slachiewicz commented on PR #295: URL: https://github.com/apache/flink-connector-kafka/pull/295#issuecomment-5471408634
Heads-up on overlap with #289 ([FLINK-40137](https://issues.apache.org/jira/browse/FLINK-40137), approved and ready to merge): it extracts the reader-recovery gating state from `DynamicKafkaSourceEnumerator` — including `pendingReportedSplitsByReader` — into a dedicated `ReaderRecoveryGate` class. This PR writes that field directly (constructor restore, `addReader` merge, snapshot getter), so the two will conflict. Once #289 lands, the rebase here should be mechanical: restore becomes `readerRecoveryGate.recordReportedSplits(...)` per restored reader, and the snapshot path needs a non-draining accessor on the gate (it currently only exposes `drainReportedSplits`). The extraction should actually make the merge-on-downscale semantics here easier to unit-test. Also worth noting for coordination: #279 adds a new `DynamicKafkaSourceEnumeratorRecoveryTest.java` with the same filename this PR modifies. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
