RE: Proposal for marshalling property type information

"Dan Brotsky" <dbrotsky@Adobe.COM> wrote:
> I also like Julian's proposal and would be glad to see it
> incorporated into 2518.  But there are a few questions related to
> live properties that I'm hoping Julian and others would comment on:
>
> 1. I work on a number of servers that have specialized live
> ("computed" in the deltaV sense) properties for workflow tracking.
> It seems that we could use the extended PROPFIND to indicate to
> clients the datatype of those properties, but Julian only shows an
> example where the client has indicated the datatype.  Were live
> properties expected to obey the same extension rule?  If so we might
> want to clarify this and add an example.

I suggest we say that servers MAY return type information for live
proeprties.

> 2.  Some of my servers implement "type-restricted" live properties
> which are essentially dead properties whose values are restricted to
> a certain datatype.  These servers will reject PROPPATCH requests
> that use the wrong datatype whether or not the client has declared a
> datatype in the PROPPATCH.  Julian's proposal shows an example of a
> 422 response when the PROPPATCH-declared datatype doesn't match the
> datatype of the value; I would also like to use such a response when
> the value's datatype doesn't match the PROPFIND-shown (and enforced)
> datatype.  How does this strike people?

Great.

Tim

Received on Tuesday, 26 June 2001 06:09:10 UTC