SEZ9 commented on PR #12301: URL: https://github.com/apache/seatunnel/pull/12301#issuecomment-5724456773
Thanks for the clear correction, @201811510411lw — that resolves the confusion about where the candidate lives. Status of this PR as I understand it from the thread: the head here is still `48264cda6a`, and the two blocking findings raised earlier remain open on this branch — (1) the `SupportSinkDataPartition` check in `SinkExecuteProcessor` evaluates on the `MultiTableSink` wrapper, so the partitioner is never constructed on a real Paimon-on-Flink job, and (2) the partitioner would still crash on the zero-field schema-control rows coming out of `BroadcastSchemaSinkOperator`. The fixes for both now live in #12366 (commit `5de24b8cb` on top of `dev` at `1650a04cd`), which retains the `SupportSinkDataPartition` / `SinkDataPartitioner` SPI and `PaimonFixedBucketPartitioner` from here with co-author credit. Also agreed that the earlier read-through of `c56a0f96c` is not an approval of #12366; it will get its own review. Concrete asks so we don't end up with two implementations of the same #12243 fix: 1. Please coordinate with the original author of this PR and state explicitly in both threads which PR will carry the fix forward. If it is #12366, this one should be marked superseded and closed rather than left with an outdated branch; if it is this one, the fixes need to be rebased onto this branch. 2. On #12366, please post the remote CI result once it finishes, and confirm whether the Paimon Docker E2E was run (you noted it compiled but did not run locally) — the wiring-layer regression coverage is the piece we most want to see exercised end-to-end. 3. Keep the scope as you described: routing through the real `MultiTableSink` path, `schema_subtask_id` routing for control rows, and the MiniCluster/E2E coverage. Global-commit recovery, new checkpoint-state serializers, and the writer/committer rework should stay out and come as separate PRs if they are still wanted. I will not do a fresh full review here unless new commits land on this branch specifically; the review effort moves to #12366. <!-- streview-comment:1136 --> -- 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]
