kazantsev-maksim commented on PR #4744: URL: https://github.com/apache/datafusion-comet/pull/4744#issuecomment-5836253150
Thanks for catching that, @sunchao. You were completely right — the `ShortCircuitBinaryExpr` implementation was accidentally missing from the previous commit. I have now properly committed and pushed all the Rust and physical planning changes to the branch. To address your point about asserting native execution and prevent any silent JVM codegen fallback, I have also explicitly disabled the codegen dispatcher (`spark.comet.exec.scalaUDF.codegen.enabled=false`) in the regression tests where native execution is expected. This guarantees that DataFusion's native runtime is directly exercised. All tests now pass natively without fallback, including: - Stateful expressions (`monotonically_increasing_id()` counter no longer advances when skipped, returning `[0, 1]`) - PRNG sequence preservation with `rand()` - `element_at(..., 0)` and `abs(INT_MIN)` in guarded branches - Guarded division on `[0, 1]` under ANSI mode - SQL 3VL nullable boolean conditions and nested predicates Could you please take another look when you have a chance? -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
