Two points:
1. I think it's a good thing to phrase this in terms of the BNF from
3986 and 3987. I don't think it's obvious that this piece of the spec
needs to reuse the HTML URI parser.
2. Equivalence of ASCII domain names is defined in terms of an ASCII
case insensitive comparison. Equivalence of IDN labels is defined in
terms of case insensitive comparison of their ASCII representations.
Given various troubles and instabilities in IDN space, the only IDNs
that can be put into files safely are U-labels, i.e., labels that are
mapped onto themselves when ToUnicode is applied to the result of
ToASCII.
The relevant references here are:
http://tools.ietf.org/html/rfc3490
http://tools.ietf.org/html/draft-klensin-idnabis-issues-07
You'll also want to say something about URI encoding in the path part
of the IRI reference.
Regards,
--
Thomas Roessler, W3C <[email protected]>
