DanielLeens commented on PR #12900: URL: https://github.com/apache/seatunnel/pull/12900#issuecomment-6337990142
Thanks for tracking the schema-reuse work in #12940, @hutiefang76. Keeping both layouts, the legacy byte/name contract, and the different-parent/same-leaf case in that follow-up is the right boundary; a field-name-only cache would not capture the owning schema. It remains an optional follow-up rather than a new blocker in this unchanged head. The CI fact from my prior review has changed: [the fork Build on `7c2cdf5a8df9d3122bf103c228f1ec179254cd32`](https://github.com/hutiefang76/seatunnel/actions/runs/39429128558) is now completed with failure. The inspected unit-test log fails on `LLMBooleanOutputTest.readsLegacyFactoryWrapperWithValidationDisabled` (`InvalidClassException` on `AbstractMultiCatalogTransform`); `engine-v2-it (11)` also fails `BackpressureSlowSinkIT.testCheckpointsKeepCompletingUnderSustainedBackpressure`. I checked upstream: the LLM tracking issue #12929 and fix #12938 are still open, and the deterministic-backpressure/source-lock proposals #12313 and #12316 are also open. These fixes are not in current `dev`, so syncing now is not a cure for the observed failures. Keep the Avro layout change separate; after the relevant upstream corrections land, sync and rerun the failed jobs, including the required round-trip coverage. My earlier source-level conclusion is unchanged; this is a discussion/CI follow-up, not a fresh full review or an approval. -- 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]
