[Date Prev][Date Next] [Chronological] [Thread] [Top]

RE: current control combination proposals



Hi Jim,

I think you have to be careful in assigning an error to a particular combination of controls as this error would stick even if you subsequently wanted to use the controls together.

I don't think return of errors should be prescribed or specified. The greatest direction I would give would be, say, "if a particular combination of controls could not be performed because the implementation deemed the combination to be (inappropriate?), the error protocolError should be returned."

Ron

PS I appreciate the plain text; my mail system won't allow me to respond to messages whose encoding is Arabic.

-----Original Message-----
From: owner-ietf-ldapbis@OpenLDAP.org
[mailto:owner-ietf-ldapbis@OpenLDAP.org]On Behalf Of Jim Sermersheim
Sent: Saturday, 15 May 2004 07:49
To: ietf-ldapbis@OpenLDAP.org
Subject: current control combination proposals


I'm having a hard time gauging consensus on this. Maybe distilling it to
the core changes will help:

Problem:

What behavior is to be followed when encountering a combination of
controls which is not speficied?

Proposal A:

Treat any such combination as an error.

Proposal B

Allow such combinations to be treated as an error, or to be handled in
a non-standard way.

Proposal C:

Allow certain control combinations  (including those in the problem
statement) to be considered as 'innapropriate'. This allows servers to
ignore non-critical controls and/or return unavailableCriticalExtension
for these innapropriate combinations.

Proposal D:

Same as proposal C, except allow any field of an LDAP message
(including controls) to cause the operation to be considered as
'innapropriate'. Also, stating that operation X is appropriate indicates
that all combinations of operation X are appropriate (unless otherwise
specified).

Do these capture the essence of the current proposals? If so, which do
people prefer?

For me, if we are strictly avoiding the addition of new features, I
prefer B, otherwise I prefer D.

Jim