pnowojski commented on code in PR #19723:
URL: https://github.com/apache/flink/pull/19723#discussion_r876793547
##########
flink-runtime/src/main/java/org/apache/flink/runtime/checkpoint/channel/ChannelStateWriteRequest.java:
##########
@@ -72,6 +74,51 @@ static ChannelStateWriteRequest write(
(writer, buffer) -> writer.writeOutput(info, buffer));
}
+ static ChannelStateWriteRequest write(
+ long checkpointId,
+ ResultSubpartitionInfo info,
+ CompletableFuture<List<Buffer>> dataFuture) {
+ return buildFutureWriteRequest(
+ checkpointId,
+ "writeOutputFuture",
+ dataFuture,
+ (writer, buffer) -> writer.writeOutput(info, buffer));
+ }
+
+ static ChannelStateWriteRequest buildFutureWriteRequest(
+ long checkpointId,
+ String name,
+ CompletableFuture<List<Buffer>> dataFuture,
+ BiConsumer<ChannelStateCheckpointWriter, Buffer> bufferConsumer) {
+ return new CheckpointInProgressRequest(
+ name,
+ checkpointId,
+ writer -> {
+ try {
+ List<Buffer> buffers = dataFuture.get();
+ if (buffers == null || buffers.isEmpty()) {
Review Comment:
Yes, I think there is no performance benefit for having null instead of
`Collections.emptyList()` in this case (checkpoints are happening too rarely),
while it's a bit safer (avoiding NPE)
--
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: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
