RE: Eastwood for different JDKs and deprecated methods. Don't we want to stop using deprecated methods as soon as they become deprecated in the latest JDK?

That would couple linting to JDK $latest which is not representative of what most users are running In general linting with eastwood implicitly asserts "this code can be evaluated and exercised in a real runtime" - the more runtimes the better

I've found a good deal of bugs that way over the years