TianHengZhuang commented on PR #12512: URL: https://github.com/apache/seatunnel/pull/12512#issuecomment-5902306187
Both points taken, thanks. **Issue 1 (docs wording)**: reworded all four places. You are right that runtime rejection never went away, `FactoryUtil.createAndPrepareSink` already ran the option rule at job start, so saying the check moved from runtime to `--check` time was just wrong. The entry now says the constraint is declared in `optionRule()` via `Conditions.extension(...)`, that it is enforced as part of the option-rule check before the sink opens the ClickHouse connection and resolves shards, and that `--check` / `--dry-run` run it too. The Impact bullet now leads with the exception type change (`ClickhouseConnectorException` -> `OptionValidationException`) since that is the part operators actually have to react to, and the migration guide says to match on `OptionValidationException`. Same shape in zh. The connector doc sentence is now "declared in the sink's option rule ... before the sink connects to ClickHouse", and it mentions the empty value as well, which the old text did not. **Issue 2 (empty-value test)**: `emptyDelimiterIsRejected` now asserts the message contains `file_fields_delimiter`, using the same `assertTrue` with a message-carrying supplier as the multi-character case. On CI: agreed those three are unrelated (`RestApiHttpsForTruststoreTest`, the Windows `PayPalClientTest` timing flake, and deepwiki 429 on the dead-links job). I will rerun the two unit-test jobs. -- 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]
