[
https://issues.apache.org/jira/browse/FLINK-13767?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
zhijiang updated FLINK-13767:
-----------------------------
Description:
StreamInputProcessor#processInput could return InputStatus instead of current
boolean value to keep consistent with PushingAsyncDataInput#emitNext.
For the implementation of StreamTwoInputProcessor#processInput, we could
maintain and judge the two input status together with the next selected input
index to determine the final precise status. To do so we could avoid invalid
processInput call except for the first call.
In addition, AvailabilityProvider#isFinished has the duplicated semantic with
InputStatus#END_OF_INPUT for PushingAsyncDataInput, and it is only meaningful
for PullingAsyncDataInput now. So we migrate the #isFinished method from
AvailabilityProvider to PullingAsyncDataInput.
was:
AvailabilityListener is both used in AsyncDataInput and StreamTaskInput. We
already introduced InputStatus for PushBasedAsyncDataInput#emitNext, and then
InputStatus#END_OF_INPUT has the same semantic with
AvailabilityListener#isFinished.
But for the case of AsyncDataInput which is mainly used by InputGate layer, the
isFinished() method is still needed at the moment. So we migrate this method
from AvailabilityListener to AsyncDataInput, and refactor the
StreamInputProcessor implementations by using InputStatus to judge the finished
state.
> Refactor StreamInputProcessor#processInput based on InputStatus
> ---------------------------------------------------------------
>
> Key: FLINK-13767
> URL: https://issues.apache.org/jira/browse/FLINK-13767
> Project: Flink
> Issue Type: Sub-task
> Components: Runtime / Network, Runtime / Task
> Reporter: zhijiang
> Assignee: zhijiang
> Priority: Minor
> Labels: pull-request-available
> Time Spent: 10m
> Remaining Estimate: 0h
>
> StreamInputProcessor#processInput could return InputStatus instead of current
> boolean value to keep consistent with PushingAsyncDataInput#emitNext.
> For the implementation of StreamTwoInputProcessor#processInput, we could
> maintain and judge the two input status together with the next selected input
> index to determine the final precise status. To do so we could avoid invalid
> processInput call except for the first call.
> In addition, AvailabilityProvider#isFinished has the duplicated semantic
> with InputStatus#END_OF_INPUT for PushingAsyncDataInput, and it is only
> meaningful for PullingAsyncDataInput now. So we migrate the #isFinished
> method from AvailabilityProvider to PullingAsyncDataInput.
--
This message was sent by Atlassian Jira
(v8.3.2#803003)
