Hi Emmanuel,

On 5/21/07, Emmanuel Lecharny <[EMAIL PROTECTED]> wrote:



On 5/21/07, Stefan Zoerner <[EMAIL PROTECTED]> wrote:
>
> Emmanuel Lecharny wrote:
>
> > - Compare should be accepted whatever the AT is
>
> Isn't it necessary to determine a matching rule for a compare operation?
> How do we accomplish this without an attribute type?


My guess is that if a schema is disabled, then it has been enabled once
upon a time, then all the matching rules are still OK. AT are stille present
in the server as soon they have been declared.

*But* it raises another question : what if we delete an AT ?

If the attribute type is not in the schema any longer, a compare should
> raise an error, I think.


Yes, I agree

Sorry for making the problem even more complicated :-(


If it was simple, everybody would have wrote a LDAP server ;)


Heh my thoughts exactly.  I think the best way to handle these issues is via
tooling
where we can describe the impact to the user clearly and perform various
opperations
to check if the schema change leaves the server in an inconsistent state.

Alex

Reply via email to