DanielLeens commented on PR #11932: URL: https://github.com/apache/seatunnel/pull/11932#issuecomment-5661999311
Thanks, @SEZ9 — agreed on the F7 conditional close: once the Flink/Spark follow-up issue is filed and linked here, F7 is fully closed on my side too. As before, that link is on @davidzollo as the PR author, not something I can create on his behalf. Everything else stands exactly as both of us have now independently restated, since the head is still `b6c96b97356d` with no new commit: 1. F1/F2 on `IncrementalSourceReader` — the empty-`checkpointTables` restore guard and gating `restoreCheckpointProducedType`/`restoreSchema` on the same condition. 2. F3/F6/F8 on `JdbcSink` — deterministic (or fail-fast) schema selection in `restoreWriter`, preserving the primary-key/upsert path on a restored schema without a `PrimaryKey`, and rejecting/filtering null-Xid states after an `exactly_once` flip. 3. F4/F5 on `JdbcSinkState`/`JdbcSinkWriter` — the docs/changelog note for the new checkpoint state format and confirming Serializable-safety of the persisted schema/column payloads. I'll do the full re-review as soon as a commit addressing these lands. One more thing worth flagging now rather than waiting for the next push: I just re-checked `mergeable_state` on this PR and it currently reports `DIRTY`/`CONFLICTING` against the base branch — that's new since my last check, `dev` has moved past where this branch's tip can auto-merge. @davidzollo, since you'll be pushing a commit for F1–F8 anyway, please pull the latest `dev` and resolve the conflict in the same pass; otherwise CI on any new commit won't reflect a mergeable state and we'd have to redo this check. -- 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]
