referencing IDNA2008 (and IDNA2003?)

John C Klensin klensin at
Fri Oct 22 21:42:37 CEST 2010

--On Friday, October 22, 2010 12:26 -0700 Adam Barth
<ietf at> wrote:

> It doesn't matter if X (described below) is a valid label.  If
> it's not an octet-by-octet match for something that we know is
> well-formed, we just ignore the whole cookie.

And that is why there is an explicit loophole in IDNA2008 that
permits working with A-labels (or things that look like them)
without the validation that Vint identifies.  This isn't an
exact statement of the rule but the validation is basically
required only if one expects to display the string or is doing
some other processing that actually requires validity, not just,
e.g., octet-by-octet string comparison.


More information about the Idna-update mailing list