MINOR: add missing java.util.Collections import

[harmadasg]

fix missing import

Reviewers: Bill Bejeck bbejeck@apache.org, Ken Huang s7133700@gmail.com, Chia-Ping Tsai chia7712@gmail.com

[bbejeck]

Thanks @harmadasg apologies for the mixup.

[m1a2st]

I think we should use Set.of() instead of Collections.singleton()

[bbejeck]

I think we should use Set.of() instead of Collections.singleton()

Sure, but I think we should do that in a follow-up PR since the build is currently broken

[bbejeck]

@harmadasg one thing though looking at the #23285 I don't see where it was removed and the build was passing for the merge

[bbejeck]

I think we should use Set.of() instead of Collections.singleton()

Sure, but I think we should do that in a follow-up PR since the build is currently broken

@harmadasg do you want to pick this up?

[chia7712]

@harmadasg thanks for quick fix

[harmadasg]

I think we should use Set.of() instead of Collections.singleton()

Sure, but I think we should do that in a follow-up PR since the build is currently broken

@harmadasg do you want to pick this up?

@bbejeck thanks for checking the PR. Sure I will open a new MINOR PR shortly to cover the change to Set.of().

[bbejeck]

Just to close this out Root cause was a semantic (not textual) merge conflict between two independently-green PRs:

• MINOR: Replace Collections factory methods with Java 11+ equivalents in streams #23242 removed the last existing Collections.* uses in StoreChangelogReader.java, so the unused-import cleanup dropped import java.util.Collections;.
• KAFKA-20940: Windowed changelog restore optimization is skipped whenever a checkpointed offset exists #23285 was branched before MINOR: Replace Collections factory methods with Java 11+ equivalents in streams #23242 landed and added new Collections.singleton(...) uses, inheriting the import from its older base.

#23285 was squash-rebased onto post-#23242 trunk at merge time. The rebase kept the import deletion (untouched context on #23285's side) and applied the new uses in a different region — no textual conflict, so nothing flagged it

[harmadasg]

I think we should use Set.of() instead of Collections.singleton()

@m1a2st opened #23320 to cover the review item

[chia7712]

#23285 was squash-rebased onto post-#23242 trunk at merge time. The rebase kept the import deletion (untouched context on #23285 side) and applied the new uses in a different region — no textual conflict, so nothing flagged it

@bbejeck thanks for the thorough cogitation upon the root cause 🧐. Don't worry about the build failure. It is a good-first-issue for everyone