[10:47] <stickupkid> achilleasa, https://github.com/juju/juju/pull/12253
[11:22] <achilleasa> stickupkid: what's up with the linter in ^
[11:23] <stickupkid> achilleasa, ah, it's because the first argument to printf is a var
[11:23] <stickupkid> i'll fix
[11:40] <stickupkid> achilleasa, fixed
[12:13] <stickupkid> achilleasa, I think you're right here, I might be worth ammending the patch of #12250 with your comment
[12:13] <mup> Bug #12250: Latest evolution not installable due to libcamel problems. <evolution (Ubuntu):Fix Released by seb128> <https://launchpad.net/bugs/12250>
[12:13] <stickupkid> bad bot
[12:19] <achilleasa> stickupkid: I will open a PR against 2.8 with the wait-for-service-files-to-appear helper and then forward port
[12:19] <stickupkid> wicked
[12:19] <achilleasa> there seems to be a similar issue on 2.8 as well
[12:19] <stickupkid> I didn't see your comment until I merged it
[12:19] <achilleasa> https://jenkins.juju.canonical.com/job/test-hooks-lxd/446/console: Expected one remaining reboot monitor state file after subordinate removal; got 2
[12:20] <achilleasa> removal is async (done by the deployer worker) so it may take a few moments for the file to be removed
[12:31] <achilleasa> stickupkid: got a question on 12253
[13:29] <stickupkid> achilleasa, or we drop the first line from the output?
[13:29] <stickupkid> I'm not really sure if series is valuable to check here
[13:30] <achilleasa> stickupkid: either works. This test is only interested in the overlay bits so we could cut on '---' and compare the second half
[13:30] <stickupkid> that is not a bad idea
[14:38] <stickupkid> achilleasa, updated the pr...
[14:45] <achilleasa> stickupkid: what is \f? vertical tab?
[14:45] <stickupkid> form feed
[14:45] <stickupkid> i'm abusing sed here
[14:45] <stickupkid> i.e. sed doesn't work over multiple lines, so force \n to be \f and then remove and revert
[14:47] <achilleasa> hmmm I am pretty sure it can work in multi-line mode (you can do a match and print that works across lines).
[14:47] <stickupkid> yeah, but that's a hack also
[14:47] <stickupkid> this one's not as bad imo
[14:50] <achilleasa> stickupkid: you can simplify: pipe the bundle to: "grep 'overlay.yaml' -A 200 | tail +1"
[14:51] <achilleasa> :D
[14:52] <stickupkid> achilleasa, mine ensures that you always get the right overlay, even if the bundle output changes
[14:53] <achilleasa> this one still does. it is basically a split on the separator (you can replace the grep with "---" to be more consistent) and then prints everything below that
[14:54] <achilleasa> I don't think we will ever want to output multiple overlays so that should always work, right?
[14:59] <stickupkid> achilleasa, k
[14:59] <achilleasa> stickupkid: if you still want sed though:  "| sed -n '/---/,$p' | tail +1"
[14:59] <achilleasa> this one is even more generic
[14:59] <stickupkid> the latter one is better
[22:28] <erik_elox> HOw do I get the unit public ip address using the operator framework? (self.unit.   ????)
[23:56] <wallyworld> erik_elox: you want to use the "network-get" information. See the Network class in https://github.com/canonical/operator/blob/master/ops/model.py
[23:56] <wallyworld> public ip as a concenpt is deprecated
[23:56] <wallyworld> network-get will give you ingress addresses
[23:58] <erik_elox> On AWS I get two possible addresses: (172.x.y.z) and internet facing addresses (typically something like 35.x.y.z) I need to access the former (35.x.y.z)
[23:59] <erik_elox> ... and I still don't understand from the model.py how to access the unit/host "Network" in my code.
[23:59] <wallyworld> you then need to use spaces and endpoint bindings to model what you want to do