[20:03] wgrant: regarding turnip configuration, do we want to be running turnip with twistd and a tac? [20:45] turnip ? [21:00] lifeless: some sort of root vegetable :P [21:03] deploying on vegetables is more sustainable. [21:10] but nowhere near as healthy :) [22:22] blr: Most of our other Twisted services do that. twistd can be a little limiting at times, but I don't envisage problems in this case. [22:23] I suspect/hope twistd will die out as sensible init systems become universal. [22:23] But for now it's probably a good idea. [22:25] wgrant: cool, will do that then. [22:25] wgrant: good weekend? [22:26] blr: The aircon died, but otherwise pretty good :) [22:26] You? [22:28] ugh, imagine that's not much fun. Got the winter vegetables in, alas, no turnips. [22:28] Heh [22:30] wgrant: hoping to have a chat with pitti today about his experiences with nova-compute-lxc - thomi mentioned he posted to canonical-tech last year while working on the uci-engine. [22:30] well, this evening more likely, given timezones. [22:31] blr: If you're wanting to test deployment on OpenStack, I'd get creds to our bootstack and use the real thing. [22:31] The standard for sensible juju work nowadays is local provider for most development and testing, then real OpenStack when you want to do more. [22:33] wgrant: okay, just thought it sounded appealing provisioning the entire stack locally - although at a guess, it would be painful. [22:36] blr: Oh yeah, it's certainly appealing. But I'm not sure it's worth the pain. [22:40] Ah, found the buildd-manager double-upload bug. [22:40] Yay [22:44] https://bugs.launchpad.net/launchpad/+bug/1422199 [22:44] Bug #1422199: buildd-manager can repeatedly upload builds if handleStatus fails in the wrong place [22:48] wgrant: good morning =) how is monday looking so far? [22:50] xnox: Rather good. [22:50] xnox: Well done on the launchpadlib py3k port. [22:50] wgrant: tah. Well, pitti only found half a dozen bugs in it, when he tried to port apport. [22:51] mostly around marshalling binary data around. [22:51] Heh, of course. [22:51] * xnox can't run full test-suite, as i didn't port _all_ the testsuite deps - which is like the rest of lazr.* [22:51] Yeah. Would be nice if it could run against an external lazr.restful. [22:51] plus i'm not sure test-suites would have helped here. Cause i had to port them to str/bytes world view as well. [22:52] yeah, but too crazy for me to implement =) [22:52] i find pitti to be a better testsuite. [22:52] * xnox hides