Github user shivaram commented on a diff in the pull request: https://github.com/apache/spark/pull/16154#discussion_r91145714 --- Diff: sql/core/src/main/scala/org/apache/spark/sql/api/r/SQLUtils.scala --- @@ -247,7 +247,7 @@ private[sql] object SQLUtils extends Logging { dataType match { case 's' => // Read StructType for DataFrame - val fields = SerDe.readList(dis).asInstanceOf[Array[Object]] + val fields = SerDe.readList(dis, jvmObjectTracker = null).asInstanceOf[Array[Object]] --- End diff -- The two changes below are functions used in serializing and deserializing StructType from R to JVM (when calling collect or parallelize) Again I think since StructTypes themselves can't contain other JVM references I think its fine, but I'm less sure about this. (i.e what happens when we say embed an array into the structType). Another option here is to figure out which RBackend is calling this function and use its JVMObjectTracker ? Do we want to support cases where there are multiple active RBackends ? If not we can denote an `activeRBackend` and then use its jvm object tracker in these places ?
--- If your project is set up for it, you can reply to this email and have your reply appear on GitHub as well. If your project does not have this feature enabled and wishes so, or if the feature is enabled but not working, please contact infrastructure at infrastruct...@apache.org or file a JIRA ticket with INFRA. --- --------------------------------------------------------------------- To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org For additional commands, e-mail: reviews-h...@spark.apache.org