SEZ9 commented on PR #11503:
URL: https://github.com/apache/seatunnel/pull/11503#issuecomment-5409355969

   Thanks for the detailed re-review, @DanielLeens, and for confirming that the 
`BasicType.LONG_TYPE` change in the `ff4463922272..24d4d2e92adf` diff resolves 
the compile issue you flagged.
   
   Your review appears to be cut off at "New finding — a second, distinct, 
blocking". Could you post the full details of the newly exposed NPE (the 
affected test method and a stack trace if available) so it can be addressed 
concretely?
   
   Since this round's delta is the one-line type fix, the previously raised 
findings remain open on my side:
   
   1. **PR11503-F1 (HIGH)** — legacy checkpoint restore path hardcodes a 
default tablePath, so the restored `CatalogTable`'s TableId cannot match the 
real source table.
   2. **PR11503-F2** — connectors/transforms not updated in this PR may hit 
unknown-event paths when `RestoreTableSchemaEvent` flows through during 
failover recovery.
   3. **PR11503-F3 / F6** — fail loudly on a null `changeAfter` instead of 
silently falling through to stale-schema behavior, and verify the restore 
shortcut targets the same table before replacing the current schema.
   4. **PR11503-F4** — non-atomic `clear()`+`putAll()` on a shared, 
non-concurrent map in `restoreCheckpointHistoryTableChanges`.
   5. **PR11503-F5** — the restore gate widened to any non-empty 
`checkpointTables`, so every restored CDC job takes the restore/event path even 
without any DDL.
   6. **PR11503-F7 / F8** — Javadoc for the new 
`restoreCheckpointHistoryTableChanges` SPI method, and dropping or redacting 
the INFO-level logging of the full `CatalogTable` list on restore.
   
   Once the NPE details are posted and the items above are fixed or given a 
rationale, I'm happy to do another full pass.
   
   <!-- streview-comment:552 -->


-- 
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]

Reply via email to