SEZ9 commented on issue #12834: URL: https://github.com/apache/seatunnel/issues/12834#issuecomment-6229097631
Thanks for the detailed report and reproduction. The proposed fix sounds reasonable: on a cache miss by unqualified name, prefer the type visible on the source connection's `search_path`, fall back to deterministic OID ordering when no candidate is visible, and leave the cached-name and OID lookups unchanged. Thanks as well for opening #12836 with the regression test; I'll take a look there. One question: have you seen the wrong OID surface in a running job (e.g. decoding failures or wrong column types after a new type is added), or was this observed only through the `TypeRegistry` API? That helps with prioritization and deciding whether a release note is needed. Agreed that this is independent of #12601; please keep the change scoped to the source-side `TypeRegistry` lookup so the two can be reviewed separately. <!-- streview-comment:1826 --> -- 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]
