Re: [ldapext] True filter vs. (objectClass=*)

[Hallvard B Furuseth <h.b.furuseth@usit.uio.no>]

Kurt D. Zeilenga writes:
>At 06:13 AM 1/1/2004, Hallvard B Furuseth wrote:
>>I suggest to add this statement to draft-zeilenga-ldap-t-f:
>>
>>   Whenever the "(objectClass=*)" filter will succeed (including
>>   reading the Root DSE), the "(&)" filter will also succeed.
> 
> Succeed?  As it match the DSE?  or returned by the search operation?

s/succeed/match/.

>> rfc2251 section 3.4 explicitly says the (objectClass=*) filter
>> should be used to read the root DSE.  It does _not_ say 'any filter
>> which matches the root DSE', so as currently defined the TRUE filter
>> might not work here.
> 
> This I-D does not alter any TS, including RFC 2251, which mandates a
> particular filter in a specific cases.

I didn't mean to.  Perhaps I should have put "If a server supports
True/False filters," in front of the statement.

> That is, this I-D only extends
> the general filter capabilities.  It does not alter when and where
> those general filter capabilities may be used.

I don't see why my suggestion (as clarified above) does that either.

> Hence, I suggest adding:
>   It is noted that certain search operations, such as those used to   
>   retrieve information from subschema subentries [RFC2251], require use
>   of particular filters.  This document does not change these requirements.

I'm fine with that too, if you also include this modification:

    These filters are commonly used when requesting DSA-
    specific Entries (DSEs)

..."except the root DSE"...

    which do not necessarily have 'objectClass'
    attributes.

That was the statement I was thinking of in my original message.  I
thought you meant it should be possible to read all DSEs with the "(&)"
filter.

-- 
Hallvard

_______________________________________________
Ldapext mailing list
Ldapext@ietf.org
https://www1.ietf.org/mailman/listinfo/ldapext