This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The following is an excerpt from Michael Kay's email (member-only): http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2009Jan/0041.html I think that where we describe the addition of the new property [inherited attributes], we need to specify that the addition of an inherited attribute to an element should invoke namespace fixup in the same way as adding a defaulted attribute. (In most cases, of course, the namespace will already be in scope, because it will have been declared on the element from which the attribute is inherited. But namespace fixup can still be needed if the relevant prefix has been rebound or unbound at an intermediate level.) This also means that the inherited attribute may have a different prefix from the original. The current text doesn't seem to give enough detail on how the attribute information items in [inherited attributes] are constructed; I think that there is a need to spell out the properties of these information items in the same way as we do for defaulted attributes.
The following is Sandy's response in (member-only): http://lists.w3.org/Archives/Member/w3c-xml-schema-ig/2009Jan/0042.html For [inherited attributes], I don't think we need to perform the same fixup. They are just references to other information items, just like the [parent] property points to the parent element. The key point is that members of the [inherited attributes], unlike defaulted ones, are *not* members of the [attributes]. But I do agree that when it comes to constructing a DTM for CTA, namespace bindings are needed, because [inherited attributes] are considered part of [attributes] at that point.
During its 2009-01-23 telecon, the schema WG adopted a proposal to address this issue. The proposal (along with other changes) can be found at (member-only): http://www.w3.org/XML/Group/2004/06/xmlschema-1/structures.omni.20090123.html The adopted change is to add the following sentence to the end of clause 1.1.3 of section 3.12.4: "When an attribute with a non-empty [namespace name] is copied, ·namespace fixup· may need to be performed on the resulting information set to ensure that a prefix P is bound to the [namespace name] and the [prefix] of the copied attribute is set to P." With this change, the WG believes that the issue raised in this bug report is addressed. I'm marking this RESOLVED and CLOSED accordingly.