fully-qualified domain names containing differently-encoded labels ?

Nicolas Williams Nicolas.Williams at oracle.com
Tue Oct 26 00:21:07 CEST 2010

On Mon, Oct 25, 2010 at 02:51:38PM -0700, =JeffH wrote:
> > On Mon, Oct 25, 2010 at 01:23:45PM -0700, =JeffH wrote:
> >> But in any case, it would seem that any specification of conversion
> >> operations on a multi-label domain name, that takes into account
> >> IDNA{2003,2008}, needs to properly allow for and accomodate
> >> label-by-label type diversity, yes?
> >
> > Given that neither IDNA2008 nor IDNA2003 had anything to say about
> > domain names, but only about label names, I should hope so.  You
> > should treat them only by label, and not by domain name.
> Understood. However, in various protocols, e.g. HTTP, one has to
> handle (e.g. compare) strings matching RFC3986's "reg-name"
> production, and thus must decompose said strings into labels, do the
> IDNA magic as necessary, and then re-construct as appropriate and
> compare (for example).

I don't think label "type" mixtures are a problem.  The presence of
U-labels in IDN-unaware domainname slots is a problem.  And FQDNs that
have A- and U-labels, in IDN-aware domainname slots might be indicative
of a bug.  But a priori there's nothing wrong with having a domainnames
with both, A- and U-labels (and NR-LDH labels too).


More information about the Idna-update mailing list