This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
An attribute declaration may be local to an attribute group definition, in which case its scope is specified as *absent*. But when the group is referenced in a complex type definition, the spec doesn't specify that it's scope is set to that of the complex type. There may be a similar problem for elements/element groups as well. See: http://lists.w3.org/Archives/Public/www-xml-schema- comments/2001OctDec/0209.html
*** Bug 1895 has been marked as a duplicate of this bug. ***
Resolved at May f2f (2002 [1]) as an error. Editor is instructed to draft erratum that repairs the limitations imposed by the spec, and to do so for elements and elementGroups as well. A similar issue was discussed at May 2003 F2F ([2]). [1] http://www.w3.org/XML/Group/2002/05/xml-schema-ftf-minutes#ab2b3b3c21b8 [2] http://www.w3.org/XML/Group/2003/05/xml-schema-ftf-minutes#d0e787
*** Bug 1973 has been marked as a duplicate of this bug. ***
(From comments on bug 1973) Mary Holstege and Henry S Thompson are convinced the WG agreed to set this to the enclosing group definition, fixing a long-standing bug in Structures, but we can't find any as-it-were Phase-1 agreement on this. Appears to have been classified with error-with-erratum, but HST is not convinced that this change makes sense for 1.0, needs discussion.
It appears that if we did have agreement on the classification and technical solution to this issue, then that agreement has come unstuck. Accordingly, I'm labeling this unclassified, so the WG can deal with it in the normal course of handling 1.0 errata reports.
Discussed at 2005-09-28 F2F meeting. Keep the classification of "error with corrigendum" and instruct the editor to implement the {scope}=enclosing attr group defn fix, and come back to us if that reveals a contradiction. Also, per minutes of May 2002 (see ref), check and if necessary do for element decl.scope and elt group defns