[03:07] <thumper> jam: ping?
[03:27] <jam> thumper: hey, today's a holiday in UAE, but if you have anything you want to cover quickly I'm willing to stop by for a bit
[03:27] <thumper> jam: no, we're all good
[05:06] <wallyworld> axw: i had to revert my attempt to use common hook command logic. but i did get to delete a tonne of unused code that i discovered https://github.com/juju/juju/pull/8155
[05:06] <axw> wallyworld: cool, looking
[05:08] <axw> wallyworld: seems unlikely that we'll be able to debug-hooks CAAS charms, so not sure about movint that one...
[05:09] <wallyworld> axw: why is that?
[05:10] <axw> wallyworld: are we going to run sshd in the operator pod? manage keys in it?
[05:11] <wallyworld> i was thinking we would eventually
[05:11] <axw> hmmmm
[05:11] <wallyworld> maybe we can leave things as is until then
[05:11] <wallyworld> in case we don't get to it
[05:11] <axw> wallyworld: yeah, I think just leave it where it is, and move it if we do do that
[05:11] <wallyworld> ok, sgtm
[05:12] <wallyworld> i have to pop out for a bit, will revert the move of those when i get back
[05:18] <axw> oh god, I'd forgotten about that peeker stuff
[05:18]  * axw has flashbacks
[05:51] <axw> wallyworld: can/should we backport https://github.com/juju/juju/pull/8149 to the 2.3 branch? it has some changes to non-test code
[05:52] <axw> wallyworld: same q for https://github.com/juju/juju/pull/8154, though that is a test-only change
[06:44] <wallyworld> axw: yeah, i was thinking we might. we will have a 3..1 soon enough
[06:44] <wallyworld> 2.3.1
[06:47] <axw> wallyworld: in that case, here's something I prepared earlier: https://github.com/juju/juju/pull/8157
[06:48] <wallyworld> axw: great. it looks the same as the devel pr from memory, so should be good to go
[06:49] <axw> wallyworld: actually can you please take a look at https://github.com/juju/juju/pull/8156, then I'll add it to the backport to save another PR
[06:49] <wallyworld> just looking at that one
[06:49] <axw> ta
[06:50] <wallyworld> axw: yeah, lgtm, thanks for fixing
[06:51] <axw> wallyworld: np. for CAAS I'm currently adding app/charm config watching to the API and worker
[06:51] <axw> then we can do the initial config-changed hook once we've got hook bits
[06:52] <wallyworld> sounds good. i need to get this runner shite sorted before we can move much more on the other front
[06:52] <axw> yup
[06:52] <wallyworld> we can also look to add the container spec data model
[06:52] <wallyworld> i'm thinking a separate collection, keyed on application global id
[06:52] <wallyworld> and a watcher etc
[06:53] <wallyworld> so provisioner can watch for changes and react acccordingly
[07:01] <wallyworld> axw: those other changes pushed, so when you get a chance, PTAL
[07:03] <axw> wallyworld: LGTM
[07:03] <wallyworld> ta
[12:31] <rogpeppe1> anyone know what might be going on here?
[12:31] <rogpeppe1> % juju destroy-controller -y jaas-local
[12:31] <rogpeppe1> ERROR getting controller environ: validating cloud spec: "empty" auth-type not supported
[12:32] <rogpeppe1> i can talk to the controller OK
[12:32] <rogpeppe1> i just wanna remove it...
[12:41] <rogpeppe1> ah, looks like fff2d6ff710ead222741eb61b7fbee3510eab769 broke it
[12:41] <rogpeppe1> axw: ping
[12:41] <rogpeppe1> (unlikely!)
[12:45] <rogpeppe1> in case anyone sees the above later, I worked around it by commenting out the test in the source code and rebuilding juju: http://paste.ubuntu.com/26080377/