Re: WG Last Call: draft-ietf-ldapext-ldapv3-dupent-03.txt

["Jim Sermersheim" <JIMSE@novell.com>]

>>> "Kurt D. Zeilenga" <Kurt@OpenLDAP.org> 6/14/00 10:22:45 PM >>>
<snip>
>>In that case, can't the client just infer whether it's getting duplicates back or not based on the exsistence of duplicated DN's and/or the presence of multi-valued attr's where none should be?
>
>No, because the result code might be sizeLimitExceeded after
>one of many dups were returned.  (I assume this is how the
>control interacts with sizeLimits... this should be specified).

Bah... fine.

<snip>
>>No, it may not be clear, but the intent here is that this is an "all or nothing" control. Either the control is supported across the entire set of results or it's not.
>
>So, the server must first run through all the entries to determine
>if it able to honor the control before returning a single entry.

Before you introduced the notion of the control being able to handle partial support, the thought was that the server can either support the control or not. The scenario above (timeLimit or sizeLimit exceeded) does present a problem though. I'm going to add the partial support stuff, so this discussion is moot now.

<snip>
agreed

Thanks.

Jim