I also like option a:
- Allow for implementations that will NEVER try to evaluate a
syntactically invalid policy at the time a Request is received:
such implementations will be exempted from the tests that have
syntactically invalid policies.
- If an implementation MAY EVER try to evaluate a syntactically
invalid policy at the time a Request is received: specify that
the Decision is to be Indeterminate and the StatusCode Value is
to be "urn:...:syntax-error"
I will modify the Conformance Tests that currently generate an
"Indeterminate" by a type incompatibility to use either a missing
MustBePresent attribute or a divide-by-zero, except in the cases
where I was explicitly testing the handling of an invalid policy
file.
I can use a validating XML parser to do the XML syntax checking
of all the Conformance Test policies, but I do not have a way of
checking type correctness at this time. If someone could make a
pass through the Conformance Test *Policy*.xml files to check for
type correctness, I would appreciate it.
Anne
On 4 December, bill parducci writes: Re: [xacml] IIC012: syntax-error or processing-error?
> From: bill parducci <>
> To: "'XACML TC '" <>
> Subject: Re: [xacml] IIC012: syntax-error or processing-error?
> Date: Wed, 04 Dec 2002 10:28:46 -0800
>
>
> Polar Humenn wrote:
> > For the case when you have a single malformed policy to be evaluated
> > against a request, the answer is just simply undefined. The implementer
> > can choose. Therefore, there should not be a "CONFORMANCE" test for it.
>
> 'undefined' needs to manifest itself somehow in the decision.
>
> back to the run-time vs. precompiled topic for a sec..
>
> supposing that there is a use case/conformance test that has a mis-typed (or otherwise malformed) policy: under the precompiled model, this policy will simply be thrown out upon being communicated to the system; under the run-time model it will generate a status code 'other than OK' and issue a decision of INDETERMINATE. with this in mind we can: (a) allow for both of these scenarios to have it owns conformance criteria; (b) take a stand on which evaluation mechanism is supported; (c) consider it out of scope and allow for unlimted of decision/status combinations.
>
> i personally think that option 'a' is worth considering because option 'b' forces us to make an implementation decision and option 'c' is a bit loose for my taste.
>
> b
>
>
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>
>
--
Anne H. Anderson Email:
Sun Microsystems Laboratories
1 Network Drive,UBUR02-311 Tel: 781/442-0928
Burlington, MA 01803-0902 USA Fax: 781/442-1692
