At 07:42 PM 6/7/2001, Clif Harden wrote:
>This may or may not be a problem, depends on your point of
>view I guess, but I do not think it is right.
>
>When we parse the objectClasses may or must attribute definitions
>it appears that we are including these attribute names in the
>in the list with the true attributeType definitions. This is okay
>as long as there are true attributeType definitions to match the
>may and must attributes.
>
>However this is not always the case, depending on the schema(s)
>that are being used. My netscape DS has many "may contain" attributes
>that do not have matching attributeType definitions.
>(I found this out trouble shooting this problem.)
>Trying to extract data for these "bogus" attributeType(s) causes a minor
>problem, there is no data for them.
>
>IMHO, when we ask for the attributes we should get only the
>true attributeType definitions that are in the schema.
Clients should be prepared for a number of schema oddities.
In addition to the above, you find that not all LDAPsyntaxes
or matchingRules mentioned in attributeType descriptions
are published (ala OpenLDAP 2.0). A client should assume
that if something is not published that it's not supported.
Kurt
|
