W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > April to June 1998

RE: Circular References

From: Jim Whitehead <ejw@ics.uci.edu>
Date: Fri, 8 May 1998 13:38:43 -0700
To: w3c-dist-auth@w3.org
Message-ID: <002e01bd7ac1$4ac8eac0$d115c380@galileo.ics.uci.edu>
The risk from circular references for the server comes when operations on
the server "follow" the referential member to its destination, and end up in
a loop.  However, my understanding of the sense of the room at the LA IETF
and from previous discussions on this topic is that people are opposed to
having the server follow referential members.  So, for example, a PROPFIND
of Depth infinity used to list the members of a collection hierarchy would
only list the existence of the referential members, but would not follow
these referential members into any collections to which they might
potentially refer.

Based on this, I see no risks to the server from circular references, since
the server will not encounter these loops in its processing.

A will client which is proactively requesting the destination of a
referential member of a collection could run into situations where it
encounters circular reference loops.  As Judy points out, these loops are
inevitable since external members can point off to servers outside the
control of the originating server. Since these circular links can occur, and
are infeasible to prevent, I recommend we permit circular loops, and add
language to the protocol draft warning client implementors of the potential
existence of these loops.

- Jim

> -----Original Message-----
> From: Judith Slein [mailto:slein@wrc.xerox.com]
> Sent: Friday, April 24, 1998 1:31 PM
> To: w3c-dist-auth@w3.org
> Subject: Circular References
> Based on the minutes of the April 2 WebDAV meeting, it sounds as if there
> needs to be more discussion of circular references in the context of the
> advanced collections requirements.
> Once we support referential members of collections, there is potential for
> circular references unless we do something to prevent it.  Earlier DAV
> drafts explicitly stated that servers are not required to detect circular
> references.  It is probably unrealistic to expect a server to prevent
> circular references, since references can cross servers.
> What do people think?
> Name:		Judith A. Slein
> E-Mail:		slein@wrc.xerox.com
> Phone:  	(716) 422-5169
> Fax:		(716) 422-2938
> Xerox Corporation
> Mail Stop 105-50C
> 800 Phillips Road
> Webster, NY 14580
Received on Friday, 8 May 1998 16:43:25 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:01:13 UTC