Normalize x-kubernetes-* fields for outputs

[rquitales]

Proposed changes

This PR normalizes outputs so that any x-kubernetes-* fields are stored in their respective underscored forms. When creating a CRD with Pulumi, we are able to set the x-kubernetes-* fields correctly during create/update (except in Java). However, if we attempt to read this field as an output, it would fail since the stored state is not normalized to the underscored forms. In other languages (Go/Python/TS), this was fine - but in Java, this fails as the builder pattern we use for codegen is unable to set the x-kubernetes-* fields since we only have setters for the underscored versions.

This also fixes the issue where we cannot read the values of x_kubernetes_* fields from outputs.

This PR adds additional unit tests, expands on the existing nodejs e2e test to verify that we can now read x_kubernetes_* fields, and creates a new Java e2e test to ensure we can successfully create CRDs that contain x_kubernetes_* fields.

Related issues (optional)

Closes: #3325

[github-actions]

Does the PR have any schema changes?

Looking good! No breaking changes found.
No new resources/functions.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 41.24%. Comparing base (e065015) to head (15631b8).
Report is 3 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #3348      +/-   ##
==========================================
+ Coverage   41.11%   41.24%   +0.12%     
==========================================
  Files          85       85              
  Lines       12741    12740       -1     
==========================================
+ Hits         5238     5254      +16     
+ Misses       7113     7096      -17     
  Partials      390      390              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[blampe]

This PR normalizes outputs so that any x-kubernetes-* fields are stored in their respective underscored forms

To clarify -- we currently store dashes in state, and will continue to store dashes, right? So underscores are only ever exposed to the SDK?

[rquitales]

This PR normalizes outputs so that any x-kubernetes-* fields are stored in their respective underscored forms

To clarify -- we currently store dashes in state, and will continue to store dashes, right? So underscores are only ever exposed to the SDK?

I believe that it is necessary to store underscores for it to be exposed to the SDK. This should not be a breaking change for users.

[EronWright]

Complementary to: #2128

[EronWright]

Looks great to me.