Test cases expect this behavior, so I think we should keep it.
In particular a capability typeXXX with "allFeaturs" set, is expected to cover
all features whether or not they exist in the type system (according to existing
test cases).
-Marshall
On 8/22/2010 4:56 PM, Marshall Schor wrote:
> If you have a result spec which specifies typeXXX:featYYY, and that type
> and/or
> feature doesn't exist in the type system,
>
> 1) no error is indicated.
> 2) a query to the result spec: containsFeature("typeXXX:featYYY") returns
> true.
>
> Do people think this is OK?
>
> -Marshall
>
>
