[GitHub] [spark] yaooqinn commented on a change in pull request #32373: [SPARK-35248] Spark shall load system class first in IsolatedClientLoader

[GitBox <gi...@apache.org>]

yaooqinn commented on a change in pull request #32373:
URL: https://github.com/apache/spark/pull/32373#discussion_r623541690



##########
File path: sql/hive/src/main/scala/org/apache/spark/sql/hive/HiveUtils.scala
##########
@@ -398,7 +398,7 @@ private[spark] object HiveUtils extends Logging {
         case childFirst: ChildFirstURLClassLoader =>
           childFirst.getURLs() ++ allJars(Utils.getSparkClassLoader)
         case urlClassLoader: URLClassLoader =>
-          urlClassLoader.getURLs ++ allJars(urlClassLoader.getParent)
+          allJars(urlClassLoader.getParent) ++ urlClassLoader.getURLs

Review comment:
       I am 100% sure this change will 'guarantee' the class loading order. Because in `ChildFirstURLClassLoader` we re-constructor it by setting the parent classloader to null and load classes in an order we explicitly set.




-- 
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.

For queries about this service, please contact Infrastructure at:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org