RE: controls

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

If a control X behaved differently depending on whether control Y appeared
before control Z, it would be better to have a data item in control X so
that control X(data=0) and control X(data=1) represented the two behaviours
than to require an implementation to guess which behaviour is required based
on the ordering of other controls.

So I would prefer to see the position as "no control should rely on the
actual order of other controls included in the same request."

Ron.

-----Original Message-----
From: Jim Sermersheim [mailto:jimse@novell.com]
Sent: Friday, 4 October 2002 4:21 AM
To: Ramsay, Ron; Kurt@OpenLDAP.org
Cc: ietf-ldapbis@OpenLDAP.org
Subject: RE: controls


We could state that the ordering in the sequence is not significant
unless specified by a control specification.

>>> "Kurt D. Zeilenga" <Kurt@OpenLDAP.org> 10/02/02 08:13PM >>>
At 05:40 PM 2002-10-01, Ramsay, Ron wrote:
>Actually, I think the stand should be that order is not important.

I don't see any reason why we should take a stand here,
especially given that any future extension specification
can override any stand we might take. 

We should, however, clarify whether the default ordering
semantics is "set" or "sequence" so that authors and readers
of control extension specifications don't have to guess
as to what semantics apply in absence of an explicit
statement as to what semantics apply.

Kurt