Don't copy live properties

At 08:07 AM 12/23/98 PST, Slein, Judith A wrote:
>.. the spec's directives about how to treat live properties on
>COPY are broken ...  The spec says that servers SHOULD copy live properties
>as live, identically behaving properties (that is, they wouldn't necessarily
>have the same value at the destination); if they cannot keep the properties
>live at the destination, they MUST copy them as dead properties or else fail
>the request.

>For some properties, where the value could be seriously misleading, it is
>better not to copy them at all than to copy them as dead properties.

I agree, this is broken.  As you say, copying lockdiscovery, for example,
is a bad idea.  

I propose we drop copying live properties altogether.  If a property is
live at the destination, the server will already implement it (it will be
"copied" as an "identically behaving" property for free.)  if it's not
live, it's better to not copy it at all.

Can we drop this from the spec (8.8.2)?

It's also bogus to specify copying a live property "octet for octet", since
there is no canonical serialized form of a WebDAV property.  So it's
meaningless to talk about an octet for octet  copy.

Proposed re-wording:

8.8.2     COPY for Properties

   The following section defines how properties on a resource are
   handled during a COPY operation.

   Live properties SHOULD be duplicated as identically behaving live
   properties at the destination resource.  If a property cannot be
   copied live, then it MUST NOT be copied.

   The propertybehavior XML element can specify that properties are
   copied on best effort, that all live properties must be successfully
   copied or the method must fail, or that a specified list of live
   properties must be successfully copied or the method must fail. The
   propertybehavior XML element is defined in section 12.12.


Objections?

Received on Thursday, 31 December 1998 15:44:55 UTC