hutiefang76 commented on PR #12506: URL: https://github.com/apache/seatunnel/pull/12506#issuecomment-5864024173
Thanks for tracing the upstream-database fallback. I kept the strict behavior and added the explicit error during DuckDB save-mode initialization, including IGNORE. The factory regression now covers an inherited mydb name, checks that the existing main table is untouched, and verifies the migration with database=main. Both language guides and incompatible-changes now describe this, the reserved aliases, and the discovery limit. Catalog matching is case-insensitive and shares the alias rule. I also checked the neighboring paths: CREATE TABLE was still dropping the catalog qualifier, and the catalog injected user=duckdb/password placeholders that can clash with a simultaneous writer connection. Both now have deterministic regressions; the connection mismatch also reproduced directly with JDBC 1.3.1. JDK 8 and 17 each pass module verify: 40 executed tests, no failures/errors, plus one PG/S3 test skipped. The local DuckLake and SQLite extensions were enabled for the actual lake tests. I made the default-CI gating and the earlier distributed evidence explicit in the description. -- 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]
