Kurt Zeilenga wrote:

Declarations of schema elements should be published as specified in RFCs. [..] If caseIgnoreListMatch is not supported, than the MatchingRule definition for caseIgnoreListMatch should not be published.

I suspected an answer like this. I've changed my code to also look up whether the matching rule description itself exists but loose subschema definitions like this makes it really hard to implement smarter schema-aware clients. :-/

These kind of things would be interesting subjects of interop testing.

I note that the specification does allow publication of unimplemented descriptions.

Even worse. :-(

Ciao, Michael.