Re: Specifying dn as sort key in SSS control

[Ed Reed <ED_REED@novell.com>]

Case Ignore, not Case Exact.

----------------------
Ed Reed, Technologist
Novell, Inc.
+1 801 861-3320

>>> "Jim Sermersheim" <JIMSE@novell.com> 11/06/1998 17:27:15 >>>
Good point, I would say that the attribute values of the RDN component should be converted to a string (as in 2.4 of rfc2253) and compared using Case Ignore (or Exact?) Ordering Match.  In the event that the values don't convert to strings, they sort to the end.


>>> David Boreham <dboreham@netscape.com> 11/6/98 4:29:33 PM >>>

Jim Sermersheim wrote:

> I'm trying to understand how to (if possible) specify that the sort key be the rdn (the leftmost component of the name) of an entry.
>
> I tried to see ways to use what's been done with the ":dn" keyword in 2254.  Nothing struck me as appropriate.  The other would be, to come up with an AttributeDescription option like dn or rdn.  So "cn;rdn" or "cn;dn", would mean, use cn but consider the entry's dn also.  When an entry doesn't have a cn, dn would be used exclusively.
>
> Is there an existing way?  A better way?

This sounds problematic. The rdn's of two sibling
entries don't need to be the same attribute (type).
Hence, they might have different syntax, hence you'd be
comparing apples to pears in trying to sort.