- From: Lisa Dusseault <lisa@xythos.com>
- Date: Fri, 25 May 2001 09:27:47 -0700
- To: "Julian F. Reschke" <julian.reschke@greenbytes.de>, <ietf-dav-versioning@w3.org>
> > > <D:propfind xmlns:D="DAV:"> > > > <D:all-dead-prop/> > > > <D:checked-in/> > > > <D:checked-out/> > > > <D:version-name/> > > > </D:propfind> > > > > > > <D:propfind xmlns:D="DAV:"> > > <D:all-dead-prop/> > > <D:prop> > > <D:checked-in/> > > <D:checked-out/> > > <D:version-name/> > > </D:prop> > > </D:propfind> > > Could you explain? > > Old servers will ignore the "include" element -- a new client > will be aware > that is was ignored because the additionally selected properties will not > turn up anywhere in the multistatus response. An old client will never use > the "include" element, therefore there'll be no interoperatibility issues. I don't understand what you're saying about the "include" element. Nevertheless, I do have strong logic for putting the property names under some container, and that's to distinguish property names from non-property names. "all-dead-prop" is not a property name. "checked-in" is. If they are glommed together in a list, then that's treating "all-dead-prop" as if it was a property name. That prevents the server from ever having a property called "all-dead-prop", since the XML element with that name was used the way you're proposing. lisa
Received on Friday, 25 May 2001 12:29:23 UTC