[
https://issues.apache.org/jira/browse/FLINK-2125?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14802964#comment-14802964
]
ASF GitHub Bot commented on FLINK-2125:
---------------------------------------
Github user fhueske commented on a diff in the pull request:
https://github.com/apache/flink/pull/1077#discussion_r39749641
--- Diff:
flink-staging/flink-streaming/flink-streaming-core/src/main/java/org/apache/flink/streaming/api/functions/source/SocketTextStreamFunction.java
---
@@ -117,12 +122,13 @@ private void streamFromSocket(SourceContext<String>
ctx, Socket socket) throws E
continue;
}
- if (data == delimiter) {
- ctx.collect(buffer.toString());
- buffer = new StringBuilder();
- } else if (data != '\r') { // ignore carriage
return
- buffer.append((char) data);
+ buffer.append(charBuffer, 0, readCount);
+ String[] splits =
buffer.toString().split(delimiter);
--- End diff --
`String.split()` and `String.replace()` create new String objects which
must be garbage collected. This adds quite some overhead, because these
functions are called very often. The previous implementation was operating on a
byte-level and avoiding the creation of new objects. It would be good if we
could preserve this behavior.
> String delimiter for SocketTextStream
> -------------------------------------
>
> Key: FLINK-2125
> URL: https://issues.apache.org/jira/browse/FLINK-2125
> Project: Flink
> Issue Type: Improvement
> Components: Streaming
> Affects Versions: 0.9
> Reporter: Márton Balassi
> Priority: Minor
> Labels: starter
>
> The SocketTextStreamFunction uses a character delimiter, despite other parts
> of the API using String delimiter.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
