=== Ursinha-afk is now known as Ursinha [13:01] Yellow Slug Video of the Day: http://www.kqed.org/quest/television/science-on-the-spot-banana-slugs-unpeeled [13:17] benji, the URL itself kinda scares me [13:17] heh [13:18] for a nature segment it is a little violent sounding, but be comforted that no banana slugs were harmed during the filming [13:19] it has comforting guitar music [13:19] "the unsung hero of this forest comunity" [13:20] "eating fallen plants and even dead animals" [13:20] (more comforting guitar music [13:20] ) [13:20] Yellow Squad -- we poop dirt [13:21] It's great contribution is that it enriches the soil (presumably be defecating) [13:21] hey, and they are hung hermaphodites [13:21] mm, the largest male sex organ...but a hemaphrodite... [13:21] lol [13:21] gary_poster: i think we hit play at the same time [13:21] I think so too :-) [13:22] mm, comforting music has switched to piano.strings [13:23] mm...slime... [13:23] gary_poster: did you see bug 761257. :( [13:23] <_mup_> Bug #761257: Advanced subscription overlay does not handle unsubscribing teams < https://launchpad.net/bugs/761257 > [13:23] gary_poster, benji :) [13:24] I did not bac...this must be about direct subscriptions, yeah? [13:25] bac benji danilos, will look at kanban in a munite, call in 5-ish [13:25] minute [13:26] gary_poster: yes. bug title updated. [13:27] thanks bac. I'll adda backlog defect card too [13:27] gary_poster: i put a card in FW1 [13:27] oh ok cool [13:27] good enough [13:27] thanks [13:37] https://dev.launchpad.net/LEP/BetterBugSubscriptionsAndNotifications/Testing/EditingRound2/Slide1 [13:46] anybody seen the south park episode about a "dream inside a dream within a dream"? :) [13:46] gary's mention reminded me of that [14:15] benji, my impression was that you thought you were comfortable without further discussion. If you do want to talk, just lemme know. [14:17] gary_poster: a call would be good [14:17] benji, cool [14:17] benji, call me when you are ready [16:34] just encountered one of those integration errors we have been rightly worried about. I wrote a JS test for a change, and it failed, and then fixed the code, and all tests passed...and then actually tried using the thing for some QA and I had screwed up the server request we make. :-/ [17:20] ...it appears to be at least two bugs in one... [17:20] not related to my change :-/ [17:21] but I can't verify my fix via QA until they are fixed [17:21] apparently patching a team bug doesn't work in some cases that I don't understand yet [17:21] and apparently deleting a newly created filter doesn't work in some cases that I don't understand yet [17:23] patching a team filter I mean [17:24] maybe a security problem? i.e., you have permission to create a team filter but not update/delete it? [17:25] yes, that's the first problem--Unauthorized--but pretty weird since we have been able to edit team filters on the other page. Have not found source of that one yet. The deleting a newly created filter I see and can fix, I think. Doing that now, and then will look at Unauthorized [17:31] ah, interesting. it is something specific to rosetta admins [17:33] danilo, what are you doing still here? :-) [17:33] gary_poster, heh, trying to get a few more tests done, got side-tracked with crashing desktop :( [17:33] danilos gotcha :-( [17:34] danilos don't work too late. [17:34] gary_poster, I won't, just a few more of these and I'll be handing off [17:34] ok cool [17:37] oh, weird [17:38] this person, Foo Bar, #1 is listed in the JS as an admin of Rosetta Administrators but that is not listed in https://launchpad.dev/~name16/+participation , and # 2 was able to create a filter for that team, but not delete or patch it. [17:39] oh, weirder [17:39] person is a member of the team that owns the team, but not an active member [17:40] that scenario sinks our battleship [17:41] and that's a valid scenario [17:42] if weird [17:42] so it looks like either the code we have that inserts teams-of-which-you-are-an-administrator needs to intersect with teams-of-which-you-are-a-member [17:43] or if you are an administrator then you should be able to patch and delete a subscription [17:43] I lean towards the second solution, though I wonder how it will pan out to implement [17:47] * gary_poster is going to take a break, though will be around [17:51] gary_poster, oh, I've seen that with rosetta-admins, forgot to add a card; basically, it should not be listed in administratedTeams [17:51] danilos, you think? [17:51] I gave the two options above [17:51] you think not administrated is better than allowing them to patch/delete because they are owners? [17:51] gary_poster, well, all the web service calls fail, perhaps that [17:52] except they can create [17:52] gary_poster, I haven't thought closely about it, to be honest :) [17:52] :-) [17:52] it would probably be easier to not list them [17:52] so under the circumstances, maybe that's the right thing to do for right now [17:53] gary_poster, yeah, we can always file a bug to let someone else handle it later :)) [17:54] woo hoo! :-) [18:50] gary_poster, I am out, you've got email in your inbox, my branch is up at lp:~danilo/launchpad/bug728370-action-display and I left you with the hardest tests to write (those for on-click handler :)) [18:50] gary_poster, other than that particular test, I think it's pretty complete [18:50] danilos, lol, cool thanks [18:50] have a great weekend! [18:50] cheers, so do you! [19:02] bac, a branch when you have a moment: https://code.launchpad.net/~gary/launchpad/bug761798/+merge/57922 [19:02] gary_poster: ok [19:02] thanks [19:03] gary_poster: working on a large branch now but you're next [19:03] thanks bac, no rush [19:34] gary_poster: what do you thing about this scaffolding for the direct actions? http://pastebin.ubuntu.com/594581/ [19:34] * gary_poster looks [19:35] I need to add tests and fill in the unmute action next or if someone's ready to contribute an action I can push this to a branch for them to work from. [19:37] danilos handed off to me, so I'm booked now; sounds like bac is also booked for awhile yet. Tests sound like the next thing to work on, yeah [19:37] * gary_poster still looking [19:40] benji, still looking, but first thought worth sharing: there are two variations for unsubscribe in the mockups: unsubscribe, and unsubscribe with warning (because there are other subscriptions). That will need to be modeled somehow (and I don't think it is yet). [19:41] You could do that with another node... [19:41] kinda sorta [19:41] doesn't quite fit in with the current setup [19:42] ah, I missed the with-warning flavor; I think I can add that without too much trouble [19:42] cool [19:43] benji, looks like a good start to me. Cool, yay, thanks! :-) [19:43] very good [19:43] (by good start, I mean, good scaffolding, not good start on the scaffolding) ;-) [19:48] (yep, I understood) === Ursinha is now known as Ursinha-afk [20:22] gary_poster: i'm a little confused by all of the indirection in your branch wrt clean_up. [20:22] where is the actual fix to the early closing problem? [20:24] bac, maybe I should have tried harder to clean it up. er. The actual fix is that clean_up can't be called after it initially calls the handler. The handler has to call it on success, eventually. [20:24] The extra indirection is because sometimes success isn't true success unrtil after two web calls. [20:24] Let me look at it again and see if I see a way to simplify... [20:25] ok, so line 114 of the diff is the good part [20:26] bac, well, we have to do it in two scenarios. So, yeah, 114, and also... [20:26] gary_poster: i wasn't implying it was unnecessarily confusing...just that i was confused [20:26] also 60 [20:27] no, np bac, I kind of thought that it maybe could have been simplified but this worked and I wanted to get back to the other task. Actually, I think clean_up can be defined as a module-level function and not passed around [20:27] That would make things cleaner. [20:28] Let me try that. [20:28] gary_poster: it would, as it seems to be a hot potato atm [20:28] yeah [20:35] bac, I did this: http://pastebin.ubuntu.com/594603/ . Pushed to public branch and waiting on diff change... [20:36] gary_poster: i think that's much clearer [20:36] cool, thank you bac [20:42] bac, I considered your add_subscription_overlay approach but I felt it would have been more fragile: when code changed the local variable, it would have needed to be sure that it also changed the namespace version in order for the tests to be reliable. I felt that always getting the overlay from the namespace was a cheap cost for that reduction in fragility. Therefore, fwiw, I intend to leave it as it is. [20:43] thank you for the review! [20:47] gary_poster: not to belabor the point, but wouldn't the test break if the names were changed out from under it? [20:47] it is a idiom we use in that module [20:47] bac, it depends on how that name were set, and even if it is cleaned up across test runs [20:47] it might very well work now [20:48] getting it from the namespace also simply didn't bother me in the slightest [20:49] I'm guessing you don't feel differently enough to make a point of it, since you didn't do so in the review. Had you, I would have complied, because even though I feel comfortable with what I did, I don't feel strongly about it either. [21:18] gary_poster: i'm going to take the dog for a walk. will be back later. have a good weekend [21:18] you too bac. bye