Re: [xml] Schema validity failure for valid document

[Kasimier Buchcik <kbuchcik 4commerce de>]

Hi,

Daniel Veillard wrote:
> On Mon, Jan 10, 2005 at 04:47:33PM +0100, Kasimier Buchcik wrote:
>
> > Hi,
> >
> > Daniel Veillard wrote:

> > > Well, there is only one string in the regexp automata.
> > > xmlRegExecPushString2() coalesce pairs. I'm not sure duplicating
> > > the API is a good thing, either the info function expand coalesced
> > > sctrings or not, but having 2 function calls sounds a bit messy to me.
> >
> > OK, so I'll format the local name and the namespace on the schema side.
> > The problem I saw was that the regex engine formats the local name
> > and the namespace name, to squeeze it into one string using
> > xmlAutomataNewTransition2; extracting those values on the schema side
> > would mean that the format is no longer encapsulated in the regex
> > code. But if this is OK with you, I'll do it.
>
>
>    yeah, if needed we can provide a public splitting function. I though
> about this a bit more and I think the current API is better because one
> doesn't need to make deallocation of object as the result of calling this
> function, all string returned are comming directly from the regexp. If
> we split one will need to deallocate the chunks produced because they
> possibly don't exist in the regexp.

I tried an initial implementation.
There is a problem with negated namespaces in wildcards:
Example:

errRep1.xsd
(the "##other" namespace indicator in <any> is of importance here):

<?xml version="1.0"?>
<xsd:schema xmlns:xsd="http://www.w3.org/2001/XMLSchema";
  targetNamespace="http://FOO";
  xmlns="http:FOO"
  elementFormDefault="qualified">
  <xsd:element name="foo">
    <xsd:complexType>
      <xsd:sequence>
        <xsd:element name="a"/>
        <xsd:element name="b" minOccurs="0" maxOccurs="unbounded"/>
        <xsd:any minOccurs="0" maxOccurs="1" namespace="##other" 
processContents="lax"/>
        <xsd:element name="c"/>
      </xsd:sequence>
    </xsd:complexType>
  </xsd:element>
</xsd:schema>

errRep1_0.xml:
<?xml version="1.0"?>
<foo xmlns="http://FOO";
  xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance";
  xsi:schemaLocation="http://FOO errRep1.xsd">
  <a/>
  <b/>
  <x/>
  <c/>
</foo>

P:\libxml2-lab\tests\2005-01-10>xmllint --noout --schema errRep1.xsd 
errRep1_0.xml
DEBUG terminal: 0
DEBUG nbval: 0
Element 'c': This element is not expected. Expected is .
errRep1_0.xml fails to validate

Note that nbval is zero here.
Changing the negated namespace (##other) to "http://BAR";, I get:

P:\libxml2-lab\tests\2005-01-10>xmllint --noout --schema errRep1.xsd 
errRep1_0.xml
DEBUG terminal: 0
DEBUG nbval: 3
Element 'x': This element is not expected. Expected is one of { 
("http://FOO";, "b"), ("http://BAR";, any), ("http://FOO";, "c") }.
errRep1_0.xml fails to validate

This is due to the fact that negated namespaces are build using
an automaton approach in xmlSchemaBuildAContentModel:

  deadEnd = xmlAutomataNewState(ctxt->am);
  ctxt->state = xmlAutomataNewTransition2(ctxt->am,
    start, deadEnd, BAD_CAST "*", wild->negNsSet->value, type);
  ctxt->state = xmlAutomataNewTransition2(ctxt->am,
    start, NULL, BAD_CAST "*", BAD_CAST "*", type);
  xmlAutomataNewEpsilon(ctxt->am, ctxt->state, end);

The namespace is let through and then caught with a dead-end.

Any ideas?

One workaround I see, would be to add a special negating character to
the namespace name if calling xmlAutomataNewTransition2, and augment
xmlRegStrEqualWildcard to handle negations:
Example:
  "*|~http://FOO"; - the tilde indicates a negation

But this would be a non-automaton approach, plus I don't know if it's
too hackish.

Regards,

Kasimier