DanielLeens commented on PR #11885: URL: https://github.com/apache/seatunnel/pull/11885#issuecomment-5601152258
@SEZ9 @li3zhi4 confirmed — I re-fetched my 2026-09-08 review body directly via the API (not the rendered view) and it is complete end to end, including the `2026-09-08T05:07:10Z` paragraph and the full F1-F8 mapping SEZ9 asked about. Nothing is missing server-side; this looks like the same markdown-rendering artifact I ran into on another PR's thread this same week (a long review body apparently trips up some clients' renderer around certain inline-code/heading combinations, not GitHub's stored content). @li3zhi4's repost above is accurate — it matches my original text verbatim on all five points. @SEZ9, to directly restate what you flagged as still needing a pointer, now confirmed against the actual diff at `befad345c` (not just the summary): - **F1/F6**: no separate checkpoint serializer exists to "wire" in this codebase — confirmed by the absence of any `PendingSplitsStateSerializer`-style class under `connector-cdc-base`. The stop offset is a plain field on the `Serializable` `IncrementalPhaseState`, picked up automatically by the engine's default `DefaultSerializer` (plain Java object serialization). `IncrementalPhaseStateTest` exercises exactly that round-trip plus the legacy-field-absent (pre-upgrade checkpoint) case. Resolved. - **F2**: `resolvedStopOffset == null` inside `createIncrementalSplit()` is the single resolution point; `HybridSplitAssigner` only creates incremental splits after snapshot completion, so there's exactly one resolution per assignment, shared by every split it produces. Resolved. - **F5**: `resolveLatestStopOffsetWithRetry()` in `IncrementalSplitAssigner.java` retries up to 3 attempts with 300ms/600ms backoff before wrapping and throwing. Resolved. - **F3**: `runLatestStopStartupMode` waits for row id=1 at the sink, then defensively asserts row id=2000 is NOT yet present before proceeding — a runtime-verified structural gate, not just a documented claim. Resolved. - **F4/F7/F8**: `docs/en|zh/connectors/source/MySQL-CDC.md` documents the snapshot-completion semantics for `latest` with a worked example, and the PR description names the enumerator-side mechanism (`IncrementalSplitAssigner.createIncrementalSplit()`). Resolved. No new commit since I last reviewed — the head is still `befad345c`. Happy to hear if either of you still sees an open item after checking the reposted/re-fetched text above against the current diff yourselves. -- 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]
