On Wed, 15 Jan 2025 13:21:30 GMT, Jaikiran Pai <j...@openjdk.org> wrote:

> Can I please get a review of this test-only change which updates the 
> `test/jdk/java/net/DatagramSocket/InterruptibleDatagramSocket.java` to match 
> the specified behaviour of `DatagramSocket.receive()` method?
> 
> This test was introduced in Java 14 
> (https://bugs.openjdk.org/browse/JDK-8233018) and that version of Java 
> through Java 18, behave differently when a DatagramSocket associated with a 
> DatagramChannel, when blocked in a receive() is interrupted. In those 
> versions, a `SocketException` gets thrown with `ClosedByInterruptException` 
> as the cause of that `SocketException`. 
> 
> Starting Java 19, the specification of DatagramSocket.receive() has been 
> updated to clarify the expectations of this scenario. Given those updates, 
> the DatagramSocket.receive() is expected to throw a (top level) 
> `ClosedByInterruptException` when associated with a `DatagramChannel`.
> 
> The change in this PR updates the test code to no longer expect a 
> `SocketException` to contain the `ClosedByInterruptException`. With this 
> change the test continues to pass both with platform threads as well as 
> virtual threads. The test will no longer be skipped when the main() is 
> launched through a virtual thread.

test/jdk/java/net/DatagramSocket/InterruptibleDatagramSocket.java line 87:

> 85:                 throw e;
> 86:             }
> 87:             System.out.println("Got expected: " + e);

Should the check on the exception be a little more focused than blindly 
accepting any SocketException when the socket is interruptible?

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/23135#discussion_r1916790398

Reply via email to