[01:01] <bigjools> lazyPower: configure the existing proxy ACLs on the region
[03:38] <jtv> Why is authbind documented as required, but not listed among the required packages!?
[03:41] <jtv> Oh it's not documented as required, things just break when you don't have it.
[03:47] <jtv> ...And seemingly only less often when you do install it.  What gives?
[04:27] <bigjools> jtv: oh dear, packaging problem or dev deps?
[04:36] <jtv> This is in a branch.
[05:01] <lazyPower> bigjools: well i think what fixed it was adding ppa.ubuntu.com to the squid-deb-proxy config
[05:01] <lazyPower> i didn't know that squid-deb-proxy needed help out of the box to do ppa installations. hazmat turned me on to this fix
[05:02] <bigjools> lazyPower: yeah, it's not configured for ppas by default, I reckon it should be.
[05:02] <lazyPower> i think its a good default configuration ootb - but it wasn't readily apparent to me what needed to happen to fix it
[05:02] <lazyPower> i was just seeing the symptom, and didnt understand the fix was to hax on squid
[05:02] <lazyPower> i blamed maas's templates as that was the workaround I had discovered in the hot seat to fix it
[05:02] <lazyPower> but that's clearly not ideal
[05:03] <bigjools> jtv: authbind is a dependency of the cluster controller in the package
[05:04] <jtv> bigjools: if it helps development though, it probably ought to be in required-packages.  Does it?
[05:05] <bigjools> jtv: yeah prob should be in there
[05:05] <bigjools> I rarely run from dev these days since I made package building easier
[05:19] <jtv> bigjools: cheap karma available — https://code.launchpad.net/~jtv/maas/dev-authbind/+merge/231669 & https://code.launchpad.net/~jtv/maas/bug-1359517/+merge/231670
[05:20] <bigjools> jtv: BASH
[05:20] <bigjools> whyyyyyyy
[05:21] <jtv> Because tcsh is sooo 20th-century.
[05:21] <bigjools> jtv: does it print line numbers out or any context?
[05:21] <jtv> (Sorry.  Unix joke.  Most people here are probably too young to understand it, or too experienced to find it funny.)
[05:21] <bigjools> I am in the latter camp
[05:21] <jtv> bigjools: it shows the filename and the line of code.
[05:22] <bigjools> can you put a -n in
[05:22] <jtv> Probably.
[05:22] <bigjools> line num is very useful
[05:22] <jtv> I shall strive to build this new feature.  Just for you.
[05:23] <bigjools> jtv: I was thinking
[05:23] <bigjools> we could have just overridden exception to call error in the logger
[05:24] <jtv> Bad boy.  For penance, type "exit" in the REPL.
[05:51] <bigjools> jtv: grah, there's a bad problem in packaging and I cannot work out WTF is going wrong.  When uprading, it often leaves the cluster celery down.
[05:51] <bigjools> I thought it was a race when rabbit is restarted, but I took that out of the equation and it still screws up
[06:19] <bigjools> jtv: https://code.launchpad.net/~julian-edwards/maas/fix-migration-error/+merge/231674
[06:20] <bigjools> I don't have a test - only noticed the problem when buggering around with migrations but I've no idea how to recreate the IntegrityError since I can't see how it would happen, yet it does!
[06:57] <lifeless> bigjools: you could run all but one then deliberately mess up the schema
[07:08] <bigjools> lifeless: it's not a bug in the schemas, it's a data migration
[07:08] <bigjools> and hello!
[07:51] <lifeless> bigjools: hi :)
[07:51] <lifeless> bigjools: so just saying that you can fiddle in the middle to make your test have no room to wiggle
[08:25] <bigjools> lifeless: you're a poet and you don't know it
[13:12] <jtv> blake_r: region-image-downloader approved with comments.
[13:12] <blake_r> jtv: thanks
[14:47] <roaksoax> blake_r: nice! so I also saw there's a form.. is it usable? :)
[14:48] <blake_r> roaksoax: next branch will enable it for the api
[14:48] <blake_r> roaksoax: then on to the view
[14:48] <roaksoax> blake_r: awesome!