Convert Guava Sets.filter

[lredor]

What's changed?

• Add conversion for Sets.filter (as described in Sets.filter is not migrated, which creates a conflict with the PreferJavaUtilPredicate recipe #897)

Anything in particular you'd like reviewers to focus on?

First contributions, so I’m not sure I’ve mastered all aspects of OpenRewrite yet.

Anyone you would like to review specifically?

@timtebeek

Checklist

• I've added unit tests to cover both positive and negative cases
• I've read and applied the recipe conventions and best practices
• I've used the IntelliJ IDEA auto-formatter on affected files

[timtebeek]

Near perfect already, great to see, thanks! One question before we merge as we don't want to potentially change the behavior for folks relying on special handling for SortedSet.

[timtebeek]

There's a slight concern here where Sets has special handling for SortedSet, whereas the the replacement always uses toSet, which uses HashSet. Should we conditionally use .toCollection(TreeSet::new) ?

[lredor]

You're right. I added a case for that.

[timtebeek]

This case is tricky; we can't really deduce from the type system whether we'd end up in this conditional of special handling:
https://github.com/google/guava/blob/21cdcf6d9a49c1a498acb422471d8d71f9219865/guava/src/com/google/common/collect/Sets.java#L1190-L1194

[timtebeek]

I'd hope that to be rare enough to perhaps still merge the current work, but there is a risk here which could only be mitigated by data flow analysis to figure out if it might be a SortedSet going into this filter(Set, Predicate) method. 🤔

[timtebeek]

Either way thanks for the quick rework; would love to hear your thoughts on this edge case.

[timtebeek]

Thanks again! I think the risk is low here for folks relying on that instanceof SortedSet check, so we'll merge this as is with the test documented to fail.