Restore exception compatibility in TcpListener.EndAccept*** #41745
Conversation
|
Tagging subscribers to this area: @dotnet/ncl |
| { | ||
| return TaskToApm.End<TResult>(asyncResult); | ||
| } | ||
| catch (SocketException ex) when (ex.SocketErrorCode == SocketError.OperationAborted) |
There was a problem hiding this comment.
There could be other cases that cause this exception.
I think we should check explicitly for !_active before changing to ODE. That's what other routines here do.
There was a problem hiding this comment.
I'm wondering what are those other cases?
Anyways, I changed the line to catch (SocketException) when (!_active). Want to be equivalent with the old code.
There was a problem hiding this comment.
I'm wondering what are those other cases?
Possibly there aren't any. But it seems better to explicitly check _active, rather than just assume this is the only way that exception can happen.
antonfirsov
added
the
tenet-compatibility
|
/azp run runtime-libraries outerloop |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
@antonfirsov I didn't suggest any pushing out of 5.0 -- I just used correct milestone for the PR - which is in master and hence goes into 6.0. Once we have it merged, we can start porting it to 5.0 RC2 branch. |
|
/azp run runtime |
|
Azure Pipelines successfully started running 1 pipeline(s). |
|
/backport to release/5.0-rc2 |
|
Started backporting to release/5.0-rc2: https://github.com/dotnet/runtime/actions/runs/243045182 |
ghost
locked as resolved and limited conversation to collaborators
#40476 changed
EndAcceptSocketandEndAcceptTcpClientto utilize the Task-basedSocket.AcceptAsyncimplementation instead of the APM code path.This however introduced a breaking change for the exception being thrown when
listener.Stop()has been called, closing the underlying socket between the calls toBeginAcceptandEndAccept.According to the old code and the documentation, we should throw
ObjectDisposedExceptionhere.Fixes #41585
/cc @geoffkizer