On Wed, 8 Feb 2023 12:19:05 GMT, Mark Sheppard <[email protected]> wrote:
> The IdleTimeoutTask is synchronizing on the idleConnections object, while you > are using a re-entrant lock for your change, so each is using a separate > independent lock for its synchronization --- shouldn't both processing blocks > be co-ordinates via the same lock mechanism ? It's a good observation and will taken care of once Darragh implements Jaikiran's suggestion here: https://github.com/openjdk/jdk/pull/12413#discussion_r1100020301 ------------- PR: https://git.openjdk.org/jdk/pull/12413
