zhang-arvin commented on PR #11866: URL: https://github.com/apache/seatunnel/pull/11866#issuecomment-6051905612
@DanielLeens Thanks for the deep trace against the actual jars — you're right, and I've corrected the record rather than defend the claim. I independently reproduced your finding: with the pinned versions (`opengauss-jdbc-5.1.0-og` registers as `org.opengauss.Driver`, accepts only `jdbc:opengauss:`/`jdbc:dws:iam:`; `postgresql-42.4.3` registers as `org.postgresql.Driver`, accepts only `jdbc:postgresql:`), there is neither a class-name collision nor a URL ambiguity, so the original #10242 mechanism does not apply to what ships here. Actions taken: 1. **Root-cause claim corrected** — rewrote the PR description so it no longer represents this as the root-cause fix for #10242. It is now framed as defensive hardening of driver selection (relevant when a driver genuinely collides by class name, e.g. an older non-`-og` `opengauss-jdbc`, or other classloading/ordering non-determinism), with the issue link marked informational. 2. **Real-driver test added (Issue 1)** — `RealJdbcDriverAcceptsUrlTest` uses the actual `org.postgresql.Driver` and `org.opengauss.Driver` classes (not the synthetic stubs): asserts each only accepts its own URL prefix, and that they register under distinct class names. If a future driver upgrade reintroduces a real collision, this fails and forces a re-check of the premise. Local: `Tests run: 3, Failures: 0`. Pushed head `970cbcacd` (also rebased onto the current head; prior content unchanged). -- 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]
