[00:34] <babbageclunk> externalreality: around for standup?
[00:53] <externalreality> babbageclunk, I sorry was in the kitchen and totally lost track.
[00:53] <babbageclunk> externalreality: no worries, it was pretty low key.
[00:54] <babbageclunk> externalreality: did you make progress on your action change?
[00:55] <externalreality> babbageclunk, I was making track but @thumper gave me an idea that I hadn't thought of which makes the code a lot simpler.
[00:55] <externalreality> Let the uniter grab the machine lock and then when it finds out that it doesn't need the lock, it simply releases it.
[00:55] <externalreality> wtf didn't I think of that
[00:57] <externalreality> I was attempting to restructure the code so that the uniter had knowledge of whether or not it needed the machine lock earlier
[00:57] <externalreality> the former suggestion is indeed much easier and works well
[01:09] <babbageclunk> externalreality: oh, nice
[02:10] <babbageclunk> menn0, thumper: could someone take a look at this? https://github.com/juju/juju/pull/7590
[02:11] <babbageclunk> It's not the full fix for the sub-sub thing, but it fixes the part that I broke in the first change.
[02:11]  * babbageclunk goes for a run while it's nice and sunny.
[03:56] <menn0> babbageclunk: looking
[03:57] <babbageclunk> menn0: thanks!
[04:07] <blahdeblah> Quick Q: If I run "juju config myapp --file path/to/myconfig.yaml" and myconfig.yaml only contains values for 1 config option and myapp has 10, will it wipe out the other 9, or just ignore them and set the 1?
[04:07] <blahdeblah> The docs on this aren't clear
[04:11] <menn0> blahdeblah: it just updates the configuration mentioned in the yaml file
[04:11] <menn0> blahdeblah: the others are left alone
[04:11] <menn0> I just tried it to be sure
[04:11] <blahdeblah> sweet - thanks menn0
[04:27] <menn0> babbageclunk: done
[04:28] <babbageclunk> menn0: cheers
[05:14] <babbageclunk> menn0: damn, I meant to have a discussion about whether we should be squashing commits before merging at the sprint. Do you know whether there's a consensus there?
[05:35] <menn0> babbageclunk: I'm pretty sure there will be widespread disagreement
[05:35] <babbageclunk> :(
[05:39] <menn0> babbageclunk: FWIW, I prefer a small number of logic commits per PR (something like 1 to 5), where each has a proper commit message
[05:39] <menn0> logical
[05:39] <menn0> I tend to do some squashing and commit message rewording before proposing a PR
[05:40] <menn0> my pet hate is seeing stupid commit messages that just say "now works" or "addressed review comments". it really sucks when you run into them while trying to diagnose an issue.
[05:42] <babbageclunk> menn0: I tend not to do the squashing until after review - otherwise you sometimes get someone saying "actually, could you split that one change out into a separate PR" and find that you've already squashed it.
[07:41] <mup> Bug #1702236 opened: Juju 1.25.10 is running hooks prior to additional network interfaces being up <canonical-bootstack> <juju-core:New> <https://launchpad.net/bugs/1702236>
[21:35] <babbageclunk> Morning everyone!
[22:05] <thumper> morning]
[23:49] <babbageclunk> menn0, thumper: I felt bad about skipping a flaky test so I fixed another one: https://github.com/juju/juju/pull/7596