Anne van Kesteren wrote: > I couldn't find definitions of "subdomain" and "label" elsewhere. http://tools.ietf.org/html/rfc3696 (+ errata) is a nice intro, what you want are "LDH" (letter-digit-hyphen) labels permitted in "host" names, not labels containing arbitrary octets. Frank
On Fri, 28 Sep 2007 17:00:01 +0200, Frank Ellermann <nobody <at> xyzzy.claranet.de> wrote: > Anne van Kesteren wrote: > >> I couldn't find definitions of "subdomain" and "label" elsewhere. > > http://tools.ietf.org/html/rfc3696 (+ errata) is a nice intro, what > you want are "LDH" (letter-digit-hyphen) labels permitted in "host" > names, not labels containing arbitrary octets. I'm not sure that RFC helps as it's informational only. Would using ireg-name be appropriate? (We do not want to allow IP addresses for now.) -- -- Anne van Kesteren <http://annevankesteren.nl/> <http://www.opera.com/>
Anne van Kesteren wrote: > Would using ireg-name be appropriate? If you'd want IRIs excluding domain-literals. But your draft is about HTTP, isn't it ? For HTTP you need URIs, or RFC 3986 reg-name, not RFC 3987 ireg-name. An HTTP header uses Latin-1. You could say that clients are supposed to translate an ireg-name into a reg-name as explained in RFC 3987 3.1, that would be IDNA punycode, not simply percent-encoded UTF-8. Frank
At 18:58 07/09/29, Anne van Kesteren wrote: >On Fri, 28 Sep 2007 17:00:01 +0200, Frank Ellermann ><nobody <at> xyzzy.claranet.de> wrote: >> Anne van Kesteren wrote: >> >>> I couldn't find definitions of "subdomain" and "label" elsewhere. >> >> http://tools.ietf.org/html/rfc3696 (+ errata) is a nice intro, what >> you want are "LDH" (letter-digit-hyphen) labels permitted in "host" >> names, not labels containing arbitrary octets. > >I'm not sure that RFC helps as it's informational only. Would using >ireg-name be appropriate? (We do not want to allow IP addresses for now.) ireg-name doesn't cut up the name into labels, because it is intended (similar to reg-name in RFC 3986) to cover other registries than just the DNS, if and when these would be used in new schemes. I suggest that you look at the IDNA RFCs if you want to get a definition of a label including non-ASCII characters. IDNA definitely deals with individual labels, because that's where punicode and the xn-- prefix and all that stuff happen. Regards, Martin.s #-#-# Martin J. Du"rst, Assoc. Professor, Aoyama Gakuin University #-#-# http://www.sw.it.aoyama.ac.jp mailto:duerst <at> it.aoyama.ac.jp
On Sat, 29 Sep 2007 12:31:54 +0200, Martin Duerst <duerst <at> it.aoyama.ac.jp> wrote: > At 18:58 07/09/29, Anne van Kesteren wrote: > I suggest that you look at the IDNA RFCs if you want to get a definition > of a label including non-ASCII characters. IDNA definitely deals with > individual labels, because that's where punicode and the xn-- prefix > and all that stuff happen. Ah, it seems RFC 3490 defines "internationalized domain name" which can replace "subdomain" above. "internationalized label" can replace "label". Not sure why I didn't use those from the start. However, for HTTP it seems the ToASCII variant should be required (not just allowed) as HTTP doesn't do Unicode. Thanks. Also thanks to Frank Ellerman for pointing out the limitation of HTTP that has escaped me (and the rest of the people reviewing; I begin to wonder how many there are...). -- -- Anne van Kesteren <http://annevankesteren.nl/> <http://www.opera.com/>
RSS Feed