This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The assertions for the test fn:avg((xs:float("-3.4028235E38"),xs:float("-3.4028235E38"))) accept the value -3.4028235E38 as valid. However, -3.4028235E38 is an xs:double and thus not equal to xs:float("-3.4028235E38"). Even casting the input sequence to double would not return xs:double("-3.4028235E38"), but xs:double(xs:float("-3.4028235E38")) which is xs:double("-3.4028234663852885E38") So it should allow xs:float("-3.4028235E38") instead, and perhaps not allow -3.4028235E38
I think you are right, the return type must be xs:float here. Since -3.4028235E38 ne xs:double('-3.4028235E38'), the test result should probably have the expression xs:float("-3.4028235E38") instead.
> -3.4028235E38 ne xs:double('-3.4028235E38') Oops, that should have read: -3.4028235E38 ne xs:float('-3.4028235E38')
I also think you are right. I have made the change as suggested and committed it to cvs.