RE: WG Consensus: Abandon Operation

["Ramsay, Ron" <Ron.Ramsay@ca.com>]

Yes, why can't you add to your bulleted list that

* update operations are not expected to be targets for the abandon operation
- if your server wishes to honour them, you should be aware that there is no
response defined in LDAP for an abandoned operation so the state of the DIT
will be indeterminate.

Ron.

-----Original Message-----
From: Kurt D. Zeilenga [mailto:Kurt@OpenLDAP.org]
Sent: Thursday, 1 August 2002 17:44
To: ietf-ldapbis@OpenLDAP.org
Subject: WG Consensus: Abandon Operation


The message summarizes the WG discussions of Abandon
operations.

There is no evidence that any implementation returns or
expects the return of a response for an abandoned operation.
The fact that no specific result code indicating abandonment
was defined in the LDAP technical specification is a clear
indication that no response was to be returned in the case
of abandonment.

While some are concerned that design of Abandon operation is
flawed (as it doesn't provide a clear indication of
abandonment), redesign of the Abandon operation is beyond
our scope of work.  This work has been undertaken on an
individual basis (see draft-zeilenga-ldap-cancel).

The chairs believe there is consensus that the
Abandon operation and message ID reuse (in regards to the
Abandon operation) sections should be clarified to clearly
indicate:
	* Servers are not required to honor the Abandon
	  request;
	* The Abandon operation has no response;
	* The abandoned operation has no response; and
	* Message ids of the Abandon operation and the
	  abandoned operation cannot be reused unless
	  a subsequent bind operation completes.

The Editor is directed to produce a revision containing
clarifications consistent with this consensus.  The WG
should review the clarifications once offered to ensure
that they do reflect this consensus.

The chairs draw no conclusion as to the applicability of
the Abandon operation to abandon update operations.  The
chairs suggest that those who believe that some statement
should be added to the technical specification regarding
use of the Abandon operation to abandon update operations
should offer a specific proposal for WG review.

-- Kurt & Bob