Re: Order of dead properties

In general, WebDAV does not take advantage of the ordering of XML 
elements when a single element occurs multiple times inside the parent 
element.  As well as allowing any order for the properties of a 
resource (live or dead), the server may put resources in any order in a 
Depth 1 or Depth Infinity PROPFIND response.  Servers aren't required 
to take the order of property changes in a PROPPATCH request into 
effect, either.  In some cases the spec explicitly says that order 
isn't important (e.g. section 8.1, order of resources in PROPFIND 
response) but in other cases it's silent.

Note that DTDs do make XML element order clear but only for elements 
with different names.  When elements might occur multiple times (as 
with 'response' in 'multistatus') or when the parent element content is 
ANY, the DTD doesn't say anything about ordering those elements.

   <!ELEMENT multistatus (response+, responsedescription?) >
    <!ELEMENT response (href, ((href*, status)|(propstat+)),
    responsedescription?) >
   <!ELEMENT propstat (prop, status, responsedescription?) >
   <!ELEMENT prop ANY >

We'd have to add extraneous text specifically to say what additional 
ordering is required, and in the absence of that text agents must 
accept any ordering.

Lisa

On Mar 8, 2005, at 4:00 AM, <Heiko.Weber@softwareag.com> wrote:

>
> Hello everyone,
>
> has anything been said about the order of dead properties so far? When 
> a
> PROPFIND with DAV:allprop is issued on a resource, should a server
> return the dead properties in the order in which they where defined, or
> should the order be arbitrary? Since the element names have to be 
> unique
> for one resource, I guess it wouldn't make much sense to interpret any
> logic into the order of the property elements, but since the order of
> elements in XML documents is defined, some applications might expect 
> the
> properties in the order in which they were inserted.
>
> Any ideas on that?
>
> thanks,
>     Heiko
>
> ===
> Software AG
> Tamino Development
>
>

Received on Tuesday, 8 March 2005 15:21:33 UTC