On Tue, Mar 6, 2012 at 2:43 PM, Carsten Bormann <c...@tzi.org> wrote: > Given Bill's draft, we may not have to argue this, but: > > On Mar 6, 2012, at 01:15, Brian E Carpenter wrote: > >> The ABNF does not describe the produced (encoded) >> URI. > > That is certainly *not* my understanding of RFC 3986. > > (It is true that a number of real-word implementations allow pct-encoded in > more places than the ABNF does. But being liberal in what you accept is a > rather usual way of non-compliance. Challenge: Find a real-world URI not > matched by the 3986 ABNF!) > Not sure what you mean by a "real-world URI", but http://[fe80::206:98ff:fe00:232%tap0] *used* to work in Firefox, and probably still does in 3.6.x. They ripped out the capability specifically because it wasn't supported by the ABNF in 3986.
-K- > Grüße, Carsten > > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > ipv6@ietf.org > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------