Replace Kustomize with a new hypershift CLI

[ironcladlou]

Kustomize was being used for:

1. Installation
2. Example cluster generation

In these narrow use cases, it's not clear Kustomize makes things easier, and it
also propagates dependencies downstream as we move into the CI system. This
commit replaces Kustomize entirely by providing a new hypershift CLI tool which
knows how to install the application and also generate example clusters.

This approach should enable a more portable integration point with CI and also
remove dependencies for users wanting to kick the tires of HyperShift. Here are
some other ways users can take advantage:

1. Use the hypershift binary directly through a release or (potentially) go get
2. Run hypershift using a container engine
3. Run hypershift in an OpenShift cluster using oc run

Another side effect of the move to a 100% native Go solution is that the fixtures
representing things like an example cluster can be more clearly shared with the e2e
test infrastructure, which this commit also implements.

As a further simplification and cleanup, the external Ginkgo binary is removed
and the go test tool is now the entrypoint for all testing.

Lastly, this commit adds support for looking up the OCP release image
so that reliable defaults can be used in all contexts (while providing easy
flags to override it.)

[openshift-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: ironcladlou

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Details

Needs approval from an approver in each of these files:

• OWNERS [ironcladlou]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[csrwng]

Hey @ironcladlou just a couple of comments.
I imagine that the automatic typing of resources for serialization will be done in a follow-up?

[ironcladlou]

@csrwng

I imagine that the automatic typing of resources for serialization will be done in a follow-up?

are you referring to the TODO in the serializer (https://github.com/openshift/hypershift/pull/16/files#diff-362c8c955f0738a540549420caa66fd217d29fd3b422b6ee45c1da2b4fdfbf91R22)? If so, I think we should do a followup to get to the bottom of it. The possible fixes in the TODO drag in a lot of dependencies, and after talking with David I think there's either some misunderstanding or bug anyway because what we're already doing should work.

[sjenning]

This is working for me. Thanks @ironcladlou 👍

[ironcladlou]

@csrwng had the good idea to default the OCP image based on a lookup to https://openshift-release.apps.ci.l2s4.p1.openshiftapps.com/api/v1/releasestream/4-stable/latest which removes the value of baking any sort of OCP image version into the operator itself (as he pointed out, we should be agnostic about the possible deployable versions anyway). So I'm going to make that change now.

[csrwng]

/lgtm