Nice clean up Chris. Looks fine. Regards, Sean.
On 09/06/16 20:49, Chris Hegarty wrote:
This test has been seen to fail intermittently. There is an assumption in the test that the client-side will fill the outgoing TCP buffer by writing 1 megabyte of data, allowing for the server-side to close the connection out from under it. This is the crux of the test. By slowing down the writes on the client-side we can eliminate this assumption. The remainder of the changes just modernise the code, and remove many unnecessary synchronization points. http://cr.openjdk.java.net/~chegar/8041924/ https://bugs.openjdk.java.net/browse/JDK-8041924 -Chris.