NAMING CONVENTION

The compiler can be instructed to use a particular naming convention in the generated code. A number of widely-used conventions can be selected using the --type-naming and --function-naming options. A custom naming convention can be achieved using the --type-regex, --accessor-regex, --one-accessor-regex, --opt-accessor-regex, --seq-accessor-regex, --modifier-regex, --one-modifier-regex, --opt-modifier-regex, --seq-modifier-regex, --parser-regex, --serializer-regex, --enumerator-regex, and --element-type-regex options.

The --type-naming option specifies the convention that should be used for naming C++ types. Possible values for this option are knr (default), ucc, and java. The knr value (stands for K&R) signifies the standard, lower-case naming convention with the underscore used as a word delimiter, for example: foo, foo_bar. The ucc (stands for upper-camel-case) and java values a synonyms for the same naming convention where the first letter of each word in the name is capitalized, for example: Foo, FooBar.

Similarly, the --function-naming option specifies the convention that should be used for naming C++ functions. Possible values for this option are knr (default), lcc, and java. The knr value (stands for K&R) signifies the standard, lower-case naming convention with the underscore used as a word delimiter, for example: foo(), foo_bar(). The lcc value (stands for lower-camel-case) signifies a naming convention where the first letter of each word except the first is capitalized, for example: foo(), fooBar(). The java naming convention is similar to the lower-camel-case one except that accessor functions are prefixed with get, modifier functions are prefixed with set, parsing functions are prefixed with parse, and serialization functions are prefixed with serialize, for example: getFoo(), setFooBar(), parseRoot(), serializeRoot().

Note that the naming conventions specified with the --type-naming and --function-naming options perform only limited transformations on the names that come from the schema in the form of type, attribute, and element names. In other words, to get consistent results, your schemas should follow a similar naming convention as the one you would like to have in the generated code. Alternatively, you can use the --*-regex options (discussed below) to perform further transformations on the names that come from the schema.

The --type-regex, --accessor-regex, --one-accessor-regex, --opt-accessor-regex, --seq-accessor-regex, --modifier-regex, --one-modifier-regex, --opt-modifier-regex, --seq-modifier-regex, --parser-regex, --serializer-regex, --enumerator-regex, and --element-type-regex options allow you to specify extra regular expressions for each name category in addition to the predefined set that is added depending on the --type-naming and --function-naming options. Expressions that are provided with the --*-regex options are evaluated prior to any predefined expressions. This allows you to selectively override some or all of the predefined transformations. When debugging your own expressions, it is often useful to see which expressions match which names. The --name-regex-trace option allows you to trace the process of applying regular expressions to names.

The value for the --*-regex options should be a perl-like regular expression in the form /pattern/replacement/. Any character can be used as a delimiter instead of /. Escaping of the delimiter character in pattern or replacement is not supported. All the regular expressions for each category are pushed into a category-specific stack with the last specified expression considered first. The first match that succeeds is used. For the --one-accessor-regex (accessors with cardinality one), --opt-accessor-regex (accessors with cardinality optional), and --seq-accessor-regex (accessors with cardinality sequence) categories the --accessor-regex expressions are used as a fallback. For the --one-modifier-regex, --opt-modifier-regex, and --seq-modifier-regex categories the --modifier-regex expressions are used as a fallback. For the --element-type-regex category the --type-regex expressions are used as a fallback.

The type name expressions (--type-regex) are evaluated on the name string that has the following format:

[namespace ]name[,name][,name][,name]

The element type name expressions (--element-type-regex), effective only when the --generate-element-type option is specified, are evaluated on the name string that has the following format:

namespace name

In the type name format the namespace part followed by a space is only present for global type names. For global types and elements defined in schemas without a target namespace, the namespace part is empty but the space is still present. In the type name format after the initial name component, up to three additional name components can be present, separated by commas. For example:

http://example.com/hello type

foo

foo,iterator

foo,const,iterator

The following set of predefined regular expressions is used to transform type names when the upper-camel-case naming convention is selected:

/(?:[^ ]* )?([^,]+)/\u$1/

/(?:[^ ]* )?([^,]+),([^,]+)/\u$1\u$2/

/(?:[^ ]* )?([^,]+),([^,]+),([^,]+)/\u$1\u$2\u$3/

/(?:[^ ]* )?([^,]+),([^,]+),([^,]+),([^,]+)/\u$1\u$2\u$3\u$4/

The accessor and modifier expressions (--*accessor-regex and --*modifier-regex) are evaluated on the name string that has the following format:

name[,name][,name]

After the initial name component, up to two additional name components can be present, separated by commas. For example:

foo

dom,document

foo,default,value

The following set of predefined regular expressions is used to transform accessor names when the java naming convention is selected:

/([^,]+)/get\u$1/

/([^,]+),([^,]+)/get\u$1\u$2/

/([^,]+),([^,]+),([^,]+)/get\u$1\u$2\u$3/

For the parser, serializer, and enumerator categories, the corresponding regular expressions are evaluated on local names of elements and on enumeration values, respectively. For example, the following predefined regular expression is used to transform parsing function names when the java naming convention is selected:

/(.+)/parse\u$1/

See also the REGEX AND SHELL QUOTING section below.

TYPE MAP

Type map files are used in C++/Parser to define a mapping between XML Schema and C++ types. The compiler uses this information to determine the return types of post_* functions in parser skeletons corresponding to XML Schema types as well as argument types for callbacks corresponding to elements and attributes of these types.

The compiler has a set of predefined mapping rules that map built-in XML Schema types to suitable C++ types (discussed below) and all other types to void. By providing your own type maps you can override these predefined rules. The format of the type map file is presented below:

namespace <schema-namespace> [<cxx-namespace>]
{
  (include <file-name>;)*
  ([type] <schema-type> <cxx-ret-type> [<cxx-arg-type>];)*
}
  

Both <schema-namespace> and <schema-type> are regex patterns while <cxx-namespace>, <cxx-ret-type>, and <cxx-arg-type> are regex pattern substitutions. All names can be optionally enclosed in " ", for example, to include white-spaces.

<schema-namespace> determines XML Schema namespace. Optional <cxx-namespace> is prefixed to every C++ type name in this namespace declaration. <cxx-ret-type> is a C++ type name that is used as a return type for the post_* functions. Optional <cxx-arg-type> is an argument type for callback functions corresponding to elements and attributes of this type. If <cxx-arg-type> is not specified, it defaults to <cxx-ret-type> if <cxx-ret-type> ends with * or & (that is, it is a pointer or a reference) and const <cxx-ret-type>& otherwise. <file-name> is a file name either in the " " or < > format and is added with the #include directive to the generated code.

The # character starts a comment that ends with a new line or end of file. To specify a name that contains # enclose it in " ". For example:

namespace http://www.example.com/xmlns/my my
{
  include "my.hxx";

  # Pass apples by value.
  #
  apple apple;

  # Pass oranges as pointers.
  #
  orange orange_t*;
}
  

In the example above, for the http://www.example.com/xmlns/my#orange XML Schema type, the my::orange_t* C++ type will be used as both return and argument types.

Several namespace declarations can be specified in a single file. The namespace declaration can also be completely omitted to map types in a schema without a namespace. For instance:

include "my.hxx";
apple apple;

namespace http://www.example.com/xmlns/my
{
  orange "const orange_t*";
}
  

The compiler has a number of predefined mapping rules that can be presented as the following map files. The string-based XML Schema built-in types are mapped to either std::string or std::wstring depending on the character type selected with the --char-type option (char by default).

namespace http://www.w3.org/2001/XMLSchema
{
  boolean bool bool;

  byte "signed char" "signed char";
  unsignedByte "unsigned char" "unsigned char";

  short short short;
  unsignedShort "unsigned short" "unsigned short";

  int int int;
  unsignedInt "unsigned int" "unsigned int";

  long "long long" "long long";
  unsignedLong "unsigned long long" "unsigned long long";

  integer "long long" "long long";

  negativeInteger "long long" "long long";
  nonPositiveInteger "long long" "long long";

  positiveInteger "unsigned long long" "unsigned long long";
  nonNegativeInteger "unsigned long long" "unsigned long long";

  float float float;
  double double double;
  decimal double double;

  string std::string;
  normalizedString std::string;
  token std::string;
  Name std::string;
  NMTOKEN std::string;
  NCName std::string;
  ID std::string;
  IDREF std::string;
  language std::string;
  anyURI std::string;

  NMTOKENS xml_schema::string_sequence;
  IDREFS xml_schema::string_sequence;

  QName xml_schema::qname;

  base64Binary std::auto_ptr<xml_schema::buffer>
               std::auto_ptr<xml_schema::buffer>;
  hexBinary std::auto_ptr<xml_schema::buffer>
            std::auto_ptr<xml_schema::buffer>;

  date xml_schema::date;
  dateTime xml_schema::date_time;
  duration xml_schema::duration;
  gDay xml_schema::gday;
  gMonth xml_schema::gmonth;
  gMonthDay xml_schema::gmonth_day;
  gYear xml_schema::gyear;
  gYearMonth xml_schema::gyear_month;
  time xml_schema::time;
}
  

The last predefined rule maps anything that wasn't mapped by previous rules to void:

namespace .*
{
  .* void void;
}
  

When you provide your own type maps with the --type-map option, they are evaluated first. This allows you to selectively override predefined rules.

REGEX AND SHELL QUOTING

When entering a regular expression argument in the shell command line it is often necessary to use quoting (enclosing the argument in " " or ' ') in order to prevent the shell from interpreting certain characters, for example, spaces as argument separators and $ as variable expansions.

Unfortunately it is hard to achieve this in a manner that is portable across POSIX shells, such as those found on GNU/Linux and UNIX, and Windows shell. For example, if you use " " for quoting you will get a wrong result with POSIX shells if your expression contains $. The standard way of dealing with this on POSIX systems is to use ' ' instead. Unfortunately, Windows shell does not remove ' ' from arguments when they are passed to applications. As a result you may have to use ' ' for POSIX and " " for Windows ($ is not treated as a special character on Windows).

Alternatively, you can save regular expression options into a file, one option per line, and use this file with the --options-file option. With this approach you don't need to worry about shell quoting.

DIAGNOSTICS

If the input file is not a valid W3C XML Schema definition, xsd will issue diagnostic messages to STDERR and exit with non-zero exit code.

BUGS

Send bug reports to the xsd-users@codesynthesis.com mailing list.