[01:17] <wgrant> lifeless: "- Issue #7902: When using explicit relative import syntax, don't try implicit relative import semantics."
[01:17] <wgrant> lifeless: Changed in 2.6.6.
[01:18] <wgrant> lifeless: I presume that's it, but it doesn't really make sense...
[01:20] <wgrant> sinzui: Oh, nice.
[01:20] <wgrant> sinzui: I'm glad someone finally did that.
[01:20] <sinzui> wgrant, I am fixing the private-by-default case when bugs are created by email
[01:21] <wgrant> Ah
[03:20] <StevenK> wgrant: Did you see my branch of evil?
[03:38] <wgrant> StevenK: I did.
[03:38] <wgrant> lifeless: You fail at dropping that table.
[03:45] <StevenK> wgrant: And I don't get called a horrible person? :-)
[03:45] <wgrant> Only if you don't check the test counts afterwards.
[03:46] <StevenK> Ah yes, I will ec2 *test* it and compare the numbers to buildbot
[03:46] <wgrant> Compare the numbers with a previous ec2 run, preferably of the same rev. I don't trust buildbot to report the same numbers.
[03:47] <StevenK> Same base rev?
[03:47] <wgrant> Right.
[03:47] <wgrant> ie start an ec2 run of the branch and not the branch at the same time.
[03:47] <wgrant> compare numbers.
[03:47] <wgrant> profit
[03:47] <wgrant> It's going to be fairly easy for a single test to be missed here.
[03:47] <StevenK> Was exactly the plan
[03:48]  * StevenK pulls out r13677 of devel into a branch
[03:54]  * StevenK races ec2test
[04:23] <StevenK> Other than bin/ec ls lying, I have two instances running
[04:31] <LPCIBot> Yippie, build fixed!
[04:31] <LPCIBot> Project devel build #970: FIXED in 5 hr 51 min: https://lpci.wedontsleep.org/job/devel/970/
[05:04] <nigelb> Morning!
[06:43] <wgrant> nigelb: Finished Launchpad yet?
[06:44] <nigelb> wgrant: Wha?
[06:44] <nigelb> You mean finished the launchpad patch I was working on?
[06:44] <StevenK> Dear ec2. Please finish quicker.
[06:44] <wgrant> No, finished lots of Launchpad patches that make Launchpad finished.
[06:45] <nigelb> Hah. I'm not you. :D
[06:45] <StevenK> wgrant: And then what will you do when you're back from holidays? :-P
[06:46] <nigelb> Its a 3-day weekend for me, so I intend to finish what I was working on.
[06:46] <nigelb> I didn't realize being new to YUI would be slightly complicated :)
[06:46] <StevenK> nigelb: Did you see my mega-branch of evil?
[06:46] <wgrant> StevenK: Another three branches of bugtask cleanup!
[06:46] <wgrant> Then that saga should be over.
[06:46] <wgrant> After almost 15 branches.
[06:46] <StevenK> Heh
[06:47] <StevenK> wgrant: But, it's utterly awesome now!
[06:47] <StevenK> wgrant: Next, remove conjoinment
[06:47] <wgrant> Die.
[06:47] <StevenK> :-D
[06:47] <nigelb> StevenK: No! What does it do? Kick wgrant away? :P
[06:47] <StevenK> nigelb: https://code.launchpad.net/~stevenk/launchpad/kill-test-suite-dead/+merge/71353
[06:49] <nigelb> StevenK: All I have to say is what adeuring said "Blessed be those who remove obsolete code.
[06:49] <nigelb> "
[06:50] <nigelb> StevenK: Also, there can be no greater pleasure than seeing code removed
[07:13]  * StevenK tries to page his LaTeX skills back in
[07:58] <kb9vqf> Quick question: My poppy/ directory keeps filling up with files that are never purged
[07:59] <kb9vqf> I figured I could safely delete poppy/failed/* and poppy/rejected/*, but what about the files in poppy/accepted/*?
[07:59] <kb9vqf> can I delete them?
[07:59] <wgrant> kb9vqf: You can, yes.
[07:59] <kb9vqf> ok
[08:00]  * kb9vqf doesn't want to mess up his nicely running QuickBuild system :)
[08:00] <kb9vqf> wgrant: Just to make sure, I can completely purge poppy/rejected/* without affecting librarian or any other Launchpad systems?
[08:00] <wgrant> We delete everything after a week, just in case.
[08:01] <wgrant> Everything in the poppy upload dirs is unrequired once process-upload runs.
[08:01] <kb9vqf> darn copy + paste
[08:01] <kb9vqf> ok
[08:01] <kb9vqf> (obviously I meant poppy/accepted/* above)
[08:01] <wgrant> Yup.
[08:01] <kb9vqf> thanks!
[08:01] <kb9vqf> I was wondering where all my disk space was wandering off to
[08:02] <StevenK> Poppy keeps looging "OM NOM NOM"
[08:03] <StevenK> Sigh. scriptactivity for request_daily_build is still daily
[08:03] <StevenK> And still on loganberry
[08:03] <wgrant> StevenK: Is that a problem?
[08:03] <wgrant> It's probably the ackeeness that is a problem.
[08:04] <wgrant> I'm not sure if it was moved, though...
[08:04] <StevenK> It's part of bzrsyncd, I thought?
[08:04] <wgrant> Shouldn't be, but maybe.
[08:04] <StevenK> It runs every 15 minutes, so perhaps the checking should be more often
[08:04] <wgrant> So it is.
[08:04] <wgrant> It shouldn't need to be bzrsyncd, though.
[08:04] <StevenK> 15
[08:04] <StevenK> 15? 10? I don't remember.
[08:05] <wgrant> */15, yes.
[08:05] <StevenK> wgrant: I should get my answer from the two ec2 instances in the next few minutes ...
[08:06] <wgrant> You could also diff test lists without ec2, but I like to see if ec2 works too :)
[08:07] <StevenK> I have at least four test failures, too
[08:08] <StevenK> But it was better than the previous incantation of this branch, which had 500.
[08:08] <wgrant> Hah
[08:09] <StevenK> r13685 of devel => 15,607 tests.
[08:23] <StevenK> I am disappoint. 15,575 tests.
[08:23] <StevenK> But only 8 errors.
[08:25] <wgrant> StevenK: Try bin/test --list-tests and diff?
[08:26] <StevenK> I threw the subunit streams through subunit-ls
[08:26] <wgrant> Or that.
[08:26] <StevenK> Looks like test_bzrutils needs reverting, it repeats tests
[08:29] <StevenK> Before revert: Ran 10 tests with 0 failures and 4 errors in 0.909 seconds.
[08:29] <StevenK> After: Ran 46 tests with 0 failures and 0 errors in 4.804 seconds.
[08:30] <wgrant> Which is interesting, because that's 4 more tests than were missing.
[08:30] <StevenK> Heh, yes.
[08:31] <StevenK> The subunit streams didn't show tests which errored, either.
[08:34] <StevenK> wgrant: Right, another file reverted and it said 102 before, and 99 after.
[08:34] <StevenK> So now we're at -1
[08:35] <wgrant> --list-tests diff
[08:35] <StevenK> Which is canonical.launchpad.database.tests.test_oauth.BaseOAuthTestCase.test__get_store_should_return_the_main_master_store
[08:36] <StevenK> wgrant: http://pastebin.ubuntu.com/664825/
[08:40] <StevenK> wgrant: I think those tests are skipped/disabled by fiddling with the test loader.
[08:47] <StevenK> wgrant: Using --list-tests and diffing: http://pastebin.ubuntu.com/664832/
[08:47] <StevenK> The first hunk is a red herring
[08:47] <wgrant> Ah, and the rest are too.
[08:47] <StevenK> Really?
[08:48] <wgrant> They're all testrunner* config name differences, aren't they?
[08:48] <wgrant> Remember that each testrunner process creates a new testrunner and testrunner-appserver config with the PID in the name.
[08:48] <StevenK> Except that my branch has more of them
[08:48] <StevenK> Double, in fact
[08:48] <wgrant> Indeed, but that may just mean you ran and killed the testrunner badly.
[08:48] <StevenK> Which I probably did
[08:49] <StevenK> wgrant: So shall I commit, push and lp-land this monstrosity?
[08:49] <wgrant> If that's the whole diff now, I suspect so.
[08:49] <StevenK> That's the whole diff from --list-tests
[11:12] <wgrant> StevenK: You never said there was Perl involved!
[11:12] <wgrant> StevenK: This changes everything.
[13:30] <StevenK> wgrant: You didn't read the MP?
[13:33] <nigelb> heh
[13:34] <nigelb> All the more reasons to be "extra" evil
[13:34] <StevenK> nigelb: I pasted the *Perl script* in the description of the MP!
[13:39] <nigelb> StevenK: I saw :)
[13:40] <nigelb> StevenK: This shows how much attention wgrant paid to the MP.
[13:43] <StevenK> Haha
[13:45] <nigelb> Btw, you both are up late. On a Saturday! Why!
[13:47] <StevenK> I was entertaining people for dinner, had to clean up, and are now winding down
[13:49] <nigelb> :)
[19:48] <lifeless> wgrant: select * from pg_class where relname like '%lookup%';
[19:48] <lifeless>  entitlement_lookup_idx |         2200
[19:49] <lifeless> (1 row)
[22:25] <wgrant> StevenK: I don't tend to read MP descriptions until the end of the review, because it may influence how I treat the diff.
[22:27] <wgrant> lifeless: :( why is Ensemble treating LP like Ubuntu does -- its own namespace to stomp all over?
[23:08]  * jelmer waves to wgrant
[23:08] <wgrant> Morning jelmer.
[23:09] <jelmer> wgrant: How is Ubuntu stomping over Launchpad namespaces? Most things either seem to live under /ubuntu/ or have a ubuntu- prefix?
[23:10] <wgrant> jelmer: Nowadays they're pretty good, yeah.
[23:10] <wgrant> They used to not be :)
[23:10] <jelmer> hah, ok
[23:10] <wgrant> https://launchpad.net/~motu https://launchpad.net/~techboard are old examples.
[23:10] <wgrant> https://launchpad.net/~developer-membership-board is a new onew.
[23:11] <wgrant> https://launchpad.net/~mono
[23:11] <wgrant> That sort of thing.
[23:12] <jelmer> gotcha