[00:46] <hpidcock> thumper:  https://github.com/juju/juju/pull/10660 LGTM
[00:46] <thumper> hpidcock: thanks, still fixing some bugs in it though, just worker tests
[00:50] <hpidcock> happy to look at it again when your done, but worked well for me on lxd and k8s, though I accidentally deployed something into the controller model and didn't get a log file, which looks to be what you wanted
[02:54] <thumper> hpidcock: yeah, that was planned
[02:54] <thumper> I'm just sitting here a little confused
[02:54] <thumper> I had test failures which I didn't understand
[02:54] <thumper> it seems like the test is flakey
[02:54] <thumper> but I can't tell why
[02:54] <thumper> I don't like flakey tests
[02:55] <hpidcock> flakey tests make me sad
[02:55] <babbageclunk> I love 'em!
[02:55] <thumper> although the tests now pass...
[02:55] <hpidcock> babbageclunk: keeps things exciting I guess :P
[02:55] <babbageclunk> problem solved!
[02:55] <thumper> they are also insanely long tests
[02:56] <thumper> I think there may be a bug in the reporting of the tests
[02:56] <thumper> that would explain the weird test results
[02:56]  * thumper pokes
[03:10]  * thumper waits for things to pass
[04:38] <hpidcock> https://github.com/juju/testing/pull/147 small one for someone
[08:41] <nammn_de> someone want to take a quick look?  https://github.com/juju/charm/pull/293, one-liner
[08:49] <stickupkid> na
[08:49]  * stickupkid just fat fingered that
[08:50] <nammn_de> timClicks: thx for the quick look
[08:50] <timClicks> nammn_de: np only 21:50 here
[08:51] <stickupkid> jam, this is interesting, I was playing around with github actions for windows tests, but we fail because we don't have a windows 2019 server for the os/series package. That's fine, the issue is that the windows tests aren't run, we should fix that! https://github.com/juju/os/pull/13
[08:52] <jam> stickupkid: any other gaps than just 2019? There seems to be some big gaps in the dates
[08:52] <stickupkid> jam, i'm unsure, probably
[08:53] <stickupkid> jam, weirdly not - 2012, 2016, 2019
[08:53] <stickupkid> jam, seems like they have long LTS support
[08:54] <stickupkid> jam, they do have a lot of annual releases, but we probably don't support that directly
[08:54] <stickupkid> jam, this is my reference - not the best tbh - https://en.wikipedia.org/wiki/List_of_Microsoft_Windows_versions
[08:59] <ociuhandu> On windows different versions should be treated differently, i.e. one can not consider Windows 2019 Datacenter same with Windows 2019 Standard
[08:59] <ociuhandu> imho there should be different entries for those in the definition (e.g. win2019dc / win2019std / win2019hv and so on)
[09:01] <ociuhandu> also, there is windows 2019 (which is the LTS version) and there are the frequent releases that get only 6 months support and are numbered based on year and month of release (i.e. 1809 or 1903 and so on)
[09:03] <stickupkid> we should probably window the windows versions like the linux versions tbh
[09:04] <stickupkid> ociuhandu, we only use these for if we support the OS, we don't deploy to windows directly and only the juju client runs on the OS, so I don't think we need to change.
[10:00] <stickupkid> manadart, i'm back on cmr-migrations, i'm unsure about the attach method from your playground example
[10:00] <nammn_de> timClicks: still around?
[10:00] <stickupkid> manadart, it inverts the dependency which causes us not to be able to move the migration to a new folder
[10:01] <stickupkid> manadart, https://play.golang.org/p/GC1tbWd272e <- playground in question
[10:03] <manadart> stickupkid: Can we not export the parts that allow attachment?
[10:03] <stickupkid> manadart, this is the problem "func (m RemoteEntitiesMigration) Attach(mig *ExportStateMigration) {"
[10:04] <stickupkid> manadart, the ExportStateMigration knows too much info i.e. *State
[10:04] <stickupkid> manadart, if we make it an interface, that means we then have to make shims for complex types, which we may have too
[10:05] <stickupkid> manadart, let me have a think
[10:05] <stickupkid> manadart, we're sooooo close
[10:06] <manadart> stickupkid: OTP with Atos. Let's HO in a mo'.
[10:06] <stickupkid> manadart, sure
[10:57] <nammn_de> hey guys are we/the customers using `juju list-clouds --format yaml` in a automated way? Im thinking of adding an additional key on top for bug: https://bugs.launchpad.net/juju/+bug/1826957
[10:57] <mup> Bug #1826957: YAML produced by list-clouds is not acceptable to update-cloud <bitesize> <papercut> <update-clouds> <juju:Triaged by nammn> <https://launchpad.net/bugs/1826957>
[10:59] <nammn_de> else i could change the yaml parser from `juju update-cloud ..` to accept values without the highest key being `clouds: `
[11:01] <manadart> stickupkid: I am good to go here if you've time.
[11:01] <stickupkid> manadart, sure
[11:01] <stickupkid> do it
[11:02] <manadart> stickupkid: Hanging out in daily.
[11:02] <stickupkid> nammn_de, isn't it a case of you can't update localhost because it's built in
[11:03] <nammn_de> stickupkid: for me it works to update localhost if i modify the yaml file. Update in general does not work. Try to follow the steps from here: https://bugs.launchpad.net/juju/+bug/1826957 maybe we mean something different
[11:03] <mup> Bug #1826957: YAML produced by list-clouds is not acceptable to update-cloud <bitesize> <papercut> <update-clouds> <juju:Triaged by nammn> <https://launchpad.net/bugs/1826957>
[11:36] <achilleasa> jam: is there a recommended pattern for extracting lengthy code from inside buildTxn blocks?
[11:37] <jam> achilleasa: its just a function, so have helper functions. They tend to be called "doStuffOps"
[11:37] <achilleasa> jam: makes sense. thanks
[11:59] <nammn_de> rick_h: got a min to talk about the bug you just replied?
[12:00] <rick_h> nammn_de:  otp atm, will ping when free
[12:01] <nammn_de> Okay, im grabbing lunch we can talk after daily if that works for you :)
[12:01] <nammn_de> rick_h: ^
[12:52] <nammn_de> what is the https://github.com/juju/juju/blob/9039181e06f6988b1c62d3e71d28fbc031c0d32f/juju/osenv/old_home.go#L19 "juju dir" used for? "home/<name>/.juju"
[14:53] <nammn_de> stickupkid: im going to add an acceptance test into /tests/suites . Should i add this under the folder " cli" as this relates to the commandline?
[14:55] <stickupkid> nammn_de, yeah, we can always move it
[14:55] <stickupkid> "coverage: 100.0% of statements"
[14:55] <stickupkid> this is a nice thing :D
[15:23] <nammn_de> stickupkid: how do you run the suite? Feel like I am overlooking something important. `./main.sh test_static_analysis_go` does not work for me. Did we document that somewhere?
[15:24] <stickupkid> nammn_de, yeah, because you're running a task there not a suite
[15:25] <nammn_de> stickupkid: ups, meant that. just wanna run a task but seems not to work for me regarding permissions
[15:25] <stickupkid> nammn_de, run `./main.sh static_analysis test_static_analysis_go`
[15:25] <nammn_de> but thats how I run a task, right?
[15:25] <stickupkid> also ./main -h
[15:25] <stickupkid> ./main.sh -h
[15:26] <stickupkid> see the "Examples" section
[15:26] <nammn_de> ahh  you need to define the <folder> and <file>
[15:26] <stickupkid> nammn_de, yeah, similar to how other test things work
[15:26] <stickupkid> nammn_de, go test <folder> -check.f <test>
[15:27] <nammn_de> stickupkid: got it
[15:27] <stickupkid> nice
[15:46] <nammn_de> stickupkid: any readme/quick function to create a model? Could find destroy_model from examples?
[15:47] <stickupkid> nammn_de, juju add-model :D
[15:47] <stickupkid> nammn_de, ho?
[15:47] <nammn_de> stickupkid: oh makes sense :D Thought maybe we did add some convenience things
[15:47] <nammn_de> stickupkid: sure!
[15:47] <nammn_de> stickupkid: probably faster to get into
[15:48] <stickupkid> nammn_de, yeah, just use normal juju cli
[23:37] <hpidcock> https://github.com/juju/juju/pull/10669 PR for the person who reads this :) It's a short one.
[23:37] <timClicks> urgh why do I get notifications?
[23:40] <hpidcock> timClicks: thanks :D