RFR 8224477 [13] java.net.Socket::setOption - implementation mismatch to spec

Chris Hegarty chris.hegarty at oracle.com
Mon May 27 15:48:11 UTC 2019


Alan,

On 27/05/2019 12:50, Alan Bateman wrote:
> ..
> This looks a good cleanup. There may be an argument to have the default 
> implementations of get/setOption throw NPE when name is null.

Good point. Done.

> Does isServer need to be package-private? I didn't spot any usages in 
> PSI/PDSI.

Not yet. It is still used by the unix PlatformSocketImpl, when
creating new instances. A future cleanup opportunity, possibly.

> For SocketChannelImpl/ServerSocketChannelImp then I'd prefer to use the 
> same check that we have in the new NioSocketImpl to test that the value 
> type matches the SocketOption::type. We can do the same check in 
> DatagramChannelImpl.setOption.

Yes. Done.

> The tests looks good. A suggestion for the TestImplSpec methods is to 
> rename them to something like TestDefaultBehavior as I think that is 
> better describes what they test.

Done.

Updated webrev:
   https://cr.openjdk.java.net/~chegar/8224477/webrev.04/

-Chris


More information about the net-dev mailing list