Test-infra: flags introduced by dependencies?

[hongkailiu]

What should be cleaned up or changed:
flags are introduced by dependencies.
Ideally, we may want to do some refactoring in the dependency's repo.
Provide any links for context:

$ git status
On branch master
Your branch is up to date with 'origin/master'.

nothing to commit, working tree clean
$ git log --oneline -1
2ebe6361e (HEAD -> master, origin/master, origin/HEAD) Merge pull request #14192 from petr-muller/proper-bz-resoluitons
$ go mod why github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/
# github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1
k8s.io/test-infra/prow/apis/prowjobs/v1
github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1
$ GOPROXY=https://proxy.golang.org go mod vendor
$  grep -irn "flag.String(" ./vendor/github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/
./vendor/github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/git_resource.go:33:    gitImage = flag.String("git-image", "override-with-git:latest",
./vendor/github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/artifact_pvc.go:30:    bashNoopImage = flag.String("bash-noop-image", "override-with-bash-noop:latest", "The container image containing bash shell")
./vendor/github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/gcs_resource.go:30:    gsutilImage              = flag.String("gsutil-image", "override-with-gsutil-image:latest", "The container image containing gsutil")
./vendor/github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/build_gcs_resource.go:30:      buildGCSFetcherImage = flag.String("build-gcs-fetcher-image", "gcr.io/cloud-builders/gcs-fetcher:latest",
./vendor/github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/build_gcs_resource.go:32:      buildGCSUploaderImage = flag.String("build-gcs-uploader-image", "gcr.io/cloud-builders/gcs-uploader:latest",
./vendor/github.com/tektoncd/pipeline/pkg/apis/pipeline/v1alpha1/cluster_resource.go:32:        kubeconfigWriterImage = flag.String("kubeconfig-writer-image", "override-with-kubeconfig-writer:latest", "The container image containing our kubeconfig writer binary.")

Those flags are added via var section, such as git-image. Any cmd depending on it will have to do workaround like flag.NewFlagSet to exlude them.

$ grep -irn "flag.NewFlagSet" ./prow/cmd/ | grep -v "_test.go"
./prow/cmd/deck/main.go:282:    o := gatherOptions(flag.NewFlagSet(os.Args[0], flag.ExitOnError), os.Args[1:]...)
./prow/cmd/mkpj/main.go:197:    fs := flag.NewFlagSet(os.Args[0], flag.ExitOnError)
./prow/cmd/hook/main.go:96:     o := gatherOptions(flag.NewFlagSet(os.Args[0], flag.ExitOnError), os.Args[1:]...)

[stevekuznetsov]

@cjwagner @fejta @Katharine @clarketm

This is another reason we should reconsider the tekton dependencies :|

[stevekuznetsov]

$ docker run -it gcr.io/k8s-prow/peribolos:v20190911-918053970 --help
Usage of /app/prow/cmd/peribolos/app.binary:
  -bash-noop-image string
        The container image containing bash shell (default "override-with-bash-noop:latest")
  -build-gcs-fetcher-image string
        The container image containing our GCS fetcher binary. (default "gcr.io/cloud-builders/gcs-fetcher:latest")
  -build-gcs-uploader-image string
        The container image containing our GCS uploader binary. (default "gcr.io/cloud-builders/gcs-uploader:latest")

boo!

[fejta]

Yeah that's definitely not great. Tekton is at least one thing that prevents us from doing minor module upgrades (since it transitively depends on k8s.io/kubernetes)

Filed an issue about this over in the pipeline repo.

/cc @cjwagner @bobcatfish @wbrefvem

Thoughts?

[cjwagner]

This isn't something we can fix on our end short of removing the Prow + Pipelines integration right? If thats the case it would be great to see this fixed by the Pipelines project, especially since we're probably not the only ones with these problems.

[stevekuznetsov]

We can (and IMO should) just copy-paste the types we need from tekton into the third_party/ folder and snip our dependencies. It's causing us so many headaches importing them right now.

[fejta-bot]

Issues go stale after 90d of inactivity.
Mark the issue as fresh with /remove-lifecycle stale.
Stale issues rot after an additional 30d of inactivity and eventually close.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle stale

[fejta-bot]

Stale issues rot after 30d of inactivity.
Mark the issue as fresh with /remove-lifecycle rotten.
Rotten issues close after an additional 30d of inactivity.

If this issue is safe to close now please do so with /close.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/lifecycle rotten

[fejta-bot]

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

[k8s-ci-robot]

@fejta-bot: Closing this issue.

In response to this:

Rotten issues close after 30d of inactivity.
Reopen the issue with /reopen.
Mark the issue as fresh with /remove-lifecycle rotten.

Send feedback to sig-testing, kubernetes/test-infra and/or fejta.
/close

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository.