You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kafka.apache.org by "Apurva Mehta (JIRA)" <ji...@apache.org> on 2017/05/18 17:47:04 UTC

[jira] [Updated] (KAFKA-5030) Suggested clean-ups from message format V2 review

     [ https://issues.apache.org/jira/browse/KAFKA-5030?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Apurva Mehta updated KAFKA-5030:
--------------------------------
    Labels: exactly-once  (was: )

> Suggested clean-ups from message format V2 review
> -------------------------------------------------
>
>                 Key: KAFKA-5030
>                 URL: https://issues.apache.org/jira/browse/KAFKA-5030
>             Project: Kafka
>          Issue Type: Sub-task
>          Components: clients, core, producer 
>            Reporter: Ismael Juma
>              Labels: exactly-once
>             Fix For: 0.11.0.0
>
>
> A few random clean-ups raised during the review of the message format V2 PR:
> 1. We should use magic or message format version consistently. It seems like we are moving towards the former.
> 2. We should consider moving Log.analyzeAndValidateRecords to LogValidator so that the validation logic is all in the same class.
> 3. MemoryRecords.toString should include batch information, it currently only includes records information.
> 4. Record.hasMagic should be hasCompatibleMagic to make the behaviour clearer. For version 2, it returns true if the passed magic is greater than or equal to 2.
> 5. Once we move to Java 8, move AbstractRecordBatch's methods to RecordBatch (i.e. default methods) and remove AbstractRecordBatch.
> 6. Consider moving some of the down-conversion logic from KafkaApis to FetchRequest so that it can be more easily unit testable.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)