The class javax.net.ssl.SSLSocketFactory defines a method createSocket(Socket s, String host, int port, boolean autoClose), which returns an instance of SSLSocket.
I'm trying to understand this method, so my question is: why does the method take a host parameter? Shouldn't the host be read from the plain socket's underlying getInetAddress().getHostName(), or am I missing something?
Following the implementation in SSLSocketFactoryImpl, it would return new SSLSocketImpl, as in here you can see the following snippet of code
if (peerHost == null || peerHost.isEmpty()) {
boolean useNameService =
trustNameService && conContext.sslConfig.isClientMode;
useImplicitHost(useNameService);
} else {
conContext.sslConfig.serverNames =
Utilities.addToSNIServerNameList(
conContext.sslConfig.serverNames, peerHost);
}
(Note that peerHost is same as the host parameter)
So i think the point here is that if you don't provide the hostname, it would take a little more effort to lookup to do some stuff to get the hostname - You can follow useImplicitHost(useNameService) to see the logic in there in case the host is missing