SEZ9 commented on PR #11912: URL: https://github.com/apache/seatunnel/pull/11912#issuecomment-6393034951
Thanks for the detailed walkthrough against head `d3a8a348ddb0`. That maps cleanly onto the earlier points: - Eager applier creation (F1/F2/F7): lazy creation in `createSchemaChangeApplier(TablePath)`, with no appliers built in `createWriter()` / `restoreWriter()`. - Nondeterministic resolution / inconsistent keying (F3): logical-key lookup with a `resolvePhysicalSinkTable(...)` fallback that throws an `IOException` on ambiguous physical matches, and the applier created for the resolved physical path. - Identifier safety (F4): the `SchemaChangeApplier.apply()` javadoc now covers untrusted identifiers and idempotent re-delivery. - Writer contract (F6): fail-fast in `FlinkSink` when a sink declaring coordinated support creates a writer that does not implement `SupportSchemaRefreshSinkWriter`. - Docs and upgrade note (F5/F8): the sink-connector-development, schema-evolution, and incompatible-changes docs in both languages. I'll re-check each of these against the diff for `d3a8a348ddb0` and close out the findings that are covered. <!-- streview-comment:2021 --> -- 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]
