You are viewing a plain text version of this content. The canonical link for it is here.

Posted to jira@kafka.apache.org by GitBox <gi...@apache.org> on 2020/12/17 17:46:58 UTC

[GitHub] [kafka] hachikuji commented on pull request #9689: KAFKA-10740; Replace OffsetsForLeaderEpochRequest.PartitionData with automated protocol

hachikuji commented on pull request #9689:
URL: https://github.com/apache/kafka/pull/9689#issuecomment-747594813


   @dajac Yeah, I'm definitely in favor of the improvements you have suggested, but I'm not sure they 100% address the concern. The case of `Optional` is an interesting example. A field might become optional because it is not supported in older version of the protocol. That optionality leaks into the code even when we can guarantee that internal data structures will have the field present. Anyway, I still think we should be thinking about these conversions when they make sense. Probably down the Fetch paths (including the replica fetcher) is the biggest opportunity for performance improvements. I'm just highlighting some downsides.


----------------------------------------------------------------
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