KYinXu opened a new pull request, #5680: URL: https://github.com/apache/texera/pull/5680
### What changes were proposed in this PR? Pin the runtime reconfiguration contract of `MapOpDesc` — the abstract base class for map-family operator descriptors (e.g. `ProjectionOpDesc`, `TypeCastingOpDesc`, `DictionaryMatcherOpDesc`). Its only override is `runtimeReconfiguration`, which always succeeds, dispatches to the new descriptor's `getPhysicalOp`, ignores the old descriptor, and returns no state-transfer function. Any drift in that policy would silently break mid-run operator edits for every map-family operator. No production-code changes. | Spec | Source Class | Tests | | --- | --- | --- | | `MapOpDescSpec` | `MapOpDesc` | 4 | Spec file name follows the `<srcClassName>Spec.scala` one-to-one convention. `MapOpDesc` is abstract, so the spec exercises it through a minimal test-only concrete subclass (`RecordingMapOpDesc`) that returns a recognizable sentinel `PhysicalOp` from `getPhysicalOp` and records the `workflowId` / `executionId` arguments passed in. **Behavior pinned — `MapOpDesc.runtimeReconfiguration`** | Surface | Contract | | --- | --- | | `runtimeReconfiguration` | returns `Success(...)`, never `Failure`, regardless of inputs | | Result's `PhysicalOp` | same instance as `newOpDesc.getPhysicalOp(workflowId, executionId)` (reference identity) | | Result's `StateTransferFunc` | `None` — map operators are stateless and need no state transfer across reconfiguration | | `oldOpDesc` argument | ignored — two different `oldOpDesc` instances (`DummyOpDesc` vs `ProjectionOpDesc`) produce identical results when `newOpDesc` is held constant | | Distinct `workflowId` / `executionId` | propagate through to `newOpDesc.getPhysicalOp` (the same arguments must reach the new desc's call) | ### Any related issues, documentation, discussions? Closes #5664 ### How was this PR tested? Unit tests were added here, tested through local verification: - `sbt "WorkflowOperator/testOnly org.apache.texera.amber.operator.map.MapOpDescSpec"` — all 4 tests passed - `sbt scalafmtCheckAll` — clean ### Was this PR authored or co-authored using generative AI tooling? Generated-by: Composer 2.5 Fast -- 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]
