SEZ9 commented on PR #11206:
URL: https://github.com/apache/seatunnel/pull/11206#issuecomment-6513685802

   Thanks for the continued work on this. A short status update scoped to the 
open review points.
   
   **CI at head `03973e5cbd3d7f215f80aa2e38e0fe9f053c00ab`:** [Signal pull 
request 
update](https://github.com/apache/seatunnel/actions/runs/40642775355/job/122796019439)
 is `SUCCESS`, but 
[Build](https://github.com/apache/seatunnel/runs/122809387163) is `FAILURE`, 
with `MultiTableSinkWriterTest.testRuntimeRegistrationFailurePolicy:234` 
failing across multiple unit-test lanes, and the engine/MySQL CDC lanes also 
unsuccessful. Since this PR changes runtime table registration, please treat 
that test failure as a possible regression rather than infrastructure noise. 
Could you share the assertion message / stack trace and which policy branch the 
test exercises? That would help determine whether it relates to the 
unsupported-sink `CreateTableEvent` handling (F2) or the `IOException` 
retry/idempotency contract (F5), or is something else. The P0/P1 items already 
noted earlier in this thread also remain open, so a fix to the registration 
path alone may not be enough to get exact-head CI green
 .
   
   **Still open from the previous review:**
   - **F1 / F4 / F8 (identifier handling and target scope):** I still need to 
see in-code enforcement for names flowing into generated sink SQL and the 
`${database_name}_${table_name}` template, strict routing/target-scope 
validation for runtime-created tables, and a decision on the 
`[A-Za-z_][A-Za-z0-9_$]*` allowlist excluding digit-leading MySQL table names. 
Please point me at the relevant changes or confirm they are pending.
   - **F3 (`scan.newly-added-table.enabled` defaulting to `true`):** please 
either flip the default to `false` or document the upgrade impact for existing 
wildcard CDC jobs with a rationale.
   - **F6 (`MultiTableState` format change):** please describe how an old 
checkpoint restores and what happens on version rollback.
   - **F7:** the JDBC sink docs still need the `exactly_once = false` / 
`generate_sink_sql = true` constraints.
   
   Once CI is green on the exact head and the items above are addressed or 
answered, I'll do a final pass.
   
   <!-- streview-comment:2133 -->


-- 
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