A server name indication extension will be sent whenever the name is
> recognizable by the matches. I did not check for the types of
cipher suites. I think it is the proper approach because although for
anonymous cipher cuties, there is not certificates, but the ssl context
may be different, so it is still can be regarded as the server do something
> different related to the specified SNI.
True!
According to effective java,
Item 15? It doesn't specifically talk about final methods, but I think
I see where you are coming from since this class was originally non-final.
> I would prefer to use final keyword for
> the new methods. I did not see clear requirements that customers
> need to override the methods. So I would like a stricter restriction
> for the new methods in case of any mis-use. Does it make sense to you?
Yes. I can't really think of why not, other than for potential
extendability.
Brad