Dear Mike, That's a good catch.
If we want to "save the spirit of the spec" while stretching its letter reasonably, my interpretation of "The schema import may [...] declare that target namespace to be the default element/type namespace" is that a schema import that declares a default element/type namespace also literally constitutes a "default element/type namespace declaration" in an extended sense (in addition to the "vanilla" default element/type namespace declarations of 4.14). As a consequence, my interpretation of "A Prolog may contain at most one default element/type namespace declaration [err:XQST0066]." would speak in favor of throwing XQST0066 for the query below. Does this reasoning make sense? Kind regards, Ghislain > On 11 Nov 2020, at 09:25, Michael Kay <m...@saxonica.com> wrote: > > I can't see any rule that makes it an error to write > > import schema default element namespace "http://example.org/abc";; > declare default element namespace "http://example.org/xyz";; > > This looks to me like an oversight in the spec. > > I think Saxon simply uses whichever one comes last. > > Any thoughts? > > Michael Kay > Saxonica > _______________________________________________ > talk@x-query.com > http://x-query.com/mailman/listinfo/talk _______________________________________________ talk@x-query.com http://x-query.com/mailman/listinfo/talk
