SEZ9 commented on PR #12079:
URL: https://github.com/apache/seatunnel/pull/12079#issuecomment-5788485766

   Thanks for the detailed walkthrough of `d8045271fcf6` against `6dcbe452af3`. 
I haven't been able to verify the current diff myself yet, so before I sign off 
on the earlier findings I'd like to confirm a few things against the actual 
changes:
   
   - **F1 (duplicate output IDs):** Releasing dependents once per distinct 
output ID rather than once per producer is the right approach. Could you point 
me at the diff hunk and the tests covering the duplicate-output case so I can 
confirm?
   - **F2 / F4 (omitted `plugin_input`):** Distinguishing an omitted input from 
an explicit empty list sounds correct. Is the last-inserted-table fallback now 
applied for omitted inputs regardless of transform count (i.e. no longer 
limited to single-transform jobs)? Same ask for a pointer to the change and 
tests.
   - **F3 / F5 (shared scheduler):** Good to hear there is a single 
implementation used by both dry-run validation and the runtime parser. For F3 
specifically, please confirm that the scheduler and the validation step now 
resolve input IDs through the same code path.
   - **F6 (single-transform self-cycle):** Does a single transform that lists 
its own output as `plugin_input` now fail dry-run with a cycle error, 
consistent with the runtime parser? A test for that case would settle it.
   - **F7 (docs):** Is there a docs/upgrade-note change describing the new 
fail-fast rejection of cyclic/unresolved transform graphs? I don't see one 
referenced.
   - **F8 (helper duplication / unused `getOutputId()`):** Were the redundant 
static helpers removed in the consolidation, and is `getOutputId()` now used or 
dropped?
   
   Happy to approve once I can see the diff for these.
   
   <!-- streview-comment:1257 -->


-- 
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]

Reply via email to