caching normalized values

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

As part of this work, I'm thinking it best to rework
the schema engine a bit.  In particular, I plan on
removing syntax "normalizers" in favor of just using
matching rule "normalizers".  This should make the
code much simpler and, hopefully, reduce confusion
as to where normalization should be done.

For now, all the new code is behind SLAP_NVALUES.

>To avoid having to repeatedly normalize stored values,
>I'm going to extend the Attribute structure to hold
>them.  Presently I'm thinking it best not to store
>them on disk.
>
>I'm also think of reworking indexer routines to
>expect the input to be normalized.  But I might save
>this for another day.

Also for another day... extensibleMatch indexing...


Kurt