[09:39] <danilos> gmb, hi, do you have any idea why we don't have structural subscriptions on ISourcePackage (distroseries-specific, as compared to IDistributionSourcePackage, which we do have them on)?
[09:41] <gmb> Danilos: not a clue, sorry.
[09:42] <danilos> gmb, ok, I've added a card to investigate and chat later on with Gary
[09:42] <gmb> Ok
[10:00] <danilos> gmb, just to confirm something else, if I am subscribed to a "primary" bug, I *won't* get email from a duplicate that I have nothing to do with, right? (i.e. I didn't file it, I was never subscribed...)
[10:02] <gmb> danilos: That's correct, yes.
[10:02] <gmb> At least I *hope* it is :)
[10:02] <gmb> (There's actually a flag for this somewhere in the code; if we flipped it we could spam everyone with notifications from duplicates)
[10:04] <danilos> gmb, ah, I like flipping flags! :)
[10:05] <gmb> danilos: Just remember to tell everyone that it was someone else's idea so that they get all the shit for it.
[10:05] <gmb> Note: NOT MINE.
[10:05] <danilos> gmb, naturally!
[10:06] <gmb> :)
[11:55]  * danilos -> lunch
[12:59] <bac> danilos: re: ISourcePackage, I'm unsure why they are not IStructuralSubscriptionTargets either
[13:29] <gary_poster> bac benji danilos gmb looking at board; meeting in 2
[13:29] <gmb> Yup
[13:30] <gary_poster> bac & benji I don't see you on Skype; maybe it lies
[13:30] <gary_poster> there's bac
[13:31]  * benji restarts skype
[13:31] <danilos> gary_poster, I usually just try calling people in such cases and it suddenly starts seeing people
[13:31] <benji> gary_poster: now?
[13:35] <danilos> gary_poster, nobody hears me?
[13:35] <gary_poster> danilos I guess not :-/
[13:35] <danilos> gary_poster, please re-invite me
[13:36] <danilos> gmb, btw, grep for "setStyle" as well
[13:37] <gmb> danilos: Indeed, thanks.
[14:09] <benji> gary_poster: I just summarized the PQM situation in -dev
[14:22] <gmb> benji: In bug 750561 are you talking about adding subscriptions via the overlay or on the +subscriptions page?
[14:22] <_mup_> Bug #750561: Spinner for adding/editing subscriptions <story-better-bug-notification> <Launchpad itself:Triaged> < https://launchpad.net/bugs/750561 >
[14:23] <benji> gmb: overlay (although both probably need it)
[14:23] <gmb> Okay, I'll take a look, see how easy it'll be.
[14:24] <gmb> gary_poster: w.r.t to the CSS stuff, I think it needs more thought, i.e. should we also accomodate setStyle() calls? What about styles that are very specific to only one element? that kind of thing.
[14:24] <gmb> I'll update the bug.
[14:26] <gary_poster> gmb, yeah, that's the kind of thing I was thinking about too
[14:26] <gary_poster> if there's not a clear action to do then I return to questioning the bug, but maybe that's being...antu-progress
[14:26] <gary_poster> anti
[14:27] <gmb> Well, I'll add my thoughts to the bug... maybe we should take this to the launchpad-dev list and make it a Policy decision.
[14:27] <gary_poster> so gmb, are you up for a 1400 line review, since you are offering reviews so generously? :-)  I can probably figure out some way of dividing it up, and we probably should not continue to set this precedent...
[14:28] <gmb> gary_poster: Yes, I can fit that in. actually should take me up to my slightly-earlier-than-normal EoD, so it works well.
[14:31] <gary_poster> ack gmb thanks.  I'll go run make an mp now
[14:31] <gary_poster> quickly. :-)
[14:31] <gmb> Cool
[14:39] <gary_poster> benji, you comfortable with what wgrant is suggesting?
[14:40] <benji> gary_poster: yep; I managed to get the conflict in my working copy and it's an easy fix, so I think we'll be good
[14:40] <gary_poster> benji cool.  So getting it on db-devel will be mildly annoying
[14:40] <gary_poster> you'll either need to use bzr land with a new MP to db-devel
[14:41] <benji> gary_poster: I'm confused; this is targeted at devel
[14:41] <gary_poster> was the conflict in devel or db-devel?
[14:41] <benji> I'm pretty sure it was devel; double checking
[14:41] <gary_poster> k
[14:42] <gary_poster> if it is devel then nm, yeah, sraightforward
[14:42] <benji> gary_poster: grrrrrrr
[14:42] <gary_poster> ?
[14:42] <benji> it was db-devel
[14:42] <gary_poster> yeah, figured
[14:42] <gary_poster> so that means it is in devel but didn't make it through to db-devel
[14:42] <gary_poster> so that's back to what I was saying
[14:43] <gary_poster> you have three options
[14:43] <gary_poster> 1) make an mp and bzr land.  that seems heaviweight if this is only JS tests *and* windmill is turned off.
[14:43] <gary_poster> eh
[14:43] <gary_poster> that was supposed to be "make an mp and ec2 land"
[14:44] <gary_poster> 2) make an mp and bzr land.  That's easier, and uses tools that are fairly user friendly.  it is still heavierweight than you need because you don't really need a review, but practically it might be the easiest, unless you are comfortable with...
[14:45] <gary_poster> 3) use bzr pqm-submit directly.  This has an incantation-y feel to me, but does not require an mp.
[14:45] <gary_poster> if you have not used it much, I think it would probably just frustrate you
[14:45] <gary_poster> So my recommendation is #2, or #3 if you want to.
[14:46] <benji> I've done pqm-submit a little so I'll give that a shot
[14:46] <gary_poster> ok cool
[14:50] <benji> gary_poster: so let me get this straight, even though my branch was originally targeted to devel, since we have a conflict with the merge of db-devel into devel (after my branch is applied), I'm going to directly merge my branch into db-devel so when the automated merge happens it will work
[14:50] <gary_poster> benji, uh, I don't think so.  Though I admit to being hazy about at least one aspect of it as well.
[14:50] <gary_poster> So...
[14:51] <gary_poster> You targeted your branch to devel
[14:51] <gary_poster> It merged to devel (right?)
[14:51] <gary_poster> When being automatically merged to db-devel in conflicted
[14:51] <gary_poster> it
[14:52] <gary_poster> wgrant resolved it, maybe?  By removing the revision entirely?  This is the part I know I'm hazy about.
[14:52] <gary_poster> Now you need to get it back in db-devel
[14:52] <gary_poster> but devel is fine
[14:52] <gary_poster> benji, that's my understanding, such as it is
[14:57] <benji> gary_poster: re. "devel is fine", why don't I see the merge of that revision when I do a bzr log?
[14:58] <gary_poster> gmb, diff is being generated, but here it is.  My cover letter was kinda lame, but I was in a rush.  Of course, please feel free to ask questions here or even request a call if goals are not clear.  Thank you!
[14:58] <gary_poster> https://code.launchpad.net/~yellow/launchpad/bug728370/+merge/56942
[14:59] <gmb> gary_poster: Okay, thanks. I'll see whether I can make head or tail of it... (Though your cover letter is about 900% longer than some I've been given on large branches).
[14:59] <gary_poster> heh
[14:59] <gary_poster> ok
[15:00] <gary_poster> benji, you are right that your new tests don't appear to be on devel.  Investigating further...
[15:01] <gary_poster> benji, would you forward me the conflict email please?
[15:02] <gary_poster> benji, whatever happened, which I don't understand either, you are right that you need to land this on devel
[15:03] <benji> gary_poster: forwarded
[15:05] <gary_poster> benji, my suspicion is that the conflict you forwarded me is for the ss-whatever branch
[15:06] <gary_poster> and that you had a separate one for the branch you are looking at now
[15:06] <gary_poster> my further suspicion is that wgrant has resolved the ss-whatever branch into db-devel conflict
[15:06] <gary_poster> so now, somewhere in your inbox, there is a conflict email about the new test test branch (whatever-2)
[15:07] <gary_poster> and now that you have resolved it, things will return to be sunny and shiny
[15:07] <gary_poster> after you submit it to devel
[15:07] <gary_poster> (which you could just do with bzr land
[15:07] <gary_poster> )
[15:07] <gary_poster> (because it has an approved MP pointed to the right destination already)
[15:08] <gary_poster> (and you don't need tests)
[15:08] <gary_poster> (I have to go blow my nose now)
[15:14] <benji> ok, lp-land fired off, we'll see how it goes
[15:22] <danilos> gary_poster, btw, I keep forgetting about the visa, I'll have to ask about that as well
[15:23] <gary_poster> danilos, ? ireland?
[15:23] <danilos> gary_poster, right
[15:23] <danilos> gary_poster, (sorry for just blurting half-sentence out :)
[15:23] <gary_poster> :-) np.  is there something I need to do?
[15:23] <danilos> (and doing it again)
[15:24] <danilos> gary_poster, nope, other than ask me if I've booked my travel and arranged everything re a visa :)
[15:24] <gary_poster> Hey, Danilo...have you booked your travel and arranged everything re a visa?
[15:24] <gary_poster> :-D
[15:24]  * gary_poster will add it to my mtg notes
[15:25] <danilos> gary_poster, whoa, I totally forgot about that, thanks for the reminder :)
[15:25] <gary_poster> lol
[15:25] <danilos> gary_poster, and thanks :)
[15:25] <gary_poster> np
[15:32] <benji> gary_poster: I /think/ everything is right with the world.
[15:32] <gary_poster> benji, now *that* is the best news I've heard in ages :-)
[15:32] <benji> heh
[15:33] <gary_poster> but even within the realm of your branch, very good :-)
[15:44] <gmb> gary_poster: r=me with some tweaks. Nice branch - very readable and I understood the concept, which helped, I think. I've got a couple of concerns about the readability of a couple of nested for loops, but it's nothing huge.
[15:45] <gary_poster> cool gmb, thank you!  I'll take a look
[15:49] <gary_poster> gmb, you have...12 min or hour and 12 min?  either way, I did a bit of investigation on staging for bug 753000, which you and I discussed yesterday.  Here are some results.  I think some of them might indicate problems, but others (like the distribution) may indicate that my query was naive.
[15:49] <gary_poster> https://pastebin.canonical.com/45864/
[15:49] <_mup_> Bug #753000: NotOneError caused by duplicate stuctural subscriptions <merge-deactivate> <oops> <story-better-bug-notification> <Launchpad itself:Triaged> < https://launchpad.net/bugs/753000 >
[15:49] <gary_poster> but if anything comes to mind, there it is :-)
[15:50] <gmb> gary_poster: I have 11 minutes now, so I'll take a quick sken and come back to it later if needs be.
[15:50] <gary_poster> no worries on coming back.  Go have a vacation!
[15:52] <gmb> gary_poster: So, if I'm reading those results right, there are 819 subs to distros which have duplicates, 6 product subs, 1 project sub (who is user 82, I wonder?) and 58 sourcepackage subs, right?
[15:53] <gary_poster> (user 2, project 82)
[15:53] <gmb> Ah,.
[15:53] <gmb> Yes.
[15:53] <gmb> So, who's user 2? (Mark, maybe?)
[15:53] <gmb> Anyway...
[15:54] <gary_poster> yes, I think you are right
[15:54] <gmb> gary_poster: It might be useful to see how many dupes each subscription actually has, but I don't know if that's worth worrying about at this point.
[15:54] <gmb> Hmm.
[15:54] <gary_poster> I think the distribution query *might* be broken
[15:55] <gary_poster> I wonder if, for instance, source packages also distinguish by distribution
[15:55] <gary_poster> so 819 might be just wrong
[15:55] <gary_poster> but the others look "real"
[15:55] <gmb> gary_poster: Hmm, possibly. You could re-run the sourcepackage query with the distro in there too to see what happens.
[15:55] <gmb> gary_poster: Right, I agree.
[15:56] <gary_poster> If I were doing it right, as I may do later, I would look at the joins defined in structuralsubscription.py and dupe them
[15:56] <gmb> Yeah.
[15:56] <gary_poster> but this seemed a reasonable naive approximation
[15:56] <gmb> gary_poster: Ideally I'd say "let's write a test to reproduce this" but I fear we'd be groping in the dark a bit.
[15:56] <gary_poster> right
[15:56] <gary_poster> we have no idea how to dupe
[15:56] <gary_poster> we merely know that the data looks bad
[15:56] <gmb> So.
[15:57] <gmb> My proposition would be:
[15:57] <gmb> 1. clean up the data
[15:57] <gmb> 2. add a constraint
[15:57] <gmb> 3. Wait for the OOPSes to roll in (if any)
[15:57] <gary_poster> (multiple constraints, probably)
[15:57] <gmb> 4. Fix the problem in the tree if 3.
[15:57] <gmb> Right.
[15:58] <gary_poster> yeah agree.  I think I'll need to include Robert on this, since he will want to know and might even want to change the plan somehow
[15:58] <gary_poster> OK, I'll refine this later
[15:58] <gary_poster> Thanks for looking
[15:58] <gary_poster> I think it's time for you turn off IRC now
[15:58] <gary_poster> :-)
[15:58] <gary_poster> Have a great 2 weeks!
[15:58] <gmb> gary_poster: np. I think (Robert, Stuart) is the usual pairing for this kind of stuff.
[15:58] <gary_poster> sure
[15:58] <gmb> gary_poster: Your wish is my command, sir.
[15:58] <gary_poster> :-)
[15:58]  * gmb -> exeunt, in pursuit of a Mustang.
[15:58] <gary_poster> lol
[16:00] <gary_poster> user 2 is lifeless, fwiw :-)
[17:19] <gary_poster> lunching & volunteering, back in hour or so
[19:41] <benji> gary_poster: I know more about YUI 3 positioning than I ever thought possible.  I'm working on the make-the-overlay-start-out-higher-on-the-page card.  How's this for the position of the overlay: http://i.imgur.com/wl7e1.png (it's aligned with the top of the side portlet and will scroll into positoin if you're scrolled down when you click on the subscribe link.
[19:41] <benji> there's not much to the code:
[19:41] <benji> http://pastebin.ubuntu.com/591400/
[19:42] <gary_poster> benji, sounds good.  We'll want a UI review for the whole shebang at the end, so for now I Think we can try to improve with impunity.  kinda.
[19:43] <gary_poster> the scrollIntoView makes me the most "nervous" from a UI perspective, but I'm not worried about it
[19:43] <benji> Our chief tool is impunity!  And jslint.  Amongst our tools are...
[19:43] <gary_poster> :-)
[21:27] <gary_poster> bac, small JS branch review? https://code.launchpad.net/~gary/launchpad/bug754958/+merge/57010
[21:27] <bac> gary_poster: sure
[21:28] <gary_poster> thank you
[21:37] <gary_poster> benji, I cannot dupe bug 750571
[21:37] <_mup_> Bug #750571: If status "Unknown" is unchecked and the structural subscription overlay is closed, it will be unchecked when reopened. <story-better-bug-notification> <Launchpad itself:Triaged> < https://launchpad.net/bugs/750571 >
[21:37] <gary_poster> could you confirm that it is fixed, or give me better instructions?
[21:37] <benji> looking
[21:38] <gary_poster> well, "better"...help me understand better :-)
[21:38] <gary_poster> So far I have "fixed" one card by discovering it was already addressed, I wouldn't mind making it two ;-)
[21:40] <gary_poster> bac or benji, do you have anything to point me to if I want to show a spinner for actions
[21:41] <gary_poster> like add, edit, delete
[21:41] <bac> no, i've never added a spinner
[21:41] <gary_poster> I know we do it, but I don't know how, and can't think of where
[21:41] <gary_poster> ok
[21:41] <bac> gmb seemed to think it was easy
[21:41] <gary_poster> yeah :-)
[21:41] <bac> grep spinner?
[21:41] <gary_poster> too bad he's gone for two weeks ;-)
[21:41] <gary_poster> yeah ok
[21:42] <bac> gary_poster: pay for his transatlantic wifi and ask him to do it tomorrow
[21:42] <benji> gary_poster: I think you "fixed" this one too.
[21:42] <gary_poster> heh
[21:42] <gary_poster> benji, sweet :-)
[21:42] <gary_poster> thanks for doublechecking.  I'll handle the paperwork
[22:02] <gary_poster> benji for bug 750567 that you are working on, do you happen to know what part of the code you have to change yet?  (I'd rather not provoke a conflict unnecessarily)
[22:02] <_mup_> Bug #750567: Structural subscription overlay doesn't instantly disappear. <story-better-bug-notification> <Launchpad itself:Triaged> < https://launchpad.net/bugs/750567 >
[22:02] <gary_poster> I'm working on the spinner
[22:03] <gary_poster> so the full story should probably be click...spinner...close.  The way that the handlers work, there's a reasonable chance that we won't step on each other but I thought I'd check
[22:08] <benji> gary_poster: the change for that was really small and is done if you want to see: https://code.launchpad.net/~benji/launchpad/bug-750573-move-overlay/+merge/56999
[22:08] <gary_poster> cool thank you
[22:10] <gary_poster> gotcha--"add_subscription_overlay.on('submit', clean_up);" + "add_subscription_overlay.hide();"
[22:10] <gary_poster> thanks