W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > July to September 2001

RE: A query

From: Clemm, Geoff <gclemm@rational.com>
Date: Tue, 24 Jul 2001 12:14:39 -0400
Message-ID: <3906C56A7BD1F54593344C05BD1374B103B61C15@SUS-MA1IT01>
To: WebDAV <w3c-dist-auth@w3.org>

   From: sathyanarayanan_s [mailto:sathyanarayanan_s@infy.com]

   Consider a set up like below :

			   S1	W1	S2

   S1 and S2 are two normal servers, while W1 is a webserver with the DAV

   What should happen if I want to stream the file from S2 to S1 via W1 ?
   Would the following transaction be performed by W1 ? 

   COPY /~D1/F1 HTTP/1.1
   Host: www.S2.com
   Destination: http://www.S1/D2/F1

No, the transaction would be performed by www.S2.com (since that is
the host you sent it to).  W1 is never mentioned in this request,
and therefore is not involved in processing it.

   If no, then what should be transaction to be sent across to W1?

Some have asked for a Source: header for COPY to allow you to copy
to a WebDAV server (the Destination: header only lets you COPY
from a WebDAV server).  If both a Source and Destination header were
supported at the same time by a server, you could get the functionality
you are asking for.

But note that a non-WebDAV server does not have collections, so even
if you tried to copy something that looked like a collection (e.g.
copying /x/y, when there is a resource /x/y/foo.html), unless you
are copying from a WebDAV server, you would just copy the content of
whatever GET would return from /x/y, and not a collection (or any
properties, since only WebDAV resources have properties).

I would probably vote against such a Source header for COPY, because
of these edge cases around copying a collection, but I could be swayed (:-).

Received on Tuesday, 24 July 2001 12:08:13 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:01:23 UTC