[
https://issues.apache.org/jira/browse/FLINK-13762?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
zhijiang updated FLINK-13762:
-----------------------------
Description:
Currently StreamOneInputProcessor and StreamTwoInputSelectableProcessor have
separate implementations of ForwardingValueOutputHandler. Especially for the
implementation in StreamTwoInputSelectableProcessor, it couples the internal
input index logic which would be a blocker for the following unification of
StreamTaskInput/Output.
We could realize a unified ForwardingValueOutputHandler for both
StreamOneInput/ TwoInputSelectableProcessor, and it does not consider different
inputs to always consume StreamStatus. Then we refactor the implementation of
StreamStatusMaintainer for judging the status of different inputs internally
before really emitting the StreamStatus.
was:
Currently StreamOneInputProcessor and StreamTwoInputSelectableProcessor have
separate implementations of ForwardingValveOutputHandler. Especially for the
implementation in StreamTwoInputSelectableProcessor, it couples the internal
input index logic which would be a blocker for the following unification of
StreamTaskInput/Output.
We could realize a unified ForwardingValveOutputHandler for both
StreamOneInput/ TwoInputSelectableProcessor, and it does not consider different
inputs to always consume StreamStatus. Then we refactor the implementation of
StreamStatusMaintainer for judging the status of different inputs internally
before really emitting the StreamStatus.
> Integrate the implementation of ForwardingValveOutputHandler for
> StreamOne/TwoInputProcessor
> --------------------------------------------------------------------------------------------
>
> Key: FLINK-13762
> URL: https://issues.apache.org/jira/browse/FLINK-13762
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Task
> Reporter: zhijiang
> Assignee: zhijiang
> Priority: Minor
>
> Currently StreamOneInputProcessor and StreamTwoInputSelectableProcessor have
> separate implementations of ForwardingValueOutputHandler. Especially for the
> implementation in StreamTwoInputSelectableProcessor, it couples the internal
> input index logic which would be a blocker for the following unification of
> StreamTaskInput/Output.
> We could realize a unified ForwardingValueOutputHandler for both
> StreamOneInput/ TwoInputSelectableProcessor, and it does not consider
> different inputs to always consume StreamStatus. Then we refactor the
> implementation of StreamStatusMaintainer for judging the status of different
> inputs internally before really emitting the StreamStatus.
--
This message was sent by Atlassian JIRA
(v7.6.14#76016)