DanielLeens commented on PR #12314: URL: https://github.com/apache/seatunnel/pull/12314#issuecomment-6397570634
Thanks for the contribution. The CI fact in my previous review has changed: the exact-head [fork Build](https://github.com/DanielLeens/seatunnel/actions/runs/39984445999) is now **completed with failure**, rather than queued/in progress. This updates the old guidance; I am not posting a second source review on the unchanged head `7ee96e15bd9242a8fc88120df4c239e7bd6fab77`. The [inspected failed job](https://github.com/DanielLeens/seatunnel/actions/runs/39984445999/job/120574848238) reports `PreconditionViolationException: None of the supporting TestTemplateInvocationContextProviders ... provided a non-empty stream`. Specifically: HbaseIT (22 errors); inspected transform-v2 part-2, all-connectors-it-2, all-connectors-it-6, and MySQL CDC job samples have the same empty TestTemplate-provider signature. There is an upstream fix already in latest dev: [#12895](https://github.com/apache/seatunnel/pull/12895), commit `b65fe82b040c203ce669db83431916108f3e866e`. It adds `ContainerSelectionCondition` to skip a suite/method when every engine selected by `RUN_ENGINES` is explicitly disabled, instead of passing an empty invocation stream to JUnit. A genuinely empty container discovery still remains an error. I checked that this head does not contain that commit, and this PR does not change that shared E2E framework. The comparison is `diverged`, ahead 3 / behind 13 against dev, with no reported conflict. For this demonstrated upstream test-selection failure, please **sync latest dev and rerun the affected failed jobs first**. Do not copy unrelated framework changes into this PR or add dependencies to chase the old failure. Correctly skipped unsupported-engine cases are not runtime test passes; inspect the refreshed run for any other failures and preserve the relevant engine coverage. My previous Issue 1 still needs the requested proof or guard for the terminal/shutdown race. This CI correction does not clear that code-side question; both gates must be addressed. A successful current-head required Build and write-capable maintainer approval remain separate merge gates. Daniel has READ permission; this follow-up is not an effective approval. Thanks for keeping the patch focused. -- 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]
