[00:08] <wallyworld> hpidcock: lgtm, ta
[00:09] <wallyworld> thumper: do you know the status of the open PRs for branches etc - are they due to land prior to rc1 being built?
[00:10] <thumper> wallyworld: rick_h said that they had landed what needed to be landed and had a hash to use
[00:10] <thumper> so I'd say that nothing should land that isn't a bug fix
[00:10] <thumper> until we are past RC1
[00:11] <wallyworld> yup, no worries, just checking
[00:11] <wallyworld> we're just landing the race fix and that's it from us
[02:21] <wallyworld> timClicks: fyi i updated k8s release info in https://discourse.jujucharms.com/t/new-features-and-changes-in-juju-2-7/2268
[02:21] <wallyworld> anastasia will need to do some cloud updates
[02:22] <wallyworld> as the examples are now out of date
[02:22] <timClicks> wallyworld: awesome, thanks for adding details there
[02:22] <wallyworld> i also will add extra info the the referenced pod spec v2 post
[02:22] <wallyworld> as i need to add custom resources
[02:23] <wallyworld> there's a lot in this release
[02:23] <timClicks> I'll get the cloud stuff updated tomorrow
[02:23] <wallyworld> i also made a slight correction tojuju-run
[02:23] <wallyworld> k8s doesn't support juju ssh
[02:23] <wallyworld> you need to kubectl exec in directly
[02:23] <timClicks> ah okay, ty
[02:24] <wallyworld> not that hard to add but ran out of time :-(
[02:24] <wallyworld> you would think as a user it would be there
[02:24] <wallyworld> but not yet
[02:24] <wallyworld> there is a valid workaround k8s folks know how to use
[02:25]  * timClicks nods
[02:25] <timClicks> we could take out the help output and change it to something else..
[02:25] <timClicks> we could take out the help output and change the example to something else..
[02:26] <hpidcock> should we also update documentation re action id not being uuids? https://jaas.ai/docs/working-with-actions
[02:27] <timClicks> hpidcock: thanks for volunteering - scroll to the bottom of the page, click the link
[02:28] <hpidcock> hah should keep my mouth shut
[02:28] <timClicks> somewhat more seriously, if you would like to add a comment on that page i'll make sure that gets done
[02:29] <timClicks> I'm trying to block out some time this arvo for some other content
[02:29] <timClicks> i realise that it feels a little bit scary to directly edit the docs yourself
[02:29] <hpidcock> oh I'm happy to edit it
[02:30] <hpidcock> just a shame we don't version the documents by release.
[02:31] <timClicks> i think there's history there
[02:31] <thumper> hpidcock: we used to
[02:31] <thumper> hpidcock: but not any more
[02:47] <hpidcock> wallyworld: did we want to remove the developer-model feature flag requirement for k8s actions and exec?
[02:49] <wallyworld> hpidcock: ah bollocks, i didn't realise we still had that turned on
[02:50] <wallyworld> yeah, let's get up a pr
[02:50] <hpidcock> I'll make a PR?
[02:50] <hpidcock> kk
[02:58] <wallyworld> thumper: i have one to add for 2.7.1 https://github.com/juju/juju/pull/10848
[02:59] <wallyworld> thumper: also, we need to land this for rc1 as a feature flag was left in https://github.com/juju/juju/pull/10849
[03:00] <wallyworld> make you wonder how much non dev testing we got
[03:00] <thumper> wallyworld: ok
[03:00] <thumper> land it and update the email
[03:00] <wallyworld> i guess it's just for juju exec and most people would have tried juju call
[03:00] <hpidcock> wallyworld: https://github.com/juju/juju/pull/10849 I think it was just exec/run that was checking for the ff.
[03:00] <wallyworld> yeah
[03:00] <wallyworld> it was
[03:01] <hpidcock> so I'm ok to land?
[03:01] <wallyworld> hpidcock: yeah, i just hit $$merge$$
[03:01] <wallyworld> good catch
[03:02] <hpidcock> oops
[03:02] <hpidcock> double $$merge$$
[03:02] <wallyworld> it won't mind
[03:02] <hpidcock> I guess we are testing jenkins then ;)
[05:06] <wallyworld> jam: one thing - application relation data isn't propagated between models for cmr i don't think, so that will need to be a known linitation for 2.7.0
[05:06] <wallyworld> there's params structs that need updating, plus apis etc
[05:07] <jam> wallyworld: sure. we can target that as a 2.7.1 if you like.
[05:08] <wallyworld> i think we should, same with remaining model migration work etc
[05:26] <hpidcock> wallyworld: how do I upgrade an IAAS controller using a local jujud?
[05:26] <hpidcock> do I just pass --build-agent?
[05:38] <wallyworld> hpidcock: yup
[06:02] <hpidcock> wallyworld: I think I hit the bug kelvinli_ hit where the caas model is upgraded but the workloads aren't redeployed. I think the caasunitprovisioner's application worker doesn't fire when the model version changes, not sure what the solution to this would be. Possibly the caasunitprovisioner worker should pass a chan into the application worker so it can trigger on model version change?
[06:04] <hpidcock> I'll see if I can smash out a PR to fix it.
[06:05] <wallyworld> hpidcock: quick chat about it?
[06:05] <hpidcock> stdup
[06:28] <kelvinli_> yeah, we need to change the init container using the updated jujud version
[06:29] <kelvinli_> not sure if the init container will be re-run if the image path changed
[09:56] <stickupkid> manadart, https://github.com/juju/juju/pull/10844 Fancy giving this a review?
[09:56] <manadart> stickupkid: Sure.
[09:58] <stickupkid> ta
[10:37] <manadart> stickupkid: Couple of UX things.
[10:37] <stickupkid> noice
[10:38] <stickupkid> manadart, good shout on -n 0
[10:46] <nammn_de> stickupkid: wanna take a look at branch completion? Removed the Option flags and added `--branch < tab > --> branch  https://github.com/juju/juju/pull/10845
[10:47] <stickupkid> nammn_de, sure
[10:48] <nammn_de> stickupkid: and this one? https://github.com/juju/juju/pull/10836 its about the charm dep we recently changed. With this some tests are failing because how we place and version those charm tests. Not important for now but might wanna take a look and discuss
[10:48] <nammn_de> more information and this issue with suggestions how to solve. But not that important for now
[10:51] <nammn_de> stickupkid: ahhh  I need some feedback on this comment https://github.com/juju/juju/pull/10836#issuecomment-548414256
[10:51] <nammn_de> i reformated the comment, makes it more readable
[10:51] <stickupkid> ah ok
[10:53] <nammn_de> stickupkid: if read. Does it make sense to you?
[10:53] <stickupkid> nammn_de, thinking
[10:53] <stickupkid> i don't like the first or the last
[10:54] <stickupkid> the test sucks
[10:54] <stickupkid> that's what the test does
[10:54] <stickupkid> suck
[10:54] <nammn_de> stickupkid: haha
[10:54] <nammn_de> yeah
[10:54] <nammn_de> and additionally its difficult because that they are under git and using version regarding the charm thing. Makes all more complex as the tests in general already are :D
[10:55] <stickupkid> nammn_de, ho?
[10:58] <stickupkid> manadart, got a sec as well?
[10:58] <manadart> stickupkid: Yeah.
[11:06] <nammn_de> stickupkid: sure sorry wasnt here
[12:06] <stickupkid> manadart, updated my PR
[12:12] <manadart> stickupkid: OK, gimme 2.
[15:35] <nammn_de> stickupkid rick_h: coming back to the branch completion cli thingi https://github.com/juju/juju/pull/10845
[15:35] <nammn_de> Not even sure if we want those flags or not, just opened pr in case to have it
[15:36] <stickupkid> nammn_de, kewl
[15:41] <rick_h> nammn_de:  my feedback would be to make it looks like the rest of tab completion tbh
[15:44] <nammn_de> rick_h: IMO I think that its would fit that way more with the rest of the tabl completion. E.g. juju switch <tab> has no flags.  --controller, --unit ect. as well not.  BUT juju expose and other similiar commands have both. The "flags/option" and the values
[16:32] <hml> review anyone?  quick pr to fix nw-export-bundle test: https://github.com/juju/juju/pull/10852
[16:32] <hml> ci test
[17:46] <hml> stickupkid: ta!
[17:47] <nammn_de> am I looking at things wrong? This fails for me for now reason. https://github.com/juju/utils/blob/master/fs/copy.go#L27  Why do we expect there to be an error before we continue?
[17:47] <nammn_de> *no
[17:49] <nammn_de> ahh I see it creates the folder and does not want to exist, ufff
[17:54] <rick_h> guild 2.7 branch is created and pushed
[17:54] <rick_h> have fun!
[17:55] <rick_h> guild to be clear, that means develop is "unlocked". I'll send an email
[17:55] <stickupkid> we should retarget all our stuff to 2.7 right and merge forward to develop
[17:55] <stickupkid> rick_h, ^
[17:55] <rick_h> stickupkid:  hmmm...actually...so I was hoping to not touch the rc1 space in case we need a fix/release
[17:56] <stickupkid> rick_h, but at some point we'd need to back port?
[17:56] <rick_h> stickupkid:  yea, so thinking if the idea for now would be land on develop and keep track of what we need to backport
[17:56] <rick_h> after we cut .0
[17:56] <stickupkid> rick_h, sweet
[17:56] <rick_h> I'll put that in the email
[19:57] <thumper> morning
[23:45] <babbageclunk> anyone knowledgeable about shell quoting want to help me with a shell quoting problem?
[23:46] <babbageclunk> my heuristic of changing quotes and adding/removing backslashes has reached its limit
[23:54] <hpidcock> babbageclunk: is this for templating a script or are you trying to process arguments passed in?
[23:54] <wallyworld> need a +1 on version bump pr for 2.7 branch https://github.com/juju/juju/pull/10853
[23:55] <hpidcock> wallkworld done
[23:55] <hpidcock> wallyworld!
[23:55] <babbageclunk> hpidcock: I'm trying to use the wait_for function in a shell ci test
[23:55] <babbageclunk> wallyworld!
[23:56] <hpidcock> use heredoc?
[23:56] <wallyworld> ta
[23:57] <wallyworld> and for develop https://github.com/juju/juju/pull/10854
[23:57] <babbageclunk> the weird bit is that it works if I ask for the current workload status (in jq) but not if I ask for the message.
[23:58] <babbageclunk> doing my head in
[23:58] <hpidcock> can you paste it?
[23:58] <babbageclunk> hpidcock: jump in standup so I can talk you through it?
[23:59] <hpidcock> sure