SEZ9 commented on PR #12048: URL: https://github.com/apache/seatunnel/pull/12048#issuecomment-5691639338
Thanks for the update in e5ffcd0f1, @goutamadwant. **F2 (ServicesResourceTransformer)** — agreed with your reading: the root POM already supplies `ServicesResourceTransformer` for the shade goal, so no duplicate child configuration is needed for the relocated com.azure / io.netty / org.apache.qpid.proton / Jackson service providers. Your note that the freshly shaded JAR loads the relocated Jackson service providers in isolation matches that. I'm treating F2 as resolved. Also noted that the move from `f3db7dfd3de2` to `e5ffcd0f19dc` is a `dev` sync with no changes to the connector module, its e2e module, docs, or the shared registration files, so nothing else in the previous review scope has moved this round. Remaining asks before I can approve: 1. **Event Hubs emulator E2E** — as you mentioned, it was skipped in the previous run and still needs a fresh run. Please post back once it has actually executed (not skipped) on the current head, since it is the only thing exercising the shaded JAR end-to-end. 2. **F1 (retention-trimmed checkpoint position → crash loop)** — still open. Please either add an operator escape hatch (e.g. a configurable fallback when the checkpointed sequence number is no longer retained) or document explicitly why a hard failure is the intended behavior. 3. **F3 (connection-string masking on every output path)** — please confirm the option is flagged sensitive, and that the EntityPath-rejection error message and split/config `toString` never echo the raw SAS key. 4. **F4 (unpinned transitive Netty/Jackson/Proton in the shaded JAR)** — please pin these through the parent's dependency management (or explain why the Azure SDK-resolved versions must be kept as-is) so they stay visible for future CVE remediation. 5. **F5 (EntityPath rejection blocks hub-scoped SAS policies)** — please either accept EntityPath (validating it matches the configured hub) or document the least-privilege limitation. 6. **F6 (`prefetch_count >= max_batch_size`)** — this cross-field check should fail at job submission rather than at reader open on the worker. 7. **F7 (no partition rediscovery)** and **F8 (no reconnect/retry tuning)** — if these are intentionally out of scope for this PR, please state that in the docs' limitations section and open follow-up issues; otherwise let me know your plan. A short per-item status (fixed / deferred with rationale) in your next comment would make the final pass quick. <!-- streview-comment:1086 --> -- 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]
