srowen commented on code in PR #40808:
URL: https://github.com/apache/spark/pull/40808#discussion_r1190496413
##########
core/src/main/scala/org/apache/spark/network/netty/NettyBlockTransferService.scala:
##########
@@ -51,6 +51,7 @@ import org.apache.spark.util.Utils
private[spark] class NettyBlockTransferService(
conf: SparkConf,
securityManager: SecurityManager,
+ serializerManager: SerializerManager,
Review Comment:
I think I'm asking something narrower, and maybe still misinformed.
Right now you've basically lifted "new JavaSerializer(conf)" into the
caller, where it's wrapped in a SerializerManager, then passed back in.
NettyBlockTransferService, for instance, already takes 'conf' as an argument.
What's different if you make the JavaSerializer and SeralizerManager inside the
class, from 'conf'? I'm just missing how it's different.
I see that this isn't quite the case in a few tests though, OK. Maybe that's
reason enough for this setup, OK.
--
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]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
