DanielLeens commented on issue #11651: URL: https://github.com/apache/seatunnel/issues/11651#issuecomment-5186334725
Thanks for breaking these benchmark misses into concrete pattern-level knowledge gaps. This reads as a focused knowledge-layer enhancement rather than a runtime bug, and the scope is reasonable because it stays in skills / examples / prompt guidance instead of changing SeaTunnel execution semantics. The strongest part of the proposal is that the three gaps are specific and measurable: 1. conditional routing with mutually exclusive predicates; 2. PostgreSQL-CDC-specific prerequisites and naming rules; 3. wide-DAG wiring after mid-chain transforms. Before this turns into a PR, please keep the boundary explicit: - content-only changes in the CLI knowledge layers and prompt templates; - before/after benchmark evidence on the affected task subset from #11553; - concrete examples that show the exact failure modes being corrected, not only the final "good" config. If you keep it in that shape, this looks like a useful incremental feature to continue. -- 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]
