SEZ9 commented on PR #11885: URL: https://github.com/apache/seatunnel/pull/11885#issuecomment-6159409276
Thanks @li3zhi4 for the point-by-point mapping against `7f17332b7a` — that is exactly the information I was missing. I'm not marking anything resolved from the mapping alone; I'll verify each item against the changed files at `7f17332b7a` before closing it. A few things I'll be looking at specifically, and where a pointer from you would help: - **F1/F6:** whether the round-trip in `IncrementalPhaseStateTest` exercises the same serialization path the enumerator uses for its checkpoint state, so the `stopOffset` field is proven to survive a real checkpoint/restore. - **F2:** the `resolvedStopOffset == null` guard addresses the pre-completion split case, but the finding also covered multiple splits assigned around the resolution point ending up with different stop positions. Which test or code path rules that out? - **F3:** how the e2e test establishes that the snapshot is still active when it asserts row 2000 is not visible (e.g. row count relative to `snapshot.split.size=20`), so the negative assertion cannot pass trivially on a fast run. - **F5, F4/F7/F8:** the retry/backoff and the EN/ZH `MySQL-CDC.md` updates look like they cover these; I'll confirm against the diff. Once I've checked those, I'll update the findings accordingly. <!-- streview-comment:1732 --> -- 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]
