- From: Jeremy Carroll <jjc@hpl.hp.com>
- Date: Fri, 21 Dec 2007 15:19:12 +0000
- To: Andrea Perego <andrea.perego@uninsubria.it>
- CC: public-powderwg@w3.org
Andrea Perego wrote: >>> mailto:jeremy.carroll@hp.com (([^:/?#]+):)?(//([^:/?#@]*)(:([0-9]+))?)?([^?#]*)(\?([^#]*))? (([^:/?#]+):) matches mailto: (//([^:/?#@]*)(:([0-9]+))?)? does not match any sequence stating j so (//([^:/?#@]*)(:([0-9]+))?)? matches the empty string ([^?#]*) then matches the remainder and (\?([^#]*))? matches the empty string. ahh, you've left off the fragment in the regex. Also see section 3 of RFC 3986, the example at the bottom of page 15 shows a path example:animal:ferret:nose in the URI urn:example:animal:ferret:nose === You may choose to be totally clear whether resources identified by URIs with fragments (so-called secondary resources) are or are not potential members of the RS. On my, admittedly rather hurried reading, I was assuming that a resource identified by http://example.org/foo#frag would be included in <ResourceSet> <includeHosts>example.org</includeHosts> </ResourceSet> but I can see that one might argue against (and I can believe I missed the part where it is clearly stated that it is not) Jeremy
Received on Friday, 21 December 2007 15:20:03 UTC