[00:46] <lifeless> wgrant: thanks
[00:55] <StevenK> Hmmm, looks like celery is causing the hang.
[00:55] <StevenK> If I kill the celery main process the test runner actually exits at least.
[01:27] <nigelb> lifeless: looks like I won't be seeing you soon afterall.
[01:28] <StevenK> nigelb: No NZ for you?
[01:29] <nigelb> looks like.
[01:29] <nigelb> Can't say for sure though, sorting out what's next.
[01:36] <mwhudson> :(
[01:42] <sinzui> StevenK, wgrant, wallyworld_ the maintainers of https://launchpad.net/akiban-server cannot access branches
[01:43] <wallyworld_> private branches or any?
[01:43] <wgrant> They're all private
[01:43] <wgrant> sinzui: How'd you discover this?
[01:43] <sinzui> https://code.launchpad.net/akiban-server/+activereviews gives a 403. I can see the page, but I see nothing in it since there is no BVP that covers me
[01:44] <sinzui> wgrant, arielweil reported it to me
[01:44] <sinzui> the issue started in the last 3 hours
[01:44] <wgrant> sinzui: Cannot access branches, or cannot access +activereviews?
[01:44] <sinzui> I hope he will join us in a minute
[01:44] <wgrant> The latter has various known bugs that the recent changes may have exacerbated.
[01:44] <sinzui> the page will not load
[01:44] <wgrant> But it's a very very different situation from not being able to access the branches.
[01:45] <sinzui> it is like something we iterated over raised an Forbidden error
[01:45] <wgrant> I'm trying to track down the OOPS
[01:45] <wgrant> Bug #900431 is what I was thinking of
[01:45] <_mup_> Bug #900431: branch visibility queries do not consider visibility of stacked on branches <403> <branches> <disclosure> <privacy> <sharing> <Launchpad itself:Triaged> < https://launchpad.net/bugs/900431 >
[01:45] <sinzui> ha, I was looking at that a few minutes ago
[01:47] <wgrant> https://oops.canonical.com/oops.py/?oopsid=OOPS-8ccd6bd99a7dab48e98b2f44b941fff9
[01:47] <wgrant> Ah, that's actually another akiban project
[01:47] <wgrant> But same issue, probably.
[01:48] <sinzui> wgrant, I expect that. They have projectgroup-level rules that are applied to all the projects
[01:48] <sinzui> should I set of project-specific rules for them?
[01:49]  * sinzui ponders the swift death of project groups
[01:49] <wgrant> https://oops.canonical.com/oops.py/?oopsid=OOPS-2c591db5bd212de7766bf7afcbcbf591
[01:49] <wgrant> For akiban-server
[01:50] <sinzui> I see
[01:50] <sinzui> users cannot see each other's branches because the owner attr is private
[01:51] <sinzui> well not sufficient permission
[01:51] <wgrant> Well
[01:51] <wgrant> They should hold launchpad.View
[01:51] <wgrant> Note that this is a newish branch
[01:51] <wgrant> Just a couple of days old.
[01:51] <wgrant> Can you see who is subscribed?
[01:51] <wgrant> I guess not.
[01:51] <wgrant> Oh
[01:51] <wgrant> Oh no
[01:51] <wgrant> Hm
[01:51] <wgrant> Actually, codehosting should only create branches over XML-RPC...
[01:52] <wgrant> So outdated codehosting code shouldn't matter.
[01:52] <wgrant> Right?
[01:52] <sinzui> I don't think that is a big concern at the moment
[01:52] <wgrant> Hm?
[02:14] <lifeless> nigelb: :( customs?immigration? job?
[02:14] <nigelb> lifeless: immigration
[02:15] <lifeless> nigelb: ugh :(
[02:15] <lifeless> nigelb: did they say why?
[02:16] <nigelb> Mostly, they couldn't get in touch with my current or previous employer.
[02:16] <nigelb> I don't even know what they tried.
[02:17] <nigelb> I have to wait till Mountain View wakes up to figure out what's next.
[02:23] <lifeless> ah, thats very interesting.
[02:23] <lifeless> would they reevaluate if they could ?
[02:24] <nigelb> I suppose they willl.
[02:24] <sinzui> nigelb, Sorry to hear that. The US Government makes it very hard to hide talent. They make companies and hiring managers to a lot of work because the US doesn't believe skill is needed for any job
[02:24] <nigelb> sinzui: This is NZ, not US
[02:24] <lifeless> sinzui: s/hide/hire/ ?
[02:24] <sinzui> NZ is better I think/
[02:24] <sinzui> yes hire
[02:25] <sinzui> though I did think to try and hide on H1 visa employee I had. I hate to see good people go
[02:28] <ajmitch> nigelb: I hope you can reapply & get approved
[02:45] <nigelb> ajmitch: yeah, I hope so too.
[04:05] <wgrant> wallyworld_: There were disturbingly few failures from the trigger removal. I've sent it back after fixing them.
[04:06] <StevenK> How few, out of interest?
[04:06] <wgrant> 23
[05:22] <StevenK> wgrant: I don't get why I get test hangs with celery :-(
[05:22] <StevenK> I ran bin/test under strace after lunch and I'm still waiting for my eyes to stop bleeding.
[05:23] <wgrant> Yeah, it happens. Make sure all existing test processes, rabbits, etc. are dead.
[05:24] <lifeless> after 2 hours bleeding you'd hope they are already.
[05:26] <StevenK> wgrant: http://pastebin.ubuntu.com/1083954/ it doesn't hang with the diff as it is, but if you uncomment the branch_invisible_filter line or the branch unsubscribe loop in run(), it does
[05:32] <nigelb> Every time I'm about to expire from a team, I twich at the blatant lies the email says.
[05:32] <StevenK> nigelb: You and me both.
[05:33] <nigelb> StevenK: How fugly is it to fix?
[05:33] <nigelb> I might be tempted to spin LP up in a VM for this :/
[05:34] <StevenK> I keep getting stuck on wording
[05:35] <nigelb> My problem is the "One more email" bit.
[05:35] <lifeless> of course, the real irony here is that noone except a few team admins value that setup at all.
[05:36] <wgrant> lifeless: Yeah, our new team creation forms go to some lengths to hide that, and we're removing the autorenew option because it's stupid.
[05:37] <wgrant> We can't really remove the self-renewing, but we're going to conceal it.
[05:37] <lifeless> autorenew is == nonexpiry
[05:37] <lifeless> so I'm +1 on that
[05:37] <nigelb> I also want to say "I know. Keep quiet".
[05:37] <lifeless> self renewing, I'm fairly sure a rigorous set of interviews with stakeholders and users could remove
[05:38] <lifeless> I'm glad you're going to make it less prominent in the short term
[05:38] <wgrant> lifeless: Ubuntu uses it extensively to try to prevent teams from accumulating cruft.
[05:38] <lifeless> yes
[05:38] <lifeless> Does it work?
[05:38] <nigelb> no :P
[05:39] <wgrant> Apparently it does :)
[05:39] <nigelb> ...
[05:39] <StevenK> Haha, you told him
[05:40] <nigelb> Did I give him bad news?
[05:40] <nigelb> Destroy his dream?
[05:40] <nigelb> :P
[05:41] <StevenK> That's okay, I'm sure he can come up with a good dream, half implement it and then have wgrant rewrite it.
[05:41] <nigelb> Or have you delete it.
[05:51] <spm> heya nigelb
[05:57] <StevenK> wgrant: Sigh, http://pastebin.ubuntu.com/1083975/ is the actual diff
[06:14] <StevenK> wgrant: So, those hangs were caused by one typo, and one referencing person_id instead of personID.
[06:14] <StevenK>  /wrists
[06:19] <wgrant> StevenK: Yeah
[06:19] <wgrant> StevenK: This is the problem with running all the tests under celery
[06:19] <wgrant> When we really should have more direct tests that are about 20x faster.
[06:20] <StevenK> wgrant: So, I wanted to chat about it, anyway.
[06:26] <StevenK> wgrant: I wonder if would be worth trying to nail the two queries into one
[06:29] <wgrant> StevenK: Bugs and branches?
[06:29] <StevenK> wgrant: Yeah
[06:34] <wgrant> StevenK: You like to tempt fate.
[06:34] <wgrant> I would not.
[06:34] <StevenK> wgrant: Heh
[06:34] <StevenK> wgrant: I can put up this branch at it stands then, and you can see if you vomit?
[06:35] <wgrant> Yes.
[06:35] <wgrant> I shall bring my own bucket.
[06:35] <StevenK> Is it a red bucket?
[06:44] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/teach-rasj-about-branches/+merge/114106
[06:58] <wgrant> stop-using-legacy-bug-access => devel     [OK]       (up for 2:52:53) i-86ee57fe
[06:58] <wgrant> remove-legacy-bug-access => devel         [OK]       (up for 2:42:03) i-9ed069e6
[06:58] <wgrant> Yay
[06:58] <nigelb> hey spm :)
[06:59] <StevenK> wgrant: http://pastebin.ubuntu.com/1084037/ per your comments
[07:05] <wgrant> StevenK: Sounds good.
[07:09] <StevenK> lifeless: Last use of IMemcacheClient is about to get removed.
[07:10] <StevenK> I'm not sure if we want to drop it completly or keep it around for garbo jobs.
[07:11] <wgrant> Keep it for now.
[07:11] <wgrant> The infrastructure which allowed people to easily abuse it is gone, so it's not particularly problematic to keep the remains around.
[07:12] <StevenK> Until we have something else, I guess.
[07:12] <wgrant> We'll hopefully make use of memcache soon
[07:12] <wgrant> Just in ways that make sense.
[07:13] <StevenK> To do what?
[07:13] <wgrant> Make pages lightning fast, perhaps.
[07:13] <wgrant> Previously we've used memcache to take pages from timing out to near-acceptably glacial.
[07:13] <wgrant> With no invalidation
[07:14] <wgrant> Which is a completely inappropriate way to use it
[07:14] <wgrant> memcached it great at taking pages from fast to lightning fast.
[07:14] <wgrant> s/it/is/
[07:14]  * StevenK races ec2
[07:17] <StevenK> I wish AWS could start up an instance in a few seconds rather than ~ 180
[07:19] <lifeless> StevenK: e.g. load in json blobs for entire pages, then do cheap business logic validation and render
[07:20] <StevenK> lifeless: Right, so your plans to rip it out completly are on ice?
[07:26] <lifeless> StevenK: I never put forward such a plan
[07:27] <lifeless> StevenK: I ripped out inappropriate uses
[07:27] <lifeless> StevenK: and said we definitely couldn't remove it entirely while it was used
[07:27] <lifeless> but also that there are valid uses for it still
[07:27] <lifeless> I guess that that is ambiguous at best
[07:27] <lifeless> sinzui has expressed interest in entirely destroying it ;)
[07:49] <adeuring> good morning
[08:16] <stub> wgrant: The problem we found with memcache is there isn't much we can actually cache in the webapp.
[08:29] <cjwatson> wgrant: what did you mean by "it'd be nice if we could fix BPB/PU permissions around copies"?
[08:31] <cjwatson> BPBs seem to already have code to check whether SPRs have been copied to a visible archive, much as I just added to PU
[08:35] <wgrant> stub: There's *tonnes*. Just not in the way we were doing it before.
[08:35] <wgrant> cjwatson: Right, what you've done with PU matches BPB
[08:35] <wgrant> cjwatson: So it'll do for now. But neither is correct.
[10:45] <danilos> gmb, hi, I've got a simple MP up (that's fixing qa-bad landing and thus includes the full patch that's already been reviewed): https://code.launchpad.net/~danilo/launchpad/bug-1021196/+merge/114157
[10:45] <gmb> danilos, Looking.
[10:46] <danilos> gmb, thanks, fwiw, new changes up at https://pastebin.canonical.com/69731/
[10:47] <gmb> danilos, Approved.
[10:48] <danilos> gmb, woohoo, thanks
[10:50] <jam> dpm: did anything happen with opening Q translations?
[10:51] <danilos> gmb, do I need to use any special flags to mark the rollback as cleaned up now? (I assume not, but just checking)
[10:51] <gmb> danilos, Not that I know of, no.
[10:51] <danilos> cool, thanks
[10:53] <dpm> hi jam, sorry for not having come back to you earlier. I'm working on the Ubuntu app showdown right now, and I won't be able to do any translations work until the end of the week. Let me come back to you then.
[10:53] <jam> dpm: is there another Ubuntu-translations person? or is it just you?
[10:53] <jam> (*I* don't need them open, but I figure Ubuntu wants to open them pretty soon)
[10:53] <jam> Also, I was wondering if we would want to re-start the copy in case there were new P translations since we did the copy
[10:55] <dpm> jam, anyone in the ubuntu-translations-coordinators team can open translations, but there are still some pending tasks, such as manually approving/blocking templates in the Q translations imports queue that should probably be done. I listed them on the e-mail I sent to the ubuntu-translations-coordinators list and to you before going on holiday
[10:58] <dpm> but I'll get onto it by the end of the week, it's just that the app showdown is a higher priority right now
[11:59] <jam> cjwatson: any chance you got to the filenames list yet?
[12:03] <cjwatson> sorry, not quite yet
[13:19] <Laney> how hard is the LoC delta policy? :-)
[13:58] <rick_h_> Laney: pretty solid, no credit atm?
[13:59] <Laney> rick_h_: nah, this is only my second change. But it's only 30 lines. I'll find it.
[13:59] <rick_h_> Laney: ok thanks. Looks ok otherwise, left a comment on your question on the test adjustment
[13:59] <Laney> ty
[14:01] <nigelb> Laney: ONE OF US, ONE OF US!
[14:01] <nigelb> :D
[14:01] <nigelb> (though I haven't written LP code in a long time:( )
[14:03] <Laney> DO IT.
[14:10] <cjwatson> kill off another doctest, that's usually a good candidate :)
[14:10] <cjwatson> (where by "kill off" I mean "rewrite as unit tests" of course)
[15:28] <cjwatson> wgrant: I think it should be safe to look at https://code.launchpad.net/~cjwatson/launchpad/queue-api-fix-urls/+merge/113776 again now; the PU security changes have landed and I fixed it to redirect everything through the webapp
[15:47] <jcsackett> rick_h_, gmb: can one of you look at https://code.launchpad.net/~jcsackett/launchpad/hidden-comment-count-error/+merge/114219
[15:49] <rick_h_> jcsackett: loading up
[15:49] <jcsackett> thanks.
[16:02] <rick_h_> jcsackett: so this /home/jc/wtf.html file?
[16:02] <rick_h_> just for sanity checking hte tests?
[16:03] <jcsackett> rick_h_: that should have been deleted. :-P
[16:03] <jcsackett> one sec.
[16:04] <jcsackett> ok, that line is gone and change is being pushed. sorry, rick_h_.
[16:04] <rick_h_> jcsackett: k, also, not following the out of order logic. Shouldn't it be adding more time to the hidden comment so that the last one comes before it?
[16:04] <jcsackett> no, because i need a visible comment before and after the hidden, but i need to move the hidden out of it's usual spot.
[16:05] <jcsackett> so, i push the last one way out and push the hidden one slightly less out in time.
[16:05] <rick_h_> ah ok, so the moving of the last assures there's something after, so this should drop -5 to -2 and the last stays last
[16:05] <jcsackett> yup.
[16:05] <rick_h_> ok gotcha.
[16:06] <rick_h_> jcsackett: ok r=me
[16:06] <jcsackett> thanks, rick_h_.
[17:30] <jcsackett> sinzui: free to talk a bit?
[17:34] <sinzui> jcsackett, in a few minutes yes, and only for a few minutes
[17:34]  * sinzui has meeting at 2:00
[17:34] <jcsackett> dig. i'll get on google+ post haste.
[17:36] <jcsackett> sinzui: set up. you get the invite?
[17:42]  * cjwatson wonders what the record is for number of branches attached to a single-task bug
[19:31] <lifeless> cjwatson: what ever it is, its too many.
[21:32] <maxb> Could a member of ~launchpad nuke ~registry's bugmail subscription here, please: https://bugs.launchpad.net/unity-foundations/+subscriptions
[21:33]  * sinzui loos
[21:33] <sinzui> looks
[21:34] <maxb> ugh, actually, there are lots
[21:34] <maxb> https://bugs.launchpad.net/~registry/+structural-subscriptions
[21:34] <sinzui> holy #!!!
[21:35] <sinzui> someone deleted a team without cleanup
[21:38] <sinzui> The subscriptions are gone.
[21:39] <sinzui> I am checking branches
[22:17] <sinzui> http://people.canonical.com/~curtis/lp-milestone/report.html
[22:18] <sinzui> ^ StevenK, wgrant, wallyworld_
[23:48] <wgrant> StevenK: You'll want to confirm that with jtv.
[23:49] <StevenK> Yeah
[23:49] <StevenK> wgrant: I'm looking at https://code.launchpad.net/~cjwatson/launchpad/queue-api-fix-urls/+merge/113776 , and I'm concerned by lines 215-219
[23:53] <wgrant> StevenK: I haven't looked at the recent changes in that branch yet. Let me see.
[23:53] <wgrant> win 45
[23:54] <wgrant> StevenK: What worries you?
[23:55] <wgrant> StevenK: It looks pretty good to me.
[23:55] <wgrant> The lack of traversal to binaries is odd, but understandable.
[23:55] <StevenK> wgrant: SPRF.one()
[23:56] <wgrant> StevenK: It's restricting by filename
[23:56] <wgrant> We have bigger problems if there's multiple SPRFs in one SPR with the same filename.
[23:56] <StevenK> Ah