[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: [ldapext] Dynamic group draft



Hallvard B Furuseth wrote:
> Jaimon Jose writes:
>> Howard Chu wrote:
>>> Fair enough. That just brings me back to my earlier point that there
>>> should be an actual syntax defined for URLs.
>> Thinking more about this, I think, I like the idea.  If I have a large
>> number of URLs stored, my searches can be more efficient if I can index
>> on each of the URL elements.
> 
> You need special matching rules too for that, and support for indexing
> of extended filters like "(memberQueryURL:LDAPURLScopeMatch:=sub)".
> 
> Alternatively, you could define a "normalized" form of LDAP URLs where
> all '?'s are present, and then a sort of substring match, so
> (memberQueryURL=ldap:///c=foo??*) would match 'ldap:///c=foo'.
> A '*' should only match one URL component.
> 
> 
> BTW, keep in mind that if the draft makes use a new syntax and new
> matching rules, it can't be implemented (correctly) unless the syntax
> and matching rules are implemented as well.  So it might be a bad idea
> to make them very complicated.
> 
> Still, with a really useful draft which pushes for their implementation,
> there may be a real hope that such a syntax will get widely implemented,
> so we can make use of it for other URLs too:-)

An (LDAP) URL is a composite datum, so a syntax for the definition of an
URL-valued attribute should definitely rely on separate syntax
definitions for each component, according to RFC4516.  The DN portion is
a clear example.

I would have no objection to defining a URL syntax; this would really
avoid issues like syntax errors detection at run-time, resulting in
incorrect handling of URLs that would essentially go unnoticed.  Also,
it would be much more efficient to normalize URLs once for all, easing
the parse phase.  URL-valued attributes would be based on this syntax,
and, of course, they would have nothing to do with the current
labeledURI attribute.

However, I believe this should go in a separate draft.  The fact that a
brand new syntax is used in the dynamic group draft would not be a
weakness in my opinion: regular (I mean: read) clients wouldn't have to
deal with it.

p.

p.



Ing. Pierangelo Masarati
OpenLDAP Core Team

SysNet s.n.c.
Via Dossi, 8 - 27100 Pavia - ITALIA
http://www.sys-net.it
------------------------------------------
Office:   +39.02.23998309
Mobile:   +39.333.4963172
Email:    pierangelo.masarati@sys-net.it
------------------------------------------


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