How is the intention to change config?

[openldap <openldap@ayni.com>]

Hi LDAPers

[cvs@myhost ~]$ rpm -q openldap-servers
openldap-servers-2.3.39-1.fc8
[cvs@myhost ~]$

As far as i have understood, the openldap config is moving towards 
beeing fully LDAP integrated, i.e. you will change the config using 
ldapmodify, you will add to the config using ldapadd, you will view the 
config using ldapsearch.

That created me a problem: i had to add a  schema. what i did: i created 
a new olcInclude object into cn=config. This object, however, was not 
stable, it got lost at the next slapd restart.

When i consulted this discussion list, i was informed that adding a new 
olcInclude object is not the correct way to add a new schema. But i was 
not informed on how the intention is to do it.

I imagine that another way would be to convert the schema from the 
schema file manually/directly into olcAtttributeTypes, olcObjectClasses, 
olcObjectIdentifier and insert them  into the cn=schema,cn=config. This, 
however, would be a very error prone action.

Also, the handling of the access rights in the future is far from clear.

When i consulted the openldap docu, i found put me in front of a 
fait-accompli: that is how it is, help yourself.

could someone shed some light on the evolution process from file-based 
config into LDAP-based config?

thanks in advance.

suomi