[01:26] <StevenK> wgrant: https://code.launchpad.net/~stevenk/launchpad/use-timeout-for-checkwatches/+merge/147586
[01:41] <wgrant> StevenK: Have you tested locally that it actually works?
[01:42] <StevenK> Haha
[01:43] <StevenK> You tell funny jokes
[01:43] <StevenK> Now I have to remember how to do that
[01:49] <StevenK> wgrant: 2013-02-11 01:48:53 INFO    Updating 1 watches for 1 bugs on http://bugzilla.abisource.com
[01:53] <wgrant> StevenK: But does the timeout work?
[01:56] <StevenK> wgrant: I've also discovered mantis and trac also call their own urllib2.open, so I've changed them too
[01:56] <StevenK> I'll check the timeout itself after lunch
[01:56] <wgrant> Great
[02:19] <StevenK> 2013-02-11 02:19:13 INFO    Error updating http://bugzilla.abisource.com/: http://bugzilla.abisource.com: <urlopen error timed out>
[02:24] <StevenK> 2013-02-11 02:23:26 INFO    Updating 1 watches for 1 bugs on http://tracker.ardour.org
[02:24] <StevenK> 2013-02-11 02:23:56 INFO    Error updating http://tracker.ardour.org: http://tracker.ardour.org: <urlopen error timed out>
[02:24] <StevenK> (ardour is Mantis)
[03:17] <StevenK> wgrant: Are you happy enough with that?
[03:17] <StevenK> That's XMLRPC and Mantis
[03:17] <StevenK> I can dig for Trac and old Bugzilla if you wish
[03:19] <wgrant> StevenK: Seems fine to me
[03:19] <wgrant> No further changes?
[03:19] <StevenK> Let me push them up
[03:20] <StevenK> I'm tempted to refactor Mantis and Trac so there is one callsite for urlopen in base
[03:21] <wgrant> That seems like a good idea, and it shouldn't be too hard
[03:22] <StevenK> Current idea: Set a self.url_opener on both Mantis and Trac and teach urlopen in the base class to use it if it's set
[03:23] <StevenK> For bonus points, set a variable to the function and call it as a function
[03:35] <StevenK> wgrant: The MP is updated
[03:37] <wgrant> StevenK: Can you just default url_opener to urllib2.urlopen?
[03:38] <StevenK> How does that help? I can't call .open on that?
[03:39] <wgrant> Ah, true.
[03:39] <wgrant> r=me
[03:59] <wgrant> StevenK: You lose!
[04:01] <StevenK> Blink
[04:01] <StevenK> I hate the checkwatches tests
[04:08] <StevenK> -        def raise404(request, data):
[04:08] <StevenK> +        def raise404(request, data, timeout=None):
[05:10]  * StevenK peers at these branch vocab tests
[05:10] <StevenK> There is a test to make sure a vocab doesn't return branches owned by a OPEN or DELEGATED team
[05:34] <wgrant> StevenK: That sounds like the series branch picker
[05:38] <StevenK> Which is on product
[05:39] <StevenK> Since that deals with IProduct or IProductSeries
[05:54] <StevenK> Right, thanks to some hideous code, the branch vocab tests pass
[05:54] <StevenK> Except my last lot of refactoring broke them
[06:09] <StevenK> wgrant: OMG, these changes work
[06:10] <wgrant> :)
[06:12] <StevenK>  6 files changed, 196 insertions(+), 199 deletions(-)
[06:17]  * StevenK accidently deletes IBranchCollection.search
[06:17] <StevenK>  6 files changed, 197 insertions(+), 445 deletions(-)
[08:53] <adeuring> good morning
[12:20] <wgrant> StevenK: Oh
[12:20] <wgrant> StevenK: You still lose at buildbot
[12:20] <wgrant> StevenK: Look at those test counts
[12:21] <wgrant> There's one very long error which apparently breaks the subunit stream
[12:21] <wgrant> lp.bugs.tests.test_bugtracker.TestMantis and checkwatches.txt, from my ec2 run
[12:21] <StevenK> Bleh
[12:21] <StevenK> I did wonder why the test count was quite low
[12:23] <StevenK> wgrant: So I managed to destroy another ec2 run of yours?
[12:25] <wgrant> StevenK: Yup
[12:25] <wgrant> And I've only run two batches of them in the last month...
[12:25] <wgrant> you sabotaged both!
[12:28]  * StevenK adds another stroke to his "wgrant's ec2 runs I've doomed by accident" board
[12:29] <wgrant> Lies
[12:33] <StevenK> Probably shouldn't be too bad to sort out
[12:43] <StevenK> wgrant: Testfix landing.
[12:48] <wgrant> StevenK: Thanks