-
Notifications
You must be signed in to change notification settings - Fork 510
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Charmhub bases #12886
Charmhub bases #12886
Conversation
Although the API still accepts platform for refresh API, it's actually not the right way to do it. Instead we should push bases all the way through the stack. That way we know that this is actually future proofed against any API changes. The annihilation of platfrom from charmhub client.
The charmhub client has changed to ensure that it always talks about bases. This means that all consumers/users of the client also needs updating.
We should always validate a channel comming in so that we know how to deal with it correctly.
The following ensures that we take into account the bases for download.
if err != nil { | ||
return corecharm.Platform{}, errors.Trace(err) | ||
} | ||
series, err := coreseries.VersionSeries(track) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I still think we need to check how CentOS and Windows charms will appear as bases. Seems to me we are unlikely to get win10
or centos7
as a track. Something like name: "windows"
and track: "10"
would be the correct representation.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
There are currently no way to tell, as we don't have any pre-existing charms to test.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Working as expected for this iteration.
|
#12909 Merge 2.9 #12827 Unsubscribe from hub when closing state pool #12829 Correct default bootstrap-timeout value displayed in help. #12840 Constraint tags can be used for pod affinity #12842 Fix upgrade series agent version handling #12794 Add disk provisioning customization #12845 Restore space support for manual machines #12839 Support merging of netplan configs #12853 Add display type for network-get results #12854 Fix for LP1921557 sni in Juju login. #12850 Use Base in Charmhub packge and its response structures. #12858 Ensure assess-upgrade-series does not report started prematuremly #12860 Removed logging from core annotations. #12861 Fixes bug where empty error can happen in storage #12865 Update Pebble version to include new files API #12866 Workaround for k8s dashboard URL with k8s client proxy #12862 Fix/lp 1923051 #12867 Fix/lp 1923561 #12870 Use channel logic in charm library #12873 Add support for setting pod affinity topology key #12874 Use Patch instead of Update for ensuring ingress resources #12831 Integration fixes #12879 Ensure refresh uses version #12864 bug: fix for bootstrap fail on vsphere 7 + multi network #12883 Initial work to allow juju trust for sidecar charms #12884 Fix ssh with sidecar charms and containers. #12886 Charmhub bases #12881 Use charm pkg updates #12889 Ignore projected volume mounts when looking up juju storage #12890 Fix passing empty string container name to unit params #12893 Add CLA checker GH action and remove codecov push action #12897 Use production charmhub endpoint #12887 Resource validation error #12888 Ensure we validate the model target #12898 Remove usage of systems package from CAAS application provisioner #12899 CAAS bundle deployments #12900 Bump up Pebble version to include user/group in list-files #12901 charm Format helper #12902 charm Iskubernetes helper #12903 Display ... for really long k8s app versions in status #12904 Filter out more full registry paths for app version in status #12905 Fix k8s bundle deploys with v2 charms #12906 Register resource-get for containeragent binary Conflicts mostly due to charm v8 vs v9 imports. The other one was due to changes to dashboard CLI. ``` # Conflicts: # api/common/charms/common.go # api/common/charms/common_test.go # apiserver/facades/client/application/application.go # apiserver/facades/client/application/charmstore.go # apiserver/facades/client/application/update_series_mocks_test.go # apiserver/facades/client/charms/client.go # apiserver/facades/client/charms/convertions.go # apiserver/facades/client/machinemanager/types_mock_test.go # apiserver/facades/controller/caasoperatorprovisioner/provisioner.go # cmd/juju/application/deployer/bundlehandler_test.go # cmd/juju/application/refresh_test.go # cmd/juju/application/refresher/refresher_mock_test.go # cmd/juju/dashboard/dashboard.go # core/charm/strategies_mock_test.go # core/model/model.go # core/model/model_test.go # go.mod # go.sum # resource/resourceadapters/charmhub.go # scripts/win-installer/setup.iss # service/agentconf_test.go # snap/snapcraft.yaml # state/charm.go # state/migration_export.go # state/state.go # version/version.go # worker/caasfirewallerembedded/applicationworker.go # worker/caasfirewallerembedded/applicationworker_test.go ``` ## QA steps See PRs
Although the API still accepts a platform for refresh API, it's actually
not the right way to do it. Instead, we should push bases all the way
through the stack. That way we know that this is actually future-proofed
against any API changes.
The following code just does that.
QA steps
$ juju bootstrap lxd test $ juju deploy hello-juju $ juju deploy facundo-snappass-test $ juju deploy ubuntu-juju-qa --channel 2.0/edge