This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
These tests use <assert-serialization> to check the results. One value in the result is an xs:double 0, but I think that xs:double -0 is also valid. Perhaps the tests could be tweaked to using assert-serialization.
I believe that -0 cannot appear in this result. The test case is: declare default element namespace "http://www.w3.org/XQueryTestOrderBy"; <results> { for $x in /DataValues/NegativeNumbers/orderData order by ($x * -1) descending return ($x * -1e0) (:force to xs:double:) } </results> This test case uses orderdata2, so that the source document is not validated. The source document contains an xs:untypedAtomic -0 value. The return multiplication first casts this to xs:double -0 and then does the multiplication, returning xs:double 0.
Agreed. It was my test harness, trying to avoid loading the same document more than once, which mean it was using the validated version.