Rename client-java-api to *-legacy

[yue9944882]

No description provided.

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: yue9944882

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

The pull request process is described here

Needs approval from an approver in each of these files:

• OWNERS [yue9944882]

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

[brendandburns]

Do we really need to move all of the fluent files? I think we just need to build for both?

Some more details on this would be useful.

Thanks!

[yue9944882]

Do we really need to move all of the fluent files? I think we just need to build for both?

hmm. i'm not sure if we can build fluent for both. i removed fluent related changes in PR so we can try generate for both in a follow-up PR

[brendandburns]

Sorry, I'm thinking about this some more. Do you think it would be easier to maintain this as a branch in git, rather than duplicating the code on master?

[yue9944882]

Sorry, I'm thinking about this some more. Do you think it would be easier to maintain this as a branch in git, rather than duplicating the code on master?

that's definitely viable but i'm slight worried if it's as convenient to maintain as a legacy module. additionally, putting it into a separate branch will sort of double the effort for code generation for every kubernetes release because we will need to open (or automate) 2 separate PRs upon every release. do you see any downside of having a legacy module?

[brendandburns]

Aren't we going to have to do double code generation runs anyway? Or are you suggesting that we update the code generation script to just run twice?

I personally would vote for the branch approach, it's worked reasonably well in the Javascript client. We setup dependabot etc to send PRs to both repos.

I'm sort of ok with the copy everything approach, but I worry it's going to be a lot of duplicated code that is going to make the repository messy and also make all of the CI/CD runs twice as long.

Also, since presumably we're going to eventually stop supporting the legacy branch except for critical fixes, it feels like we'll just eventually have to delete all the legacy code which will make it harder to use the legacy library.

It also feels simpler to implement since we can just branch, with the current state that works with java8, and then convert main to be java11.

But ultimately, if you feel strongly, since you are doing the work, I defer to your preference.

[yue9944882]

@brendandburns gotcha, let's move those to a separate branch then. i will file another PR after closing this!

[brendandburns]

Sounds good. Thank you!