RE: ModifyDN -- subtrees or not?

["Kurt D. Zeilenga" <Kurt@OpenLDAP.org>]

At 09:10 AM 2001-09-19, Jim Sermersheim wrote:
>Due to the amount of confusion regarding this issue. I suggest that either a general and very concrete statement is made that explicitly allows a server to refuse to service any operation for any reason,

I believe that some general statement could be made in section
3.1. Protocol Model.  It should note that a request is only
a request to perform, not a mandate to perform, an operation.


>or that each operation be reworded to replace "will" language with something along the lines of "attempts to".

Yes, and not just in this particular case, but in other such cases.

>I suppose the first choice can be done by adding or referencing an applicability statement, but that imposes an extra level of indirection that the reader must jump through.

I would suggest that the paragraph be added to the end of Section
2 noting that this specification is a Technical Specification and
that Applicability Statements should be separately specified when
appropriate.

I also note that Section 2 likely should be an Introduction and a
more abstract (and brief) Abstract be added.


>Jim
>
>>>> "Kurt D. Zeilenga" <Kurt@OpenLDAP.org> 9/17/01 11:57:40 AM >>>
>At 03:28 AM 2001-09-17, Chris Harding wrote:
>>My suggestion is that you clarify whether the RFC requires servers to implement Move Subtree (and other operations that are the subjects of protocol requests).
>
>While I think the language used could be improved, I believe
>the technical specification is reasonable clear that a server
>can refuse to successfully complete any operation for any
>reason and is only obligated to return an appropriate response.
>
>An implementation is free to return a non-successful result in
>response to any and all subtree move (rename) requests.  This may
>be due to mapping onto a directory system which cannot support
>this request [RFC 2251, 3.3].  I note that many LDAP servers
>do not support subtree move (rename).
>
>Kurt