Github user aweisberg commented on a diff in the pull request:
https://github.com/apache/cassandra/pull/224#discussion_r187434827
--- Diff: src/java/org/apache/cassandra/batchlog/BatchlogManager.java ---
@@ -490,16 +497,16 @@ private static int gcgs(Collection<Mutation>
mutations)
{
private final Set<InetAddressAndPort> undelivered =
Collections.newSetFromMap(new ConcurrentHashMap<>());
- ReplayWriteResponseHandler(Collection<InetAddressAndPort>
writeEndpoints, long queryStartNanoTime)
+ ReplayWriteResponseHandler(Replicas writeReplicas, long
queryStartNanoTime)
{
- super(writeEndpoints,
Collections.<InetAddressAndPort>emptySet(), null, null, null,
WriteType.UNLOGGED_BATCH, queryStartNanoTime);
- undelivered.addAll(writeEndpoints);
+ super(writeReplicas, ReplicaList.of(), null, null, null,
WriteType.UNLOGGED_BATCH, queryStartNanoTime);
+ Iterables.addAll(undelivered, writeReplicas.asEndpoints());
--- End diff --
This is a case where it seems like we could call a Collections2.transform
version and not have to create a new list. One thing to keep in mind is that
Collections2.transform makes contains and remove slow because it's O(n) no
matter what the underlying collection is.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
