huanghanyu-vungle opened a new pull request, #57371: URL: https://github.com/apache/spark/pull/57371
### What changes were proposed in this pull request? This PR marks `XPathExtract`, `CallMethodViaReflection`, and `UserDefinedGenerator` as stateful Catalyst expressions. It also adds focused regression tests verifying that `freshCopyIfContainsStatefulExpression()` returns a distinct expression instance while preserving the expression's children and configuration. ### Why are the changes needed? These expressions retain mutable runtime state, including an XPath evaluator, a reusable reflection argument buffer, and generator projections/converters. They currently inherit `stateful = false`, so independent query executions can reuse the same expression instance. Concurrent executor tasks may then race on the shared state and silently produce incorrect results. Declaring the expressions stateful makes Catalyst create a fresh, uninitialized expression for each execution path using its existing fresh-copy mechanism. ### Does this PR introduce _any_ user-facing change? Yes. Concurrent query executions no longer risk incorrect results caused by shared mutable state in these expressions. There is no SQL API or syntax change. ### How was this patch tested? Added fresh-copy regression coverage to `XPathExpressionSuite`, `CallMethodViaReflectionSuite`, and `GeneratorExpressionSuite`. ```bash ./build/sbt "catalyst/testOnly org.apache.spark.sql.catalyst.expressions.xml.XPathExpressionSuite org.apache.spark.sql.catalyst.expressions.CallMethodViaReflectionSuite org.apache.spark.sql.catalyst.expressions.GeneratorExpressionSuite" ./build/sbt "catalyst/scalastyle" ``` The three suites passed 27 tests, and Catalyst Scalastyle processed 705 files with no findings. ### Was this patch authored or co-authored using generative AI tooling? Generated-by: OpenAI Codex (GPT-5) -- 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]
