SEPURI-SAI-KRISHNA commented on PR #12495:
URL: https://github.com/apache/seatunnel/pull/12495#issuecomment-5970963755
The `Build` check on this PR is red, and the cause is a known open defect on
`dev` rather than anything in this diff. Details so nobody has to re-derive
them.
Run `36437284233` on head `ade4e41c8`, job `engine-v2-it (8,
ubuntu-latest)`. The whole module reports **205 tests, 0 failures, 1 error, 7
skipped**, and the single error is:
```
[ERROR] testStreamJobCancelResolvesWhenWorkerCrashesBeforeCancelAck Time
elapsed: 67.156 s <<< ERROR!
org.awaitility.core.ConditionTimeoutException
at
SplitClusterFaultToleranceIT.assertEventuallyCanceled(SplitClusterFaultToleranceIT.java:557)
at
SplitClusterFaultToleranceIT.testStreamJobCancelResolvesWhenWorkerCrashesBeforeCancelAck(:449)
Caused by: org.opentest4j.AssertionFailedError: expected: <CANCELED> but
was: <FAILED>
```
That is #12353, open since 2026-09-16 and assigned to @zhangshenghang. The
reporter reproduced it on unmodified `dev` at base
`6ee0c374478a7c8be090e3eb33895027c5ac2ffd` on JDK 17, and @SEZ9 corroborated
the identical signature on JDK 8 and Linux, so it is neither JDK nor OS
specific. The test itself is regression coverage added by #12030 for a defect
that was never fixed.
The fix is #12311, which resolves a lost `CANCELING` vertex to `CANCELED`
instead of `FAILED`. **It is still open.** So this leg will keep failing on any
PR whose path filters reach `engine-v2-it` until #12311 merges, and a rerun
here will not clear it.
This PR cannot be the cause either way. It touches eleven files, all under
`seatunnel-transforms-v2`, and not one of them is in `seatunnel-engine` or
`seatunnel-e2e`. The change is confined to making keyword and type-name lookups
use `Locale.ROOT`, so it does not reach the Zeta cancellation state machine.
The other non-success in that run, `kudu-connector-it (11)`, is cancelled
rather than failed.
Two things I can do if either helps:
1. Rebase onto current `dev`. The branch is 31 commits behind, so this is
worth doing regardless, though it will not change the `engine-v2-it` result
while #12311 is unmerged.
2. Leave it and wait for #12311.
I checked the overlap with my #12605, which also edits
`SystemFunction.java`: the two do not collide. #12605 works at lines 105 to 198
in the INT cast path, this PR at line 275 in the `TRUE`/`FALSE` string branch,
so they can land in either order.
--
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]
