I would disagree with using a String. There are too much room for error IMO. I do agree that resolving the address is not the best approach either. Would it make any sense to write an "IP Address" class that at least validates the IP?
On Tue, Dec 23, 2008 at 9:05 AM, Ashish <paliwalash...@gmail.com> wrote: >>> The reason why I am a little confused is, InetAddress.getByAddress API >>> actually resolves the address which can hit the performance. ApacheDS >>> uses this approach. >>> >> You can have a look at the current implementation of this codec in the >> Directory project. It's written on top of MINA. >> >> http://svn.apache.org/repos/asf/directory/apacheds/trunk/protocol-dhcp/ >> > I have already looked at it and that's why debate started. ApacheDS > uses InetAddress approach > and I just thought about using a simple String. > On a similar not xid is being as int in ApacheDS DhcpMessageDecoder, > which is an unsigned int > as per RFC (4 bytes). Since ApacheDS is well tested, I am sticking to > the ApacheDS > implementation as the guideline. > > thanks >
