Re: [ws-caf] security considerations

[Simeon Greene <simeon.m.greene@Oracle.com>]

Doug,
I don't quite see how this addition will be a violation of the Unique 
Particle Attribution constraint.  Adding the wsu:id attribute to the 
ContextType in wsctx.xsd can be done as follows:
<xs:complexType name="ContextType">
        <xs:sequence>
            <xs:any namespace="##other" processContents="lax" 
minOccurs="0" maxOccurs="unbounded"/>
            <xs:element name="context-identifier" type="xs:anyURI"/>
            <xs:element name="context-service" type="ref:ServiceRefType" 
minOccurs="0"/>
            <xs:element name="type" type="xs:anyURI" minOccurs="1"/>
            <xs:element name="context-manager" type="ref:ServiceRefType" 
minOccurs="0"/>
            <xs:element name="parent-context" type="tns:ContextType" 
minOccurs="0" maxOccurs="1"/>
        </xs:sequence>
        <xs:attribute name="timeout" type="xs:int" use="optional"/>
          <xs:attribute ref="wsu:id"/>
    </xs:complexType>
I am assuming here that wsu:id is a global attribute in the wsu 
namespace (and of course wsu.xsd has been imported).  If it isn't global 
then a new attribute will have to be added as such:
       <xs:attribute name="id" type="wsu:IDType" use="optional"/>
(assuming that wsu:IDType is a simple type defined in the wsu namespace).

A violation of the Unique Particle Attribution constraint could happen 
if we had an anyAttribute declaration in the ContextType... but we don't 
(it won't clash with the any element declaration).  In any case , the 
schema spec says that this constraint is non-normative. 

Regards,
Sim

Doug Bunting wrote:

> This proposed addition (if you mean to the schema) will itself run 
> into some XML Schema constraints: The addition of any particular 
> attribute in a ##other namespace (such as wsu:id) will make the schema 
> invalid according to the Unique Particle Attribution (§3.8.6)[1] 
> constraint.  This is described in detail in Appendix H of XML Schema 
> part 1[2].  The Attribute Wildcard Intersection description[3] adds 
> some support for this position.
>
> Text describing use of this attribute in the specification would, of 
> course, avoid such problems.
>
> thanx,
>     doug
>
> [1] http://www.w3.org/TR/2001/REC-xmlschema-1-20010502/#cos-nonambig
> [2] http://www.w3.org/TR/2001/REC-xmlschema-1-20010502/#non-ambig
> [3] http://www.w3.org/TR/2001/REC-xmlschema-1-20010502/#cos-aw-intersect
>
>
> On 10-Sep-04 11:47, Greg Pavlik wrote:
>
>> yes, this is a proposed addition.
>>
>> Doug Bunting wrote:
>>
>>> Greg,
>>>
>>> A minor point...
>>>
>>> On 10-Sep-04 09:40, Greg Pavlik wrote:
>>>
>>>> The ContextType schema includes an optional attribute, wsu:Id, 
>>>> which is 
>>>
>>>
>>>
>>>
>>> Actually, the schema does not explicitly mention any attribute in 
>>> another namespace.  In other words, the text may describe this 
>>> particular option but the schema does not.  I suggest "The 
>>> ContextType schema allows use of an optional..." with a similar 
>>> change for the context-identifier mention later in the same paragraph.
>>>
>>> thanx,
>>>     doug
>>>
>>
>>