Github user mjsax commented on a diff in the pull request:
https://github.com/apache/flink/pull/1398#discussion_r45735762
--- Diff:
flink-contrib/flink-storm-examples/src/main/java/org/apache/flink/storm/util/BoltFileSink.java
---
@@ -18,20 +18,23 @@
package org.apache.flink.storm.util;
import backtype.storm.task.TopologyContext;
+import org.apache.flink.core.fs.FSDataOutputStream;
+import org.apache.flink.core.fs.FileSystem;
+import org.apache.flink.core.fs.Path;
import java.io.BufferedWriter;
-import java.io.FileWriter;
import java.io.IOException;
+import java.io.OutputStreamWriter;
import java.util.Map;
/**
- * Implements a sink that write the received data to the given file (as a
result of {@code Object.toString()} for each
+ * Implements a sink that writes the received data to the given file (as a
result of {@code Object.toString()} for each
* attribute).
*/
public final class BoltFileSink extends AbstractBoltSink {
private static final long serialVersionUID = 2014027288631273666L;
- private final String path;
+ private final Path path;
private BufferedWriter writer;
--- End diff --
No. People will have spout/bolt code they do not want to touch when running
it in Flink. Thus there code will be written in the same way -- and so should
be example be. Otherwise, we deliver the impression they need to change there
code -- but they don't. Thus, we implement the example Spout/Bolts in a pure
Storm way. Of course, if somebody develops a new Spout/Bolt with Flink in mind,
your approach makes sense. However, this is not the main focus (it would be
even better if the code new stuff Flink native in embedded mode, instead of
develop Spout/Bolts which are Flink tailored).
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
