Dr. Stefan Schimanski

Indeed, great work @negz!

Replaced by https://github.com/kubernetes-sigs/controller-runtime/pull/3019.

For those reading, this is currently a little outdated. https://github.com/kubernetes-sigs/controller-runtime/pull/2726 has a changed design proposed by @alvaroaleman. Will come back soon to both PRs.

Linked the updated implementation https://github.com/kubernetes-sigs/controller-runtime/pull/3019.

I already discussed this with @sbueringer and others: I went ahead earlier this week to implement the design outside of controller-runtime: https://github.com/multicluster-runtime/multicluster-runtime. As written in the [upper comment](https://github.com/kubernetes-sigs/controller-runtime/pull/2746#discussion_r1953484877), everything but...

@thockin there was one more topic @Jefftree told us you talked about: where to declare the strategy. Can you elaborate on your opinion? This is the most tricky aspect of...

> The problem with that is that different components could ask for different strategies, and it becomes indeterminate who "wins". It does not become indeterminate. The components are different versions...

If the reflection based, generic equality is not good enough (e.g. because of private values), just register into `reflect.Equalities` with your custom callback. Handling private fields by ignoring them feels...

@tlwr I don't think the proposed steps are a good solution. The behaviour to fail on private fields is intensional because it is defensive. If there are private fields, the...

I don't think these resources should be claimable. It is not clear to me what the semantics would even be.