Re: [xml] XSD Schema validation issues with boolean type.

[Bas Driessen <bas driessen xobas com>]

On Mon, 2006-02-20 at 03:39 -0500, Daniel Veillard wrote:
> On Mon, Feb 20, 2006 at 01:29:32PM +1000, Bas Driessen wrote:
> > Hello,
> > 
> > I am using the latest version of libxml2 (2.6.23) and have an issue with
> > validation boolean type fields using "xmllint --schema". If I have an
> > XML file as follows:
> > 
> > <TestRoot xmlns="urn:Test-schema"
> > xmlns:xsd="http://www.w3.org/2001/XMLSchema"
> > xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
> > xsi:schemaLocation="urn:Test-schema Test.xsd">
> >   <Boolean1>true</Boolean1>
> >   <Boolean2>xtrue</Boolean2>
> >   <Boolean3>   true   </Boolean3>
> >   <Boolean4/>
> > </TestRoot>
> > 
> > My understanding is that only Boolean1 should validate OK and the other
> > 3 cases should fail. My test.xsd is as follows:
> > 
> > <xs:schema xmlns="urn:Test-schema"
> > xmlns:xs="http://www.w3.org/2001/XMLSchema" id="Test" version="1"
> > targetNamespace="urn:Test-schema" elementFormDefault="qualified">
> > <!-- Root -->
> >   <xs:element name="TestRoot">
> >     <xs:complexType>
> >       <xs:sequence>
> >         <xs:element name="Boolean1" type="xs:boolean"/>
> >         <xs:element name="Boolean2" type="xs:boolean"/>
> >         <xs:element name="Boolean3" type="xs:boolean"/>
> >         <xs:element name="Boolean4" type="xs:boolean"/>
> >       </xs:sequence>
> >     </xs:complexType>
> >   </xs:element>
> > </xs:schema>
> > 
> > The output of xmllint however is as follows:
> > 
> > test.xml:4: element Boolean2: Schemas validity error : Element
> > '{urn:Test-schema}Boolean2': 'xtrue' is not a valid value of the atomic
> > type 'xs:boolean'.
> > test.xml fails to validate
> > 
> > So it only picks up the problem in situation 2, while situation 3 and 4
> > is not being picked up as a problem.
>
>   http://www.w3.org/TR/xmlschema-2/#boolean
>
> there is a 'whiteSpace' constraining facet, so I think Boolean3 might validate
> For Boolean4 this looks like a bug to me. Also note that '1' and '0' are 
> accepted values for this type. 
> Kasimier certainly has a more precise idea about this. I'm surprized this
> kind of problem isn't caught by the huge OASIS datatype regression tests.

Thanks for that Daniel. Yes, just confirmed that the values 0 and 1 validate OK while any other numeric value like (2,3,-0,-1 etc) fail as expected, so no issues there. 
Will wait for comments from Kasimier.

Bas.