On Fri, Jun 28, 2013 at 10:33:12PM +0000, Templin, Fred L wrote: > Hi Fernando, > > > Considering the above, I guess I'm in the camp of "avoid fragmentation > > where possible". However, I don't think I'd go as far as deprecating > > it. > > I was originally thinking this way too, but now I am thinking > that deprecation of an old method is often the first step in > the path to adoption of a new method that avoids the old issues. > There have been examples in the past (e.g., deprecation of site > locals leading to ULAs), and I don't see why it would be any > different in this case.
While browsing the draft I wondered if applications do handle the EPROTO errno case for writes on a socket with a recevied icmp error parameter-problem correctly today. I doubt that recommending not implementing reassembly would be beneficial. Greetings, Hannes -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------