XML Schema Support


	Introduction

Xerces-C++ includes an implementation of the W3C XML Schema specification, a recommendation of the Worldwide Web Consortium available in three parts: XML Schema: Primer, XML Schema: Structures and XML Schema: Datatypes. We consider this implementation complete except for the limitations outlined below.


	Limitations

In certain complex content models specifying large values for the minOccurs or maxOccurs attributes may result in poor performance and/or large amount of memory being allocated by the parser. In such situations large values for minOccurs should be avoided, and unbounded should be used instead.
The parser treats local elements in the same scope with the same name and namespace as one element declaration and does not differentiate between them.


	Interpretation of Areas that are Unclear or Implementation-Dependent


	keyref

We have interpreted the specs as requiring <keyref> Identity Constraints to refer to <key> or <unique> identity constraints within the scope of the elements to which the <keyref> is attached. This interpretation is at variance with the Schema Primer, which contains an example with a <keyref> declared on an element used inside the element of its corresponding <key>.


	out-of-bound float values

For float data, the specification does not explicitly prescribe how to deal with out-of-bound data. Xerces-C++ converts these values as shown below:

Values in range	Values converted
less than -2²⁴ * 2¹⁰⁴ (approx -3.402823669e+38)	-INF
greater than -1 * 2^-149 (approx -1.401298464e-45) and less than -0	-0
greater than +0 and less than +1 * 2^-149 (approx +1.401298464e-45)	+0
greater than +2²⁴ * 2¹⁰⁴ (approx 3.402823669e+38)	+INF

The effect of this conversion would invalidate an instance data, for example, "1.1e-46", of a data type derived from float, with minExclusive value '+0', since "1.1e-46" is converted to "+0", which is the same as the minExclusive.


	out-of-bound double values

Similarly, Xerces-C++ converts double values as shown below (the values depend on the system specific values of DBL_MAX and DBL_MIN):

Values in range	Values converted
less than -DBL_MAX (approx -1.7976931348623158e+308)	-INF
greater than -DBL_MIN (approx -2.2250738585072014e-308) and less than -0	-0
greater than +0 and less than +DBL_MIN (approx +2.2250738585072014e-308)	+0
greater than +DBL_MAX (approx +1.7976931348623158e+308)	+INF


	Usage

Below is an example that shows how to turn on schema processing in DOMParser (default is off). Note that you must also turn on namespace support (default is off) for schema processing.


	// Instantiate the DOM parser. XercesDOMParser parser; parser.setDoNamespaces(true); parser.setDoSchema(true); parser.parse(xmlFile);

Usage in SAXParser is similar, please refer to the SAXCount sample program for further reference.

Below is an example that shows how to turn on schema processing in SAX2XMLReader (default is on). Note that namespace must be on (default is on) as well.


	SAX2XMLReader* parser = XMLReaderFactory::createXMLReader(); parser->setFeature(XMLUni::fgSAX2CoreNameSpaces, true); parser->setFeature(XMLUni::fgXercesSchema, true); parser->parse(xmlFile);


	Associating Schema Grammar with instance document

Schema grammars can be associated with instance documents in three ways.


	Specifying Schema Grammar through attributes in the instance document

If schema grammar was not specified externally through methods, then each instance document that uses XML Schema grammars must specify the location of the grammars it uses by using an xsi:schemaLocation attribute if they use namespaces, and xsi:noNamespaceSchemaLocation attribute otherwise.

Here is an example with no target namespace:


	<?xml version="1.0" encoding="UTF-8"?> <personnel xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:noNamespaceSchemaLocation='personal.xsd'> ... </personnel>

Here is an example with a target namespace. Note that it is an error to specify a different namespace in xsi:schemaLocation attribute than the target namespace defined in the Schema.

<?xml version="1.0" encoding="UTF-8"?>
<personnel xmlns="http://my.com"
          xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
	   xsi:schemaLocation="http://my.com personal.xsd http://my2.com test2.xsd">
...
</personnel>


	Specifying Schema Grammar through method calls

An application developer may associate schemas with instance documents through methods setExternalSchemaLocation if they use namespaces, and setExternalNoNamespaceSchemaLocation otherwise. (For SAX2XMLReader, use the properties: "http://apache.org/xml/properties/schema/external-schemaLocation" and "http://apache.org/xml/properties/schema/external-noNamespaceSchemaLocation")

Here is an example with no target namespace:

// Instantiate the DOM parser.
XercesDOMParser parser;
parser.setDoNamespaces(true);
parser.setDoSchema(true);
parser.setExternalNoNamespaceSchemaLocation("personal.xsd");
parser.parse("test.xml");

// Instantiate the SAX2 XMLReader.
SAX2XMLReader* parser = XMLReaderFactory::createXMLReader();
XMLCh* propertyValue = XMLString::transcode("personal.xsd");
ArrayJanitor<XMLCh> janValue(propertyValue);

parser->setProperty(
       XMLUni::fgXercesSchemaExternalNoNameSpaceSchemaLocation,
       propertyValue);
parser.parse("test.xml");

Here is an example with a target namespace. Note that it is an error to specify a different namespace in setExternalSchemaLocation than the target namespace defined in the Schema.

// Instantiate the DOM parser.
XercesDOMParser parser;
parser.setDoNamespaces(true);
parser.setDoSchema(true);
parser.setExternalSchemaLocation(
  "http://my.com personal.xsd http://my2.com test2.xsd");
parser.parse("test.xml");

// Instantiate the SAX2 XMLReader.
SAX2XMLReader* parser = XMLReaderFactory::createXMLReader();
XMLCh* propertyValue = XMLString::transcode(
  "http://my.com personal.xsd http://my2.com test2.xsd");
ArrayJanitor<XMLCh> janValue(propertyValue);

parser->setProperty(
       XMLUni::fgXercesSchemaExternalSchemaLocation,
       propertyValue);
parser.parse("test.xml");


	Pre-parsing and Caching Schema Grammar

An application developer may also pre-parse and cache the grammar corresponding to the XML document namespace as described in the Pre-parsing Grammar and Grammar Caching section of the Xerces-C++ Programming Guide.