Github user JoshRosen commented on a diff in the pull request:
https://github.com/apache/spark/pull/6361#discussion_r31438253
--- Diff:
core/src/main/scala/org/apache/spark/util/SerializableBuffer.scala ---
@@ -21,12 +21,16 @@ import java.io.{EOFException, IOException,
ObjectInputStream, ObjectOutputStream
import java.nio.ByteBuffer
import java.nio.channels.Channels
+import com.esotericsoftware.kryo.{Kryo, KryoSerializable}
+import com.esotericsoftware.kryo.io.{Input, Output}
+
/**
* A wrapper around a java.nio.ByteBuffer that is serializable through
Java serialization, to make
* it easier to pass ByteBuffers in case class messages.
*/
private[spark]
-class SerializableBuffer(@transient var buffer: ByteBuffer) extends
Serializable {
+class SerializableBuffer(@transient var buffer: ByteBuffer)
+ extends Serializable with KryoSerializable {
--- End diff --
I think that one gotcha is that users can configure Kryo to require that
classes are registered and throw an error if they're not.
It might be nice to try running some of the basic Kryo jobs with that
setting enabled (I think it's called something like `registrationRequired`, if
you want to grep for that).
---
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
