RFR: 8344629: SSLSocketNoServerHelloClientShutdown test timeout [v4]

[Artur Barashev]

On Tue, 3 Dec 2024 22:44:44 GMT, Bradford Wetmore <wetmore at openjdk.org> wrote:

> I was moving too fast last week and didn't catch that you had removed **both** `Socket.SoTimeOut`s. I thought you were still trying to minimize the client/server interaction time in one place. The default `Socket` timeout is likely just fine.
> 
> So yes, let's pull the `CountDownLatch`, as we no longer need the lockstep control, and I think it's good to go.
> 
> One minor suggestion, you might also add to the class comment (lines 53-55) that you're using the `SSLEngine` to prevent the client from receiving/responding to the server's ServerHello, and thus force the generation of the plaintext shutdown.

Client read timeout is irrelevant because we are not reading to this socket, only writing from it. So removing it is basically a no-op. I've updated the comment.

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

PR Comment: https://git.openjdk.org/jdk/pull/22263#issuecomment-2518143180