jolshan commented on a change in pull request #11331:
URL: https://github.com/apache/kafka/pull/11331#discussion_r744138204
##########
File path: core/src/main/scala/kafka/server/FetchSession.scala
##########
@@ -163,18 +173,35 @@ class CachedPartition(val topic: String,
mustRespond
}
- override def hashCode: Int = Objects.hash(new TopicPartition(topic,
partition), topicId)
+ /**
+ * We have different equality checks depending on whether topic IDs are used.
+ * This means we need a different hash function as well. We use name to
calculate the hash if the ID is zero and unused.
+ * Otherwise, we use the topic ID in the hash calculation.
+ *
+ * @return the hash code for the CachedPartition depending on what request
version we are using.
+ */
+ override def hashCode: Int = if (topicId != Uuid.ZERO_UUID) (31 * partition)
+ topicId.hashCode else
+ (31 * partition) + topic.hashCode
def canEqual(that: Any): Boolean = that.isInstanceOf[CachedPartition]
+ /**
+ * We have different equality checks depending on whether topic IDs are used.
+ *
+ * This is because when we use topic IDs, a partition with a given ID and an
unknown name is the same as a partition with that
+ * ID and a known name. This means we can only use topic ID and partition
when determining equality.
+ *
+ * On the other hand, if we are using topic names, all IDs are zero. This
means we can only use topic name and partition
+ * when determining equality.
+ */
override def equals(that: Any): Boolean =
that match {
case that: CachedPartition =>
this.eq(that) ||
(that.canEqual(this) &&
Review comment:
Confirmed this was a strange quirk from 4 years ago
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: jira-unsubscr...@kafka.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
