Hi devs - I saw https://github.com/apache/pinot/issues/7716, which caught my eye due to activity in the Flink project to remove Scala dependencies. I was curious why the

pinot-kafka

(0.9 and 2.0) plugins needed Scala, since my use of Kafka clients never triggered Scala dependencies. From the issue above, it’s all about being able to start up a Kafka server (

KafkaDataServerStartable

). And that only seems to be needed by test code, or the

tools

subproject when starting a server for quickstart. In my experience bundling test/demo code into core jars often leads to this type of dependency problem, so it seems better to have this as a test dependency, with a separate demo jar that provides Kafka server startup for quickstart.