Refactor e2e test by using some common functions

[shipwright-io]

Refactor e2e test by using some common functions

[qu1queee]

I spoke with @zhangtbj on this, I think being able to specify which strategies to run during e2e tests isn´t ideal, e2e tests should run everything. In the meantime, we can live with a minor patch from our side to disable the two strategies we cannot run successfully(buildah and s2i).

However, I think we would like to have the flexibility to avoid running the following pieces during e2e tests, due to our specific tenant environments and the way where we run components:

• Avoid running the framework.AddToFrameworkScheme , see lines
• Avoid running the ctx.InitializeClusterResources, see lines
• Avoid running the e2eutil.WaitForOperatorDeployment, see lines
• Avoid running the createClusterBuildStrategy calls for all strategies, see lines
• Avoid running the tekton task verification , see lines , this is because we cannot check the tekton resource on the tenant namespace.

All of the above can be done with a new env variable, I can do this easily(I think), @sbose78 pls let us know what u think. With this we will only patch to skip those two strategies, and once we resolved the issues, we will not longer need to patch anything. Also, this will not break your setup.

[qu1queee]

@sbose78 let me know about the above.

[sbose78]

If I understand this correctly,

1. There are specific strategies that we would want skipped.

2. There are specific things we would not want to validate. Example, tekton resources.

Can we make our tests resilient enough to understand that specific validations may be skipped?

Which user do you run these tests as in your internal environment?

1. In my opinion we should create "environment templates" in our upsteam CI for running tests in such environments.
2. Make our tests smarter to skip validating Specific things based on the environment.

[zhangtbj]

Agree.

Because our Production system is ready and deployed at the beginning (build controller and cluster build strategies). So every time, we just need to update the controller and re-run the e2e tests.

• So require to deploy the builder controller and cluster build strategies is not good for our production system upgrade.
• If I deploy a test env and already deploy above required resources, I would like to run the e2e tests with a similar parameter --skip-deploy to JUST run the e2e tests directly.

[zhangtbj]

But ....

It issues is used to extract the current test code as the common functions, so that each test can reuse and share together.

I think later, we can create a new issue to separate these two requirements.

[zhangtbj]

This basic requirement is resolved by using this PR:
#120

Close this first and will open the advanced issue for the new requirement.

Thanks!