DanielLeens commented on PR #11206: URL: https://github.com/apache/seatunnel/pull/11206#issuecomment-6190334215
Thanks @SEZ9. Taking your points in order, checked against `9007896511517e4b874142cfe704102861bbc444`. **F3 (`scan.newly-added-table.enabled` default): already flipped to `false`.** `MySqlIncrementalSourceOptions.java:65-71` declares `defaultValue(false)`, and `docs/en/connectors/source/MySQL-CDC.md` (option table row and the "Newly Added Tables" section) states the default `false` retains only the checkpointed table set and that `true` must be set explicitly. So existing wildcard jobs keep their checkpoint-restore behavior after upgrade and no upgrade note is needed for this option. Sorry this was easy to miss in my status comment; it was in the first bullet group of the follow-up on this head. **F1 / F8 test ask: agreed.** `MySqlSchemaTest` currently covers an unsafe name with a hyphen (`db1.unsafe-table`). I will add cases with a backtick and a `;`, plus a deserializer-level test asserting that a table whose identifier fails validation produces no `CreateTableEvent` and never reaches the sink. The MySQL-CDC doc already describes the enforced shape (ordinary unquoted identifier containing at least one letter, underscore or dollar sign); I will also add that all-numeric names are rejected and skipped with a WARN. **F2, F4, F5, F6, F7:** as in my previous reply these are not done yet. F2 (Javadoc on both `supportsNewlyCreatedTable` declarations saying a runtime `CreateTableEvent` is never silently dropped and the job fails when the sink has not opted in), F7 (explicit `exactly_once = false` in the en and zh JDBC docs) and your F6 request (a restore test with an old-format `MultiTableState`) will land in the next code push together with F4 and F5. **CI.** The fork Build on this head has two unrelated red groups: `all-connectors-it-2` (`CassandraIT` `testCassandra` and `testCassandraMultiTableSource`, container variant `[4]`, in `connector-cassandra-e2e`; this PR touches no Cassandra code) and `engine-v2-it` (`SplitClusterFaultToleranceIT.testStreamJobCancelResolvesWhenWorkerCrashesBeforeCancelAck`, a worker-loss cancel timing issue in the engine fault-tolerance suite). I will merge current `dev` with the next push and share the CI outcome then. I will ping once it is up. -- 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]
