This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
In the Sun ElemDecl test set, tests <test set="ElemDecl" group="valueconstraint00501m1" name="Positive"/> <test set="ElemDecl" group="valueconstraint00501m2" name="Positive"/> the essence of the test is to check that the fixed value declared for an element is valid against the definition of the type invoked using xsi:type. The rules state that the canonical lexical representation of the fixed value must be valid against this type. In these two tests the fixed value is a floating point value. Although Part 2 defines a "canonical representation" for float and double, it is not truly canonical, in that multiple representations are permitted. Specifically, an implementation can return 9.999999776482582E-3 as the canonical representation of the value 0.01, and this will violate the pattern facet, causing the instance to be invalid.
If we agree with Mike's analysis, I guess this test gets deleted
The WG decided to mark this one as "result is undefined in 1.0" but the instance is valid in 1.1.