zentol commented on pull request #15741:
URL: https://github.com/apache/flink/pull/15741#issuecomment-831198755


   > For example, I think I could make the test pass if schedule completely 
ignores the delay and runs the result directly. Hence, there is no longer a 
guard preventing such a regression.
   
   How so? It would call the wrong method on the `MainThreadExecutable`, 
failing the test. Now if you break the actual scheduling behavior in the 
`AkkaInvocationHandler` then this can indeed happen, but that reveals further 
issues in these tests (i.e., that they are reliant to a specific `RpcService` 
implementation).
   
   > I fear that we are actually losing test coverage with these changes
   
   We do, of the `AkkaInvocationHandler`. I can try adding a test for the 
actual scheduling behavior, but I don't think we should revert the proposed 
changes.


-- 
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]


Reply via email to