1526 2005-07-09 07:26:27 +0000 $x cast as xdt:anyAtomicType 2005-09-29 11:32:48 +0000 1 1 1 Unclassified XPath / XQuery / XSLT Functions and Operators 1.0 Last Call drafts PC Windows XP CLOSED FIXED P2 normal --- 1 mike ashok.malhotra public-qt-comments oldest_to_newest 4462 0 mike 2005-07-09 07:26:28 +0000 (Raised by Frans Englich on the xsl-list at mulberrytech) We don't seem to have any rule that says what should happen if you write $x cast as xdt:anyAtomicType One would think it should fail, because xdt:anyAtomicType is abstract - except that we no longer seem to say anywhere that xdt:anyAtomicType is abstract. Saxon 8.4 allows this expression through, because if ($x instance of T) is true then it treats ($x cast as T) as a no-op. Michael Kay 4506 1 mrys 2005-07-11 19:15:59 +0000 It should not be allowed. This could be one reason to re-introduce the notion of abstract types (together with xs:NOTATION). 4529 2 ashok.malhotra 2005-07-12 21:21:49 +0000 The WGs decided on the joint call on July 12, 2005 to add wording to section 17.1 of the F&O harmonizing with the wording in the XQuery/XPath documents to clarify that casting is not permitted to xdt:anyAtomic type and error XPST0080 is raised if it is attempted.