OCM descriptor as CRD

[Skarlso]

What this PR does / why we need it:

Which issue(s) this PR fixes:
Fixes #35

Special notes for your reviewer:

Release note:

[gardener-robot]

@Skarlso Thank you for your contribution.

[Skarlso]

Hmm, hmm. I'm a bit perplexed as to the structure based on the v2, v3 schemas. The question here is whether to embed the Object or not. But if we embed the Object, all the values will have an apiVersion which is not something that will work nicely in Kubernetes.

So, maybe we stay with an approximation instead of a full 1:1 mapping.

[mandelsoft]

Is it possible to put the CRD stuff into a some sub directory?
It is quite prominent, so far. And the structure does not really meet the actual project structure.

Basically, instead of using own types for the various elements the CRD should use the
already existing binding types. Is this possible?

Or alternatively, we need a transfer function, that provides a component descriptor (according to this library) from the K8S resource, and the opposite direction, putting a CD into an K8S resource.

This should be based on the internal structure (package compdesc), not the serialization versions. This would support a smooth usage together with the rest of the lib.

[Skarlso]

Or alternatively, we need a transfer function, that provides a component descriptor (according to this library) from the K8S resource, and the opposite direction, putting a CD into an K8S resource.

Yep, that's how most projects do it. If you have some kind of internal structure you usually do a transformation between that and the k8s representation.

Is it possible to put the CRD stuff into a some sub directory?
It is quite prominent, so far. And the structure does not really meet the actual project structure.

This is the convention for any CRD in any project by the CRD generator codes. We can put it elsewhere of course, but that's not really intuitive, not to mention that some tooling like code-generators might not work appropriately.

I suggest we follow the standards in this case as it's a well-trodden path.

[gardener-robot]

@yitsushi, @In-Ko You have pull request review open invite, please check

[Skarlso]

This is now ready for review. It got a lot simpler as its reused internal, already existing definitions and types. 🎉

[gardener-robot]

@In-Ko You have pull request review open invite, please check