Re: implementing redefines - additional test

> Although we redefine "dummy.xsd", which neither contains, nor
> includes/imports the "base" type, this type was somehow redefined;
> Is this the expected behaviour or a bug?

Stricly speaking, that's a bug.

Section 4.2.2 [1] says:

        The definitions within the <redefine> element itself are 
restricted to be
        redefinitions of components from the <redefine>d schema document,
        in terms of themselves. 

and SRC 'Individual Component Redefinition' [2] says, among other things:

        In all cases there must be a top-level definition item of the 
appropriate
        name and kind in the <redefine>d schema document.

> If the behaviour is expected, then what's the sense of the
> schemaLocation attribute of a <redefine>, if we just need to
> <include> the component to be able to redefine it?

Conceptually, a redefine is just an include and then an edit of one or 
more components.   Your example leads me to say that perhaps the spec, as 
written, is too strict...in that I don't really see a reason why the 
component that is redefined must exist in the redefined schema 
document...it should just have to exist in the redefining schema (which it 
normally does by virture of the fact that a redefine is an include _and 
then an edit_.)

pvb

pvb

[1] http://www.w3.org/TR/xmlschema-1/#modify-schema
[2] http://www.w3.org/TR/xmlschema-1/#src-expredef