Particles: particlesZ001 and XSD 1.0

[nwellnhof]

I can't make sense of the expected result of the particlesZ001 test in the MS Particles test set under XSD 1.0. Relevant parts of the test group:

      <annotation>
         <documentation>TEST :3.9.1 The Particle Schema Component [ check length of element information items ] : Schema Component Constraint: Particle Derivation OK (Elt:All/Choice/Sequence -- RecurseAsIfGroup) rule is ambiguous</documentation>
         <documentation>Invalid restriction which becomes valid in XSD 1.1 - MHK</documentation>
      </annotation>
      <schemaTest name="particlesZ001">
         <expected validity="valid"/>
      </schemaTest>
      <instanceTest name="particlesZ001.i">
         <expected validity="invalid" version="1.0"/>
         <expected validity="valid" version="1.1"/>
      </instanceTest>

The test involves particle restriction which has changed in XSD 1.1. Without going into the gory details, the primary issue is that under XSD 1.0, only the instance is expected to be invalid, while I'd expect the schema to be invalid. After all, what's tested is a Schema Component Constraint. So I'd expect (without changing the 1.1 outcome):

      <schemaTest name="particlesZ001">
         <expected validity="invalid" version="1.0"/>
         <expected validity="valid" version="1.1"/>
      </schemaTest>
      <instanceTest name="particlesZ001.i">
         <expected validity="valid"/>
      </instanceTest>

This would match many similar tests that are invalid under 1.0 and became valid in 1.1.

[michaelhkay]

I think you're probably right.

(Saxon has never attempted to implement the 1.0 subsumption rules, we have always followed 1.1 here even if other 1.1 features are not enabled.)