NEWS

Version 4.2.0

  * In this version we have switched to the build2 build system. For the
    step-by-step instructions on how to build XSD from source on all the
    major platforms, see:

    https://codesynthesis.com/products/xsd/doc/install-build2.xhtml

  * In this version we have changed the default C++ standard that is used
    by the generated code from C++98 to C++11. While you can still request
    C++98 with the --std=c++98 option, note that this is the last release
    that supports the C++98 standard and the next release will require C++11
    or later (in fact, building the XSD compiler itself already requires
    C++11 or later). Note also that the --std option now recognizes the
    c++14, c++17, c++20, and c++23 additional values.

  * This version will be the last release that supports Xerces-C++ earlier
    than 3.1.0 and the next release will require Xerces-C++ 3.1.0 or later.

  * New --file-list-only option that allows only writing the list of C++
    files that would be generated without actually generating them.

  * The --file-list option now recognize the `-` value as a request to write
    to stdout.

  * New --dep-file option that allows specifying the dependency file name.
    It also allows writing the dependency information to stdout by specifying
    `-` as this option's value.

  * This version contains a large number of bug fixes and minor improvements
    that have accumulated over the years.

C++/Tree

  * Support for abstract XML Schema types. The corresponding C++ classes now
    have the _clone() member function declared pure virtual which prevents
    the construction of instances of such types.

    Note that if after upgrading to this version you start getting C++
    compile errors in the generated code with diagnostics saying that a
    type cannot be instantiated because _clone() is pure virtual, then
    it's a strong indication that this type (or its base) is polymorphic
    and needs to be marked as such (see --polymorphic-type* options).

  * New base_string() accessors in the xml_schema::{string,uri} types that
    return the underlying string.

  * Support for `ucc` (upper-camel-case) value in the --function-naming
    option.

  * New C++/Tree `secure` example (xsd-examples/cxx/tree/secure/) shows how
    to perform more secure XML parsing by disabling the XML External Entity
    (XXE) Processing. See the accompanying README file for details.

Version 4.1.0

  * This version was never released. It was skipped due to a minor versioning
    scheme adjustment in the pre-release component when migrating to build2.

Version 4.0.0

  * Xerces-C++ 2-series (2.8.0 and earlier) is no longer supported.

  * Visual Studio 2003 (7.1) is no longer supported.

  * HP aCC3 (HP-UX/PA-RISC) is no longer supported.

  * Oracle/Berkeley DB XML support has been removed since it no longer
    supports the Xerces-C++-based document access.

  * New option, --std, specifies the C++ standard that the generated code
    should conform to. Valid values are c++98 (default) and c++11.

    The C++ standard affects various aspects of the generated code that are
    discussed in more detail in mapping-specific documentation (guides and
    manuals). Overall, when C++11 is selected, the generated code relies on
    the move semantics and uses std::unique_ptr instead of deprecated
    std::auto_ptr. See also the documentation for the --std option in the
    XSD compiler command line manual (man pages).

  * New option, --fat-type-file, triggers the placement of code corresponding
    to global elements into type files instead of schema files in the file-
    per-type mode. This option is primarily useful when trying to minimize
    the amount of object code that is linked to an executable by packaging
    compiled generated code into a static (archive) library.

C++/Tree

  * Support for ordered types. C++/Tree flattens nested compositors which
    sometimes can result in the loss of element ordering information that
    could be significant to the application. Ordered types address this
    problem. For more information, refer to Section 2.8.4, "Element Order"
    in the C++/Tree Mapping User Manual.

  * Support for mixed content in ordered types. For more information, refer
    to Section 2.13, "Mapping for Mixed Content Models" in the C++/Tree
    Mapping User Manual.

  * xml_schema::type represents anyType content as a DOM fragment, similar
    to wildcards. For more information, refer to Section 2.5.2, "Mapping
    for anyType" in the C++/Tree Mapping User Manual.

  * xml_schema::simple_type represents anySimpleType content as a text
    string. For more information, refer to Section 2.5.3, "Mapping for
    anySimpleType" in the C++/Tree Mapping User Manual.

  * Improved streaming example that now provides better XML namespace
    handling and supports streaming parsing and serialization at multiple
    document levels.

  * New option, --generate-dep, triggers the generation of the make
    dependency files (.d) for the generated C++ files. Other options
    controlling dependency generation are: --generate-dep-only,
    --dep-phony, --dep-target, --dep-suffix, and --dep-regex. For
    details on this functionality, refer to the XSD compiler command
    line manual (man pages).

  * New option, --suppress-assignment, suppresses the generation of copy
    assignment operators for complex types. If this option is specified,
    the copy assignment operators for such types are declared private and
    left unimplemented.

  * Binary representation now stores string-based enumerations as integer
    values corresponding to C++ enumerators instead of string literals.

  * Binary representation now pools polymorphic type-id strings in an
    implicit string pool. The string pool support can also be used to
    pool strings in other situations. For example, you can implement
    string insertion/extraction operators for your stream to pool all
    strings. This can be useful if your documents contain a large number
    of repetitive strings.

  * New option, --polymorphic-plate, allows the creation of multiple
    polymorphic map plates in the same application. For details, refer
    to the XSD compiler command line manual (man pages).

  * To get the DOM association in the copy of an object model tree one
    now needs to explicitly pass the xml_schema::flags::keep_dom flag as
    the second argument to the copy constructor or clone() function.

Version 3.3.0

  * New option, --char-encoding, allows you to specify the character encoding
    that should be used in the generated code. Valid values for the 'char'
    character type are 'utf8' (default), 'iso8859-1' (new), 'lcp' (Xerces-C++
    local code page), and 'custom' (provides support for custom encodings).
    Note that if you use a non-default character encoding and include some
    libxsd headers (e.g., xsd/cxx/xml/string.hxx) directly, then you will
    need to first include the correct xsd/cxx/xml/char-<enc>.hxx header,
    where <enc> is iso8859-1, lcp, etc. This mechanism replaces the
    XSD_USE_LCP macro.

    For the wchar_t character type the only valid value for this option is
    'auto' and the encoding is automatically selected between UTF-16 and
    UTF-32, depending on the wchar_t type size.

  * When the XSD compiler is built with Xerces-C++ 3.1.0 or later, the
    handling  of multiple imports for the same namespace is enabled. Before,
    all subsequent imports for a namespace were ignored which caused errors
    in some schemas. Note that if your application has XML Schema validation
    enabled, then you will also need to build it with Xerces-C++ 3.1.0 or
    later to take advantage of this feature.

  * Automatic mapping for the urn-style XML namespaces. The last component
    in the urn name is used to derive the C++ namespace name.

  * New option, --schema-file-regex, in combination with the existing
    --type-file-regex, can be used to place the generated files into
    subdirectories or to resolve file name conflicts in the file-per-
    type mode (--file-per-type).

  * Warning id's have changed to start with a letter identifying the
    component issuing the warning. F - compiler frontend, D - compiler
    driver, P - C++/Parser mapping, T - C++/Tree mapping.

  * Strings used to match regular expressions supplied with the
    --namespace-regex and --anonymous-regex options now include the file
    component for the schema being compiled.

  * The XSD_NO_EXPORT macro can be used to omit code generated with the
    --export/import-maps and, for C++/Tree, --generate-xml-schema options
    during C++ compilation. This may be useful if you would like to use
    the same generated code across multiple platforms.

 C++/Tree

  * New option, --generate-element-type, triggers the generation of types
    instead of parsing/serialization functions for root elements. This
    is primarily useful to distinguish object models with the same root
    type but with different root elements. For more information, refer
    to the messaging example and Section 2.9.1, "Element Types" in the
    C++/Tree Mapping User Manual. To support the customization of the
    element type naming the --element-type-regex option has been added.
    See the NAMING CONVENTION section in the compiler command line manual
    (man pages) for details.

  * New option, --generate-element-map, triggers the generation of a root
    element map. The element map allows uniform parsing and serialization
    of multiple root elements. This option can only be used together with
    --generate-element-type. For more information, refer to the messaging
    example and Section 2.9.2, "Element Map" in the C++/Tree Mapping
    User Manual.

  * Prior to this version, if the --generate-polymorphic option is
    specified, the compiler treats all types as potentially polymorphic.
    Now by default only type hierarchies used in substitution groups and
    those explicitly declared polymorphic with the new --polymorphic-type
    option are treated as polymorphic. This results in smaller and faster
    generated code. If you would like to continue using the old behavior,
    you will need to specify --polymorphic-type-all. For more information,
    on this change see Section 2.11, "Mapping for xsi:type and Substitution
    Groups" in the C++/Tree Mapping User Manual.

  * New option, --generate-detach, triggers the generation of detach
    functions for required elements and attributes. For optional and
    sequence cardinalities the detach functions are now provided by the
    respective containers even without this option. These functions, for
    example, allow one to move sub-trees in the object model either within
    the same tree or between different trees without copying. For more
    information, refer to Section 2.8 "Mapping for Local Elements and
    Attributes" in the C++/Tree Mapping User Manual.

  * New option, --export-xml-schema, causes the compiler to export/import
    types in the XML Schema namespace using the export symbol provided
    with the --export-symbol option.

  * New example, embedded, shows how to embed the binary representation of
    the schema grammar into an application and then use it to parse and
    validate XML documents.

  * New example, compression, shows how to compress an XML document during
    serialization and decompress it during parsing using the zlib library.

  * New example, custom/mixed, shows how to use type customization to parse
    and serialize mixed content.

  * The streaming example has been extended to show how to perform stream-
    oriented, partially in-memory XML processing using the C++/Tree mapping.
    With the partially in-memory parsing and serialization only a part of
    the object model is in memory at any given time. With this approach one
    can process parts of the document as they become available as well as
    handle documents that are too large to fit into memory.

  * New default/fixed value initialization code. Now the default/fixed values
    are parsed by the XSD compiler at compile time instead of the standard
    parsing code at runtime. This will allow the compilation of schemas that
    use the default/fixed values without support for XML parsing
    (--suppress-parsing option).

  * Empty XML Schema enumeration values are now mapped to the 'empty' C++
    enumerator name instead of 'cxx'.

  * XML Schema union types with members that are enumeration types are
    automatically converted to equivalent enumeration types with a union
    of all the members' enumerators.

Version 3.2.0

  * New option, --disable-warning, disables printing of a warning with
    the specified id. Specifying 'all' for the warning id disables all
    warnings.

  * New options, --export-maps and --import-maps, provide support for
    splitting a polymorphic type hierarchy across several Win32 DLLs.
    See the compiler command line manual (man pages) for details.

 C++/Tree

  * During serialization the generated code automatically assigns
    generic prefixes (p1, p2, etc) to XML namespaces used in the
    vocabulary and for which no custom prefix-namespace mapping
    was provided via the xml_schema::namespace_infomap argument.
    The xml_schema::namespace_infomap argument in the serialization
    functions is now default-initialized to an empty map. The
    xml_schema::no_namespace_mapping and xml_schema::xsi_already_in_use
    exceptions have been removed.

  * New example, performance, measures the performance of parsing and
    serialization. This example also shows how to structure your code
    to achieve the maximum performance for these two operations.

  * New example, xpath, shows how to use the C++/Tree mapping together
    with XPath.

  * New options, --one-accessor-regex, --opt-accessor-regex,
    --seq-accessor-regex, --one-modifier-regex, --opt-modifier-regex,
    and --seq-modifier-regex, allow specification of transformations
    for accessor and modifier function names for elements and attributes
    with specific cardinalities. For more information see the NAMING
    CONVENTION section in the compiler command line manual (man pages).

  * Support for comparison (--generate-comparison) and printing
    (--generate-ostream) of polymorphic object models.

  * New serialization flag, xml_schema::flags::dont_pretty_print,
    disables extra spaces and new lines that make the resulting XML
    slightly bigger but easier to read.

  * New example, custom/double, shows how to customize parsing and
    serialization code for the xsd:double XML Schema built-in type.
    It can be used as a guide on how to customize built-in XML Schema
    types that are mapped to fundamental C++ types.

  * Support for fractionDigits and totalDigits facets in serialization
    of types derived from xsd:decimal.

  * New set of compile-time macros that control how the xsd:float,
    xsd:double, and xsd:decimal types are serialized. The following
    macros control the format:

    XSD_CXX_TREE_FLOAT_FIXED
    XSD_CXX_TREE_FLOAT_SCIENTIFIC
    XSD_CXX_TREE_DOUBLE_FIXED
    XSD_CXX_TREE_DOUBLE_SCIENTIFIC

    The following macros control the precision:

    XSD_CXX_TREE_FLOAT_PRECISION_MAX
    XSD_CXX_TREE_FLOAT_PRECISION
    XSD_CXX_TREE_DOUBLE_PRECISION_MAX
    XSD_CXX_TREE_DOUBLE_PRECISION
    XSD_CXX_TREE_DECIMAL_PRECISION_MAX
    XSD_CXX_TREE_DECIMAL_PRECISION

    If the *_PRECISION_MAX macro is defined then the maximum number of
    potentially significant decimal digits that the type can represent
    is used. Otherwise, if the *_PRECISION macro is defined then its
    value is used. By default the precision is set to the number of
    decimal digits that the type can represent without change. For
    more information on these options, refer to the following paper:

    http://www.open-std.org/JTC1/SC22/WG21/docs/papers/2005/n1822.pdf

    The old macro, XSD_FP_ALL_DIGITS, that was equivalent to defining
    all three *_PRECISION_MAX macros has been removed.

    An alternative to using these macros is to customize the floating
    point type as shown in the custom/double example.

  * An additional constructor is generated in situations where a type
    contains one or more required element of complex type (that is,
    it itself contains elements or attributes). In this constructor,
    initializers for such elements are passed as std::auto_ptr and the
    newly created instance is directly initialized with and assumes
    ownership of the pointed to objects. This constructor is a logical
    addition to the non-copying modifiers that were introduced in the
    previous version.

  * Extra conversion operators in the fundamental_base class template
    which is used to emulate inheritance from fundamental types are now
    disabled by default since they cause problems on several compilers.
    To enable them compile your code with the XSD_TREE_EXTRA_FUND_CONV
    macro defined.

 C++/Parser

  * New options, --generate-xml-schema and --extern-xml-schema, trigger
    generation of the mapping for the XML Schema namespace to a separate
    header file and inclusion of that header into other generated header
    files instead of generating the necessary declarations inline,
    respectively. See the compiler command line manual (man pages) for
    details.

  * New example, performance, measures the performance of XML parsing.
    This example also shows how to structure your code to achieve the
    maximum performance for this operation.

  * Type map files can now include comments. A comment starts with #
    and ends with a new line or end of file. To specify a name that
    contains # enclose it in "".

  * In type map files the optional argument type now defaults to the
    return type if the return type ends with * or & (that is, it is
    a pointer or a reference) and 'const return type&' otherwise.

  * The interface for polymorphic parsing has been simplified. Calling the
    *_parser() functions multiple times to specify several parsers is no
    longer supported. Instead you need to pass the xml_schema::parser_map
    object which contains the parsers. For more information refer to
    Section 5.4, "Support for Polymorphism" in the C++/Parser Mapping
    Getting Started Guide.

  * The use of virtual inheritance has been reduced which results in a
    much smaller object code size (more than factor of 2 on some tests)
    and faster C++ compilation with less RAM used.

  * The low-level Expat-specific parsing API (parse_begin() and parse_end())
    has been extended to provide XML and XML Schema error translation to
    exceptions or error handler calls. See Section 7.2, "Expat Document
    Parser" in the C++/Parser Mapping Getting Started Guide for more
    information.

Version 3.1.0

  * New option, --file-per-type, triggers generation of a separate set
    of C++ files for each type defined in XML Schema. This compilation
    mode is primarily useful when some of your schemas cannot be compiled
    separately or have cyclic dependencies which involve inheritance.
    Other new options that are useful in this compilation mode are
    --type-file-regex, --type-file-regex-trace, and --file-list. See the
    compiler command line manual (man pages) for more information.

  * New option, --options-file, allows additional command line options
    to be provided in files, with one option per line.

  * New option, --reserved-name, allows insertion of additional names
    with optional replacements to the list of names that should not be
    used as identifiers. See the compiler command line manual (man pages)
    for details.

  * New options, --location-map, --location-regex, and
    --location-regex-trace, allow re-mapping of schema locations
    specified in the include and import elements without modifying the
    schema files. See the compiler command line manual (man pages) for
    more information.

  * New option, --guard-prefix, allows specification of a prefix that
    should be added to generated header inclusion guards.

  * New option, --file-list, triggers creation of a file with a list of
    generated C++ files. This option is primarily useful in the file-per-
    type compilation mode (--file-per-type) to create a list of generated
    C++ files, for example, as a makefile fragment. Other new options
    that are useful with --file-list are --file-list-prologue,
    --file-list-epilogue, and --file-list-delim. See the compiler command
    line manual (man pages) for more information.

  * Support for the upcoming Xerces-C++ 3.0.0 release.

 C++/Tree

  * New option, --generate-intellisense, triggers generation of workarounds
    for IntelliSense bugs in Visual Studio 2005 (8.0). When this option is
    used, the resulting code is slightly more verbose. IntelliSense in
    Visual Studio 2008 (9.0) does not require these workarounds. Support
    for IntelliSense in Visual Studio 2003 (7.1) is improved with this
    option but is still incomplete.

  * New options, --type-naming and --function-naming, allow specification
    of the type and function naming conventions that should be used in the
    generated code. Supported values for --type-naming are: knr (K&R), ucc
    (upper-camel-case), and java. Supported values for --function-naming
    are: knr (K&R), lcc (lower-camel-case), and java. For more information
    see the NAMING CONVENTION section in the compiler command line manual
    (man pages).

  * New options, --type-regex, --accessor-regex, --modifier-regex,
    --parser-regex, --serializer-regex, and --enumerator-regex, allow
    specification of transformations for type, accessor function,
    modifier function, parsing function, serialization function, and
    enumerator names in order to produce the generated code using a
    custom naming convention. For more information see the NAMING
    CONVENTION section in the compiler command line manual (man pages).

  * Generated list classes now provide a complete set of constructors and
    conform to the standard C++ sequence interface.

  * String-based types now provide two extra constructors that expect a
    C string and std::string as their arguments. This allows direct
    initialization of string-based types from string literals.

  * New implementations of the XML Schema date/time types (date, dateTime,
    duration, gDay, gMonth, gMonthDay, gYear, gYearMonth, and time) that
    represent the information in the numerical form.

  * New binary serialization examples: binary/boost, which shows how to
    save/load the object model to/from a custom format using the Boost
    serialization library as an example, and binary/xdr, which shows how to
    save/load the object model to/from XDR (eXternal Data Representation)
    binary format using the XDR API provided as part of Sun RPC.

  * The non-copying modifier functions can now be used to assemble object
    models from scratch. For more information see Section 4.4, "Creating
    the Object Model from Scratch" in the C++/Tree Mapping Getting Started
    Guide as well as Section 2.8, "Mapping for Local Elements and Attributes"
    in the C++/Tree Mapping User Manual.

  * Doxygen documentation was added to the XSD runtime for the built-in XML
    Schema types, exceptions, etc. This allows linking of the generated
    documentation to the XSD runtime documentation using the Doxygen tags
    mechanism. The Doxygen configuration file for the XSD runtime is
    provided in the documentation/cxx/tree/reference/ directory.

  * Support for customization of anyType. Because anyType is a base type
    for every generated type, customizing it allows one to implement custom
    functionality that spans the entire type system. See the comments
    example in the examples/cxx/tree/custom/ directory.

  * New option, --omit-default-attributes, triggers generation of extra
    checks that exclude attributes with default and fixed values from the
    serialized XML documents.

  * The parsing functions that used to read from DOMInputSource were changed
    to use InputSource to ease support of Xerces-C++ 3 and 2 series in the
    same code base.

  * The parsing function that used to parse DOMDocument* was changed to
    parse xml_schema::dom::auto_ptr<DOMDocument>& instead. If the keep_dom
    and own_dom flags are specified then this parsing function resets the
    passed automatic pointer and the returned object model assumes
    ownership of the DOM document. xml_schema::dom::auto_ptr is a simple
    automatic pointer for Xerces-C++ DOM with the same interface as
    std::auto_ptr.

  * The xml_schema::tree_node_key DOM user data key was moved to
    xml_schema::dom::tree_node_key.

 C++/Parser

  * New option, --generate-polymorphic, triggers generation of polymorphism-
    aware code. This option should be used on XML vocabularies which use
    xsi:type and/or substitution groups. For more information see Section
    5.4, "Support for Polymorphism" in the C++/Parser Mapping Getting
    Started Guide we well as the polymorphism and polyroot examples in the
    examples/cxx/parser/ directory.

  * The date/time types (date, dateTime, gDay, gMonth, gMonthDay, gYear,
    gYearMonth, and time) now represent time zone in the numerical form.

  * In order to support parsing of polymorphic XML documents, the signatures
    of the start_* functions (_start_element, _start_any_element, and
    start_root_element) have changed to include a third argument of type
    const ro_string<C>*. This argument contains the resolved type name and
    namespace in case the xsi:type attribute was specified.

Version 3.0.0

  * Anonymous type morphing (automatic type naming) is now performed by
    default in both mappings. The --morph-anonymous option does not have
    any effect but is preserved for backwards compatibility. A new option,
    --preserve-anonymous, disables anonymous type morphing. This option is
    useful together with --show-anonymous if you want to make sure your
    schemas do not have any anonymous types.

  * A number of bugs fixed in both C++/Tree and C++/Parser mappings.

 C++/Tree

  * The new C++/Tree Mapping Getting Started Guide is available in the
    documentation/cxx/tree/guide/ directory.

  * The type definitions for local elements and attributes in the form
    name::type have been changed to name_type. For example, an element
    bar in type foo with maxOccurs="unbounded" used to have its iterator
    type defined as foo::bar::iterator. With this change it becomes
    foo::bar_iterator. Furthermore, the container type name for sequence
    elements has changed from foo::bar::container to foo::bar_sequence
    and for optional elements and attributes from foo::bar::container
    to foo::bar_optional. This is a backwards incompatible change and
    may require application code adjustments (the C++ compiler will
    pinpoint the affected places).

  * New option, --generate-doxygen, triggers generation of documentation
    comments suitable for extraction by the Doxygen documentation system.
    Documentation from annotations is added to the comments if present in
    the schema.

  * New option, --generate-wildcard, triggers generation of the new
    wildcard (any and anyAttribute) mapping. This mapping represents the
    content matched by wildcards as DOM fragments. For more information on
    the new mapping see Section 2.12, "Mapping for any and anyAttribute"
    in the C++/Tree Mapping User Manual as well as the wildcard example in
    the examples/cxx/tree/ directory.

  * New option, --generate-comparison, triggers generation of comparison
    operators (== and !=) for complex types. Comparison is performed
    memberwise.

  * Support for the RPC XDR binary stream in addition to ACE CDR.

  * New constructor is generated for complex types with ultimate bases
    that are simple types and can be default-initialized. This constructor
    includes initializers for all required members but omits the initializer
    for the base type. See Section 2.7, "Mapping for Complex Types" in the
    C++/Tree Mapping User Manual for more information.

  * Support for polymorphic binary serialization and extraction. Note that
    the semantics of the --generate-insertion and --generate-extraction
    options has changed. See the the compiler command line manual (man
    pages) for details.

  * New parsing function with the DOMDocument* argument and the own_dom
    flag allow the tree to assume the ownership of the DOM document
    being parsed when DOM association is requested (keep_dom flag).
    See the C++/Tree Mapping User Manual for more information.

  * New example, multiroot, shows how to handle XML vocabularies with
    multiple root elements.

  * New example, caching, shows how to parse several XML documents while
    reusing the underlying XML parser and caching the schemas used for
    validation.

  * The mapping of built-in XML Schema type decimal has changed from
    long double to double. The old mapping can be obtained by providing
    a custom mapping for this type.

  * The xml_schema::errors type which is used in the xml_schema::parsing
    and xml_schema::serialization exceptions has been renamed to
    xml_schema::diagnostics and extended to include warnings in addition
    to errors.

  * Serialization operators now clear the element being serialized to from
    existing child nodes and attributes (except for special attributes such
    as prefix-namespace mappings, etc.).

  * Improved built-in type parsing, including support for normalization and
    whitespace collapsing.

  * Optimizations for the generated code size and compilation time,
    including space optimizations for polymorphic parsing and
    serialization. Optimizations for XML parsing speed.

 C++/Parser

  * The C++/Parser mapping have been significantly redesigned. See the new
    Getting Started Guide in documentation/cxx/parser/guide/ for details.

  * The new C++/Parser Mapping Getting Started Guide is available in the
    documentation/cxx/parser/guide/ directory.

  * The mapping now provides parser implementations for all built-in XML
    Schema types. See Chapter 6, "Built-In XML Schema Type Parsers" in
    the C++/Parser Mapping Getting Started Guide for more information.

  * The mapping now supports automatic generation of sample parser
    implementations and a test driver. The --generate-noop-impl option
    triggers generation of a sample implementation with empty function
    bodies. The --generate-print-impl option triggers generation of a
    sample implementation that prints the data stored in XML to STDOUT.
    The --generate-test-driver option trigger generation of a test driver.
    For more information on this feature see the compiler command line
    manual (man pages) and the generated example in the examples/cxx/parser/
    directory. Other relevant options include: --force-overwrite,
    --root-element-first, --root-element-last, and --root-element.

  * New example, wildcard, shows how to parse the XML data matched by
    XML Schema wildcards (any and anyAttribute).

  * The xml_schema::document parser has been extended with overridable
    virtual functions start_root_element and end_root_element to support
    parsing of XML vocabularies with multiple document roots. See the
    multiroot example in the examples/cxx/parser/ directory for more
    information.

  * The xml_schema::errors type which is used in the xml_schema::parsing
    exception has been renamed to xml_schema::diagnostics and extended to
    include warnings in addition to errors.

Version 2.3.1

  * The compiler is now capable of translating multiple schemas with
    one invocation.

  * New option, --sloc-limit, allows one to limit the amount of the
    generated code.

  * New option, --proprietary-license, instructs the compiler not to
    include the GPL banner in each generated file. Instead a short
    notice about a required proprietary license is generated. You
    should not use this option unless you have obtained a proprietary
    license from Code Synthesis.

  * The default encoding for the 'char' character type is now UTF-8.
    To get the previous behavior (local code page via the Xerces-C++
    transcode functions) define the XSD_USE_LCP preprocessor macro
    when compiling your source code.

  C++/Tree

    * The --parts option has been improved to split generated code more
      evenly by analyzing the complexity of the generated schema constructs.

    * Ability to customize serialization, std::ostream, and binary
      insertion/extraction operators. See examples/cxx/tree/custom/wildcard
      for an example on how to handle XML Schema wildcards (xsd:any and
      xsd:anyAttribute) by customizing the parsing constructor and
      serialization operators.

    * Optimizations for the run-time memory consumption.

    * Optimizations for space in the generated code.

    * Number of bug fixes.

  C++/Parser

    * Proper handling of an xsd:any nested content. Nested elements,
      attributes, and text are reported via _any_* hooks of the current
      parser.

    * Number of bug fixes, mostly in the generated validation code.


Version 2.3.0

    * Name conflicts across type inheritance hierarchies are now detected
      and resolved via name escaping.

  C++/Tree

    * New option, --suppress-parsing, suppresses generation of the parsing
      constructors and functions. This can be used to minimize the generated
      code footprint when parsing from XML is not used.

    * New option, --generate-forward, triggers generation of a forward
      declaration header file for types defined in the schema. A set of
      --fwd-* options that control the resulting file name as well as
      prologue and epilogue code are available.

    * New option, --generate-xml-schema, triggers generation of the mapping
      for the XML Schema namespace to a separate header file. See the man
      pages for details and examples/cxx/tree/custom/calendar for an example.

    * New option, --extern-xml-schema, triggers inclusion of a header
      file for the XML Schema namespace instead of generating the
      necessary declarations inline. See the man pages for details and
      examples/cxx/tree/custom/calendar for an example.

    * New options, --custom-type and --custom-type-regex, instruct the
      compiler to use custom C++ type for a type defined in the schema.
      The standard mapping can still be generated (with a different name)
      usually to be used as a base. Built-in XML Schema types can be
      customized using this mechanism. See the man pages for details and
      examples/cxx/tree/custom/* for examples.

    * The generated parsing constructors and serialization operators have
      been changed to use the Xerces-C++ DOM elements and attributes
      instead of the internal wrapper types. This should provide easier
      integration with other code and libraries that use the Xerces-C++
      DOM types such as Berkeley DB XML.

    * New example, examples/cxx/tree/dbxml, shows how to use the C++/Tree
      mapping on top of the Berkeley DB XML database.

  C++/Parser

    * Validation of the attribute structure in the generated code.

    * Validation of the character content models including mixed content in
      the generated code.

    * Validation of the built-in XML Schema types.

    * Optimizations for space and time in the generated code. In particular
      data coping during parsing and validation was significantly reduced.


Version 2.2.0

    * Detection of a version mismatch between the generated code and
      the runtime.

  C++/Tree

    * Escaping of a global element name that conflicts with a global type
      name. This is a backwards-incompatible change. Previous versions
      map them to the same name.

    * New options, --generate--insertion and --generate-extraction,
      trigger generation of (binary) data representation stream
      insertion and extraction operators, respectively. This allows
      one to serialize/deserialize in-memory representation to/from
      data representation streams such as XSD, CDR, etc. ACE CDR
      streams are supported out of the box (see the binary example).
      User-supplied streams can be used via an adaptation layer.

    * New serialization flag, no_xml_declaration, instructs the XML
      serialization functions to omit an XML declaration. This is useful
      for streaming serialization (see the streaming example).

    * Optimizations to reduce generated code size.


  C++/Parser

    * New options, --generate-validation and --suppress-validation,
      trigger and suppress generation of the validation code,
      respectively. The validation code is the implementation of the
      XML Schema validation in the generated code (also known as
      "perfect" parser). In this version validation of the element
      structure has been implemented.

    * New architecture for underlying XML parsers. This is a backwards-
      incompatible change. Existing applications will have to be
      modified. See examples for details.


Version 2.1.1

  C++/Tree

    * New option, --namespace-map, allows direct mapping of XML Schema
      namespaces to C++ namespaces without the use of regular expressions.

    * Further optimizations in the container code and enum mapping to
      reduce generated code size.

    * Number of bug fixes in the generated code.


  C++/Parser

    * New option, --namespace-map, allows direct mapping of XML Schema
      namespaces to C++ namespaces without the use of regular expressions.


Version 2.1.0

  * Automatic handling of forward inheritance. XML Schema allows
    inheritance from yet undefined types while it is illegal to do
    so in C++. Now the translator automatically handles forward
    inheritance by re-arranging the schema during compilation.


  C++/Tree

    * New enum mapping with support for inheritance. Enumerators are
      now parsed using binary search instead of linear search.

    * Associated DOM nodes now retain "back" pointers to tree nodes.

    * Optimizations to reduce generated code size.


  C++/Parser

    * Specialization for void. You can now use void as a hook argument
      type if you don't want to pass any data between parsers.

    * Support for re-use of implementations of base parsers in derived
      parsers using the mixin C++ idiom. See the examples/cxx/parser/mixin
      for more information.

    * Support for uninitialized parser. If you don't provide a parser
      for element/attribute, that element/attribute will be ignored
      during parsing.


Version 2.0.0

  * New cardinality calculator. This improves support for schemas that
    use complex structures with repeated elements, e.g.,

    <complexType name="Type">
      <choice>
        <sequence>
          <element name="a" type="string"/>
          <element name="c" type="string"/>
        </sequence>
        <sequence>
          <element name="b" type="string"/>
          <element name="c" type="string"/>
        </sequence>
      </choice>
    </complexType>


  * New identifier escaping code. With this feature xsd generates proper
    code for schemas that use the same name for an element and an attribute
    in the same type or use several elements/attributes with different
    qualified names but with the same local name, e.g.,

    <!-- base.xsd -->
    <schema xmlns="http://codesynthesis.com/xmlns/test/foo"
            targetNamespace="http://codesynthesis.com/xmlns/test/foo">

      <element name="foo" type="int"/>
    </schema>

    <schema xmlns="http://codesynthesis.com/xmlns/test/bar"
	    xmlns:f="http://codesynthesis.com/xmlns/test/foo"
            targetNamespace="http://codesynthesis.com/xmlns/test/bar">

      <import namespace="http://codesynthesis.com/xmlns/test/foo"
              schemaLocation="base.xsd"/>

      <element name="foo" type="string"/>

      <complexType name="Foo">
        <sequence>
          <element ref="foo"/>
          <element name="foo" type="long"/>
          <element ref="f:foo"/>
          <element ref="f:foo"/>
        </sequence>
        <attribute name="foo" type="string"/>
      </complexType>
    </schema>


  C++/Tree

    * New option, --generate-polymorphic, triggers generation of
      polymorphism-aware code. Before this release xsd used to always
      generate polymorphism-aware code. However, it appears to be quite
      wasteful in terms of the generated code size (up to 40%). You will
      now need to explicitly specify this option if you use substitution
      groups or xsi:type. A warning is issued if this option is not
      specified but the schema makes use of substitution groups.

    * New options, --root-element-first, --root-element-last,
      --root-element-all, --root-element-none, and --root-element, control
      generation of parsing and serialization functions. With these options
      you can avoid generating extra code for global elements that are not
      document roots. See the man pages for details.

    * New options, --parts and -parts-suffix, allows you to split generated
      source code into a number of parts. This is useful when translating
      large, monolithic schemas and a C++ compiler is not able to compile
      the resulting source code at once (usually due to insufficient memory).

    * New option, --generate-default-ctor, triggers generation of default
      constructors even for types that have required members. Required
      members of an instance constructed using such a constructor are not
      initialized and accessing them results in undefined behavior. Thanks
      to Jean-Francois Dube <jf at magnu.polymtl.ca> for suggesting this
      feature.

    * New option, --generate-from-base-ctor, triggers generation of
      constructors that expect an instance of a base type followed by all
      required members. Thanks to Jean-Francois Dube <jf at magnu.polymtl.ca>
      for suggesting this feature.

    * Information scopes for attributes and elements with default/fixed values
      now define the public static default_value function which allows one to
      obtain the default/fixed value for the element/attribute. Thanks to
      Dave Moss <david.r.moss at selex-comm.com> for suggesting this feature.

    * MSVC 7.1 has a limit on the length of the "if else if" chain. This
      results in ICE when compiling generated code for enumerations with
      a large number of values. This version addresses this issue. Thanks
      to Cyrille Chépélov <cyrille at chepelov.org> for reporting this and
      suggesting a fix.


  C++/Parser

    * The parser construction API has changed. Now, for element 'foo',
      the name of the parser modifier function is 'foo_parser'. Likewise,
      operator() for setting all parsers at once has been changed to the
      'parsers' function.


Version 1.9.0

  C++/Tree

    * The size modifier function in the base64_binary and hex_binary
      built-in types automatically adjusts capacity if needed.

    * More internal names (names that start with _xsd_) were made
      private or protected.

  C++/Parser

    * Typedef for the parser base in the xml_schema namespace.

  C++/Parser-E

    * C++/Parser mapping optimized for embedded systems. For now it
      is equivalent to 'cxx-parser --xml-parser expat'.


Version 1.8.0

  * Moved to the build 0.2 series.

  C++/Tree

    * Support for default and fixed values in attributes. An optional
      attribute with a default or fixed value is mapped to the One
      cardinality class instead of the Optional cardinality class.

    * Mapping for base64Binary and hexBinary has improved. Now these
      types support a basic buffer abstraction and perform automatic
      encoding and decoding.

    * Internal names are protected. We've noticed (via bug reports) a
      wide use of internal names (names that start with _xsd_) in user
      code. This is not portable and instead you should use public
      names. To prevent this from happening in the future we've made
      all internal names protected.

  C++/Parser

    * Support for Expat as the underlying XML parser in addition to
      Xerces-C++. This allows one to use the C++/Parser mapping in
      memory-constrained environments such as embedded systems. To
      select Expat instead of Xerces-C++ (default) add
      '--xml-parser expat' to the command line. At the moment only
      'char' (UTF-8) is supported as the base character type when
      Expat is selected.

    * The invalid_instance exception has been renamed to parsing.

    * Generic error_handler interface has been added in addition
      to Xerces-C++-specific DOMErrorHandler. It allows you to
      handle parsing errors and warnings without having to deal
      with Xerces-C++ specifics.

    * The default error handling behavior has changed in parsing
      functions. Instead of printing errors and warnings to STDERR,
      the errors are now collected and thrown as part of the parsing
      exception.

    * In parsing functions, the name, namespace arguments order has
      been reversed to be consistent with the one used in parsing
      hooks.

Version 1.7.0

  * Number of bug fixes in libxsd and the generated code.

  C++/Tree

    * Comprehensive XML Schema C++/Tree Mapping User Manual.

    * Basic support for union. A simple type that is defined using
      derivation by union is mapped to a C++ class that derives from
      string.

    * The _clone function has its arguments default-initialized.

    * The invalid_instance exception has been renamed to parsing.

    * Generic error_handler interface has been added in addition
      to Xerces-C++-specific DOMErrorHandler. It allows you to
      handle parsing/serialization errors and warnings without
      having to deal with Xerces-C++ specifics. See the user
      manual for more information.

    * The default error handling behavior has changed in parsing
      and serialization functions. Instead of printing errors and
      warnings to STDERR, the errors are now collected and thrown
      as part of the parsing/serialization exception. See the user
      manual for more information.

    * The optional and sequence containers now support operators ==,
      !=, <, >, <=, and >=.

    * Flags argument has been added to serialization functions. The
      only flag that is currently supported is dont_initialize.

    * Generated code cleanups.

  C++/Parser

    * Basic support for union. A simple type that is defined using
      derivation by union is mapped to a C++ class template that
      is just an alias for the generic parser. You are expected to
      override the _characters function in your implementation.

    * Properties argument to parsing functions which allows to
      programmatically specify schemas for instance document
      validation.

    * Flags argument to parsing functions. The following flags
      are supported:

      dont_validate   - do not validate instance documents
      dont_initialize - do not initialize the Xerces-C++ runtime

Version 1.6.0

  * Number of bug fixes in libxsd and the generated code.

  C++/Tree

    * Support for xsi:type and substitution groups in parsing and
      serialization. See examples/cxx/tree/polymorphism for a code
      sample.

    * Properties argument to parsing functions which allows to
      programmatically specify schemas for instance document
      validation.

    * Extra checks in parsing code which prevents construction
      of inconsistent in-memory representation from invalid
      instance documents. Should be useful when validation is
      disabled.

    * Accessors and modifier were made normal member functions.
      Before they were implemented via functors.

    * Workaround for g++-3.3 bug# 16650:

      http://gcc.gnu.org/bugzilla/show_bug.cgi?id=16650

  C++/Parser

    * All "service" functions were renamed to start with '_'.
      This should prevent hiding of service functions by
      elements/attributes with the same names.

Version 1.5.0

  * Number of bug fixes in libxsd and the generated code.

  C++/Tree

    * Basic support for inheritance-by-restriction in complex types.

    * The following parsing flags have been introduced:

      keep_dom        - keep association with underlying DOM nodes
      dont_validate   - do not validate instance documents
      dont_initialize - do not initialize the Xerces-C++ runtime

    * "Type-less content" such as mixed content models, xsd:anyType/
      xsd:anySimpleType, and xsd:any/xsd:anyAttribute is supported by
      exposing corresponding DOM nodes (see the keep_dom parsing flag).
      Note that only a subset of XML Schema xsd:any functionality is
      supported. The compiler will issue diagnostics for unsupported
      cases. See examples/cxx/tree/mixed for a code sample.

  C++/Parser

    * Support for inheritance-by-restriction in complex types.

    * "Type-less content" such as mixed content models, xsd:anyType/
      xsd:anySimpleType, and xsd:any/xsd:anyAttribute is supported
      by forwarding parsing events to a set of "unexpected" hooks.
      Note that only a subset of XML Schema xsd:any functionality is
      supported. The compiler will issue diagnostics for unsupported
      cases. See examples/cxx/parser/mixed for a code sample.

Version 1.4.0

  * Number of improvements and bug fixes in the diagnostics code.

  * libxsd has been reorganized to provide a clean split of code with
    regards to char/wchar_t use. It should be possible to use libxsd
    and the xsd-generated code on platforms that lack wchar_t support,
    such as mingw.

  C++/Tree

    * Work around for g++ bug# 23206.

    * Support for xsd:list.

    * Type/member name conflicts are auto-resolved. Such conflicts
      occur when a type and an element or attribute withing this type
      share the same name.

    * XML Schema extension, the 'refType' attribute, allows one to
      specify referenced type for xsd:IDREF and xsd:IDREFS data types.
      See examples/cxx/tree/library for details.

    * New option, --morph-anonymous, allows automatic morphing
      of anonymous types to named ones. See the man pages for
      details.

    * New option, --namespace-regex-trace, allows one to trace the
      namespace mapping process. See the man pages for details.

    * Mapping for optional elements/attributes (cardinality 0..1)
      has changed in a backwards-incompatible way. In the previous
      version you would write:

      Bar& bar = ...

      if (bar.foo.present ())  // test
      {
        Foo& foo (bar.foo ()); // get

        bar.foo (Foo (...));   // set

        bar.foo.reset ();      // reset
      }

      Now you would write it like this:

      if (bar.foo ().present ())      // test
      {
        Foo& foo (bar.foo ().get ()); // get

        bar.foo (Foo (...));          // set

        bar.foo ().reset ();          // reset
      }

      Or using the pointer notation:

      if (bar.foo ())           // test
      {
        Foo& foo (*bar.foo ()); // get

        bar.foo (Foo (...));    // set

        bar.foo ().reset ();    // reset
      }

  C++/Parser

    * Support for xsd:list.

    * Type/member name conflicts are auto-resolved. Such conflicts
      occur when a type and an element or attribute withing this type
      share the same name.

    * New option, --namespace-regex-trace, allows one to trace the
      namespace mapping process. See the man pages for details.

Version 1.3.0

  * Numerous bug fixes.

  * The XML subsystem of libxsd has been reorganized to provide
    a clean split of DOM and SAX functionalities.

  C++/Parser

    * New option, --morph-anonymous, allows automatic morphing
      of anonymous types to named ones. See the man pages for
      details.

  C++/Tree

    * Additional parser functions provide support for reading
      from std::istream.

Version 1.2.0

  C++/Parser

    * New backend that generates the C++/Parser mapping.

Version 1.1.1

  all backends

    * Bug fixes in the filesystem path handling logic.

Version 1.1.0

  C++/Tree

    * New option, --generate-serialization, triggers generation of
      serialization functions. Serialization functions convert an in-memory
      representation back to XML.

    * xsd::cxx::tree::vector has been extended to closely follow std::vector
      API. This allows you to access and modify element sequences as if they
      were of type std::vector.

    * Generated constructors from xml::attribute and xml::element are made
      explicit.

    * The library example was extended to showcase modification and
      serialization of the in-memory representation.

    * New "XML Schema C++/Tree Mapping Serialization Guide" has an in-depth
      treatment of the serialization mechanisms provided by xsd.

Version 1.0.1

  all backends

    * Improved diagnostics.

    * Bug fixes in the schema inclusion/importing logic.

  C++/Tree

    * Two new options: --include-with-brackets and --include-prefix

Version 1.0.0

  * First public release.