- From: Greg Stein <gstein@lyra.org>
- Date: Sat, 2 Jun 2001 00:03:13 -0700
- To: ietf-dav-versioning@w3.org
On Fri, Jun 01, 2001 at 02:00:04PM +0200, Stefan Eissing wrote: > > > Von: ietf-dav-versioning-request@w3.org > > [mailto:ietf-dav-versioning-request@w3.org]Im Auftrag von Clemm, Geoff > > > > [...] > > Julian: If you have a chance to do some interoperability testing > > with the "unknown element" nested inside of <allprop/>, > > that would be great, since that would give us a basis for > > fixing 23.3.1 (i.e. without introducing interoperability > > issues for existing servers). > > Ok, our tests showed that > <allprop/><include>...</include> > works with moddav, IIS and our servlet, while > <allprop><include>...</include></allprop> > works with moddav and IIS. Julian's PROPFIND > handler is throwing the book at me with 422s. To beat the horse: mod_dav uses the following algorithm in its propfind: 1) validate that the root element (if provided) is DAV:propfind 2) if no body was provided, or a DAV:allprop child is present in the root element, then the PROPFIND is an "allprop" 3) look for a DAV:propname as a child of the root element; if found, then the PROPFIND is a "propname" style 4) look for DAV:prop; if founc, then it is a "prop" style 5) else fail 6) For a prop style fetch, only the immediate child elements (whether a start element, or an empty element) of the DAV:prop element are used. Children of those elements are not examined. -- whitespace is completely ignored in the entire PROPFIND processing So... mod_dav is going to be extremely flexible with any kind of screwing around you might want to perform in the PROPFIND body :-) Cheers, -g -- Greg Stein, http://www.lyra.org/
Received on Saturday, 2 June 2001 02:57:46 UTC