when you remove the final filter, the subscription listing still remains) [16:37] gary_poster, https://code.launchpad.net/~danilo/launchpad/add-subscription-link/+merge/56392 [16:37] gary_poster, no tests yet at all, it's been hard enough getting it to a state like this, so I'll tackle that tomorrow morning [16:38] gary_poster, cheers === Ursinha is now known as Ursinha-lunch [17:52] benji: Could you quickly review https://code.launchpad.net/~gmb/launchpad/bug-1-timeout/+merge/56407 for me? [17:52] (12 lines of diff, 1 line changed) [17:53] gmb: sure [17:54] Ta [17:56] gmb: done [17:57] Thanks. [18:01] danilos, ack, thanks [18:02] gmb, you are out about now, yeah? or do you have one more hour? [18:03] gary_poster: I'm at EoD now, but I've been AFK for a bit longer than I intended this afternoon so I'll be around for a while to make up the time. I'm still happy to review your branch. [18:04] awesome gmb thanks. I'll try to get it ready for you quickly then [18:11] ok [18:40] gmb, took me long enough, sorry [18:40] https://code.launchpad.net/~gary/launchpad/muteteamsub-ui/+merge/56419 [18:40] diff is being generated now [18:40] No worries. [18:40] I'll take a look shortly [18:40] ended up being < 800 lines after all [18:44] * gary_poster feels like taking a nap, more than anything else :-/ [18:46] so long shipit! [18:47] who gets the honor of ripping the final bits out of our code base? === Ursinha-lunch is now known as Ursinha === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [19:17] bac: do you want to review the "bug includes link to subscription page" MP? [19:18] benji: is it short? if i can do it in a few minutes i can do it now otherwise it'll have to wait until after i go to the dentist [19:18] bac: yeah, pretty small; 88 lines and half of that are the tests [19:19] sure, i'll do it now [19:23] gary_poster: r=me, but you need to add some comments to the tests you've added in lib/lp/registry/javascript/tests/test_structural_subscription.js [19:24] And also to test_muted_team_member_subscription, now that I come to look back. [19:24] gmb, cool. Just top-level comments at beginning, or others in addition? [19:24] and thank you [19:24] gary_poster: Just explanatory ones at the beginning of the test stating the expected behaviour. [19:24] np. [19:25] cool, will do [19:25] have a nice evening [19:31] benji: if you update your pocket-lint 'make lint' will probably work [19:31] unfortunately i think you have to use curtis' PPA to get the latest working copy [19:33] bac: I don't know if I'm using Curtis' PPA or not, but upgrading pocket-lint fixed it [19:33] there is one piece of lint that I'll fix now [19:33] cool [19:38] gary_poster, benji: for the new links we've added for structural subscriptions we've been careful to only enable them for logged in users ('launchpad.AnyPerson'). the links for 'editsubscription' are shown to anon...but the same is true for other subscribe actions directly on a bug [19:38] so, leave as it is and stay consistent for that context if not a little gross? [19:38] * gary_poster reading. brain is slow. [19:39] right, that was my reasoning; since we show "Subscribe" and "Subscribe someone else" to everyone (and ask them to log in if they click on them), I did the same for the edit link [19:39] ... [19:39] I don't feel too strongly about it, but it seems like making the links visible to non-logged-in users is a nice discoverability bonus. [19:40] I also don't feel strongly about it, but would have leaned the other way. bac, is one noticeably easier than the other? or do you have a pref? [19:41] benji: when everything settles i think we should agree to do it one way or the other in the different contexts. for now i agree it should be consistent on that page [19:41] gary_poster: disabling is harder but minimally [19:41] but we should decide and fix all of the links on that page if we want to [19:41] i don't think it should affect benji's branch right now [19:41] agreed on not affecting [19:42] bac, I say, do not disable, add a card to raise an issue to Huw about it [19:42] that's if you are looking for a decision, which I think you are [19:42] if not, make a decision. :-) [19:43] gary_poster: that is the decision i made. [19:43] cool, thanks bac [19:46] benji: done, in the nick of time. [19:46] :) thanks [19:48] gary_poster: do we want the structural subscriptions feature flag enabled for team:yellow in production? [19:48] yes, benji [19:48] thank you [19:48] k, I'll shepherd that [19:56] gary_poster: should I grab one of the quick jobs cards next or one from the backlog? [19:56] hmm, I wonder why the card for 674422 is in Acrive/Coding [19:57] benji, I think the ones in backlog/design fit the bill best [19:57] but do what you think works [19:57] including making your own card [19:58] for something like "test the edit/delete functionality" (though look at my mute branch if you want that one) [19:58] because danilo reports that deleting the last filter no longer deletes the subscripton in the ui [19:58] which used to work [19:58] it's almost like we have no tests for it! [19:58] gary_poster: to be clear: "design" are things like the small QA issues I identified, right? [19:58] exactly [19:59] benji, 67422 is in active/coding because it is following the last pertinent card (Danilo's) out the door. AIUI that's what I'm supposed to be doing. [19:59] "test the edit/delete functionality" sounds the most important [19:59] (AIUI from flacoste) [19:59] sounds good to me benji [19:59] interesting [19:59] yeah [19:59] ask me when I don't feel tired ;-) [20:00] unfortunate that it takes up a limit spot [20:00] heh [20:00] yeah [20:00] benji, I suggest you build off of my branch. finding... [20:00] https://code.launchpad.net/~gary/launchpad/muteteamsub-ui/+merge/56419 [20:00] It has some bits that I intended to be useful for this purpose [20:01] ok, cool [20:01] and stuff that I would have refactored further had I worled on what you are looking at [20:01] worked [20:01] oh, but... [20:01] that branch is landing on db-devel :-( [20:02] so, merging will be problematic [20:02] I should have done the refactoring separately [20:02] If you want me to help with the mess lemme know [20:02] ok [20:17] gary_poster: feature flag rule is now in production and I've verified that I see the new UI bits [20:17] great thanks benji [20:28] benji, you available for a pre-imp? [20:29] gary_poster: sure === Ursinha is now known as Ursinha-afk === Ursinha-afk is now known as Ursinha [22:04] benji, instead of talking again, I'm running away, at Francis' encouragement, since I feel tired and pressured...and I don't even have a baby in my belly. :-) I'll refocus on what we talked about tomorrow. I might ask you or bac for pair programming, if we can do it in a reasonable way. But meanwhile...I need a break. :-) [22:04] bye all [22:04] gary_poster: sounds good, have a relaxing evening [22:05] if you want to relax in a brain-numbing way, I'd see if iTunes has the TV show Reaper from a few years ago [22:05] I'm going away now too. === Ursinha is now known as Ursinha-bbl