RE: ModifyDN -- subtrees or not?

["Slone, Skip" <skip.slone@lmco.com>]

I agree as well. Although we were looking at a specific instance, it seems
we've uncovered a general weakness in the spec.

 -- Skip

-----Original Message-----
From: John Strassner [mailto:john.strassner@intelliden.com] 
Sent: Thursday, August 23, 2001 9:28 AM
To: Jim Sermersheim; skip.slone@lmco.com; ietf-ldapbis@OpenLDAP.org
Subject: RE: ModifyDN -- subtrees or not?


I agree with Jim - I think that the language should be tightened.

regards,
John

-----Original Message-----
From: owner-ietf-ldapbis@OpenLDAP.org
[mailto:owner-ietf-ldapbis@OpenLDAP.org]On Behalf Of Jim Sermersheim
Sent: Wednesday, August 22, 2001 11:04 PM
To: skip.slone@lmco.com; ietf-ldapbis@OpenLDAP.org
Subject: Re: ModifyDN -- subtrees or not?


Hmm, a bigger question might be: do we need to add 2119 imperatives and
tight language to the entire doc? For example, The delete operation
similarly says "The Delete Operation *allows* a client to *request* the
removal of an entry..." To the purist, this leaves things wide open-- to the
degree that I could willy-nilly decide to not let people delete entries with
a Q in the rdn on odd-numbered days. An interoperability test could only
test that the PDU can be succesfully sent from the client and an appropriate
response recieved from the server. The Delete Response does say that the
server *will* attempt to perform the entry removal, but this is not a MUST.
(I'm sort of using a silly example to highlight the point).

As an aside, I also see that there is no note stating that the operation may
fail due to a move (subtree or otherwise) resulting in invalid containment.

Jim

>>> "Slone, Skip" <skip.slone@lmco.com> 8/22/01 11:57:36 AM >>>
An issue has recently come up in the Directory Interoperability Forum
concerning what RFC 2251 really says on the matter of using ModifyDN to move
subtrees.  As I read it, there's no clear conformance statement (a la RFC
2119) one way or another. The entire discussion in paragraph 4.9 is in plain
English (with no use of RFC 2119 keywords such as MUST).  On that topic,
draft-ietf-ldapbis-protocol-02.txt is no clearer.

It seems to me this is an issue that needs clarifying.  Any thoughts?

 -- Skip Slone