yew1eb commented on code in PR #1958:
URL: https://github.com/apache/auron/pull/1958#discussion_r2740373877
##########
spark-extension-shims-spark/src/test/scala/org/apache/auron/BaseAuronSQLSuite.scala:
##########
@@ -58,5 +58,8 @@ trait BaseAuronSQLSuite extends SharedSparkSession {
.set("spark.auron.enable", "true")
.set("spark.ui.enabled", "false")
.set("spark.sql.warehouse.dir", warehouseDir)
+ // Avoid the code size overflow error in Spark code generation.
+ .set("spark.sql.codegen.wholeStage", "false")
+ .set("spark.sql.codegen.factoryMode", "NO_CODEGEN")
Review Comment:
> This change may not ensure CI passes.
Why?
@cxzl25 , CI UT failures are triggered by codegen at
`DeserializeToObjectExec.$anonfun$doExecute$1(objects.scala:95)`(https://github.com/apache/auron/issues/1961)
– this part is not controlled by `spark.sql.codegen.xxx` configs, so setting
these params specifically for Spark 4.1 has no effect.
The odd thing is: Spark 4.1 CI UTs pass fine if we globally disable codegen
via `spark.sql.codegen.xxx` for all CI UTs.
--
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]
