[
https://issues.apache.org/jira/browse/KAFKA-1794?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14311038#comment-14311038
]
Jay Kreps commented on KAFKA-1794:
----------------------------------
[~navina] This is intentional. The whole config def system is meant to be
private, we don't want to leak that stuff. The only intended use of that class
is to get programmatic access to the property names. Can you expand on why you
want access to this stuff?
[~guozhang] I don't think we want people relying on those variables, though,
right? Ideally there would be a way to include something in both javadoc and
also as a code string, but I couldn't find a way. So right now the symbolic
names in ProducerConfig give their actual property name and all documentation
is pointed to the website. I think this makes sense, no?
> Make config and config defaults accessible to clients
> -----------------------------------------------------
>
> Key: KAFKA-1794
> URL: https://issues.apache.org/jira/browse/KAFKA-1794
> Project: Kafka
> Issue Type: Bug
> Affects Versions: 0.8.1.1
> Reporter: Navina Ramesh
>
> In the new Kafka producer API, the ProducerConfig is not accessible to the
> clients. Samza uses the ProducerConfig instance to access the defaults
> property values, which can then be used in the various helper utils. Config
> instance is accessible even without instantiating a Kafka producer.
> With the new API, there is no way to instantiate a ProducerConfig as the
> constructor is marked private. Also, it does not make the default config
> values accessible to the client without actually instantiating a
> KafkaProducer.
> Changes suggested:
> 1. Make the ProducerConfig constructor public
> 2. Make ConfigDef in ProducerConfig accessible by the client
> 3. Use public static variables for kafka config default "values"
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
