Re: Significance of name forms.

[Howard Chu <hyc@symas.com>]

Michael Ströder wrote:
> Howard Chu wrote:
> > There can only be one DIT Structure Rule for an entry, and a DIT
> > Structure Rule can only reference one nameForm. For any given entry,
> > only one
> > nameForm may be in effect.
>
> I've forgotten that the DIT structure rule can only reference a single
> name form (although I've implemented it that way).
>
> Hmm...still don't get it...reviewing my code...
>
> Are you saying that different DIT structure rules each referencing
> another name form which reference the same structural object class
> cannot have the same SUP id? SUP is also multi-valued.

Correct.

> Would this be possible?
>
> sr2 --SUP--> sr1
> sr3 --SUP--> sr1
>
> sr1 --FORM--> nf1
> sr2 --FORM--> nf2
> sr3 --FORM--> nf3
>
> nf1 --OC--> oc1
> nf2 --OC--> oc23
> nf3 --OC--> oc23
>
> (with nf2 and nf3 having different RDN attrs sets)

Nope, not allowed.

> The examples in X.521(1993) are pretty simple. But I remember having
> seen more flexible declarations when doing interop testing with some
> X.500 servers.

There is still flexibility here, since an X.500 server can define a 
distinct subschema per subtree. But within a single subtree, or within 
the scope of a single subschema, no.

In OpenLDAP we have stubs for supporting per-DB subschema, but would 
need to do a bit more work to support arbitrary per-subtree subschema.

--
  -- Howard Chu
  CTO, Symas Corp.           http://www.symas.com
  Director, Highland Sun     http://highlandsun.com/hyc/
  Chief Architect, OpenLDAP  http://www.openldap.org/project/