SEZ9 commented on PR #12042:
URL: https://github.com/apache/seatunnel/pull/12042#issuecomment-6433023278

   @nzw921rx thanks for the ping. The 2026-10-13 note reported the head as 
conflict-free at that time, but the base branch may well have moved since then, 
so I can't confirm the current mergeability state from here. Could you share 
what conflicts you're seeing (or which files), so we can confirm whether a 
rebase is needed?
   
   Separately, CI is still a gap: fork run 36327880735 concluded `cancelled` on 
its retry because the downstream matrix jobs were retained from the original 
run, and the rerun budget is exhausted. A fresh push to this PR (a rebase would 
do it) is needed to trigger a complete workflow run.
   
   While rebasing, it would be good to pick up the open review points as well, 
roughly in priority order:
   
   - **F1**: add a null guard around `sourceFlow.getAction().getSource()` in 
`isUnboundedSourceTask()`, since it runs on the master restore path.
   - **F2**: add a test exercising the `reportedPipelineRunning(..., true)` 
path against a real `CheckpointCoordinator` with a FINISHED reader group, to 
confirm checkpoints continue rather than time out.
   - **F3 / F5**: either cross-check the restored checkpoint state before 
treating a FINISHED unbounded-source group as idle-closed, or document why the 
state-only heuristic is safe; and clarify whether a FINISHED enumerator vertex 
is intentionally out of scope.
   - **F4**: read the execution state from the IMap once and reuse it for the 
recoverability decision.
   - **F6 / F7 / F8**: note the failover behavior change in docs/release notes, 
switch to `allMatch` / a single iteration, and soften the PhysicalVertex 
Javadoc to what the code can actually detect.
   
   Once a fresh push lands with a complete CI result, I'll do another pass.
   
   <!-- streview-comment:2077 -->


-- 
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