DanielLeens commented on issue #11617: URL: https://github.com/apache/seatunnel/issues/11617#issuecomment-5199226836
Thanks for doing the cross-connector audit. That follow-up is important, and it means this issue should not be treated as a MySQL-only fix anymore. The blocker you described in PR #11618 is real: if the shared incremental fetch path uses `!stopOffset.isNeverStop()` as the boundedness signal, then any CDC connector that still inherits the base `false` behavior can accidentally turn a valid `stop.mode = "never"` split into a finished split once the fetch task reaches `isFinished()`. That is a correctness risk, not just a missing polish item. So the acceptance bar for closing this issue needs to be stricter than "MySQL now stops correctly": 1. the shared bounded/unbounded contract must be made explicit rather than relying on a MySQL-only sentinel override; 2. PostgreSQL, SQL Server, Oracle, and MongoDB need real regression coverage on the same path; 3. PR #11618 should stay the main fix vehicle, but it is not ready to resolve this issue until the shared-reader semantics are safe across those connectors. Keeping this issue open for the broader cross-connector fix is the right call. -- 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]
