[
https://issues.apache.org/jira/browse/KAFKA-3046?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15247174#comment-15247174
]
Guozhang Wang commented on KAFKA-3046:
--------------------------------------
[~vesense] I'm currently working on KAFKA-3499 and I'm considering adding
another Bytes class wrapping immutable byte arrays for serde
(https://github.com/apache/kafka/pull/1229/files/9302bc18c8fa400b1fb7d4b36f29c8f0c812785c#r60079246).
Since your addition of ByteBufferSerde is not included in any releases yet, I'm
wondering if I can just replace these classes with BytesSerializer and
BytesDeserializer, since having both Bytes and ByteBuffer for serde sounds
duplicates. What's your usage in spark-sql? Does that must require a mutable
ByteBuffer, or does immutable Bytes work for you? Please let me know your
thoughts.
> add ByteBuffer Serializer&Deserializer
> --------------------------------------
>
> Key: KAFKA-3046
> URL: https://issues.apache.org/jira/browse/KAFKA-3046
> Project: Kafka
> Issue Type: New Feature
> Components: clients
> Reporter: Xin Wang
> Fix For: 0.10.0.0
>
>
> ByteBuffer is widely used in many scenarios. (eg: storm-sql can specify kafka
> as the external data Source, we can use ByteBuffer for value serializer.)
> Adding ByteBuffer Serializer&Deserializer officially will be convenient for
> users to use.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
