[
https://issues.apache.org/jira/browse/FLINK-1909?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14507002#comment-14507002
]
ASF GitHub Bot commented on FLINK-1909:
---------------------------------------
Github user rmetzger commented on a diff in the pull request:
https://github.com/apache/flink/pull/610#discussion_r28868234
--- Diff:
flink-staging/flink-streaming/flink-streaming-core/src/main/java/org/apache/flink/streaming/api/environment/StreamExecutionEnvironment.java
---
@@ -632,15 +608,18 @@ public void registerType(Class<?> type) {
* @return the data stream constructed
*/
@SuppressWarnings("unchecked")
- private <OUT> DataStreamSource<OUT> addSource(SourceFunction<OUT>
function,
- TypeInformation<OUT> outTypeInfo, String sourceName) {
+ private <OUT> DataStreamSource<OUT> addSource(SourceFunction<OUT>
function, String sourceName) {
+
+ TypeInformation<OUT> outTypeInfo;
- if (outTypeInfo == null) {
- if (function instanceof GenericSourceFunction) {
- outTypeInfo = ((GenericSourceFunction<OUT>)
function).getType();
- } else {
+ if (function instanceof GenericSourceFunction) {
+ outTypeInfo = ((GenericSourceFunction<OUT>)
function).getType();
+ } else {
+ try {
outTypeInfo =
TypeExtractor.createTypeInfo(SourceFunction.class,
function.getClass(), 0, null,
null);
+ } catch (InvalidTypesException e) {
+ outTypeInfo = (TypeInformation<OUT>) new
MissingTypeInfo("Custom source", e);
--- End diff --
Why did you change this?
I suspect the exception will now be thrown somewhere else?
> Refactor streaming scala api to use returns for adding typeinfo
> ---------------------------------------------------------------
>
> Key: FLINK-1909
> URL: https://issues.apache.org/jira/browse/FLINK-1909
> Project: Flink
> Issue Type: Improvement
> Components: Scala API, Streaming
> Reporter: Gyula Fora
> Assignee: Gyula Fora
>
> Currently the streaming scala api uses transform to pass the extracted type
> information instead of .returns. This leads to a lot of code duplication.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
