我有一个关于specificavrodeserializer的问题，因为类名，我一直假设specificavroserializer将产生specificrecord，specificavrodeserializer将能够使用它。
令我惊讶的是，如果我不设置'specific.avro.reader'属性，specificavrodeserializer尝试将字节数组作为genericrecord读取，因为类“abstractkafkaavrodeserializer”中有以下代码行(https://github.com/confluentinc/schema-registry/blob/master/avro-serializer/src/main/java/io/confluent/kafka/serializers/abstractkafkaavrodeserializer.java)..

private DatumReader getDatumReader(Schema writerSchema, Schema readerSchema) {
boolean writerSchemaIsPrimitive =
    AvroSchemaUtils.getPrimitiveSchemas().values().contains(writerSchema);
// do not use SpecificDatumReader if writerSchema is a primitive
if (useSpecificAvroReader && !writerSchemaIsPrimitive) {
  if (readerSchema == null) {
    readerSchema = getReaderSchema(writerSchema);
  }
  return new SpecificDatumReader(writerSchema, readerSchema);
} else {
  if (readerSchema == null) {
    return new GenericDatumReader(writerSchema);
  }
  return new GenericDatumReader(writerSchema, readerSchema);
}

}
我以为具体的avroDeserializer的构造函数应该设置这个，但它没有(https://github.com/confluentinc/schema-registry/blob/master/avro-serde/src/main/java/io/confluent/kafka/streams/serdes/avro/specificavrodeserializer.java)
那我还缺什么？