DanielLeens commented on PR #12218:
URL: https://github.com/apache/seatunnel/pull/12218#issuecomment-5645067952

   Thanks @Rangsh - I checked this mapping against the actual test file at the 
current head (`fee6db3b3c`, unchanged since my last review) rather than just 
taking the summary at face value, and it's accurate:
   
   - `TaskExecutionServiceTest.java:528` 
(`testDeployLocalTaskRollsBackAfterPostPublishFailureAndAllowsRedeploy`): 
redeploy assertion at line 586 calls 
`taskExecutionService.deployTask(redeployData)`.
   - `TaskExecutionServiceTest.java:613` 
(`testDeployLocalTaskRollsBackAfterPartialBlockingSubmitRejection`): redeploy 
assertion at line 746 also calls 
`taskExecutionService.deployTask(redeployData)`, with the inline comment at 
743-745 explicitly calling out that this is deliberate so the assertion 
exercises the same `executionContexts.containsKey` skip branch in 
`deployTask(TaskGroupImmutableInformation)` that #12164 is about.
   
   Both go through `deployTask(Data)`, not the internal `deployLocalTask`, so 
both hit the real production skip-check rather than bypassing it - this matches 
what I independently traced and confirmed as resolved in my last review (Issue 
1 in that round). Nothing new on the code side; both-sides checklist mapping is 
correct.
   
   One thing worth closing the loop on: the fork run for this exact head 
(`Rangsh/seatunnel` run `34558932185`) has since completed with `conclusion: 
failure` - `transform-v2-it-part-1 (11)` and `all-connectors-it-2 (8)` failed, 
and `paimon-connector-it (11)` was cancelled. None of these are topically 
related to `TaskExecutionService`/the rollback path this PR touches (no 
transform-v2, all-connectors, or paimon code is in this diff), so I don't have 
a reason to think this PR caused them, but I hadn't seen a completed result on 
this head when I last reviewed (it was still queued). Worth a rerun/check 
before a maintainer merges, just so the CI condition I flagged as the remaining 
item is actually closed out with a real green (or confirmed-unrelated) result 
rather than left at "queued."
   
   No open code-level findings from me on this PR.
   


-- 
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