Update Tinkerbell and linter, add linter and codespell to CI #20
Conversation
|
After talking to @gianarb, it seems reasonable to use Tinkerbell version which sandbox repo is using, so we keep it compatible at this level, so I need to rollback some changes here. Once sandbox repo is updated, we will update as well. |
go.mod
Outdated
| @@ -3,33 +3,68 @@ module github.com/kinvolk/terraform-provider-tinkerbell | |||
| go 1.14 | |||
|
|
|||
| require ( | |||
| cloud.google.com/go v0.73.0 // indirect | |||
| code.gitea.io/sdk/gitea v0.13.2 // indirect | |||
| github.com/Azure/azure-storage-blob-go v0.12.0 // indirect | |||
There was a problem hiding this comment.
Any idea where these indirects came from? Is it goreleaser?
If so, we should be able to avoid installing goreleaser in a way that affects this go.mod, even without a separate GH action - although, that may be an option too.
There was a problem hiding this comment.
Yeah, this is because of goreleaser mostly. I don't really like the approach with tools.go either. Let me create an issue for it.
There was a problem hiding this comment.
Created #21. I could replace tools.go with go run -mod=mod, but then we use tools version of whatever someone has installed in their GOPATH.
There was a problem hiding this comment.
Pushed 1d2377d which seems to work fine.
So tooling dependencies are managed separately from main module dependencies. Closes #21 Signed-off-by: Mateusz Gozdek <[email protected]>
This commit updates compatible Tinkerbell version to latest version used in sandbox repository at the time of writing. Latest version of sandbox repository no longer use SQL file to seed the database. Instead, Tink server runs the database migration if special environment variable is set. To do that on the test docker-compose.yml file, new container called 'tink-server-migration' is added, which runs the migration before Tink server starts. This setup follows what is used in the sandbox repository. Signed-off-by: Mateusz Gozdek <[email protected]>
With new Tinkerbell version, there are event objects created when creating Hardware, Workflows etc and they also cannot really cope with parallel requests, so we need to retry if one of them occurs. Reported upstream issue: tinkerbell/tink#379 Signed-off-by: Mateusz Gozdek <[email protected]>
And fix new warnings. Signed-off-by: Mateusz Gozdek <[email protected]>
To make running linter easier together with editors integration and CI systems. Signed-off-by: Mateusz Gozdek <[email protected]>
Signed-off-by: Mateusz Gozdek <[email protected]>
No description provided.