RFR[8241072]: 'Reimplement Legacy DatagramSocket API'

[Chris Hegarty]

> On 30 Mar 2020, at 19:27, Patrick Concannon <patrick.concannon at oracle.com> wrote:
> ...
> 
> http://cr.openjdk.java.net/~pconcannon/8241072/webrevs/webrev.01/

This looks very good, a testament to all the prior cleanup work that has already been done.

Just a few comments:

- DatagramSocket::getChannel now stands out a little, since it is one of the few instance methods has still has a non-delegating body. It could optionally delegate, or if not, then NetMulticastSocket doesn’t necessarily need to override it.

- In DatagramSocket::createDelegate, "enable broadcast if possible” - Possibly due to refactoring, but I cannot reconcile this with the old implementation.

  - The createDelegate's @return comment is a bit hard to parse; suggest: "@return {@code null} if {@code bindaddr == NO_DELEGATE}, OTHERWISE returns a delegate for the requested {@code type}"

- The set of socket options is now per datagram/multicast socket instance, rather than on the class of the socket. I don’t think that this is an issue, just an observation and a confirmation that it is deliberate.

-Chris.