Shibboleth Developers

[Bradley Schwoerer <>]

The code below doesn't work as intended and I will include a working 
patch to SC-132.

-Bradley


On 12/24/10 12:04 AM, Bradley Schwoerer wrote:
> I submitted a bug about an NPE in MappedAttributeDefinition when the 
> dependency attribute is null 
> (https://bugs.internet2.edu/jira/browse/SC-132).  I have a question 
> about desired functionality though.
>
> In code it is setup to allow for multiple dependency attributes 
> mapping to one new attribute.  That is straight forward.  What is not 
> straight forward is the use of default value.  If I have three 
> dependency attributes like this instead of a single multivalued 
> attribute:
>
> StudentRole = Withdrawn
> EmployeeRole = Staff
> AffiliateRole =
>
> And a rule setup like:
>
> <DefaultValue>AFFILIATE</DefaultValue>
>
> <ValueMap>
> <ReturnValue>STUDENT</ReturnValue>
> <SourceValue>enrolled</SourceValue>
> <SourceValue>eligible</SourceValue>
> </ValueMap>
>
> <ValueMap>
> <ReturnValue>EMPLOYEE</ReturnValue>
> <SourceValue>Staff</SourceValue>
> </ValueMap>
>
> <ValueMap>
> <ReturnValue>MEMBER</ReturnValue>
> <SourceValue>enrolled</SourceValue>
> <SourceValue>eligible</SourceValue>
> <SourceValue>Staff</SourceValue>
> </ValueMap>
>
> The result would be EMPLOYEE, MEMBER, and AFFILIATE, because on the 
> third iteration of the for 'Object o' loop for the AffiliateRole 
> attribute it would add the value defined as a default value.  Is that 
> the desired behavior?  I would think that the default value rule would 
> be added only after all attributes and values return no mapped value.  
> In the above example the return would be EMPLOYEE and MEMBER instead.  
> The exception would be if the passThru="true", then it would add each 
> unmatched value so in code it would look like what is attached.  I 
> have included both a before and after instead of a diff for 
> readability of logic.
>
> -Bradley