[Patch] Webdav properties in UDSAtoms

Hamish Rodda meddie at yoyo.its.monash.edu.au
Fri May 31 12:03:34 BST 2002


On Thu, 30 May 2002 20:43, Best, Jan-Pascal van wrote:
> > If you ask me this isn't too much to ask of a computer these
> > days, especially
> > where it is only going to be doing so when special features
> > are required anyway.
>
> Not too sure about that: when syncing a calendar to an exchange server,
> we might have to do so for every appointment in the store.

I see... any ideas how much that typically is?

> Which brings
> me to something else: the current webdav stat method seems to send a
> PROPFIND to the server, with no further arguments. Shouldn't it contain
> a message body like
>   <?xml version="1.0" encoding="utf-8" ?>
>      <D:propfind xmlns:D="DAV:">
>        <D:allprop/>
>      </D:propfind>
> as per example 8.1.2 from RFC2518? This would also open the door
> for YAM (yet another metadata ;) to instruct HTTPProtocol to
> query for just a selected number of arguments, instead of all
> arguments. The default might even be to query for just the arguments
> that are used in davParsePropstats. What do you think?

I think you are onto something... I have to do a similar thing to fix a bug 
with servers who don't like our not sending a content-length: 0 anyway. I'll 
switch it to only asking for the required data.

In fact, here's an idea: the davRequestResponse should contain the request 
XML, or if blank, we will send <allprop/>.

The only thing to consider now is: when we are not requesting <allprop/>, 
should the additional propfind request be sent as a separate request? Or 
maybe as two <prop> blocks in the same <propfind>, if that works?

Cheers,

Hamish.




More information about the kfm-devel mailing list