[
https://issues.apache.org/jira/browse/SPARK-60535?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ASF GitHub Bot updated SPARK-60535:
-----------------------------------
Labels: pull-request-available (was: )
> Investigate Non-ANSI nested-class assertion in the MergeRows codegen
> regression test
> ------------------------------------------------------------------------------------
>
> Key: SPARK-60535
> URL: https://issues.apache.org/jira/browse/SPARK-60535
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Affects Versions: 5.0.0
> Reporter: Xinyao Zhang
> Priority: Minor
> Labels: pull-request-available
>
> Parent: SPARK-60531.
> The Non-ANSI master run on 2026-10-23 fails WholeStageCodegenSuite's
> "SPARK-60403: MergeRows compiles with its function spilled to a nested class"
> with SPARK_ANSI_SQL_MODE=false (JDK 17, Scala 2.13).
> At WholeStageCodegenSuite.scala:3215 in that run, code.exists(c =>
> qualifiedCall.findFirstIn(c.body).nonEmpty) is false. This assertion checks
> for a qualified mergeProcessRow call through a nested-class instance. The
> failure alone does not demonstrate a generated-code compilation error.
> Evidence: https://github.com/apache/spark/actions/runs/39883743642
> Tested commit: fce7d51a3760953097fe34df8aa9a4f5326c74ff
> The test was introduced by https://github.com/apache/spark/pull/59653, which
> resolved SPARK-60403.
> Investigate whether Non-ANSI code generation changes the test's assumed code
> shape, or exposes a remaining implementation defect. Local reproduction and
> root cause are not yet established. Verify the regression coverage with ANSI
> mode disabled and enabled.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
