zhijiangW commented on issue #10029:  [FLINK-14553][runtime] Respect 
non-blocking output in StreamTask#processInput
URL: https://github.com/apache/flink/pull/10029#issuecomment-547733858
 
 
   Thanks for the review and good suggestions @pnowojski !
   
   Concerning of refactor of `getAvailableFuture()` and `isAvailable()`, I 
explained my thought in inline comments.
   
   I remembered there were some discussions for this issue before. But now I 
find it is not very convenient to judge the available state via checking 
whether the future is completed or not in some cases. Because the returned 
`CompleteFuture` is mainly used for callback action, but if we only want to 
know the current state and do not need the requirement of monitoring the state 
for callback in some cases, then the previous way is heave-weight to some 
extent.
    
   I agree it would bring some overheads for outside implementations like 
`SourceReader`. But if we can explain the semantics of two methods clearly, 
then they can be orthogonal.
   
   > I think this PR can not be merged until we fix/migrate the back-pressure 
monitor to use the availability future?
   
   Exactly, it seems not much conflicts to prelaunch this PR, so I focus on it 
firstly and the back-pressure monitor would be done by yingjie.

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
[email protected]


With regards,
Apache Git Services

Reply via email to