Fw: problem with conformance tests

[Satoshi Hada <SATOSHIH@jp.ibm.com>]

Satoshi Hada
IBM Tokyo Research Laboratory 
mailto:satoshih@jp.ibm.com
----- Forwarded by Satoshi
Hada/Japan/IBM on 2004/03/24 10:07 -----

Seth Proctor <seth.proctor@sun.com>

2004/03/24 00:39

To	Satoshi Hada/Japan/IBM@IBMJP
cc
Subject	Re: problem with conformance tests

Hi Satoshi. Thank you for your response. Could you please forward this
email
to the comment list, since I seem to be having trouble with it.

> The problem is that there is no detailed description about
> <AttributeAssignment> in the XACML 1.0 spec.
> Anyway, I'll ask Kudo-san to correct the testcases according to your
comment.

I absolutely agree that there is no detail in the spec. This is a real
issue.
The problem, of course, is that the PDP is supposed to ignore what's in
the
assignments, so it's hard to talk about the correct format :) In this case,
because AttributeAssignment _is_ an AttributeValue, I believe you have
to
follow the definition of AttributeValue to get the correct behavior. Maybe
I'll
suggest some new language for XACML 2.0 to describe this...

A separate problem here is that I'm not sure string should be used the
way
it's being used in these examples (ie, as a container for complex content).
I
raised this issue once before, but never heard any comments back. I think
I'll
try raising it again.
 
> Is there any meeting minutes on the agreement.
> I'm asking because we should mention it (if any) in the test case
description.

I can't seem to find them now, but I know I had this discussion at some
point. I'll see what I can find.


seth