atom feed12 messages in org.oasis-open.lists.xacmlRe: [xacml] wd-20 issues
FromSent OnAttachments
Paul TysonMay 26, 2011 9:47 am 
Erik RissanenMay 30, 2011 7:58 am 
Tyson, Paul HMay 31, 2011 6:07 am 
Erik RissanenMay 31, 2011 7:57 am 
Tyson, Paul HMay 31, 2011 8:39 am 
remo...@emc.comJun 6, 2011 10:08 pm 
Erik RissanenJun 9, 2011 2:26 am 
remo...@emc.comJun 10, 2011 12:51 am 
Erik RissanenJun 15, 2011 4:11 am 
rich levinsonJun 15, 2011 10:50 pm 
Erik RissanenJun 16, 2011 1:56 am 
remo...@emc.comJun 16, 2011 2:10 am 
Subject:Re: [xacml] wd-20 issues
From:Erik Rissanen (er@axiomatics.com)
Date:Jun 16, 2011 1:56:54 am
List:org.oasis-open.lists.xacml

Hi Rich,

Yes, you are right. It seems Remon did not notice the description of <AnyOf> in 5.6, which is correct. 7.7 correctly says that an empty target will always match, though the case is not listed in the table. So no action should be required on this.

Best regards, Erik

On 2011-06-16 07:50, rich levinson wrote:

Hi Erik,

On the last item, I don't understand what is going to be "fixed". Also, I do not see any inconsistency between 5.6 and 7.7. I think both situations:

* zero <AnyOf> elements * and one or more <AnyOf> elements

appear to be covered in both sections 5.6 and 7.7 in a consistent manner. Or am I missing something?

Thanks, Rich

On 6/15/2011 7:11 AM, Erik Rissanen wrote:

Remon,

See inline.

On 2011-06-10 09:51, remo@emc.com wrote:

Erik,

-----Original Message----- From: Erik Rissanen [mailto:er@axiomatics.com] Sent: Thursday, June 09, 2011 11:27 AM To: xac@lists.oasis-open.org Subject: Re: [xacml] wd-20 issues

5.29 Element<AttributeDesignator> "If the Issuer is not present in the attribute designator, then

the matching of the attribute to the named attribute SHALL be governed by AttributeId and DataType attributes alone."

- And Category.

Yes!

Also in 7.3.4 Attribute Matching.

5.48 Element<Result> "<PolicyIdentifierList> [Optional] If the ReturnPolicyIdList attribute in the<Request> is true (see section 5.42), a PDP that implements this optional feature MUST return a list of all policies which were found to be fully applicable." - This prevents the PDP from skipping evaluation of policies that cannot affect the decision. IOW, it prevents performance optimizations. This is not a big deal to me, since the feature is optional, but maybe something to note in the implementer's guide?

The intended behavior is not how you interpret it. It says "which were found..." so it's simply the list of policies the PDP worked on during evaluation. It does not mean that the PDP has to figure out which policies "might have been fully applicable if they were to be evaluated". If you have a suggestion for better wording, please post it, and I can update this while we are fixing the other issues.

7.3.7 AttributeSelector evaluation "If the DataType is not one of the primitive types listed above, then the return values shall be constructed from the nodeset in a manner specified by the of the particular DataType extension specification." - "specified by the of the" misses a crucial noun.

It should just say "in a manner specified by the particular datatype extension specification", meaning that if you extend XACML with a custom data type, you also need to define (and implement) this constructor.

7.7 Target evaluation "An empty target matches any request. Otherwise the target value SHALL be "Match" if all the AnyOf specified in the target match values in the request context." - This conflicts with 5.6 Element<Target>: "For the parent of the<Target> element to be applicable to the decision request, there MUST be at least one positive match between each<AnyOf> element of the<Target> element and the corresponding section of the<Request> element."

Yes, I will fix that. Thanks for noticing.