yzeng1618 commented on PR #10990: URL: https://github.com/apache/seatunnel/pull/10990#issuecomment-5140895395
I checked the current PR head again: `ad4ad6b8cfd5`. The four findings quoted above were based on the earlier `74ba70f` head; the current head is not unchanged from that revision. 1. **`SourceUri` sanitization** is now explicitly required in both the `SOURCE_URI` Javadoc and EN/ZH `metadata.md`: producers must remove URI user info, access tokens, signatures, and other transient authentication material before writing `SourceUri` to row options. 2. **`MetadataUtil.isMetadataField` scope**: the added keys are logical metadata keys for explicit `Metadata` transform projection. The transform still requires the input `metadataSchema` to declare the key and reads the value from row options; same-named physical columns are not classified or read as metadata. In the current repository, the direct caller is the `Metadata` transform, so this does not silently change ordinary physical columns or pipelines that do not opt into the mapping. 3. **Nullability** is now part of the enum contract. `DocumentId` and `Deleted` are non-nullable; chunk fields remain nullable specifically for document tombstones. This is covered by `KnowledgeSyncMetadataFieldTest`. 4. **Lookup cost** is now a precomputed immutable `FIELD_NAMES` set, so `isKnowledgeSyncMetadataField` does not scan `values()` per invocation. The current diff is also focused on the Knowledge Sync API/Metadata transform/docs/tests only; the prior unrelated hunks have been removed. The required Build check is green on this head. For branch freshness, the current branch is 10 commits behind `dev` (not 192). I will refresh it against the current base before final merge validation. -- 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]
