SeaSand1024 commented on PR #12391:
URL: https://github.com/apache/seatunnel/pull/12391#issuecomment-6015803276

   @DanielLeens @SEZ9 addressed on tip `563fd4e8a` (close-path + tests only; no 
#12353/#12311).
   
   **Issue 1 (Error + close-once latch):** per-lifecycle `catch (Throwable)`; 
full pass continues for non-fatal Errors (`NoClassDefFoundError` / 
`AssertionError` / `LinkageError`); early-stop only for `VirtualMachineError` / 
`ThreadDeath`; rethrow prefers `Error` over earlier `Exception`. Tests 
flipped/added accordingly 
(`testCloseAttemptsEveryCycleWhenCycleThrowsNoClassDefFoundError`, 
AssertionError case, `testCloseStopsLoopOnVirtualMachineError`, 
`testClosePrefersErrorOverEarlierException`).
   
   **Issue 2 / polish (BlockingWorker classloader):** 
`setContextClassLoader(oldClassLoader)` moved into nested `finally` so restore 
always runs.
   
   **F3 / comments / Javadoc:** end-to-end NCDFE coverage + existing 
`super.close()` failure test kept; comments on `closeSuper()` (prod Exception 
unreachable / test seam) and non-transient `closed`; `@throws` updated for the 
Error policy.
   
   **engine-v2-it (8) on `877056bf1` run `37097098821`:** failsafe shows one 
error — 
`SplitClusterFaultToleranceIT.testStreamJobCancelResolvesWhenWorkerCrashesBeforeCancelAck`
 — `expected: <CANCELED> but was: <FAILED>` (`ConditionTimeoutException` at 
`assertEventuallyCanceled`). Same cancel-vs-FAILED class as #12353; not 
close-path / lifecycle teardown. Not folded into this tip.
   
   Local: `SeaTunnelTaskStateTest` 41/0/0 + BlockingWorker fallback test green.


-- 
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