88fantasy commented on PR #12388:
URL: https://github.com/apache/seatunnel/pull/12388#issuecomment-6398002519
Thanks for the updated CI diagnosis. I have synced this Apache SeaTunnel PR
with `dev@7604a83ec802a0b148e98e557fb9184c48127988` in
`77ba7a5c1aa33d74b7c5b08fe25740c0fcc3211c`, including the upstream E2E
container-selection fix #12895 (`b65fe82b0`). The merge had no textual
conflicts; the PR diff against dev remains the original 5 files, with no
additional classification mechanism.
Current-head local validation on JDK 11:
- Full-reactor `./mvnw spotless:apply`: passed, with no tracked-file changes.
- Full-reactor `./mvnw -q -DskipTests verify`: passed; this command skips
tests.
- Focused checkpoint/restore tests: **40 passed, 0 failures, 0 errors, 0
skips** (`CheckpointCoordinatorTest`: 17; `CheckpointManagerTest`: 9;
`SubPlanRestoreStateTest`: 1; `StateTransitionCleanupTest`: 13), using:
```shell
./mvnw -B -pl seatunnel-engine/seatunnel-engine-server -am \
-Dtest=CheckpointCoordinatorTest,CheckpointManagerTest,SubPlanRestoreStateTest,StateTransitionCleanupTest
\
-Dsurefire.failIfNoSpecifiedTests=false verify
```
The [new-head
Build](https://github.com/88fantasy/seatunnel/actions/runs/40195888197) is in
progress. This is a new run after the dev sync, not a rerun of the old SHA.
Aggregate CI success is still pending; unsupported-engine skips must remain
distinct from executed runtime coverage.
I have also corrected the PR description's overly strong
FINISHED-only-means-idle-reader statement and separated current-head local
validation from the historical cluster results. The final-checkpoint completion
window and a switch while idle-reader closure awaits ACK remain unverified. No
fresh real-cluster/E2E acceptance was run locally.
For the overlap with #12042, could a maintainer confirm the preferred
classification rule and merge order? If this PR lands first, #12042's stricter
classification could be reviewed on top while preserving the closed-subtask
reconstruction here. The two competing rewrites of the SubPlan loop should not
both land unchanged. A successful current-head required Build and effective
maintainer approval are still separate merge gates.
--
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]
