Mark Laing
Mark Laing
@tomponline This is ready for review when you have time. Thanks.
I'm in favour of moving to microcluster
@troyanov we will most likely move to `slog` so it may be necessary to write a formatter for `slog` to match what you have for `zerolog`, but sounds good :)
Hi @troyanov, we haven't been able to get around to this one but I expect we will allow a logger to be set via the DaemonArgs, and made available in...
Looks good in principle - CodeQL is failing on some files that you don't appear to have changed :cry:. I'll give a more comprehensive review when tests are green and...
> the CLI client won't be able to authenticate anymore because Keycloak would still require the client secret I had a feeling this might happen :cry: The [specification](https://datatracker.ietf.org/doc/html/rfc8628) for the...
> I think a more typical set up, and probably one that should have been considered when OIDC was originally added to LXD, is to have two separate clients defined...
> Sounds good. I'll make a demo to gather additional feedback from everyone and mention that we can consider implementing the approach with two separate clients in the future. Great...
Closing as this is not planned.
@tomponline RE these lint failures: https://github.com/canonical/lxd/actions/runs/10487443271/job/29047794191#step:14:78 They seem to have appeared due to a new version of `golangci-lint`. It is to do with using non- string format values in a...