This example shows how to handle schema evolution using substitution groups.
The general idea is as follows: the initial version of the schema defines
extension points using the XML Schema substitution group mechanism. The
subsequent versions of the schema add new elements to the substitution
groups. The goal here is for applications that were built using earlier
versions of the schema to be able to handle documents corresponding to the
newer versions without validation errors or any other failures. This example
shows how to pass the unknown content through parsing and serialization so
that the output XML contains all the unknown elements. The 'ignore' example
shows how to ignore such unknown elements. 

This example uses XML Schema polymorphism in the form of substitution groups.
If you are not familiar with how to work with polymorphic object models,
refer to the C++/Hybrid Mapping Getting Started Guide as well as the 
'polymorphism' example in the examples/cxx/hybrid/ directory.

This example also customizes a generated object model type as well as parser
and serializer implementations. If you are not familiar with the customization
mechanisms, refer to the C++/Hybrid Mapping Getting Started Guide as well as
the examples in the examples/cxx/hybrid/custom/ directory.

The example consists of the following files:

transform.xsd
  The initial version of the schema that describes simple transformations,
  such as move and rotate. The schema makes it possible to add new
  transformations in subsequent versions of the schema. It also defines a
  special unknown_type type which is used to handle unknown transformations.
  Our example is built using this schema.

transform-v2.xsd
  The second version of the schema which adds the scale transformation. This
  schema is provided as an example.

transform.xml
  Sample XML document corresponding to the second version of the schema. In
  particular, it contains the scale transformation which is unknown in
  transform.xsd.

transform.hxx
transform.cxx

transform-pskel.hxx
transform-pskel.cxx
transform-pimpl.hxx
transform-pimpl.cxx

transform-sskel.hxx
transform-sskel.cxx
transform-simpl.hxx
transform-simpl.cxx
  Object model (the first pair of files), parser skeletons (the second pair),
  parser implementations (the third pair), serializer skeletons (the fourth
  pair), and serializer implementations (the fifth pair). These files are
  generated by the XSD/e compiler from transform.xsd. The --generate-parser,
  --generate-serializer, and --generate-aggregate options were used to request
  the generation of the parsing and serialization code. The 
  --generate-polymorphic option was used to request the generation of the
  polymorphism-aware code. The --custom-type, --custom-parser, and 
  --custom-serializer options were used to customize the object model class
  as well as the parser and serializer implementations for the unknown_type
  XML Schema type.

xml.hxx
xml.ixx
xml.cxx
  Simple in-memory representation for raw XML. It is used to store the content
  of unknown transformations.

unknown-type.hxx
unknown-type.cxx
  Custom unknown_type class. It uses the generated version as a base and adds
  members that store the raw XML representation of an unknown transformation
  as well as the parsed element name and namespace from the substitution group.

unknown-type-pimpl.hxx
unknown-type-pimpl.cxx
  Custom unknown_type parser implementation. It uses the implementation
  generated by the XSD/e compiler as a base and overrides the raw XML
  callbacks to parse the unknown content and store it in the customized
  unknown_type class.

unknown-type-simpl.hxx
unknown-type-simpl.cxx
  Custom unknown_type serializer implementation. It uses the implementation
  generated by the XSD/e compiler as a base and overrides the raw XML
  callbacks to serialize the unknown content stored in the customized
  unknown_type class.

driver.cxx
  Driver for the example. It first sets the substitution map callbacks that
  are used to alter the default element mapping in the parser and serializer.
  The driver then calls the parser that constructs the object model from the
  input XML file. It then prints the content of the object model to STDERR.
  Finally, the driver calls the serializer to serialize the object model
  back to XML.

To run the example on the sample XML instance document simply execute:

$ ./driver transform.xml

The example reads from STDIN if input file is not specified:

$ ./driver <transform.xml