RE: XPointer in href element

Hum... http://foo/bar#child(1,H2) and http://foo/bar#child(4,H2) look like
different HREFs to me.

			Yaron

> -----Original Message-----
> From: Jim Davis [mailto:jdavis@coursenet.com]
> Sent: Mon, May 17, 1999 5:17 PM
> To: www-webdav-dasl@w3.org
> Cc: Jim Whitehead
> Subject: XPointer in href element
> 
> 
> I'm trying to sort out how one might use an XPointer in an 
> href within a
> PROPFIND response element.
> 
> One thing concerns me.
> 
> Suppose I have two XPointers, both to the same resource, but 
> to different
> locations within it.  This might happen, for example, in a  
> DASL search
> over a set of  WebDAV resources that included XML (or XHTML) 
> resources.  It
> could certainly happen that there was more than one location 
> in a given
> resource matching the query, and it's quite likely that 
> application writers
> would want the full XPointer of the matching location, not 
> just the href of
> the resource.
> 
> But RFC 2518, section 12.9.1 says (of the response XML element)
> 
> A particular href MUST NOT appear more than once as the child 
> of a response
> XML element under a multistatus XML element.
> 
> So the question is, are two XPointers that point to the same 
> base resource
> the *same* href, for the purposes of section 12.9.1?  More 
> generally, are
> two URLs that differ only in the fragment considered the same?
> 
> If so, then how to support the kind of searching I described above?
> 
> Seems like there are two possibilities
> 
> 1) Remove the restriction in 12.9.1
> 
> 2) Require such searches to  return the base URL in the href, 
> just once,
> and the set of matching Xpointer fragments in element within it, e.g.
> something like
> 
> <D:response>
>   <D:href>http://stars.com/movies.xml</D:href>
>   <D:propstat>
>     <D:prop>
>       <D:matchinglocations>
>          <D:xpointer>CHILD(1,H2)</D:xpointer>
>          <D:xpointer>CHILD(4,H2)</D:xpointer>
>        </D:matchinglocations>
>      <X:title>Life of Brian</X:title>
>     </D:prop>
>   </D:propstat>
>  </D:response>
> 
> 
> I'd rather not try to change RFC 2518, so does anyone see a 
> problem doing
> it this way?
>

Received on Monday, 17 May 1999 20:27:54 UTC