Allow users of kafka-avro-serializer to use any Scala version.

[ewencp]

Make Kafka a provided dependency for kafka-avro-serializer. The serializer still
needs the core jar that depends on Scala, but doesn't use any Scala
code. Instead of building multiple jars for different Scala versions, we can
instead require them to provide the dependency themselves which they'll already
be doing in order to use those interfaces. Additionally, this works without the
jar available if they only use interfaces from kafka-clients, i.e. new
producer/consumer interfaces. Fixes #60.

[ewencp]

Fixing #60 turns out to be trivial since one of Maven's scopes provides for the behavior we want. If you want to verify, I applied the patch, installed locally, then used the following two files to check if failed as expected by commenting/uncommenting the kafka dependency and sections of the test program.

<!-- pom.xml -->
<project xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
         xmlns="http://maven.apache.org/POM/4.0.0"
         xsi:schemaLocation="http://maven.apache.org/POM/4.0.0
         http://maven.apache.org/maven-v4_0_0.xsd">

    <modelVersion>4.0.0</modelVersion>

    <groupId>io.confluent</groupId>
    <artifactId>testing</artifactId>
    <packaging>jar</packaging>
    <name>testing</name>
    <version>0.1-SNAPSHOT</version>

    <dependencies>
        <dependency>
            <groupId>org.apache.kafka</groupId>
            <artifactId>kafka-clients</artifactId>
            <version>0.8.2.0</version>
        </dependency>
        <dependency>
            <groupId>org.apache.kafka</groupId>
            <artifactId>kafka_2.11</artifactId>
            <version>0.8.2.0</version>
        </dependency>
        <dependency>
            <groupId>io.confluent</groupId>
            <artifactId>kafka-avro-serializer</artifactId>
            <version>0.1-SNAPSHOT</version>
        </dependency>
    </dependencies>
</project>

// src/main/java/io/confluent/testing/Test.java
package io.confluent.testing;

import java.util.Properties;

import kafka.consumer.Consumer;
import kafka.consumer.ConsumerConfig;
import kafka.javaapi.consumer.ConsumerConnector;

import org.apache.kafka.clients.producer.KafkaProducer;
import org.apache.kafka.clients.producer.ProducerConfig;
import org.apache.kafka.clients.producer.ProducerRecord;

public class Test {
    public static void main(String[] args) {
        Properties cprops = new Properties();
        cprops.put("zookeeper.connect", "localhost:2181");
        cprops.put("group.id", "foobar");
        cprops.put("consumer.id", "baz");
        ConsumerConnector consumer;
        System.out.println("Creating connector");
        consumer = Consumer.createJavaConsumerConnector(new ConsumerConfig(cprops));
        consumer.shutdown();

        try {
            Properties props = new Properties();
            props.put(ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, "io.confluent.kafka.serializers.KafkaAvroSerializer");
            props.put(ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, "io.confluent.kafka.serializers.KafkaAvroSerializer");
            props.put(ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, "localhost:9092");
            System.out.println("Creating producer");
            KafkaProducer producer = new KafkaProducer<Object,Object>(props);
            producer.send(new ProducerRecord<Object,Object>("foo", "bar")).get();
        } catch (Exception e) {
        }
    }
}

[nehanarkhede]

So provided ensures that it can be used at compile time but not for packaging?

[ewencp]

Yeah, see the description here: http://maven.apache.org/guides/introduction/introduction-to-dependency-mechanism.html#Dependency_Scope -- basically it's used during compile of the package, but anything that runs/links the application/library are expected to provide the dependency themselves.

[nehanarkhede]

Thanks! Got it.

[nehanarkhede]

+1