On Tue, 2005-02-01 at 16:36 +0200, Markku Savela wrote: > > From: Jeroen Massar <[EMAIL PROTECTED]> > > > Thus that applications should be using multiple sockets, one > > for IPv4 and one for IPv6 and one for whatever follows. > > I strongly object to this. There are other socket api's which don't > have the Unix inherited drawbacks. For such, the recommendation is > exactly the opposite: the same socket works just fine for IPv4 and > IPv6, and in unifying the application code to work for both, IPv4 > mapped address format is very useful tool.
Any documentation on this ? I do know RFC3493, not any others. > There is no reason for application to care at all whether actual > connection is over IPv4 or IPv6. There is, these are different protocols, how else are you going to address which host you are going to connect to? For instance www.kame.net port 80 results in different answers over IPv6 than the one on IPv4... If you are talking about wrappers, then these are most likely wrapping around the above already anyway, thus it matter than? Greets, Jeroen
signature.asc
Description: This is a digitally signed message part
-------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
