On Thu, Nov 17, 2011 at 12:13 AM, Carsten Bormann <c...@tzi.org> wrote:
> On Nov 17, 2011, at 13:00, Brian E Carpenter wrote: > > > Do people agree that this is a reasonable thing to do? > > Yes please! > > My ABNF is pretty rusty, so I offer the following extension to the RFC 3986 IPv6address rule only as a starting point: / "FE80::" [ *3( h16 ":" ) h16 ] [ "%" 1*4(ALPHA / DIGIT) ] If we proceed with this, should we fix the IPv6address rule to align with RFC 5952 or just make the minimal change? > I trust that the update spec will explain the "for debugging" focus and > will have appropriate warning > -- that the link identifiers are node-local names with node-local > semantics, and > -- about security considerations, e.g., that this capability might be used > to send packets to random links from malicious web pages unless properly > restricted in a browser. > > You mean "specified links", right? I still don't understand the security risk this poses, so maybe we could discuss it offline. Thanks, -K- > Grüße, Carsten > >
-------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
