icexelloss commented on PR #36499:
URL: https://github.com/apache/arrow/pull/36499#issuecomment-1629449634

   @rtpsw #36580 is not exactly what I was asking for. There are still two 
thread reading/writing the batch_index variable which can lead to race 
conditions / more complex threading model.
   
   For the sake of moving forward, I am OK with what is implemented in this PR 
- however, I would requesting adding detailed explanation of why the 
invalidation needs to happen on the input received thread (basically because 
KeyHasher class cannot detect that the top of queue batch has changed if the 
address of the next  batch happens to be the same as the top of queue one).
   
   I would still like to simply the threading model there so that only one 
thread needs to interact with the KeyHasher class. But that can be left as 
future work. 


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to