On Sun, 7 May 2023 22:17:56 GMT, David Holmes <dhol...@openjdk.org> wrote:
> This is a tricky issue to solve cleanly and minimally. If every class > extended FinalizableObject then all the implementation could go there, but we > have to try and split things across FinalizableObject and Finalizable because > some classes only implement the interface - this leads to some unfortunate > design choices. > > I think a better design for the classes that can't extend FinalizableObject > would be for them to contain a FinalizableObject, the cleanup action for > which would cleanup the host object. That could allow the removal of the > Finalizable interface and simplify the general usage patterns. But that may > be going too far for this particular PR ... ? Please, continue this discussion on the new PR: https://github.com/openjdk/jdk/pull/13884#issuecomment-1539796995. Thanks! ------------- PR Comment: https://git.openjdk.org/jdk/pull/13420#issuecomment-1539803578