There's a difference between the IBM and Saxonica XSD 1.1 tests for xs:override. The IBM tests list both the overriding and overridden schema documents in the test metadata (without distinguishing them); the Saxonica tests list only the overriding document.
The convention used in the IBM tests makes life rather difficult for test drivers. If you build a schema naively using both schema documents, you get a conflict. To alleviate matters with minimum disruption, I propose to mark any schema documents that are overridden with an attribute such as role="overridden" in the test metadata; I'll also enable role="imported | included | redefined" for other schema documents that don't need to be loaded as primary sources because they are referenced from elsewhere, and role="principal" for a schema document that is not referenced from elsewhere. (But I don't intend to use this markup everywhere, only in cases where it solves problems.)
There's a difference between the IBM and Saxonica XSD 1.1 tests for
xs:override
. The IBM tests list both the overriding and overridden schema documents in the test metadata (without distinguishing them); the Saxonica tests list only the overriding document.The convention used in the IBM tests makes life rather difficult for test drivers. If you build a schema naively using both schema documents, you get a conflict. To alleviate matters with minimum disruption, I propose to mark any schema documents that are overridden with an attribute such as
role="overridden"
in the test metadata; I'll also enablerole="imported | included | redefined"
for other schema documents that don't need to be loaded as primary sources because they are referenced from elsewhere, androle="principal"
for a schema document that is not referenced from elsewhere. (But I don't intend to use this markup everywhere, only in cases where it solves problems.)