1. I haven't thought this through, but one option would be by using generics. 
Other by extracting some common interfaces. With a base class of 
`FlinkKafkaConsumerBase`, you could probably use it in the tests with defining 
one abstract method `protected FlinkKafkaConsumerBase 
createFlinkKafkaConsumer(...)`.
2. I think we should refactor old tests and deduplicate the code. It's very 
depressing when basically one variable rename `sequenceNumbers` to 
`nextSequence ` is forcing us to add ~3000 lines of codes. 
3. I don't know, since I have no idea what's the status of 1.1 support by our 
`0.11` connector or this new `1.0` connector. If this newly added `1.0` 
connector can support without any issues Kafka `1.1`, there might be no need to 
add new module, but still this should be documented.

[ Full content available at: https://github.com/apache/flink/pull/6577 ]
This message was relayed via gitbox.apache.org for [email protected]

Reply via email to