SEZ9 commented on PR #12944: URL: https://github.com/apache/seatunnel/pull/12944#issuecomment-6485989485
Thanks for the update, @Nikk8091. On the documentation and Javadoc changes in `eba1ccb9`: the description matches what was asked for (corrected EN/ZH precedence statements, key contract documented on `MetadataProvider.datasourceMap`, `resolveDataSourceConfigs`, `mergeConfig`, and in both SPI documents). I have not yet verified the diff itself, so I'll confirm those points once I've reviewed the actual changes rather than closing them out here. Deferring the full-job round-trip short-circuit to a separate follow-up is fine while the shared helper is unmerged; please link that follow-up here once it exists so it isn't lost. On coverage: thanks for the clarification on which rows exercise the bypass path versus the rebuild paths. I'll take another look at the test when I review the diff to confirm the metadata-enabled no-datasource rows reach the reparse boundary. Also appreciated: stating plainly that `eba1ccb9` still stops at the unresolved `ConfigMapPathUtils` import and that the earlier combined-tree results against `3d6c855f` are historical only. That is the right status to record, and the compile issue remains open until the prerequisite lands. Remaining asks: 1. Once the prerequisite helper merges, rebase onto that base and confirm the module compiles standalone. 2. Rerun the metadata unit tests and `MetalakeIT.testDataSource` in CI on the rebased head and post the result here. 3. For the assertions on the test's own `datasource` fixture map, either compare against a snapshot taken before the merge call or drop those assertions — a one-line note on which you chose is enough. <!-- streview-comment:2105 --> -- 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]
