EarthChen commented on PR #16588:
URL: https://github.com/apache/dubbo/pull/16588#issuecomment-6940685405

   Closing in favor of #16585, which covers both loss paths reported in #16584.
   
   Evidence from reproducing on the `3.3` branch (base `1cf125c421`, JDK 11):
   
   1. This change only handles the saturated-pool rejection path. The issue 
also reports DATA silently discarded by `SerializingExecutor` when the backing 
executor is shut down (`runQueue.remove(removable)` without running the task) — 
that path remains leaky with this patch. Running 
`GenericHttp2ServerTransportListenerTest` from #16585 on top of this patch 
still fails 7 of 18 tests (`releasesSilentlyDiscardedDataOnClose`, 
`releasesSilentlyDiscardedDataOnRemoteCancel`, 
`queuedDataIsReleasedOnCloseBeforeExecution`, 
`closeCallbackCanWaitForConcurrentLateData`, ...).
   2. The diff bundles large unrelated reworks 
(`SerializeSecurityConfigurator`, `HttpWriteQueue`, 
`LengthFieldStreamingDecoder`) that make review harder; those should be 
separate PRs.
   3. There is no regression test for the leak.
   
   #16585 passes all 18 ownership tests plus the full `dubbo-rpc-triple` module 
with green CI. Thanks for the report — the rejection-path observation here is 
correct and is subsumed by #16585; feel free to review that PR instead.


-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to