diff options
Diffstat (limited to 'doc/cxx/tree')
| -rw-r--r-- | doc/cxx/tree/guide/guide.html2ps.in | 65 | ||||
| -rw-r--r-- | doc/cxx/tree/guide/index.xhtml | 2736 | ||||
| -rw-r--r-- | doc/cxx/tree/guide/index.xhtml.in | 2736 | ||||
| -rw-r--r-- | doc/cxx/tree/manual/index.xhtml | 6826 | ||||
| -rw-r--r-- | doc/cxx/tree/manual/index.xhtml.in | 6826 | ||||
| -rw-r--r-- | doc/cxx/tree/manual/manual.html2ps.in | 66 |
6 files changed, 0 insertions, 19255 deletions
diff --git a/doc/cxx/tree/guide/guide.html2ps.in b/doc/cxx/tree/guide/guide.html2ps.in deleted file mode 100644 index 461ffde..0000000 --- a/doc/cxx/tree/guide/guide.html2ps.in +++ /dev/null @@ -1,65 +0,0 @@ -@@html2ps { - option { - toc: hb; - colour: 1; - hyphenate: 1; - titlepage: 1; - } - - datefmt: "%B %Y"; - - titlepage { - content: " -<div align=center> - <h1><big>C++/Tree Mapping</big></h1> - <h1><big>Getting Started Guide</big></h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> -</div> - <p>Copyright © @copyright@.</p> - - <p>Permission is granted to copy, distribute and/or modify this - document under the terms of the - <a href='https://www.codesynthesis.com/licenses/fdl-1.2.txt'>GNU Free - Documentation License, version 1.2</a>; with no Invariant Sections, - no Front-Cover Texts and no Back-Cover Texts. - </p> - - <p>This document is available in the following formats: - <a href='https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/index.xhtml'>XHTML</a>, - <a href='https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/cxx-parser-guide.pdf'>PDF</a>, and - <a href='https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/cxx-parser-guide.ps'>PostScript</a>.</p>"; - } - - toc { - indent: 2em; - } - - header { - odd-right: $H; - even-left: $H; - } - - footer { - odd-left: $D; - odd-center: $T; - odd-right: $N; - - even-left: $N; - even-center: $T; - even-right: $D; - } -} - -body { - font-size: 12pt; - text-align: justify; -} - -pre { - font-size: 10pt; -} diff --git a/doc/cxx/tree/guide/index.xhtml b/doc/cxx/tree/guide/index.xhtml deleted file mode 100644 index fdaaa45..0000000 --- a/doc/cxx/tree/guide/index.xhtml +++ /dev/null @@ -1,2736 +0,0 @@ -<?xml version="1.0" encoding="iso-8859-1"?> -<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> -<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"> - -<head> - <title>C++/Tree Mapping Getting Started Guide</title> - - <meta name="copyright" content="© 2005-2023 Code Synthesis"/> - <meta name="keywords" content="xsd,xml,schema,c++,mapping,data,binding,parsing,serialization,validation"/> - <meta name="description" content="C++/Tree Mapping Getting Started Guide"/> - - <link rel="stylesheet" type="text/css" href="../../../default.css" /> - -<style type="text/css"> - pre { - padding : 0 0 0 0em; - margin : 0em 0em 0em 0; - - font-size : 102% - } - - body { - min-width: 48em; - } - - h1 { - font-weight: bold; - font-size: 200%; - line-height: 1.2em; - } - - h2 { - font-weight : bold; - font-size : 150%; - - padding-top : 0.8em; - } - - h3 { - font-size : 140%; - padding-top : 0.8em; - } - - /* Adjust indentation for three levels. */ - #container { - max-width: 48em; - } - - #content { - padding: 0 0.1em 0 4em; - /*background-color: red;*/ - } - - #content h1 { - margin-left: -2.06em; - } - - #content h2 { - margin-left: -1.33em; - } - - /* Title page */ - - #titlepage { - padding: 2em 0 1em 0; - border-bottom: 1px solid black; - } - - #titlepage .title { - font-weight: bold; - font-size: 200%; - text-align: center; - } - - #titlepage #first-title { - padding: 1em 0 0.4em 0; - } - - #titlepage #second-title { - padding: 0.4em 0 2em 0; - } - - /* Lists */ - ul.list li { - padding-top : 0.3em; - padding-bottom : 0.3em; - } - - div.img { - text-align: center; - padding: 2em 0 2em 0; - } - - /* */ - dl dt { - padding : 0.8em 0 0 0; - } - - /* Built-in table */ - #builtin { - margin: 2em 0 2em 0; - - border-collapse : collapse; - border : 1px solid; - border-color : #000000; - - font-size : 11px; - line-height : 14px; - } - - #builtin th, #builtin td { - border: 1px solid; - padding : 0.9em 0.9em 0.7em 0.9em; - } - - #builtin th { - background : #cde8f6; - } - - #builtin td { - text-align: left; - } - - /* TOC */ - table.toc { - border-style : none; - border-collapse : separate; - border-spacing : 0; - - margin : 0.2em 0 0.2em 0; - padding : 0 0 0 0; - } - - table.toc tr { - padding : 0 0 0 0; - margin : 0 0 0 0; - } - - table.toc * td, table.toc * th { - border-style : none; - margin : 0 0 0 0; - vertical-align : top; - } - - table.toc * th { - font-weight : normal; - padding : 0em 0.1em 0em 0; - text-align : left; - white-space : nowrap; - } - - table.toc * table.toc th { - padding-left : 1em; - } - - table.toc * td { - padding : 0em 0 0em 0.7em; - text-align : left; - } -</style> - - -</head> - -<body> -<div id="container"> - <div id="content"> - - <div class="noprint"> - - <div id="titlepage"> - <div class="title" id="first-title">C++/Tree Mapping</div> - <div class="title" id="second-title">Getting Started Guide</div> - - <p>Copyright © 2005-2023 Code Synthesis.</p> - - <p>Permission is granted to copy, distribute and/or modify this - document under the terms of the - <a href="https://www.codesynthesis.com/licenses/fdl-1.2.txt">GNU Free - Documentation License, version 1.2</a>; with no Invariant Sections, - no Front-Cover Texts and no Back-Cover Texts. - </p> - - <p>This document is available in the following formats: - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/index.xhtml">XHTML</a>, - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/cxx-tree-guide.pdf">PDF</a>, and - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/cxx-tree-guide.ps">PostScript</a>.</p> - - </div> - - <h1>Table of Contents</h1> - - <table class="toc"> - <tr> - <th></th><td><a href="#0">Preface</a> - <table class="toc"> - <tr><th></th><td><a href="#0.1">About This Document</a></td></tr> - <tr><th></th><td><a href="#0.2">More Information</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>1</th><td><a href="#1">Introduction</a> - <table class="toc"> - <tr><th>1.1</th><td><a href="#1.1">Mapping Overview</a></td></tr> - <tr><th>1.2</th><td><a href="#1.2">Benefits</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>2</th><td><a href="#2">Hello World Example</a> - <table class="toc"> - <tr><th>2.1</th><td><a href="#2.1">Writing XML Document and Schema</a></td></tr> - <tr><th>2.2</th><td><a href="#2.2">Translating Schema to C++</a></td></tr> - <tr><th>2.3</th><td><a href="#2.3">Implementing Application Logic</a></td></tr> - <tr><th>2.4</th><td><a href="#2.4">Compiling and Running</a></td></tr> - <tr><th>2.5</th><td><a href="#2.5">Adding Serialization</a></td></tr> - <tr><th>2.6</th><td><a href="#2.6">Selecting Naming Convention</a></td></tr> - <tr><th>2.7</th><td><a href="#2.7">Generating Documentation</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>3</th><td><a href="#3">Overall Mapping Configuration</a> - <table class="toc"> - <tr><th>3.1</th><td><a href="#3.1">C++ Standard</a></td></tr> - <tr><th>3.2</th><td><a href="#3.2">Character Type and Encoding</a></td></tr> - <tr><th>3.3</th><td><a href="#3.3">Support for Polymorphism </a></td></tr> - <tr><th>3.4</th><td><a href="#3.4">Namespace Mapping</a></td></tr> - <tr><th>3.5</th><td><a href="#3.5">Thread Safety</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>4</th><td><a href="#4">Working with Object Models</a> - <table class="toc"> - <tr><th>4.1</th><td><a href="#4.1">Attribute and Element Cardinalities</a></td></tr> - <tr><th>4.2</th><td><a href="#4.2">Accessing the Object Model</a></td></tr> - <tr><th>4.3</th><td><a href="#4.3">Modifying the Object Model</a></td></tr> - <tr><th>4.4</th><td><a href="#4.4">Creating the Object Model from Scratch</a></td></tr> - <tr><th>4.5</th><td><a href="#4.5">Mapping for the Built-in XML Schema Types</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>5</th><td><a href="#5">Parsing</a> - <table class="toc"> - <tr><th>5.1</th><td><a href="#5.1">XML Schema Validation and Searching</a></td></tr> - <tr><th>5.2</th><td><a href="#5.2">Error Handling</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>6</th><td><a href="#6">Serialization</a> - <table class="toc"> - <tr><th>6.1</th><td><a href="#6.1">Namespace and Schema Information</a></td></tr> - <tr><th>6.2</th><td><a href="#6.2">Error Handling</a></td></tr> - </table> - </td> - </tr> - - </table> - </div> - - <h1><a name="0">Preface</a></h1> - - <h2><a name="0.1">About This Document</a></h2> - - <p>The goal of this document is to provide you with an understanding of - the C++/Tree programming model and allow you to efficiently evaluate - XSD against your project's technical requirements. As such, this - document is intended for C++ developers and software architects - who are looking for an XML processing solution. For a more in-depth - description of the C++/Tree mapping refer to the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/">C++/Tree - Mapping User Manual</a>.</p> - - <p>Prior experience with XML and C++ is required to understand this - document. Basic understanding of XML Schema is advantageous but - not expected or required. - </p> - - - <h2><a name="0.2">More Information</a></h2> - - <p>Beyond this guide, you may also find the following sources of - information useful:</p> - - <ul class="list"> - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/">C++/Tree - Mapping User Manual</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/Customization_guide">C++/Tree - Mapping Customization Guide</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree - Mapping Frequently Asked Questions (FAQ)</a></li> - - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a></li> - - <li>The <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - contains a collection of examples and a README file with an overview - of each example.</li> - - <li>The <code>README</code> file in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - explains how to build the examples.</li> - - <li>The <a href="https://www.codesynthesis.com/mailman/listinfo/xsd-users">xsd-users</a> - mailing list is the place to ask technical questions about XSD and the C++/Parser mapping. - Furthermore, the <a href="https://www.codesynthesis.com/pipermail/xsd-users/">archives</a> - may already have answers to some of your questions.</li> - </ul> - - <!-- Introduction --> - - <h1><a name="1">1 Introduction</a></h1> - - <p>Welcome to CodeSynthesis XSD and the C++/Tree mapping. XSD is a - cross-platform W3C XML Schema to C++ data binding compiler. C++/Tree - is a W3C XML Schema to C++ mapping that represents the data stored - in XML as a statically-typed, vocabulary-specific object model. - </p> - - <h2><a name="1.1">1.1 Mapping Overview</a></h2> - - <p>Based on a formal description of an XML vocabulary (schema), the - C++/Tree mapping produces a tree-like data structure suitable for - in-memory processing. The core of the mapping consists of C++ - classes that constitute the object model and are derived from - types defined in XML Schema as well as XML parsing and - serialization code.</p> - - <p>Besides the core features, C++/Tree provide a number of additional - mapping elements that can be useful in some applications. These - include serialization and extraction to/from formats others than - XML, such as unstructured text (useful for debugging) and binary - representations such as XDR and CDR for high-speed data processing - as well as automatic documentation generation. The C++/Tree mapping - also provides a wide range of mechanisms for controlling and - customizing the generated code.</p> - - <p>A typical application that uses C++/Tree for XML processing usually - performs the following three steps: it first reads (parses) an XML - document to an in-memory object model, it then performs some useful - computations on that object model which may involve modification - of the model, and finally it may write (serialize) the modified - object model back to XML.</p> - - <p>The next chapter presents a simple application that performs these - three steps. The following chapters show how to use the C++/Tree - mapping in more detail.</p> - - <h2><a name="1.2">1.2 Benefits</a></h2> - - <p>Traditional XML access APIs such as Document Object Model (DOM) - or Simple API for XML (SAX) have a number of drawbacks that - make them less suitable for creating robust and maintainable - XML processing applications. These drawbacks include: - </p> - - <ul class="list"> - <li>Generic representation of XML in terms of elements, attributes, - and text forces an application developer to write a substantial - amount of bridging code that identifies and transforms pieces - of information encoded in XML to a representation more suitable - for consumption by the application logic.</li> - - <li>String-based flow control defers error detection to runtime. - It also reduces code readability and maintainability.</li> - - <li>Lack of type safety because the data is represented as text.</li> - - <li>Resulting applications are hard to debug, change, and - maintain.</li> - </ul> - - <p>In contrast, statically-typed, vocabulary-specific object model - produced by the C++/Tree mapping allows you to operate in your - domain terms instead of the generic elements, attributes, and - text. Static typing helps catch errors at compile-time rather - than at run-time. Automatic code generation frees you for more - interesting tasks (such as doing something useful with the - information stored in the XML documents) and minimizes the - effort needed to adapt your applications to changes in the - document structure. To summarize, the C++/Tree object model has - the following key advantages over generic XML access APIs:</p> - - <ul class="list"> - <li><b>Ease of use.</b> The generated code hides all the complexity - associated with parsing and serializing XML. This includes navigating - the structure and converting between the text representation and - data types suitable for manipulation by the application - logic.</li> - - <li><b>Natural representation.</b> The object representation allows - you to access the XML data using your domain vocabulary instead - of generic elements, attributes, and text.</li> - - <li><b>Concise code.</b> With the object representation the - application implementation is simpler and thus easier - to read and understand.</li> - - <li><b>Safety.</b> The generated object model is statically - typed and uses functions instead of strings to access the - information. This helps catch programming errors at compile-time - rather than at runtime.</li> - - <li><b>Maintainability.</b> Automatic code generation minimizes the - effort needed to adapt the application to changes in the - document structure. With static typing, the C++ compiler - can pin-point the places in the client code that need to be - changed.</li> - - <li><b>Compatibility.</b> Sequences of elements are represented in - the object model as containers conforming to the standard C++ - sequence requirements. This makes it possible to use standard - C++ algorithms on the object representation and frees you from - learning yet another container interface, as is the case with - DOM.</li> - - <li><b>Efficiency.</b> If the application makes repetitive use - of the data extracted from XML, then the C++/Tree object model - is more efficient because the navigation is performed using - function calls rather than string comparisons and the XML - data is extracted only once. Furthermore, the runtime memory - usage is reduced due to more efficient data storage - (for instance, storing numeric data as integers instead of - strings) as well as the static knowledge of cardinality - constraints.</li> - </ul> - - - <!-- Hello World Parser --> - - - <h1><a name="2">2 Hello World Example</a></h1> - - <p>In this chapter we will examine how to parse, access, modify, and - serialize a very simple XML document using the XSD-generated - C++/Tree object model. The code presented in this chapter is - based on the <code>hello</code> example which can be found in - the <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package.</p> - - <h2><a name="2.1">2.1 Writing XML Document and Schema</a></h2> - - <p>First, we need to get an idea about the structure - of the XML documents we are going to process. Our - <code>hello.xml</code>, for example, could look like this:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello> - - <greeting>Hello</greeting> - - <name>sun</name> - <name>moon</name> - <name>world</name> - -</hello> - </pre> - - <p>Then we can write a description of the above XML in the - XML Schema language and save it into <code>hello.xsd</code>:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:complexType name="hello_t"> - <xs:sequence> - <xs:element name="greeting" type="xs:string"/> - <xs:element name="name" type="xs:string" maxOccurs="unbounded"/> - </xs:sequence> - </xs:complexType> - - <xs:element name="hello" type="hello_t"/> - -</xs:schema> - </pre> - - <p>Even if you are not familiar with XML Schema, it - should be easy to connect declarations in <code>hello.xsd</code> - to elements in <code>hello.xml</code>. The <code>hello_t</code> type - is defined as a sequence of the nested <code>greeting</code> and - <code>name</code> elements. Note that the term sequence in XML - Schema means that elements should appear in a particular order - as opposed to appearing multiple times. The <code>name</code> - element has its <code>maxOccurs</code> property set to - <code>unbounded</code> which means it can appear multiple times - in an XML document. Finally, the globally-defined <code>hello</code> - element prescribes the root element for our vocabulary. For an - easily-approachable introduction to XML Schema refer to - <a href="http://www.w3.org/TR/xmlschema-0/">XML Schema Part 0: - Primer</a>.</p> - - <p>The above schema is a specification of our XML vocabulary; it tells - everybody what valid documents of our XML-based language should look - like. We can also update our <code>hello.xml</code> to include the - information about the schema so that XML parsers can validate - our document:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="hello.xsd"> - - <greeting>Hello</greeting> - - <name>sun</name> - <name>moon</name> - <name>world</name> - -</hello> - </pre> - - - <p>The next step is to compile the schema to generate the object - model and parsing functions.</p> - - <h2><a name="2.2">2.2 Translating Schema to C++</a></h2> - - <p>Now we are ready to translate our <code>hello.xsd</code> to C++. - To do this we invoke the XSD compiler from a terminal (UNIX) or - a command prompt (Windows): - </p> - - <pre class="terminal"> -$ xsd cxx-tree hello.xsd - </pre> - - <p>The XSD compiler produces two C++ files: <code>hello.hxx</code> and - <code>hello.cxx</code>. The following code fragment is taken from - <code>hello.hxx</code>; it should give you an idea about what gets - generated: - </p> - - <pre class="c++"> -class hello_t -{ -public: - // greeting - // - typedef xml_schema::string greeting_type; - - const greeting_type& - greeting () const; - - greeting_type& - greeting (); - - void - greeting (const greeting_type& x); - - // name - // - typedef xml_schema::string name_type; - typedef xsd::sequence<name_type> name_sequence; - typedef name_sequence::iterator name_iterator; - typedef name_sequence::const_iterator name_const_iterator; - - const name_sequence& - name () const; - - name_sequence& - name (); - - void - name (const name_sequence& s); - - // Constructor. - // - hello_t (const greeting_type&); - - ... - -}; - -std::unique_ptr<hello_t> -hello (const std::string& uri); - -std::unique_ptr<hello_t> -hello (std::istream&); - </pre> - - <p>The <code>hello_t</code> C++ class corresponds to the - <code>hello_t</code> XML Schema type. For each element - in this type a set of C++ type definitions as well as - accessor and modifier functions are generated inside the - <code>hello_t</code> class. Note that the type definitions - and member functions for the <code>greeting</code> and - <code>name</code> elements are different because of the - cardinality differences between these two elements - (<code>greeting</code> is a required single element and - <code>name</code> is a sequence of elements).</p> - - <p>The <code>xml_schema::string</code> type used in the type - definitions is a C++ class provided by the XSD runtime - that corresponds to built-in XML Schema type - <code>string</code>. The <code>xml_schema::string</code> - is based on <code>std::string</code> and can be used as - such. Similarly, the <code>sequence</code> class template - that is used in the <code>name_sequence</code> type - definition is based on and has the same interface as - <code>std::vector</code>. The mapping between the built-in - XML Schema types and C++ types is described in more detail in - <a href="#4.5">Section 4.5, "Mapping for the Built-in XML Schema - Types"</a>. The <code>hello_t</code> class also includes a - constructor with an initializer for the required - <code>greeting</code> element as its argument.</p> - - <p>The <code>hello</code> overloaded global functions correspond - to the <code>hello</code> global element in XML Schema. A - global element in XML Schema is a valid document root. - By default XSD generated a set of parsing functions for each - global element defined in XML Schema (this can be overridden - with the <code>--root-element-*</code> options). Parsing - functions return a dynamically allocated object model as an - automatic pointer. The actual pointer used depends on the - C++ standard selected. For C++11 it is <code>std::unique_ptr</code> - as shown above. For C++98 it is <code>std::auto_ptr</code>. - For example, if we modify our XSD compiler invocation to - select C++98:</p> - - <pre class="terminal"> -$ xsd cxx-tree --std c++98 hello.xsd - </pre> - - <p>Then the parsing function signatures will become:</p> - - <pre class="c++"> -std::auto_ptr<hello_t> -hello (const std::string& uri); - -std::auto_ptr<hello_t> -hello (std::istream&); - </pre> - - <p>For more information on parsing functions see <a href="#5">Chapter 5, - "Parsing"</a>.</p> - - <h2><a name="2.3">2.3 Implementing Application Logic</a></h2> - - <p>At this point we have all the parts we need to do something useful - with the information stored in our XML document: - </p> - - <pre class="c++"> -#include <iostream> -#include "hello.hxx" - -using namespace std; - -int -main (int argc, char* argv[]) -{ - try - { - unique_ptr<hello_t> h (hello (argv[1])); - - for (hello_t::name_const_iterator i (h->name ().begin ()); - i != h->name ().end (); - ++i) - { - cerr << h->greeting () << ", " << *i << "!" << endl; - } - } - catch (const xml_schema::exception& e) - { - cerr << e << endl; - return 1; - } -} - </pre> - - <p>The first part of our application calls one of the parsing - functions to parser an XML file specified in the command line. - We then use the returned object model to iterate over names - and print a greeting line for each of them. Finally, we - catch and print the <code>xml_schema::exception</code> - exception in case something goes wrong. This exception - is the root of the exception hierarchy used by the - XSD-generated code. - </p> - - - <h2><a name="2.4">2.4 Compiling and Running</a></h2> - - <p>After saving our application from the previous section in - <code>driver.cxx</code>, we are ready to compile our first - program and run it on the test XML document. On a UNIX - system this can be done with the following commands: - </p> - - <pre class="terminal"> -$ c++ -std=c++11 -I.../libxsd -c driver.cxx hello.cxx -$ c++ -std=c++11 -o driver driver.o hello.o -lxerces-c -$ ./driver hello.xml -Hello, sun! -Hello, moon! -Hello, world! - </pre> - - <p>Here <code>.../libxsd</code> represents the path to the - <a href="https://cppget.org/libxsd">libxsd</a> package root - directory. Note also that we are required to link our - application with the Xerces-C++ library because the generated - code uses it as the underlying XML parser.</p> - - <h2><a name="2.5">2.5 Adding Serialization</a></h2> - - <p>While parsing and accessing the XML data may be everything - you need, there are applications that require creating new - or modifying existing XML documents. By default XSD does - not produce serialization code. We will need to request - it with the <code>--generate-serialization</code> options:</p> - - <pre class="terminal"> -$ xsd cxx-tree --generate-serialization hello.xsd - </pre> - - <p>If we now examine the generated <code>hello.hxx</code> file, - we will find a set of overloaded serialization functions, - including the following version:</p> - - <pre class="c++"> -void -hello (std::ostream&, - const hello_t&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap ()); - - </pre> - - <p>Just like with parsing functions, XSD generates serialization - functions for each global element unless instructed otherwise - with one of the <code>--root-element-*</code> options. For more - information on serialization functions see <a href="#6">Chapter 6, - "Serialization"</a>.</p> - - <p>We first examine an application that modifies an existing - object model and serializes it back to XML:</p> - - <pre class="c++"> -#include <iostream> -#include "hello.hxx" - -using namespace std; - -int -main (int argc, char* argv[]) -{ - try - { - unique_ptr<hello_t> h (hello (argv[1])); - - // Change the greeting phrase. - // - h->greeting ("Hi"); - - // Add another entry to the name sequence. - // - h->name ().push_back ("mars"); - - // Serialize the modified object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "hello.xsd"; - - hello (cout, *h, map); - } - catch (const xml_schema::exception& e) - { - cerr << e << endl; - return 1; - } -} - </pre> - - <p>First, our application parses an XML document and obtains its - object model as in the previous example. Then it changes the - greeting string and adds another entry to the list of names. - Finally, it serializes the object model back to XML by calling - the serialization function.</p> - - <p>The first argument we pass to the serialization function is - <code>cout</code> which results in the XML being written to - the standard output for us to inspect. We could have also - written the result to a file or memory buffer by creating an - instance of <code>std::ofstream</code> or <code>std::ostringstream</code> - and passing it instead of <code>cout</code>. The second argument is the - object model we want to serialize. The final argument is an optional - namespace information map for our vocabulary. It captures information - such as namespaces, namespace prefixes to which they should be mapped, - and schemas associated with these namespaces. If we don't provide - this argument then generic namespace prefixes (<code>p1</code>, - <code>p2</code>, etc.) will be automatically assigned to XML namespaces - and no schema information will be added to the resulting document - (see <a href="#6">Chapter 6, "Serialization"</a> for details). - In our case, the prefix (map key) and namespace name are empty - because our vocabulary does not use XML namespaces.</p> - - <p>If we now compile and run this application we will see the - output as shown in the following listing:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="hello.xsd"> - - <greeting>Hi</greeting> - - <name>sun</name> - <name>moon</name> - <name>world</name> - <name>mars</name> - -</hello> - </pre> - - <p>We can also create and serialize an object model from scratch - as shown in the following example:</p> - - <pre class="c++"> -#include <iostream> -#include <fstream> -#include "hello.hxx" - -using namespace std; - -int -main (int argc, char* argv[]) -{ - try - { - hello_t h ("Hi"); - - hello_t::name_sequence& ns (h.name ()); - - ns.push_back ("Jane"); - ns.push_back ("John"); - - // Serialize the object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "hello.xsd"; - - std::ofstream ofs (argv[1]); - hello (ofs, h, map); - } - catch (const xml_schema::exception& e) - { - cerr << e << endl; - return 1; - } -} - </pre> - - <p>In this example we used the generated constructor to create - an instance of type <code>hello_t</code>. To reduce typing, - we obtained a reference to the name sequence which we then - used to add a few names. The serialization part is identical - to the previous example except this time we are writing to - a file. If we compile and run this program, it produces the - following XML file:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="hello.xsd"> - - <greeting>Hi</greeting> - - <name>Jane</name> - <name>John</name> - -</hello> - </pre> - - <h2><a name="2.6">2.6 Selecting Naming Convention</a></h2> - - <p>By default XSD uses the so-called K&R (Kernighan and Ritchie) - identifier naming convention in the generated code. In this - convention both type and function names are in lower case and - words are separated by underscores. If your application code or - schemas use a different notation, you may want to change the - naming convention used in the generated code for consistency. - XSD supports a set of widely-used naming conventions - that you can select with the <code>--type-naming</code> and - <code>--function-naming</code> options. You can also further - refine one of the predefined conventions or create a completely - custom naming scheme by using the <code>--*-regex</code> options.</p> - - <p>As an example, let's assume that our "Hello World" application - uses the so-called upper-camel-case naming convention for types - (that is, each word in a type name is capitalized) and the K&R - convention for function names. Since K&R is the default - convention for both type and function names, we only need to - change the type naming scheme:</p> - - <pre class="terminal"> -$ xsd cxx-tree --type-naming ucc hello.xsd - </pre> - - <p>The <code>ucc</code> argument to the <code>--type-naming</code> - options stands for upper-camel-case. If we now examine the - generated <code>hello.hxx</code>, we will see the following - changes compared to the declarations shown in the previous - sections:</p> - - <pre class="c++"> -class Hello_t -{ -public: - // greeting - // - typedef xml_schema::String GreetingType; - - const GreetingType& - greeting () const; - - GreetingType& - greeting (); - - void - greeting (const GreetingType& x); - - // name - // - typedef xml_schema::String NameType; - typedef xsd::sequence<NameType> NameSequence; - typedef NameSequence::iterator NameIterator; - typedef NameSequence::const_iterator NameConstIterator; - - const NameSequence& - name () const; - - NameSequence& - name (); - - void - name (const NameSequence& s); - - // Constructor. - // - Hello_t (const GreetingType&); - - ... - -}; - -std::unique_ptr<Hello_t> -hello (const std::string& uri); - -std::unique_ptr<Hello_t> -hello (std::istream&); - </pre> - - <p>Notice that the type names in the <code>xml_schema</code> namespace, - for example <code>xml_schema::String</code>, now also use the - upper-camel-case naming convention. The only thing that we may - be unhappy about in the above code is the <code>_t</code> - suffix in <code>Hello_t</code>. If we are not in a position - to change the schema, we can <em>touch-up</em> the <code>ucc</code> - convention with a custom translation rule using the - <code>--type-regex</code> option:</p> - - <pre class="terminal"> -$ xsd cxx-tree --type-naming ucc --type-regex '/ (.+)_t/\u$1/' hello.xsd - </pre> - - <p>This results in the following changes to the generated code:</p> - - <pre class="c++"> -class Hello -{ -public: - // greeting - // - typedef xml_schema::String GreetingType; - - const GreetingType& - greeting () const; - - GreetingType& - greeting (); - - void - greeting (const GreetingType& x); - - // name - // - typedef xml_schema::String NameType; - typedef xsd::sequence<NameType> NameSequence; - typedef NameSequence::iterator NameIterator; - typedef NameSequence::const_iterator NameConstIterator; - - const NameSequence& - name () const; - - NameSequence& - name (); - - void - name (const NameSequence& s); - - // Constructor. - // - Hello (const GreetingType&); - - ... - -}; - -std::unique_ptr<Hello> -hello (const std::string& uri); - -std::unique_ptr<Hello> -hello (std::istream&); - </pre> - - <p>For more detailed information on the <code>--type-naming</code>, - <code>--function-naming</code>, <code>--type-regex</code>, and - other <code>--*-regex</code> options refer to the NAMING - CONVENTION section in the <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>.</p> - - <h2><a name="2.7">2.7 Generating Documentation</a></h2> - - <p>While our object model is quite simple, real-world vocabularies - can be quite complex with hundreds of types, elements, and - attributes. For such vocabularies figuring out which types - provide which member functions by studying the generated - source code or schemas can be a daunting task. To provide - application developers with a more accessible way of - understanding the generated object models, the XSD compiler - can be instructed to produce source code with documentation - comments in the Doxygen format. Then the source code can be - processed with the <a href="http://www.doxygen.org">Doxygen</a> - documentation system to extract this information and produce - documentation in various formats. - </p> - - <p>In this section we will see how to generate documentation - for our "Hello World" vocabulary. To showcase the full power - of the XSD documentation facilities, we will first document - our schema. The XSD compiler will then transfer - this information from the schema to the generated code and - then to the object model documentation. Note that the - documentation in the schema is not required for XSD to - generate useful documentation. Below you will find - our <code>hello.xsd</code> with added documentation:</p> - - <pre class="xml"> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:complexType name="hello_t"> - - <xs:annotation> - <xs:documentation> - The hello_t type consists of a greeting phrase and a - collection of names to which this greeting applies. - </xs:documentation> - </xs:annotation> - - <xs:sequence> - - <xs:element name="greeting" type="xs:string"> - <xs:annotation> - <xs:documentation> - The greeting element contains the greeting phrase - for this hello object. - </xs:documentation> - </xs:annotation> - </xs:element> - - <xs:element name="name" type="xs:string" maxOccurs="unbounded"> - <xs:annotation> - <xs:documentation> - The name elements contains names to be greeted. - </xs:documentation> - </xs:annotation> - </xs:element> - - </xs:sequence> - </xs:complexType> - - <xs:element name="hello" type="hello_t"> - <xs:annotation> - <xs:documentation> - The hello element is a root of the Hello XML vocabulary. - Every conforming document should start with this element. - </xs:documentation> - </xs:annotation> - </xs:element> - -</xs:schema> - </pre> - - <p>The first step in obtaining the documentation is to recompile - our schema with the <code>--generate-doxygen</code> option:</p> - - <pre class="terminal"> -$ xsd cxx-tree --generate-serialization --generate-doxygen hello.xsd - </pre> - - <p>Now the generated <code>hello.hxx</code> file contains comments - in the Doxygen format. The next step is to process this file - with the Doxygen documentation system. If your project does - not use Doxygen then you first need to create a configuration - file for your project:</p> - - <pre class="terminal"> -$ doxygen -g hello.doxygen - </pre> - - <p>You only need to perform this step once. Now we can generate - the documentation by executing the following command in the - directory with the generated source code:</p> - - <pre class="terminal"> -$ doxygen hello.doxygen - </pre> - - <p>While the generated documentation can be useful as is, we can - go one step further and link (using the Doxygen tags mechanism) - the documentation for our object model with the documentation - for the XSD runtime library which defines C++ classes for the - built-in XML Schema types. This way we can seamlessly browse - between documentation for the <code>hello_t</code> class which - is generated by the XSD compiler and the <code>xml_schema::string</code> - class which is defined in the XSD runtime library. The Doxygen - configuration file for the XSD runtime is provided with the XSD - distribution.</p> - - <p>You can view the result of the steps described in this section - on the <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/hello/html/annotated.html">Hello - Example Documentation</a> page.</p> - - <!-- Chapater 3 --> - - - <h1><a name="3">3 Overall Mapping Configuration</a></h1> - - <p>The C++/Tree mapping has a number of configuration parameters that - determine the overall properties and behavior of the generated code. - Configuration parameters are specified with the XSD command line - options. This chapter describes configuration aspects that are most - commonly encountered by application developers. These include: the - C++ standard, the character type that is used by the generated code, - handling of vocabularies that use XML Schema polymorphism, XML Schema - to C++ namespace mapping, and thread safety. For more ways to configure - the generated code refer to the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>. - </p> - - <h2><a name="3.1">3.1 C++ Standard</a></h2> - - <p>The C++/Tree mapping provides support for ISO/IEC C++ 2011 (C++11) - and ISO/IEC C++ 1998/2003 (C++98). To select the C++ standard for the - generated code we use the <code>--std</code> XSD compiler command - line option. While the majority of the examples in this guide use - C++11, the document explains the C++11/98 usage difference and so - they can easily be converted to C++98.</p> - - <h2><a name="3.2">3.2 Character Type and Encoding</a></h2> - - <p>The C++/Tree mapping has built-in support for two character types: - <code>char</code> and <code>wchar_t</code>. You can select the - character type with the <code>--char-type</code> command line - option. The default character type is <code>char</code>. The - character type affects all string and string-based types that - are used in the mapping. These include the string-based built-in - XML Schema types, exception types, stream types, etc.</p> - - <p>Another aspect of the mapping that depends on the character type - is character encoding. For the <code>char</code> character type - the default encoding is UTF-8. Other supported encodings are - ISO-8859-1, Xerces-C++ Local Code Page (LPC), as well as - custom encodings. You can select which encoding should be used - in the object model with the <code>--char-encoding</code> command - line option.</p> - - <p>For the <code>wchar_t</code> character type the encoding is - automatically selected between UTF-16 and UTF-32/UCS-4 depending - on the size of the <code>wchar_t</code> type. On some platforms - (for example, Windows with Visual C++ and AIX with IBM XL C++) - <code>wchar_t</code> is 2 bytes long. For these platforms the - encoding is UTF-16. On other platforms <code>wchar_t</code> is 4 bytes - long and UTF-32/UCS-4 is used.</p> - - <p>Note also that the character encoding that is used in the object model - is independent of the encodings used in input and output XML. In fact, - all three (object mode, input XML, and output XML) can have different - encodings.</p> - - <h2><a name="3.3">3.3 Support for Polymorphism</a></h2> - - <p>By default XSD generates non-polymorphic code. If your vocabulary - uses XML Schema polymorphism in the form of <code>xsi:type</code> - and/or substitution groups, then you will need to compile - your schemas with the <code>--generate-polymorphic</code> option - to produce polymorphism-aware code. For more information on - working with polymorphic object models, refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.11">Section 2.11, - "Mapping for <code>xsi:type</code> and Substitution Groups"</a> in - the C++/Tree Mapping User Manual.</p> - - <h2><a name="3.4">3.4 Namespace Mapping</a></h2> - - <p>XSD maps XML namespaces specified in the <code>targetNamespace</code> - attribute in XML Schema to one or more nested C++ namespaces. By - default, a namespace URI is mapped to a sequence of C++ namespace - names by removing the protocol and host parts and splitting the - rest into a sequence of names with <code>'/'</code> as the name - separator.</p> - - <p>The default mapping of namespace URIs to C++ namespaces - can be altered using the <code>--namespace-map</code> and - <code>--namespace-regex</code> compiler options. For example, - to map namespace URI <code>https://www.codesynthesis.com/my</code> to - C++ namespace <code>cs::my</code>, we can use the following option:</p> - - <pre class="terminal"> ---namespace-map https://www.codesynthesis.com/my=cs::my - </pre> - - <p>A vocabulary without a namespace is mapped to the global scope. This - also can be altered with the above options by using an empty name - for the XML namespace:</p> - - <pre class="terminal"> ---namespace-map =cs - </pre> - - <h2><a name="3.5">3.5 Thread Safety</a></h2> - - <p>XSD-generated code is thread-safe in the sense that you can - use different instantiations of the object model in several - threads concurrently. This is possible due to the generated - code not relying on any writable global variables. If you need - to share the same object between several threads then you will - need to provide some form of synchronization. One approach would - be to use the generated code customization mechanisms to embed - synchronization primitives into the generated C++ classes. For more - information on generated code customization refer to the - <a href="http://wiki.codesynthesis.com/Tree/Customization_guide">C++/Tree - Mapping Customization Guide</a>.</p> - - <p>If you also would like to call parsing and/or serialization - functions from several threads potentially concurrently, then - you will need to make sure the Xerces-C++ runtime is initialized - and terminated only once. The easiest way to do this is to - initialize/terminate Xerces-C++ from <code>main()</code> when - there are no threads yet/anymore:</p> - - <pre class="c++"> -#include <xercesc/util/PlatformUtils.hpp> - -int -main () -{ - xercesc::XMLPlatformUtils::Initialize (); - - { - // Start/terminate threads and parse/serialize here. - } - - xercesc::XMLPlatformUtils::Terminate (); -} - </pre> - - <p>Because you initialize the Xerces-C++ runtime yourself you should - also pass the <code>xml_schema::flags::dont_initialize</code> flag - to parsing and serialization functions. See <a href="#5">Chapter 5, - "Parsing"</a> and <a href="#6">Chapter 6, "Serialization"</a> for - more information.</p> - - - <!-- Chapater 4 --> - - - <h1><a name="4">4 Working with Object Models</a></h1> - - <p>As we have seen in the previous chapters, the XSD compiler generates - a C++ class for each type defined in XML Schema. Together these classes - constitute an object model for an XML vocabulary. In this chapter we - will take a closer look at different elements that comprise an - object model class as well as how to create, access, and modify - object models.</p> - - <p>In this and subsequent chapters we will use the following schema - that describes a collection of person records. We save it in - <code>people.xsd</code>:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:simpleType name="gender_t"> - <xs:restriction base="xs:string"> - <xs:enumeration value="male"/> - <xs:enumeration value="female"/> - </xs:restriction> - </xs:simpleType> - - <xs:complexType name="person_t"> - <xs:sequence> - <xs:element name="first-name" type="xs:string"/> - <xs:element name="middle-name" type="xs:string" minOccurs="0"/> - <xs:element name="last-name" type="xs:string"/> - <xs:element name="gender" type="gender_t"/> - <xs:element name="age" type="xs:short"/> - </xs:sequence> - <xs:attribute name="id" type="xs:unsignedInt" use="required"/> - </xs:complexType> - - <xs:complexType name="people_t"> - <xs:sequence> - <xs:element name="person" type="person_t" maxOccurs="unbounded"/> - </xs:sequence> - </xs:complexType> - - <xs:element name="people" type="people_t"/> - -</xs:schema> - </pre> - - <p>A sample XML instance to go along with this schema is saved - in <code>people.xml</code>:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd"> - - <person id="1"> - <first-name>John</first-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>32</age> - </person> - - <person id="2"> - <first-name>Jane</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>female</gender> - <age>28</age> - </person> - -</people> - </pre> - - <p>Compiling <code>people.xsd</code> with the XSD compiler results - in three generated C++ classes: <code>gender_t</code>, - <code>person_t</code>, and <code>people_t</code>. - The <code>gender_t</code> class is modelled after the C++ - <code>enum</code> type. Its definition is presented below:</p> - - <pre class="c++"> -class gender_t: public xml_schema::string -{ -public: - enum value - { - male, - female - }; - - gender_t (value); - gender_t (const xml_schema::string&); - - gender_t& - operator= (value); - - operator value () const; -}; - </pre> - - <p>The following listing shows how we can use this type:</p> - - <pre class="c++"> -gender_t m (gender_t::male); -gender_t f ("female"); - -if (m == "female" || f == gender_t::male) -{ - ... -} - -switch (m) -{ -case gender_t::male: - { - ... - } -case gender_t::female: - { - ... - } -} - </pre> - - <p>The other two classes will be examined in detail in the subsequent - sections.</p> - - <h2><a name="4.1">4.1 Attribute and Element Cardinalities</a></h2> - - <p>As we have seen in the previous chapters, XSD generates a different - set of type definitions and member functions for elements with - different cardinalities. The C++/Tree mapping divides all the possible - element and attribute cardinalities into three cardinality classes: - <em>one</em>, <em>optional</em>, and <em>sequence</em>.</p> - - <p>The <em>one</em> cardinality class covers all elements that should - occur exactly once as well as required attributes. In our - example, the <code>first-name</code>, <code>last-name</code>, - <code>gender</code>, and <code>age</code> elements as well as - the <code>id</code> attribute belong to this cardinality class. - The following code fragment shows type definitions as well as the - accessor and modifier functions that are generated for the - <code>gender</code> element in the <code>person_t</code> class:</p> - - <pre class="c++"> -class person_t -{ - // gender - // - typedef gender_t gender_type; - - const gender_type& - gender () const; - - gender_type& - gender (); - - void - gender (const gender_type&); -}; - </pre> - - <p>The <code>gender_type</code> type is an alias for the element's type. - The first two accessor functions return read-only (constant) and - read-write references to the element's value, respectively. The - modifier function sets the new value for the element.</p> - - <p>The <em>optional</em> cardinality class covers all elements that - can occur zero or one time as well as optional attributes. In our - example, the <code>middle-name</code> element belongs to this - cardinality class. The following code fragment shows the type - definitions as well as the accessor and modifier functions that - are generated for this element in the <code>person_t</code> class:</p> - - <pre class="c++"> -class person_t -{ - // middle-name - // - typedef xml_schema::string middle_name_type; - typedef xsd::optional<middle_name_type> middle_name_optional; - - const middle_name_optional& - middle_name () const; - - middle_name_optional& - middle_name (); - - void - middle_name (const middle_name_type&); - - void - middle_name (const middle_name_optional&); -}; - </pre> - - <p>As with the <code>gender</code> element, <code>middle_name_type</code> - is an alias for the element's type. The <code>middle_name_optional</code> - type is a container for the element's optional value. It can be queried - for the presence of the value using the <code>present()</code> function. - The value itself can be retrieved using the <code>get()</code> - accessor and set using the <code>set()</code> modifier. The container - can be reverted to the value not present state with the call to the - <code>reset()</code> function. The following example shows how we - can use this container:</p> - - <pre class="c++"> -person_t::middle_name_optional n ("John"); - -if (n.present ()) -{ - cout << n.get () << endl; -} - -n.set ("Jane"); -n.reset (); - </pre> - - - <p>Unlike the <em>one</em> cardinality class, the accessor functions - for the <em>optional</em> class return read-only (constant) and - read-write references to the container instead of the element's - value directly. The modifier functions set the new value for the - element.</p> - - <p>Finally, the <em>sequence</em> cardinality class covers all elements - that can occur more than once. In our example, the - <code>person</code> element in the <code>people_t</code> type - belongs to this cardinality class. The following code fragment shows - the type definitions as well as the accessor and modifier functions - that are generated for this element in the <code>people_t</code> - class:</p> - - <pre class="c++"> -class people_t -{ - // person - // - typedef person_t person_type; - typedef xsd::sequence<person_type> person_sequence; - typedef person_sequence::iterator person_iterator; - typedef person_sequence::const_iterator person_const_iterator; - - const person_sequence& - person () const; - - person_sequence& - person (); - - void - person (const person_sequence&); -}; - </pre> - - <p>Identical to the other cardinality classes, <code>person_type</code> - is an alias for the element's type. The <code>person_sequence</code> - type is a sequence container for the element's values. It is based - on and has the same interface as <code>std::vector</code> and - therefore can be used in similar ways. The <code>person_iterator</code> - and <code>person_const_iterator</code> types are read-only - (constant) and read-write iterators for the <code>person_sequence</code> - container.</p> - - <p>Similar to the <em>optional</em> cardinality class, the - accessor functions for the <em>sequence</em> class return - read-only (constant) and read-write references to the sequence - container. The modifier functions copies the entries from - the passed sequence.</p> - - <p>C++/Tree is a "flattening" mapping in a sense that many levels of - nested compositors (<code>choice</code> and <code>sequence</code>), - all potentially with their own cardinalities, are in the end mapped - to a flat set of elements with one of the three cardinality classes - discussed above. While this results in a simple and easy to use API - for most types, in certain cases, the order of elements in the actual - XML documents is not preserved once parsed into the object model. To - overcome this limitation we can mark certain schema types, for which - content order is not sufficiently preserved, as ordered. For more - information on this functionality refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.8.4">Section - 2.8.4, "Element Order"</a> in the C++/Tree Mapping User Manual.</p> - - <p>For complex schemas with many levels of nested compositors - (<code>choice</code> and <code>sequence</code>) it can also - be hard to deduce the cardinality class of a particular element. - The generated Doxygen documentation can greatly help with - this task. For each element and attribute the documentation - clearly identifies its cardinality class. Alternatively, you - can study the generated header files to find out the cardinality - class of a particular attribute or element.</p> - - <p>In the next sections we will examine how to access and modify - information stored in an object model using accessor and modifier - functions described in this section.</p> - - <h2><a name="4.2">4.2 Accessing the Object Model</a></h2> - - <p>In this section we will learn how to get to the information - stored in the object model for our person records vocabulary. - The following application accesses and prints the contents - of the <code>people.xml</code> file:</p> - - <pre class="c++"> -#include <iostream> -#include "people.hxx" - -using namespace std; - -int -main () -{ - unique_ptr<people_t> ppl (people ("people.xml")); - - // Iterate over individual person records. - // - people_t::person_sequence& ps (ppl->person ()); - - for (people_t::person_iterator i (ps.begin ()); i != ps.end (); ++i) - { - person_t& p (*i); - - // Print names: first-name and last-name are required elements, - // middle-name is optional. - // - cout << "name: " << p.first_name () << " "; - - if (p.middle_name ().present ()) - cout << p.middle_name ().get () << " "; - - cout << p.last_name () << endl; - - // Print gender, age, and id which are all required. - // - cout << "gender: " << p.gender () << endl - << "age: " << p.age () << endl - << "id: " << p.id () << endl - << endl; - } -} - </pre> - - <p>This code shows common patterns of accessing elements and attributes - with different cardinality classes. For the sequence element - (<code>person</code> in <code>people_t</code>) we first obtain a - reference to the container and then iterate over individual - records. The values of elements and attributes with the - <em>one</em> cardinality class (<code>first-name</code>, - <code>last-name</code>, <code>gender</code>, <code>age</code>, - and <code>id</code>) can be obtained directly by calling the - corresponding accessor functions. For the optional element - <code>middle-name</code> we first check if the value is present - and only then call <code>get()</code> to retrieve it.</p> - - <p>Note that when we want to reduce typing by creating a variable - representing a fragment of the object model that we are currently - working with (<code>ps</code> and <code>p</code> above), we obtain - a reference to that fragment instead of making a potentially - expensive copy. This is generally a good rule to follow when - creating high-performance applications.</p> - - <p>If we run the above application on our sample - <code>people.xml</code>, the output looks as follows:</p> - - <pre class="terminal"> -name: John Doe -gender: male -age: 32 -id: 1 - -name: Jane Mary Doe -gender: female -age: 28 -id: 2 - </pre> - - - <h2><a name="4.3">4.3 Modifying the Object Model</a></h2> - - <p>In this section we will learn how to modify the information - stored in the object model for our person records vocabulary. - The following application changes the contents of the - <code>people.xml</code> file:</p> - - <pre class="c++"> -#include <iostream> -#include "people.hxx" - -using namespace std; - -int -main () -{ - unique_ptr<people_t> ppl (people ("people.xml")); - - // Iterate over individual person records and increment - // the age. - // - people_t::person_sequence& ps (ppl->person ()); - - for (people_t::person_iterator i (ps.begin ()); i != ps.end (); ++i) - { - // Alternative way: i->age ()++; - // - i->age (i->age () + 1); - } - - // Add middle-name to the first record and remove it from - // the second. - // - person_t& john (ps[0]); - person_t& jane (ps[1]); - - john.middle_name ("Mary"); - jane.middle_name ().reset (); - - // Add another John record. - // - ps.push_back (john); - - // Serialize the modified object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "people.xsd"; - - people (cout, *ppl, map); -} - </pre> - - <p>The first modification the above application performs is iterating - over person records and incrementing the age value. This code - fragment shows how to modify the value of a required attribute - or element. The next modification shows how to set a new value - for the optional <code>middle-name</code> element as well - as clear its value. Finally the example adds a copy of the - John Doe record to the <code>person</code> element sequence.</p> - - <p>Note that in this case using references for the <code>ps</code>, - <code>john</code>, and <code>jane</code> variables is no longer - a performance improvement but a requirement for the application - to function correctly. If we hadn't used references, all our changes - would have been made on copies without affecting the object model.</p> - - <p>If we run the above application on our sample <code>people.xml</code>, - the output looks as follows:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd"> - - <person id="1"> - <first-name>John</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>33</age> - </person> - - <person id="2"> - <first-name>Jane</first-name> - <last-name>Doe</last-name> - <gender>female</gender> - <age>29</age> - </person> - - <person id="1"> - <first-name>John</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>33</age> - </person> - -</people> - </pre> - - - <h2><a name="4.4">4.4 Creating the Object Model from Scratch</a></h2> - - <p>In this section we will learn how to create a new object model - for our person records vocabulary. The following application - recreates the content of the original <code>people.xml</code> - file:</p> - - <pre class="c++"> -#include <iostream> -#include "people.hxx" - -using namespace std; - -int -main () -{ - people_t ppl; - people_t::person_sequence& ps (ppl.person ()); - - // Add the John Doe record. - // - ps.push_back ( - person_t ("John", // first-name - "Doe", // last-name - gender_t::male, // gender - 32, // age - 1)); - - // Add the Jane Doe record. - // - ps.push_back ( - person_t ("Jane", // first-name - "Doe", // last-name - gender_t::female, // gender - 28, // age - 2)); // id - - // Add middle name to the Jane Doe record. - // - person_t& jane (ps.back ()); - jane.middle_name ("Mary"); - - // Serialize the object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "people.xsd"; - - people (cout, ppl, map); -} - </pre> - - <p>The only new part in the above application is the calls - to the <code>people_t</code> and <code>person_t</code> - constructors. As a general rule, for each C++ class - XSD generates a constructor with initializers - for each element and attribute belonging to the <em>one</em> - cardinality class. For our vocabulary, the following - constructors are generated:</p> - - <pre class="c++"> -class person_t -{ - person_t (const first_name_type&, - const last_name_type&, - const gender_type&, - const age_type&, - const id_type&); -}; - -class people_t -{ - people_t (); -}; - </pre> - - <p>Note also that we set the <code>middle-name</code> element - on the Jane Doe record by obtaining a reference to that record - in the object model and setting the <code>middle-name</code> - value on it. This is a general rule that should be followed - in order to obtain the best performance: if possible, - direct modifications to the object model should be preferred - to modifications on temporaries with subsequent copying. The - following code fragment shows a semantically equivalent but - slightly slower version:</p> - - <pre class="c++"> -// Add the Jane Doe record. -// -person_t jane ("Jane", // first-name - "Doe", // last-name - gender_t::female, // gender - 28, // age - 2); // id - -jane.middle_name ("Mary"); - -ps.push_back (jane); - </pre> - - <p>We can also go one step further to reduce copying and improve - the performance of our application by using the non-copying - <code>push_back()</code> function which assumes ownership - of the passed objects:</p> - - <pre class="c++"> -// Add the Jane Doe record. C++11 version -// -unique_ptr<person_t> jane_p ( - new person_t ("Jane", // first-name - "Doe", // last-name - gender_t::female, // gender - 28, // age - 2)); // id -ps.push_back (std::move (jane_p)); // assumes ownership - -// Add the John Doe record. C++98 version. -// -auto_ptr<person_t> john_p ( - new person_t ("John", // first-name - "Doe", // last-name - gender_t::male, // gender - 32, // age - 1)); -ps.push_back (john_p); // assumes ownership - </pre> - - <p>For more information on the non-copying modifier functions refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.8">Section - 2.8, "Mapping for Local Elements and Attributes"</a> in the C++/Tree Mapping - User Manual. The above application produces the following output:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd"> - - <person id="1"> - <first-name>John</first-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>32</age> - </person> - - <person id="2"> - <first-name>Jane</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>female</gender> - <age>28</age> - </person> - -</people> - </pre> - - <h2><a name="4.5">4.5 Mapping for the Built-in XML Schema Types</a></h2> - - <p>Our person record vocabulary uses several built-in XML Schema - types: <code>string</code>, <code>short</code>, and - <code>unsignedInt</code>. Until now we haven't talked about - the mapping of built-in XML Schema types to C++ types and how - to work with them. This section provides an overview - of the built-in types. For more detailed information refer - to <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.5">Section - 2.5, "Mapping for Built-in Data Types"</a> in the C++/Tree Mapping - User Manual.</p> - - <p>In XML Schema, built-in types are defined in the XML Schema namespace. - By default, the C++/Tree mapping maps this namespace to C++ - namespace <code>xml_schema</code> (this mapping can be altered - with the <code>--namespace-map</code> option). The following table - summarizes the mapping of XML Schema built-in types to C++ types:</p> - - <!-- border="1" is necessary for html2ps --> - <table id="builtin" border="1"> - <tr> - <th>XML Schema type</th> - <th>Alias in the <code>xml_schema</code> namespace</th> - <th>C++ type</th> - </tr> - - <tr> - <th colspan="3">fixed-length integral types</th> - </tr> - <!-- 8-bit --> - <tr> - <td><code>byte</code></td> - <td><code>byte</code></td> - <td><code>signed char</code></td> - </tr> - <tr> - <td><code>unsignedByte</code></td> - <td><code>unsigned_byte</code></td> - <td><code>unsigned char</code></td> - </tr> - - <!-- 16-bit --> - <tr> - <td><code>short</code></td> - <td><code>short_</code></td> - <td><code>short</code></td> - </tr> - <tr> - <td><code>unsignedShort</code></td> - <td><code>unsigned_short</code></td> - <td><code>unsigned short</code></td> - </tr> - - <!-- 32-bit --> - <tr> - <td><code>int</code></td> - <td><code>int_</code></td> - <td><code>int</code></td> - </tr> - <tr> - <td><code>unsignedInt</code></td> - <td><code>unsigned_int</code></td> - <td><code>unsigned int</code></td> - </tr> - - <!-- 64-bit --> - <tr> - <td><code>long</code></td> - <td><code>long_</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>unsignedLong</code></td> - <td><code>unsigned_long</code></td> - <td><code>unsigned long long</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-length integral types</th> - </tr> - <tr> - <td><code>integer</code></td> - <td><code>integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonPositiveInteger</code></td> - <td><code>non_positive_integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonNegativeInteger</code></td> - <td><code>non_negative_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>positiveInteger</code></td> - <td><code>positive_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>negativeInteger</code></td> - <td><code>negative_integer</code></td> - <td><code>long long</code></td> - </tr> - - <tr> - <th colspan="3">boolean types</th> - </tr> - <tr> - <td><code>boolean</code></td> - <td><code>boolean</code></td> - <td><code>bool</code></td> - </tr> - - <tr> - <th colspan="3">fixed-precision floating-point types</th> - </tr> - <tr> - <td><code>float</code></td> - <td><code>float_</code></td> - <td><code>float</code></td> - </tr> - <tr> - <td><code>double</code></td> - <td><code>double_</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-precision floating-point types</th> - </tr> - <tr> - <td><code>decimal</code></td> - <td><code>decimal</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">string types</th> - </tr> - <tr> - <td><code>string</code></td> - <td><code>string</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - <tr> - <td><code>normalizedString</code></td> - <td><code>normalized_string</code></td> - <td>type derived from <code>string</code></td> - </tr> - <tr> - <td><code>token</code></td> - <td><code>token</code></td> - <td>type derived from <code>normalized_string</code></td> - </tr> - <tr> - <td><code>Name</code></td> - <td><code>name</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKEN</code></td> - <td><code>nmtoken</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKENS</code></td> - <td><code>nmtokens</code></td> - <td>type derived from <code>sequence<nmtoken></code></td> - </tr> - <tr> - <td><code>NCName</code></td> - <td><code>ncname</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>language</code></td> - <td><code>language</code></td> - <td>type derived from <code>token</code></td> - </tr> - - <tr> - <th colspan="3">qualified name</th> - </tr> - <tr> - <td><code>QName</code></td> - <td><code>qname</code></td> - <td><code>xml_schema::qname</code></td> - </tr> - - <tr> - <th colspan="3">ID/IDREF types</th> - </tr> - <tr> - <td><code>ID</code></td> - <td><code>id</code></td> - <td>type derived from <code>ncname</code></td> - </tr> - <tr> - <td><code>IDREF</code></td> - <td><code>idref</code></td> - <td>type derived from <code>ncname</code></td> - </tr> - <tr> - <td><code>IDREFS</code></td> - <td><code>idrefs</code></td> - <td>type derived from <code>sequence<idref></code></td> - </tr> - - <tr> - <th colspan="3">URI types</th> - </tr> - <tr> - <td><code>anyURI</code></td> - <td><code>uri</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - - <tr> - <th colspan="3">binary types</th> - </tr> - <tr> - <td><code>base64Binary</code></td> - <td><code>base64_binary</code></td> - <td><code>xml_schema::base64_binary</code></td> - </tr> - <tr> - <td><code>hexBinary</code></td> - <td><code>hex_binary</code></td> - <td><code>xml_schema::hex_binary</code></td> - </tr> - - <tr> - <th colspan="3">date/time types</th> - </tr> - <tr> - <td><code>date</code></td> - <td><code>date</code></td> - <td><code>xml_schema::date</code></td> - </tr> - <tr> - <td><code>dateTime</code></td> - <td><code>date_time</code></td> - <td><code>xml_schema::date_time</code></td> - </tr> - <tr> - <td><code>duration</code></td> - <td><code>duration</code></td> - <td><code>xml_schema::duration</code></td> - </tr> - <tr> - <td><code>gDay</code></td> - <td><code>gday</code></td> - <td><code>xml_schema::gday</code></td> - </tr> - <tr> - <td><code>gMonth</code></td> - <td><code>gmonth</code></td> - <td><code>xml_schema::gmonth</code></td> - </tr> - <tr> - <td><code>gMonthDay</code></td> - <td><code>gmonth_day</code></td> - <td><code>xml_schema::gmonth_day</code></td> - </tr> - <tr> - <td><code>gYear</code></td> - <td><code>gyear</code></td> - <td><code>xml_schema::gyear</code></td> - </tr> - <tr> - <td><code>gYearMonth</code></td> - <td><code>gyear_month</code></td> - <td><code>xml_schema::gyear_month</code></td> - </tr> - <tr> - <td><code>time</code></td> - <td><code>time</code></td> - <td><code>xml_schema::time</code></td> - </tr> - - <tr> - <th colspan="3">entity types</th> - </tr> - <tr> - <td><code>ENTITY</code></td> - <td><code>entity</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>ENTITIES</code></td> - <td><code>entities</code></td> - <td>type derived from <code>sequence<entity></code></td> - </tr> - </table> - - <p>As you can see from the table above a number of built-in - XML Schema types are mapped to fundamental C++ types such - as <code>int</code> or <code>bool</code>. All string-based - XML Schema types are mapped to C++ types that are derived - from either <code>std::string</code> or - <code>std::wstring</code>, depending on the character - type selected. For access and modification purposes these - types can be treated as <code>std::string</code>. A number - of built-in types, such as <code>qname</code>, the binary - types, and the date/time types do not have suitable - fundamental or standard C++ types to map to. As a result, - these types are implemented from scratch in the XSD runtime. - For more information on their interfaces refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.5">Section - 2.5, "Mapping for Built-in Data Types"</a> in the C++/Tree Mapping - User Manual.</p> - - - <!-- Chapater 5 --> - - - <h1><a name="5">5 Parsing</a></h1> - - <p>We have already seen how to parse XML to an object model in this guide - before. In this chapter we will discuss the parsing topic in more - detail.</p> - - <p>By default, the C++/Tree mapping provides a total of 14 overloaded - parsing functions. They differ in the input methods used to - read XML as well as the error reporting mechanisms. It is also possible - to generate types for root elements instead of parsing and serialization - functions. This may be useful if your XML vocabulary has multiple - root elements. For more information on element types refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.9">Section - 2.9, "Mapping for Global Elements"</a> in the C++/Tree Mapping User - Manual.</p> - - - <p>In this section we will discuss the most commonly used versions of - the parsing functions. For a comprehensive description of parsing - refer to <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#3">Chapter - 3, "Parsing"</a> in the C++/Tree Mapping User Manual. For the <code>people</code> - global element from our person record vocabulary, we will concentrate - on the following three parsing functions:</p> - - <pre class="c++"> -std::[unique|auto]_ptr<people_t> -people (const std::string& uri, - xml_schema::flags f = 0, - const xml_schema::properties& p = xml_schema::properties ()); - -std::[unique|auto]_ptr<people_t> -people (std::istream& is, - xml_schema::flags f = 0, - const xml_schema::properties& p = xml_schema::properties ()); - -std::[unique|auto]_ptr<people_t> -people (std::istream& is, - const std::string& resource_id, - xml_schema::flags f = 0, - const xml_schema::properties& p = ::xml_schema::properties ()); - </pre> - - <p>The first function parses a local file or a URI. We have already - used this parsing function in the previous chapters. The second - and third functions read XML from a standard input stream. The - last function also requires a resource id. This id is used to - identify the XML document being parser in diagnostics messages - as well as to resolve relative paths to other documents (for example, - schemas) that might be referenced from the XML document.</p> - - <p>The last two arguments to all three parsing functions are parsing - flags and properties. The flags argument provides a number of ways - to fine-tune the parsing process. The properties argument allows - to pass additional information to the parsing functions. We will - use these two arguments in <a href="#5.1">Section 5.1, "XML Schema - Validation and Searching"</a> below. All three functions return - the object model as either <code>std::unique_ptr</code> (C++11) or - <code>std::auto_ptr</code> (C++98), depending on the C++ standard - selected (<code>--std</code> XSD compiler option). The following - example shows how we can use the above parsing functions:</p> - - <pre class="c++"> -using std::unique_ptr; - -// Parse a local file or URI. -// -unique_ptr<people_t> p1 (people ("people.xml")); -unique_ptr<people_t> p2 (people ("http://example.com/people.xml")); - -// Parse a local file via ifstream. -// -std::ifstream ifs ("people.xml"); -unique_ptr<people_t> p3 (people (ifs, "people.xml")); - -// Parse an XML string. -// -std::string str ("..."); // XML in a string. -std::istringstream iss (str); -unique_ptr<people_t> p4 (people (iss)); - </pre> - - - <h2><a name="5.1">5.1 XML Schema Validation and Searching</a></h2> - - <p>The C++/Tree mapping relies on the underlying Xerces-C++ XML - parser for full XML document validation. The XML Schema - validation is enabled by default and can be disabled by - passing the <code>xml_schema::flags::dont_validate</code> - flag to the parsing functions, for example:</p> - - <pre class="c++"> -unique_ptr<people_t> p ( - people ("people.xml", xml_schema::flags::dont_validate)); - </pre> - - <p>Even when XML Schema validation is disabled, the generated - code still performs a number of checks to prevent - construction of an inconsistent object model (for example, an - object model with missing required attributes or elements).</p> - - <p>When XML Schema validation is enabled, the XML parser needs - to locate a schema to validate against. There are several - methods to provide the schema location information to the - parser. The easiest and most commonly used method is to - specify schema locations in the XML document itself - with the <code>schemaLocation</code> or - <code>noNamespaceSchemaLocation</code> attributes, for example:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd" - xsi:schemaLocation="http://www.w3.org/XML/1998/namespace xml.xsd"> - </pre> - - <p>As you might have noticed, we used this method in all the sample XML - documents presented in this guide up until now. Note that the - schema locations specified with these two attributes are relative - to the document's path unless they are absolute URIs (that is - start with <code>http://</code>, <code>file://</code>, etc.). - In particular, if you specify just file names as your schema - locations, as we did above, then the schemas should reside in - the same directory as the XML document itself.</p> - - <p>Another method of providing the schema location information - is via the <code>xml_schema::properties</code> argument, as - shown in the following example:</p> - - <pre class="c++"> -xml_schema::properties props; -props.no_namespace_schema_location ("people.xsd"); -props.schema_location ("http://www.w3.org/XML/1998/namespace", "xml.xsd"); - -unique_ptr<people_t> p (people ("people.xml", 0, props)); - </pre> - - <p>The schema locations provided with this method overrides - those specified in the XML document. As with the previous - method, the schema locations specified this way are - relative to the document's path unless they are absolute URIs. - In particular, if you want to use local schemas that are - not related to the document being parsed, then you will - need to use the <code>file://</code> URI. The following - example shows how to use schemas that reside in the current - working directory:</p> - - <pre class="c++"> -#include <unistd.h> // getcwd -#include <limits.h> // PATH_MAX - -char cwd[PATH_MAX]; -if (getcwd (cwd, PATH_MAX) == 0) -{ - // Buffer too small? -} - -xml_schema::properties props; - -props.no_namespace_schema_location ( - "file:///" + std::string (cwd) + "/people.xsd"); - -props.schema_location ( - "http://www.w3.org/XML/1998/namespace", - "file:///" + std::string (cwd) + "/xml.xsd"); - -unique_ptr<people_t> p (people ("people.xml", 0, props)); - </pre> - - <p>A third method is the most useful if you are planning to parse - several XML documents of the same vocabulary. In that case - it may be beneficial to pre-parse and cache the schemas in - the XML parser which can then be used to parse all documents - without re-parsing the schemas. For more information on - this method refer to the <code>caching</code> example in the - <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package. - It is also possible to convert the schemas into a pre-compiled - binary representation and embed this representation directly into - the application executable. With this approach your application can - perform XML Schema validation without depending on any external - schema files. For more information on how to achieve this refer to - the <code>embedded</code> example in the <code>cxx/tree/</code> - directory in the <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <p>When the XML parser cannot locate a schema for the - XML document, the validation fails and XML document - elements and attributes for which schema definitions could - not be located are reported in the diagnostics. For - example, if we remove the <code>noNamespaceSchemaLocation</code> - attribute in <code>people.xml</code> from the previous chapter, - then we will get the following diagnostics if we try to parse - this file with validation enabled:</p> - - <pre class="terminal"> -people.xml:2:63 error: no declaration found for element 'people' -people.xml:4:18 error: no declaration found for element 'person' -people.xml:4:18 error: attribute 'id' is not declared for element 'person' -people.xml:5:17 error: no declaration found for element 'first-name' -people.xml:6:18 error: no declaration found for element 'middle-name' -people.xml:7:16 error: no declaration found for element 'last-name' -people.xml:8:13 error: no declaration found for element 'gender' -people.xml:9:10 error: no declaration found for element 'age' - </pre> - - <h2><a name="5.2">5.2 Error Handling</a></h2> - - <p>The parsing functions offer a number of ways to handle error conditions - with the C++ exceptions being the most commonly used mechanism. All - C++/Tree exceptions derive from common base <code>xml_schema::exception</code> - which in turn derives from <code>std::exception</code>. The easiest - way to uniformly handle all possible C++/Tree exceptions and print - detailed information about the error is to catch and print - <code>xml_schema::exception</code>, as shown in the following - example:</p> - - <pre class="c++"> -try -{ - unique_ptr<people_t> p (people ("people.xml")); -} -catch (const xml_schema::exception& e) -{ - cerr << e << endl; -} - </pre> - - <p>Each individual C++/Tree exception also allows you to obtain - error details programmatically. For example, the - <code>xml_schema::parsing</code> exception is thrown when - the XML parsing and validation in the underlying XML parser - fails. It encapsulates various diagnostics information - such as the file name, line and column numbers, as well as the - error or warning message for each entry. For more information - about this and other exceptions that can be thrown during - parsing, refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#3.3">Section - 3.3, "Error Handling"</a> in the C++/Tree Mapping - User Manual.</p> - - <p>Note that if you are parsing <code>std::istream</code> on which - exceptions are not enabled, then you will need to check the - stream state after the call to the parsing function in order - to detect any possible stream failures, for example:</p> - - <pre class="c++"> -std::ifstream ifs ("people.xml"); - -if (ifs.fail ()) -{ - cerr << "people.xml: unable to open" << endl; - return 1; -} - -unique_ptr<people_t> p (people (ifs, "people.xml")); - -if (ifs.fail ()) -{ - cerr << "people.xml: read error" << endl; - return 1; -} - </pre> - - <p>The above example can be rewritten to use exceptions as - shown below:</p> - - <pre class="c++"> -try -{ - std::ifstream ifs; - ifs.exceptions (std::ifstream::badbit | std::ifstream::failbit); - ifs.open ("people.xml"); - - unique_ptr<people_t> p (people (ifs, "people.xml")); -} -catch (const std::ifstream::failure&) -{ - cerr << "people.xml: unable to open or read error" << endl; - return 1; -} - </pre> - - - <!-- Chapater 6 --> - - - <h1><a name="6">6 Serialization</a></h1> - - <p>We have already seen how to serialize an object model back to XML - in this guide before. In this chapter we will discuss the - serialization topic in more detail.</p> - - <p>By default, the C++/Tree mapping provides a total of 8 overloaded - serialization functions. They differ in the output methods used to write - XML as well as the error reporting mechanisms. It is also possible to - generate types for root elements instead of parsing and serialization - functions. This may be useful if your XML vocabulary has multiple - root elements. For more information on element types refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.9">Section - 2.9, "Mapping for Global Elements"</a> in the C++/Tree Mapping User - Manual.</p> - - - <p>In this section we will discuss the most commonly - used version of serialization functions. For a comprehensive description - of serialization refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#4">Chapter - 4, "Serialization"</a> in the C++/Tree Mapping User Manual. For the - <code>people</code> global element from our person record vocabulary, - we will concentrate on the following serialization function:</p> - - <pre class="c++"> -void -people (std::ostream& os, - const people_t& x, - const xml_schema::namespace_infomap& map = - xml_schema::namespace_infomap (), - const std::string& encoding = "UTF-8", - xml_schema::flags f = 0); - </pre> - - <p>This function serializes the object model passed as the second - argument to the standard output stream passed as the first - argument. The third argument is a namespace information map - which we will discuss in more detail in the next section. - The fourth argument is a character encoding that the resulting - XML document should be in. Possible valid values for this - argument are "US-ASCII", "ISO8859-1", "UTF-8", "UTF-16BE", - "UTF-16LE", "UCS-4BE", and "UCS-4LE". Finally, the flags - argument allows fine-tuning of the serialization process. - The following example shows how we can use the above serialization - function:</p> - - <pre class="c++"> -people_t& p = ... - -xml_schema::namespace_infomap map; -map[""].schema = "people.xsd"; - -// Serialize to stdout. -// -people (std::cout, p, map); - -// Serialize to a file. -// -std::ofstream ofs ("people.xml"); -people (ofs, p, map); - -// Serialize to a string. -// -std::ostringstream oss; -people (oss, p, map); -std::string xml (oss.str ()); - </pre> - - - <h2><a name="6.1">6.1 Namespace and Schema Information</a></h2> - - <p>While XML serialization can be done just from the object - model alone, it is often desirable to assign meaningful - prefixes to XML namespaces used in the vocabulary as - well as to provide the schema location information. - This is accomplished by passing the namespace information - map to the serialization function. The key in this map is - a namespace prefix that should be assigned to an XML namespace - specified in the <code>name</code> variable of the - map value. You can also assign an optional schema location for - this namespace in the <code>schema</code> variable. Based - on each key-value entry in this map, the serialization - function adds two attributes to the resulting XML document: - the namespace-prefix mapping attribute and schema location - attribute. The empty prefix indicates that the namespace - should be mapped without a prefix. For example, the following - map:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].name = "http://www.example.com/example"; -map[""].schema = "example.xsd"; - -map["x"].name = "http://www.w3.org/XML/1998/namespace"; -map["x"].schema = "xml.xsd"; - </pre> - - <p>Results in the following XML document:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<example - xmlns="http://www.example.com/example" - xmlns:x="http://www.w3.org/XML/1998/namespace" - xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:schemaLocation="http://www.example.com/example example.xsd - http://www.w3.org/XML/1998/namespace xml.xsd"> - </pre> - - <p>The empty namespace indicates that the vocabulary has no target - namespace. For example, the following map results in only the - <code>noNamespaceSchemaLocation</code> attribute being added:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].name = ""; -map[""].schema = "example.xsd"; - </pre> - - <h2><a name="6.2">6.2 Error Handling</a></h2> - - <p>Similar to the parsing functions, the serialization functions offer a - number of ways to handle error conditions with the C++ exceptions being - the most commonly used mechanisms. As with parsing, the easiest way to - uniformly handle all possible serialization exceptions and print - detailed information about the error is to catch and print - <code>xml_schema::exception</code>:</p> - - <pre class="c++"> -try -{ - people_t& p = ... - - xml_schema::namespace_infomap map; - map[""].schema = "people.xsd"; - - people (std::cout, p, map)); -} -catch (const xml_schema::exception& e) -{ - cerr << e << endl; -} - </pre> - - <p>The most commonly encountered serialization exception is - <code>xml_schema::serialization</code>. It is thrown - when the XML serialization in the underlying XML writer - fails. It encapsulates various diagnostics information - such as the file name, line and column numbers, as well as the - error or warning message for each entry. For more information - about this and other exceptions that can be thrown during - serialization, refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#4.4">Section - 4.4, "Error Handling"</a> in the C++/Tree Mapping - User Manual.</p> - - <p>Note that if you are serializing to <code>std::ostream</code> on - which exceptions are not enabled, then you will need to check the - stream state after the call to the serialization function in order - to detect any possible stream failures, for example:</p> - - <pre class="c++"> -std::ofstream ofs ("people.xml"); - -if (ofs.fail ()) -{ - cerr << "people.xml: unable to open" << endl; - return 1; -} - -people (ofs, p, map)); - -if (ofs.fail ()) -{ - cerr << "people.xml: write error" << endl; - return 1; -} - </pre> - - <p>The above example can be rewritten to use exceptions as - shown below:</p> - - <pre class="c++"> -try -{ - std::ofstream ofs; - ofs.exceptions (std::ofstream::badbit | std::ofstream::failbit); - ofs.open ("people.xml"); - - people (ofs, p, map)); -} -catch (const std::ofstream::failure&) -{ - cerr << "people.xml: unable to open or write error" << endl; - return 1; -} - </pre> - - </div> -</div> - -</body> -</html> diff --git a/doc/cxx/tree/guide/index.xhtml.in b/doc/cxx/tree/guide/index.xhtml.in deleted file mode 100644 index 2f7f1e2..0000000 --- a/doc/cxx/tree/guide/index.xhtml.in +++ /dev/null @@ -1,2736 +0,0 @@ -<?xml version="1.0" encoding="iso-8859-1"?> -<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> -<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"> - -<head> - <title>C++/Tree Mapping Getting Started Guide</title> - - <meta name="copyright" content="© @copyright@"/> - <meta name="keywords" content="xsd,xml,schema,c++,mapping,data,binding,parsing,serialization,validation"/> - <meta name="description" content="C++/Tree Mapping Getting Started Guide"/> - - <link rel="stylesheet" type="text/css" href="../../../default.css" /> - -<style type="text/css"> - pre { - padding : 0 0 0 0em; - margin : 0em 0em 0em 0; - - font-size : 102% - } - - body { - min-width: 48em; - } - - h1 { - font-weight: bold; - font-size: 200%; - line-height: 1.2em; - } - - h2 { - font-weight : bold; - font-size : 150%; - - padding-top : 0.8em; - } - - h3 { - font-size : 140%; - padding-top : 0.8em; - } - - /* Adjust indentation for three levels. */ - #container { - max-width: 48em; - } - - #content { - padding: 0 0.1em 0 4em; - /*background-color: red;*/ - } - - #content h1 { - margin-left: -2.06em; - } - - #content h2 { - margin-left: -1.33em; - } - - /* Title page */ - - #titlepage { - padding: 2em 0 1em 0; - border-bottom: 1px solid black; - } - - #titlepage .title { - font-weight: bold; - font-size: 200%; - text-align: center; - } - - #titlepage #first-title { - padding: 1em 0 0.4em 0; - } - - #titlepage #second-title { - padding: 0.4em 0 2em 0; - } - - /* Lists */ - ul.list li { - padding-top : 0.3em; - padding-bottom : 0.3em; - } - - div.img { - text-align: center; - padding: 2em 0 2em 0; - } - - /* */ - dl dt { - padding : 0.8em 0 0 0; - } - - /* Built-in table */ - #builtin { - margin: 2em 0 2em 0; - - border-collapse : collapse; - border : 1px solid; - border-color : #000000; - - font-size : 11px; - line-height : 14px; - } - - #builtin th, #builtin td { - border: 1px solid; - padding : 0.9em 0.9em 0.7em 0.9em; - } - - #builtin th { - background : #cde8f6; - } - - #builtin td { - text-align: left; - } - - /* TOC */ - table.toc { - border-style : none; - border-collapse : separate; - border-spacing : 0; - - margin : 0.2em 0 0.2em 0; - padding : 0 0 0 0; - } - - table.toc tr { - padding : 0 0 0 0; - margin : 0 0 0 0; - } - - table.toc * td, table.toc * th { - border-style : none; - margin : 0 0 0 0; - vertical-align : top; - } - - table.toc * th { - font-weight : normal; - padding : 0em 0.1em 0em 0; - text-align : left; - white-space : nowrap; - } - - table.toc * table.toc th { - padding-left : 1em; - } - - table.toc * td { - padding : 0em 0 0em 0.7em; - text-align : left; - } -</style> - - -</head> - -<body> -<div id="container"> - <div id="content"> - - <div class="noprint"> - - <div id="titlepage"> - <div class="title" id="first-title">C++/Tree Mapping</div> - <div class="title" id="second-title">Getting Started Guide</div> - - <p>Copyright © @copyright@.</p> - - <p>Permission is granted to copy, distribute and/or modify this - document under the terms of the - <a href="https://www.codesynthesis.com/licenses/fdl-1.2.txt">GNU Free - Documentation License, version 1.2</a>; with no Invariant Sections, - no Front-Cover Texts and no Back-Cover Texts. - </p> - - <p>This document is available in the following formats: - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/index.xhtml">XHTML</a>, - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/cxx-tree-guide.pdf">PDF</a>, and - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/cxx-tree-guide.ps">PostScript</a>.</p> - - </div> - - <h1>Table of Contents</h1> - - <table class="toc"> - <tr> - <th></th><td><a href="#0">Preface</a> - <table class="toc"> - <tr><th></th><td><a href="#0.1">About This Document</a></td></tr> - <tr><th></th><td><a href="#0.2">More Information</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>1</th><td><a href="#1">Introduction</a> - <table class="toc"> - <tr><th>1.1</th><td><a href="#1.1">Mapping Overview</a></td></tr> - <tr><th>1.2</th><td><a href="#1.2">Benefits</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>2</th><td><a href="#2">Hello World Example</a> - <table class="toc"> - <tr><th>2.1</th><td><a href="#2.1">Writing XML Document and Schema</a></td></tr> - <tr><th>2.2</th><td><a href="#2.2">Translating Schema to C++</a></td></tr> - <tr><th>2.3</th><td><a href="#2.3">Implementing Application Logic</a></td></tr> - <tr><th>2.4</th><td><a href="#2.4">Compiling and Running</a></td></tr> - <tr><th>2.5</th><td><a href="#2.5">Adding Serialization</a></td></tr> - <tr><th>2.6</th><td><a href="#2.6">Selecting Naming Convention</a></td></tr> - <tr><th>2.7</th><td><a href="#2.7">Generating Documentation</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>3</th><td><a href="#3">Overall Mapping Configuration</a> - <table class="toc"> - <tr><th>3.1</th><td><a href="#3.1">C++ Standard</a></td></tr> - <tr><th>3.2</th><td><a href="#3.2">Character Type and Encoding</a></td></tr> - <tr><th>3.3</th><td><a href="#3.3">Support for Polymorphism </a></td></tr> - <tr><th>3.4</th><td><a href="#3.4">Namespace Mapping</a></td></tr> - <tr><th>3.5</th><td><a href="#3.5">Thread Safety</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>4</th><td><a href="#4">Working with Object Models</a> - <table class="toc"> - <tr><th>4.1</th><td><a href="#4.1">Attribute and Element Cardinalities</a></td></tr> - <tr><th>4.2</th><td><a href="#4.2">Accessing the Object Model</a></td></tr> - <tr><th>4.3</th><td><a href="#4.3">Modifying the Object Model</a></td></tr> - <tr><th>4.4</th><td><a href="#4.4">Creating the Object Model from Scratch</a></td></tr> - <tr><th>4.5</th><td><a href="#4.5">Mapping for the Built-in XML Schema Types</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>5</th><td><a href="#5">Parsing</a> - <table class="toc"> - <tr><th>5.1</th><td><a href="#5.1">XML Schema Validation and Searching</a></td></tr> - <tr><th>5.2</th><td><a href="#5.2">Error Handling</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>6</th><td><a href="#6">Serialization</a> - <table class="toc"> - <tr><th>6.1</th><td><a href="#6.1">Namespace and Schema Information</a></td></tr> - <tr><th>6.2</th><td><a href="#6.2">Error Handling</a></td></tr> - </table> - </td> - </tr> - - </table> - </div> - - <h1><a name="0">Preface</a></h1> - - <h2><a name="0.1">About This Document</a></h2> - - <p>The goal of this document is to provide you with an understanding of - the C++/Tree programming model and allow you to efficiently evaluate - XSD against your project's technical requirements. As such, this - document is intended for C++ developers and software architects - who are looking for an XML processing solution. For a more in-depth - description of the C++/Tree mapping refer to the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/">C++/Tree - Mapping User Manual</a>.</p> - - <p>Prior experience with XML and C++ is required to understand this - document. Basic understanding of XML Schema is advantageous but - not expected or required. - </p> - - - <h2><a name="0.2">More Information</a></h2> - - <p>Beyond this guide, you may also find the following sources of - information useful:</p> - - <ul class="list"> - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/">C++/Tree - Mapping User Manual</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/Customization_guide">C++/Tree - Mapping Customization Guide</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree - Mapping Frequently Asked Questions (FAQ)</a></li> - - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a></li> - - <li>The <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - contains a collection of examples and a README file with an overview - of each example.</li> - - <li>The <code>README</code> file in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - explains how to build the examples.</li> - - <li>The <a href="https://www.codesynthesis.com/mailman/listinfo/xsd-users">xsd-users</a> - mailing list is the place to ask technical questions about XSD and the C++/Parser mapping. - Furthermore, the <a href="https://www.codesynthesis.com/pipermail/xsd-users/">archives</a> - may already have answers to some of your questions.</li> - </ul> - - <!-- Introduction --> - - <h1><a name="1">1 Introduction</a></h1> - - <p>Welcome to CodeSynthesis XSD and the C++/Tree mapping. XSD is a - cross-platform W3C XML Schema to C++ data binding compiler. C++/Tree - is a W3C XML Schema to C++ mapping that represents the data stored - in XML as a statically-typed, vocabulary-specific object model. - </p> - - <h2><a name="1.1">1.1 Mapping Overview</a></h2> - - <p>Based on a formal description of an XML vocabulary (schema), the - C++/Tree mapping produces a tree-like data structure suitable for - in-memory processing. The core of the mapping consists of C++ - classes that constitute the object model and are derived from - types defined in XML Schema as well as XML parsing and - serialization code.</p> - - <p>Besides the core features, C++/Tree provide a number of additional - mapping elements that can be useful in some applications. These - include serialization and extraction to/from formats others than - XML, such as unstructured text (useful for debugging) and binary - representations such as XDR and CDR for high-speed data processing - as well as automatic documentation generation. The C++/Tree mapping - also provides a wide range of mechanisms for controlling and - customizing the generated code.</p> - - <p>A typical application that uses C++/Tree for XML processing usually - performs the following three steps: it first reads (parses) an XML - document to an in-memory object model, it then performs some useful - computations on that object model which may involve modification - of the model, and finally it may write (serialize) the modified - object model back to XML.</p> - - <p>The next chapter presents a simple application that performs these - three steps. The following chapters show how to use the C++/Tree - mapping in more detail.</p> - - <h2><a name="1.2">1.2 Benefits</a></h2> - - <p>Traditional XML access APIs such as Document Object Model (DOM) - or Simple API for XML (SAX) have a number of drawbacks that - make them less suitable for creating robust and maintainable - XML processing applications. These drawbacks include: - </p> - - <ul class="list"> - <li>Generic representation of XML in terms of elements, attributes, - and text forces an application developer to write a substantial - amount of bridging code that identifies and transforms pieces - of information encoded in XML to a representation more suitable - for consumption by the application logic.</li> - - <li>String-based flow control defers error detection to runtime. - It also reduces code readability and maintainability.</li> - - <li>Lack of type safety because the data is represented as text.</li> - - <li>Resulting applications are hard to debug, change, and - maintain.</li> - </ul> - - <p>In contrast, statically-typed, vocabulary-specific object model - produced by the C++/Tree mapping allows you to operate in your - domain terms instead of the generic elements, attributes, and - text. Static typing helps catch errors at compile-time rather - than at run-time. Automatic code generation frees you for more - interesting tasks (such as doing something useful with the - information stored in the XML documents) and minimizes the - effort needed to adapt your applications to changes in the - document structure. To summarize, the C++/Tree object model has - the following key advantages over generic XML access APIs:</p> - - <ul class="list"> - <li><b>Ease of use.</b> The generated code hides all the complexity - associated with parsing and serializing XML. This includes navigating - the structure and converting between the text representation and - data types suitable for manipulation by the application - logic.</li> - - <li><b>Natural representation.</b> The object representation allows - you to access the XML data using your domain vocabulary instead - of generic elements, attributes, and text.</li> - - <li><b>Concise code.</b> With the object representation the - application implementation is simpler and thus easier - to read and understand.</li> - - <li><b>Safety.</b> The generated object model is statically - typed and uses functions instead of strings to access the - information. This helps catch programming errors at compile-time - rather than at runtime.</li> - - <li><b>Maintainability.</b> Automatic code generation minimizes the - effort needed to adapt the application to changes in the - document structure. With static typing, the C++ compiler - can pin-point the places in the client code that need to be - changed.</li> - - <li><b>Compatibility.</b> Sequences of elements are represented in - the object model as containers conforming to the standard C++ - sequence requirements. This makes it possible to use standard - C++ algorithms on the object representation and frees you from - learning yet another container interface, as is the case with - DOM.</li> - - <li><b>Efficiency.</b> If the application makes repetitive use - of the data extracted from XML, then the C++/Tree object model - is more efficient because the navigation is performed using - function calls rather than string comparisons and the XML - data is extracted only once. Furthermore, the runtime memory - usage is reduced due to more efficient data storage - (for instance, storing numeric data as integers instead of - strings) as well as the static knowledge of cardinality - constraints.</li> - </ul> - - - <!-- Hello World Parser --> - - - <h1><a name="2">2 Hello World Example</a></h1> - - <p>In this chapter we will examine how to parse, access, modify, and - serialize a very simple XML document using the XSD-generated - C++/Tree object model. The code presented in this chapter is - based on the <code>hello</code> example which can be found in - the <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package.</p> - - <h2><a name="2.1">2.1 Writing XML Document and Schema</a></h2> - - <p>First, we need to get an idea about the structure - of the XML documents we are going to process. Our - <code>hello.xml</code>, for example, could look like this:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello> - - <greeting>Hello</greeting> - - <name>sun</name> - <name>moon</name> - <name>world</name> - -</hello> - </pre> - - <p>Then we can write a description of the above XML in the - XML Schema language and save it into <code>hello.xsd</code>:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:complexType name="hello_t"> - <xs:sequence> - <xs:element name="greeting" type="xs:string"/> - <xs:element name="name" type="xs:string" maxOccurs="unbounded"/> - </xs:sequence> - </xs:complexType> - - <xs:element name="hello" type="hello_t"/> - -</xs:schema> - </pre> - - <p>Even if you are not familiar with XML Schema, it - should be easy to connect declarations in <code>hello.xsd</code> - to elements in <code>hello.xml</code>. The <code>hello_t</code> type - is defined as a sequence of the nested <code>greeting</code> and - <code>name</code> elements. Note that the term sequence in XML - Schema means that elements should appear in a particular order - as opposed to appearing multiple times. The <code>name</code> - element has its <code>maxOccurs</code> property set to - <code>unbounded</code> which means it can appear multiple times - in an XML document. Finally, the globally-defined <code>hello</code> - element prescribes the root element for our vocabulary. For an - easily-approachable introduction to XML Schema refer to - <a href="http://www.w3.org/TR/xmlschema-0/">XML Schema Part 0: - Primer</a>.</p> - - <p>The above schema is a specification of our XML vocabulary; it tells - everybody what valid documents of our XML-based language should look - like. We can also update our <code>hello.xml</code> to include the - information about the schema so that XML parsers can validate - our document:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="hello.xsd"> - - <greeting>Hello</greeting> - - <name>sun</name> - <name>moon</name> - <name>world</name> - -</hello> - </pre> - - - <p>The next step is to compile the schema to generate the object - model and parsing functions.</p> - - <h2><a name="2.2">2.2 Translating Schema to C++</a></h2> - - <p>Now we are ready to translate our <code>hello.xsd</code> to C++. - To do this we invoke the XSD compiler from a terminal (UNIX) or - a command prompt (Windows): - </p> - - <pre class="terminal"> -$ xsd cxx-tree hello.xsd - </pre> - - <p>The XSD compiler produces two C++ files: <code>hello.hxx</code> and - <code>hello.cxx</code>. The following code fragment is taken from - <code>hello.hxx</code>; it should give you an idea about what gets - generated: - </p> - - <pre class="c++"> -class hello_t -{ -public: - // greeting - // - typedef xml_schema::string greeting_type; - - const greeting_type& - greeting () const; - - greeting_type& - greeting (); - - void - greeting (const greeting_type& x); - - // name - // - typedef xml_schema::string name_type; - typedef xsd::sequence<name_type> name_sequence; - typedef name_sequence::iterator name_iterator; - typedef name_sequence::const_iterator name_const_iterator; - - const name_sequence& - name () const; - - name_sequence& - name (); - - void - name (const name_sequence& s); - - // Constructor. - // - hello_t (const greeting_type&); - - ... - -}; - -std::unique_ptr<hello_t> -hello (const std::string& uri); - -std::unique_ptr<hello_t> -hello (std::istream&); - </pre> - - <p>The <code>hello_t</code> C++ class corresponds to the - <code>hello_t</code> XML Schema type. For each element - in this type a set of C++ type definitions as well as - accessor and modifier functions are generated inside the - <code>hello_t</code> class. Note that the type definitions - and member functions for the <code>greeting</code> and - <code>name</code> elements are different because of the - cardinality differences between these two elements - (<code>greeting</code> is a required single element and - <code>name</code> is a sequence of elements).</p> - - <p>The <code>xml_schema::string</code> type used in the type - definitions is a C++ class provided by the XSD runtime - that corresponds to built-in XML Schema type - <code>string</code>. The <code>xml_schema::string</code> - is based on <code>std::string</code> and can be used as - such. Similarly, the <code>sequence</code> class template - that is used in the <code>name_sequence</code> type - definition is based on and has the same interface as - <code>std::vector</code>. The mapping between the built-in - XML Schema types and C++ types is described in more detail in - <a href="#4.5">Section 4.5, "Mapping for the Built-in XML Schema - Types"</a>. The <code>hello_t</code> class also includes a - constructor with an initializer for the required - <code>greeting</code> element as its argument.</p> - - <p>The <code>hello</code> overloaded global functions correspond - to the <code>hello</code> global element in XML Schema. A - global element in XML Schema is a valid document root. - By default XSD generated a set of parsing functions for each - global element defined in XML Schema (this can be overridden - with the <code>--root-element-*</code> options). Parsing - functions return a dynamically allocated object model as an - automatic pointer. The actual pointer used depends on the - C++ standard selected. For C++11 it is <code>std::unique_ptr</code> - as shown above. For C++98 it is <code>std::auto_ptr</code>. - For example, if we modify our XSD compiler invocation to - select C++98:</p> - - <pre class="terminal"> -$ xsd cxx-tree --std c++98 hello.xsd - </pre> - - <p>Then the parsing function signatures will become:</p> - - <pre class="c++"> -std::auto_ptr<hello_t> -hello (const std::string& uri); - -std::auto_ptr<hello_t> -hello (std::istream&); - </pre> - - <p>For more information on parsing functions see <a href="#5">Chapter 5, - "Parsing"</a>.</p> - - <h2><a name="2.3">2.3 Implementing Application Logic</a></h2> - - <p>At this point we have all the parts we need to do something useful - with the information stored in our XML document: - </p> - - <pre class="c++"> -#include <iostream> -#include "hello.hxx" - -using namespace std; - -int -main (int argc, char* argv[]) -{ - try - { - unique_ptr<hello_t> h (hello (argv[1])); - - for (hello_t::name_const_iterator i (h->name ().begin ()); - i != h->name ().end (); - ++i) - { - cerr << h->greeting () << ", " << *i << "!" << endl; - } - } - catch (const xml_schema::exception& e) - { - cerr << e << endl; - return 1; - } -} - </pre> - - <p>The first part of our application calls one of the parsing - functions to parser an XML file specified in the command line. - We then use the returned object model to iterate over names - and print a greeting line for each of them. Finally, we - catch and print the <code>xml_schema::exception</code> - exception in case something goes wrong. This exception - is the root of the exception hierarchy used by the - XSD-generated code. - </p> - - - <h2><a name="2.4">2.4 Compiling and Running</a></h2> - - <p>After saving our application from the previous section in - <code>driver.cxx</code>, we are ready to compile our first - program and run it on the test XML document. On a UNIX - system this can be done with the following commands: - </p> - - <pre class="terminal"> -$ c++ -std=c++11 -I.../libxsd -c driver.cxx hello.cxx -$ c++ -std=c++11 -o driver driver.o hello.o -lxerces-c -$ ./driver hello.xml -Hello, sun! -Hello, moon! -Hello, world! - </pre> - - <p>Here <code>.../libxsd</code> represents the path to the - <a href="https://cppget.org/libxsd">libxsd</a> package root - directory. Note also that we are required to link our - application with the Xerces-C++ library because the generated - code uses it as the underlying XML parser.</p> - - <h2><a name="2.5">2.5 Adding Serialization</a></h2> - - <p>While parsing and accessing the XML data may be everything - you need, there are applications that require creating new - or modifying existing XML documents. By default XSD does - not produce serialization code. We will need to request - it with the <code>--generate-serialization</code> options:</p> - - <pre class="terminal"> -$ xsd cxx-tree --generate-serialization hello.xsd - </pre> - - <p>If we now examine the generated <code>hello.hxx</code> file, - we will find a set of overloaded serialization functions, - including the following version:</p> - - <pre class="c++"> -void -hello (std::ostream&, - const hello_t&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap ()); - - </pre> - - <p>Just like with parsing functions, XSD generates serialization - functions for each global element unless instructed otherwise - with one of the <code>--root-element-*</code> options. For more - information on serialization functions see <a href="#6">Chapter 6, - "Serialization"</a>.</p> - - <p>We first examine an application that modifies an existing - object model and serializes it back to XML:</p> - - <pre class="c++"> -#include <iostream> -#include "hello.hxx" - -using namespace std; - -int -main (int argc, char* argv[]) -{ - try - { - unique_ptr<hello_t> h (hello (argv[1])); - - // Change the greeting phrase. - // - h->greeting ("Hi"); - - // Add another entry to the name sequence. - // - h->name ().push_back ("mars"); - - // Serialize the modified object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "hello.xsd"; - - hello (cout, *h, map); - } - catch (const xml_schema::exception& e) - { - cerr << e << endl; - return 1; - } -} - </pre> - - <p>First, our application parses an XML document and obtains its - object model as in the previous example. Then it changes the - greeting string and adds another entry to the list of names. - Finally, it serializes the object model back to XML by calling - the serialization function.</p> - - <p>The first argument we pass to the serialization function is - <code>cout</code> which results in the XML being written to - the standard output for us to inspect. We could have also - written the result to a file or memory buffer by creating an - instance of <code>std::ofstream</code> or <code>std::ostringstream</code> - and passing it instead of <code>cout</code>. The second argument is the - object model we want to serialize. The final argument is an optional - namespace information map for our vocabulary. It captures information - such as namespaces, namespace prefixes to which they should be mapped, - and schemas associated with these namespaces. If we don't provide - this argument then generic namespace prefixes (<code>p1</code>, - <code>p2</code>, etc.) will be automatically assigned to XML namespaces - and no schema information will be added to the resulting document - (see <a href="#6">Chapter 6, "Serialization"</a> for details). - In our case, the prefix (map key) and namespace name are empty - because our vocabulary does not use XML namespaces.</p> - - <p>If we now compile and run this application we will see the - output as shown in the following listing:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="hello.xsd"> - - <greeting>Hi</greeting> - - <name>sun</name> - <name>moon</name> - <name>world</name> - <name>mars</name> - -</hello> - </pre> - - <p>We can also create and serialize an object model from scratch - as shown in the following example:</p> - - <pre class="c++"> -#include <iostream> -#include <fstream> -#include "hello.hxx" - -using namespace std; - -int -main (int argc, char* argv[]) -{ - try - { - hello_t h ("Hi"); - - hello_t::name_sequence& ns (h.name ()); - - ns.push_back ("Jane"); - ns.push_back ("John"); - - // Serialize the object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "hello.xsd"; - - std::ofstream ofs (argv[1]); - hello (ofs, h, map); - } - catch (const xml_schema::exception& e) - { - cerr << e << endl; - return 1; - } -} - </pre> - - <p>In this example we used the generated constructor to create - an instance of type <code>hello_t</code>. To reduce typing, - we obtained a reference to the name sequence which we then - used to add a few names. The serialization part is identical - to the previous example except this time we are writing to - a file. If we compile and run this program, it produces the - following XML file:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<hello xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="hello.xsd"> - - <greeting>Hi</greeting> - - <name>Jane</name> - <name>John</name> - -</hello> - </pre> - - <h2><a name="2.6">2.6 Selecting Naming Convention</a></h2> - - <p>By default XSD uses the so-called K&R (Kernighan and Ritchie) - identifier naming convention in the generated code. In this - convention both type and function names are in lower case and - words are separated by underscores. If your application code or - schemas use a different notation, you may want to change the - naming convention used in the generated code for consistency. - XSD supports a set of widely-used naming conventions - that you can select with the <code>--type-naming</code> and - <code>--function-naming</code> options. You can also further - refine one of the predefined conventions or create a completely - custom naming scheme by using the <code>--*-regex</code> options.</p> - - <p>As an example, let's assume that our "Hello World" application - uses the so-called upper-camel-case naming convention for types - (that is, each word in a type name is capitalized) and the K&R - convention for function names. Since K&R is the default - convention for both type and function names, we only need to - change the type naming scheme:</p> - - <pre class="terminal"> -$ xsd cxx-tree --type-naming ucc hello.xsd - </pre> - - <p>The <code>ucc</code> argument to the <code>--type-naming</code> - options stands for upper-camel-case. If we now examine the - generated <code>hello.hxx</code>, we will see the following - changes compared to the declarations shown in the previous - sections:</p> - - <pre class="c++"> -class Hello_t -{ -public: - // greeting - // - typedef xml_schema::String GreetingType; - - const GreetingType& - greeting () const; - - GreetingType& - greeting (); - - void - greeting (const GreetingType& x); - - // name - // - typedef xml_schema::String NameType; - typedef xsd::sequence<NameType> NameSequence; - typedef NameSequence::iterator NameIterator; - typedef NameSequence::const_iterator NameConstIterator; - - const NameSequence& - name () const; - - NameSequence& - name (); - - void - name (const NameSequence& s); - - // Constructor. - // - Hello_t (const GreetingType&); - - ... - -}; - -std::unique_ptr<Hello_t> -hello (const std::string& uri); - -std::unique_ptr<Hello_t> -hello (std::istream&); - </pre> - - <p>Notice that the type names in the <code>xml_schema</code> namespace, - for example <code>xml_schema::String</code>, now also use the - upper-camel-case naming convention. The only thing that we may - be unhappy about in the above code is the <code>_t</code> - suffix in <code>Hello_t</code>. If we are not in a position - to change the schema, we can <em>touch-up</em> the <code>ucc</code> - convention with a custom translation rule using the - <code>--type-regex</code> option:</p> - - <pre class="terminal"> -$ xsd cxx-tree --type-naming ucc --type-regex '/ (.+)_t/\u$1/' hello.xsd - </pre> - - <p>This results in the following changes to the generated code:</p> - - <pre class="c++"> -class Hello -{ -public: - // greeting - // - typedef xml_schema::String GreetingType; - - const GreetingType& - greeting () const; - - GreetingType& - greeting (); - - void - greeting (const GreetingType& x); - - // name - // - typedef xml_schema::String NameType; - typedef xsd::sequence<NameType> NameSequence; - typedef NameSequence::iterator NameIterator; - typedef NameSequence::const_iterator NameConstIterator; - - const NameSequence& - name () const; - - NameSequence& - name (); - - void - name (const NameSequence& s); - - // Constructor. - // - Hello (const GreetingType&); - - ... - -}; - -std::unique_ptr<Hello> -hello (const std::string& uri); - -std::unique_ptr<Hello> -hello (std::istream&); - </pre> - - <p>For more detailed information on the <code>--type-naming</code>, - <code>--function-naming</code>, <code>--type-regex</code>, and - other <code>--*-regex</code> options refer to the NAMING - CONVENTION section in the <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>.</p> - - <h2><a name="2.7">2.7 Generating Documentation</a></h2> - - <p>While our object model is quite simple, real-world vocabularies - can be quite complex with hundreds of types, elements, and - attributes. For such vocabularies figuring out which types - provide which member functions by studying the generated - source code or schemas can be a daunting task. To provide - application developers with a more accessible way of - understanding the generated object models, the XSD compiler - can be instructed to produce source code with documentation - comments in the Doxygen format. Then the source code can be - processed with the <a href="http://www.doxygen.org">Doxygen</a> - documentation system to extract this information and produce - documentation in various formats. - </p> - - <p>In this section we will see how to generate documentation - for our "Hello World" vocabulary. To showcase the full power - of the XSD documentation facilities, we will first document - our schema. The XSD compiler will then transfer - this information from the schema to the generated code and - then to the object model documentation. Note that the - documentation in the schema is not required for XSD to - generate useful documentation. Below you will find - our <code>hello.xsd</code> with added documentation:</p> - - <pre class="xml"> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:complexType name="hello_t"> - - <xs:annotation> - <xs:documentation> - The hello_t type consists of a greeting phrase and a - collection of names to which this greeting applies. - </xs:documentation> - </xs:annotation> - - <xs:sequence> - - <xs:element name="greeting" type="xs:string"> - <xs:annotation> - <xs:documentation> - The greeting element contains the greeting phrase - for this hello object. - </xs:documentation> - </xs:annotation> - </xs:element> - - <xs:element name="name" type="xs:string" maxOccurs="unbounded"> - <xs:annotation> - <xs:documentation> - The name elements contains names to be greeted. - </xs:documentation> - </xs:annotation> - </xs:element> - - </xs:sequence> - </xs:complexType> - - <xs:element name="hello" type="hello_t"> - <xs:annotation> - <xs:documentation> - The hello element is a root of the Hello XML vocabulary. - Every conforming document should start with this element. - </xs:documentation> - </xs:annotation> - </xs:element> - -</xs:schema> - </pre> - - <p>The first step in obtaining the documentation is to recompile - our schema with the <code>--generate-doxygen</code> option:</p> - - <pre class="terminal"> -$ xsd cxx-tree --generate-serialization --generate-doxygen hello.xsd - </pre> - - <p>Now the generated <code>hello.hxx</code> file contains comments - in the Doxygen format. The next step is to process this file - with the Doxygen documentation system. If your project does - not use Doxygen then you first need to create a configuration - file for your project:</p> - - <pre class="terminal"> -$ doxygen -g hello.doxygen - </pre> - - <p>You only need to perform this step once. Now we can generate - the documentation by executing the following command in the - directory with the generated source code:</p> - - <pre class="terminal"> -$ doxygen hello.doxygen - </pre> - - <p>While the generated documentation can be useful as is, we can - go one step further and link (using the Doxygen tags mechanism) - the documentation for our object model with the documentation - for the XSD runtime library which defines C++ classes for the - built-in XML Schema types. This way we can seamlessly browse - between documentation for the <code>hello_t</code> class which - is generated by the XSD compiler and the <code>xml_schema::string</code> - class which is defined in the XSD runtime library. The Doxygen - configuration file for the XSD runtime is provided with the XSD - distribution.</p> - - <p>You can view the result of the steps described in this section - on the <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/hello/html/annotated.html">Hello - Example Documentation</a> page.</p> - - <!-- Chapater 3 --> - - - <h1><a name="3">3 Overall Mapping Configuration</a></h1> - - <p>The C++/Tree mapping has a number of configuration parameters that - determine the overall properties and behavior of the generated code. - Configuration parameters are specified with the XSD command line - options. This chapter describes configuration aspects that are most - commonly encountered by application developers. These include: the - C++ standard, the character type that is used by the generated code, - handling of vocabularies that use XML Schema polymorphism, XML Schema - to C++ namespace mapping, and thread safety. For more ways to configure - the generated code refer to the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>. - </p> - - <h2><a name="3.1">3.1 C++ Standard</a></h2> - - <p>The C++/Tree mapping provides support for ISO/IEC C++ 2011 (C++11) - and ISO/IEC C++ 1998/2003 (C++98). To select the C++ standard for the - generated code we use the <code>--std</code> XSD compiler command - line option. While the majority of the examples in this guide use - C++11, the document explains the C++11/98 usage difference and so - they can easily be converted to C++98.</p> - - <h2><a name="3.2">3.2 Character Type and Encoding</a></h2> - - <p>The C++/Tree mapping has built-in support for two character types: - <code>char</code> and <code>wchar_t</code>. You can select the - character type with the <code>--char-type</code> command line - option. The default character type is <code>char</code>. The - character type affects all string and string-based types that - are used in the mapping. These include the string-based built-in - XML Schema types, exception types, stream types, etc.</p> - - <p>Another aspect of the mapping that depends on the character type - is character encoding. For the <code>char</code> character type - the default encoding is UTF-8. Other supported encodings are - ISO-8859-1, Xerces-C++ Local Code Page (LPC), as well as - custom encodings. You can select which encoding should be used - in the object model with the <code>--char-encoding</code> command - line option.</p> - - <p>For the <code>wchar_t</code> character type the encoding is - automatically selected between UTF-16 and UTF-32/UCS-4 depending - on the size of the <code>wchar_t</code> type. On some platforms - (for example, Windows with Visual C++ and AIX with IBM XL C++) - <code>wchar_t</code> is 2 bytes long. For these platforms the - encoding is UTF-16. On other platforms <code>wchar_t</code> is 4 bytes - long and UTF-32/UCS-4 is used.</p> - - <p>Note also that the character encoding that is used in the object model - is independent of the encodings used in input and output XML. In fact, - all three (object mode, input XML, and output XML) can have different - encodings.</p> - - <h2><a name="3.3">3.3 Support for Polymorphism</a></h2> - - <p>By default XSD generates non-polymorphic code. If your vocabulary - uses XML Schema polymorphism in the form of <code>xsi:type</code> - and/or substitution groups, then you will need to compile - your schemas with the <code>--generate-polymorphic</code> option - to produce polymorphism-aware code. For more information on - working with polymorphic object models, refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.11">Section 2.11, - "Mapping for <code>xsi:type</code> and Substitution Groups"</a> in - the C++/Tree Mapping User Manual.</p> - - <h2><a name="3.4">3.4 Namespace Mapping</a></h2> - - <p>XSD maps XML namespaces specified in the <code>targetNamespace</code> - attribute in XML Schema to one or more nested C++ namespaces. By - default, a namespace URI is mapped to a sequence of C++ namespace - names by removing the protocol and host parts and splitting the - rest into a sequence of names with <code>'/'</code> as the name - separator.</p> - - <p>The default mapping of namespace URIs to C++ namespaces - can be altered using the <code>--namespace-map</code> and - <code>--namespace-regex</code> compiler options. For example, - to map namespace URI <code>https://www.codesynthesis.com/my</code> to - C++ namespace <code>cs::my</code>, we can use the following option:</p> - - <pre class="terminal"> ---namespace-map https://www.codesynthesis.com/my=cs::my - </pre> - - <p>A vocabulary without a namespace is mapped to the global scope. This - also can be altered with the above options by using an empty name - for the XML namespace:</p> - - <pre class="terminal"> ---namespace-map =cs - </pre> - - <h2><a name="3.5">3.5 Thread Safety</a></h2> - - <p>XSD-generated code is thread-safe in the sense that you can - use different instantiations of the object model in several - threads concurrently. This is possible due to the generated - code not relying on any writable global variables. If you need - to share the same object between several threads then you will - need to provide some form of synchronization. One approach would - be to use the generated code customization mechanisms to embed - synchronization primitives into the generated C++ classes. For more - information on generated code customization refer to the - <a href="http://wiki.codesynthesis.com/Tree/Customization_guide">C++/Tree - Mapping Customization Guide</a>.</p> - - <p>If you also would like to call parsing and/or serialization - functions from several threads potentially concurrently, then - you will need to make sure the Xerces-C++ runtime is initialized - and terminated only once. The easiest way to do this is to - initialize/terminate Xerces-C++ from <code>main()</code> when - there are no threads yet/anymore:</p> - - <pre class="c++"> -#include <xercesc/util/PlatformUtils.hpp> - -int -main () -{ - xercesc::XMLPlatformUtils::Initialize (); - - { - // Start/terminate threads and parse/serialize here. - } - - xercesc::XMLPlatformUtils::Terminate (); -} - </pre> - - <p>Because you initialize the Xerces-C++ runtime yourself you should - also pass the <code>xml_schema::flags::dont_initialize</code> flag - to parsing and serialization functions. See <a href="#5">Chapter 5, - "Parsing"</a> and <a href="#6">Chapter 6, "Serialization"</a> for - more information.</p> - - - <!-- Chapater 4 --> - - - <h1><a name="4">4 Working with Object Models</a></h1> - - <p>As we have seen in the previous chapters, the XSD compiler generates - a C++ class for each type defined in XML Schema. Together these classes - constitute an object model for an XML vocabulary. In this chapter we - will take a closer look at different elements that comprise an - object model class as well as how to create, access, and modify - object models.</p> - - <p>In this and subsequent chapters we will use the following schema - that describes a collection of person records. We save it in - <code>people.xsd</code>:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:simpleType name="gender_t"> - <xs:restriction base="xs:string"> - <xs:enumeration value="male"/> - <xs:enumeration value="female"/> - </xs:restriction> - </xs:simpleType> - - <xs:complexType name="person_t"> - <xs:sequence> - <xs:element name="first-name" type="xs:string"/> - <xs:element name="middle-name" type="xs:string" minOccurs="0"/> - <xs:element name="last-name" type="xs:string"/> - <xs:element name="gender" type="gender_t"/> - <xs:element name="age" type="xs:short"/> - </xs:sequence> - <xs:attribute name="id" type="xs:unsignedInt" use="required"/> - </xs:complexType> - - <xs:complexType name="people_t"> - <xs:sequence> - <xs:element name="person" type="person_t" maxOccurs="unbounded"/> - </xs:sequence> - </xs:complexType> - - <xs:element name="people" type="people_t"/> - -</xs:schema> - </pre> - - <p>A sample XML instance to go along with this schema is saved - in <code>people.xml</code>:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd"> - - <person id="1"> - <first-name>John</first-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>32</age> - </person> - - <person id="2"> - <first-name>Jane</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>female</gender> - <age>28</age> - </person> - -</people> - </pre> - - <p>Compiling <code>people.xsd</code> with the XSD compiler results - in three generated C++ classes: <code>gender_t</code>, - <code>person_t</code>, and <code>people_t</code>. - The <code>gender_t</code> class is modelled after the C++ - <code>enum</code> type. Its definition is presented below:</p> - - <pre class="c++"> -class gender_t: public xml_schema::string -{ -public: - enum value - { - male, - female - }; - - gender_t (value); - gender_t (const xml_schema::string&); - - gender_t& - operator= (value); - - operator value () const; -}; - </pre> - - <p>The following listing shows how we can use this type:</p> - - <pre class="c++"> -gender_t m (gender_t::male); -gender_t f ("female"); - -if (m == "female" || f == gender_t::male) -{ - ... -} - -switch (m) -{ -case gender_t::male: - { - ... - } -case gender_t::female: - { - ... - } -} - </pre> - - <p>The other two classes will be examined in detail in the subsequent - sections.</p> - - <h2><a name="4.1">4.1 Attribute and Element Cardinalities</a></h2> - - <p>As we have seen in the previous chapters, XSD generates a different - set of type definitions and member functions for elements with - different cardinalities. The C++/Tree mapping divides all the possible - element and attribute cardinalities into three cardinality classes: - <em>one</em>, <em>optional</em>, and <em>sequence</em>.</p> - - <p>The <em>one</em> cardinality class covers all elements that should - occur exactly once as well as required attributes. In our - example, the <code>first-name</code>, <code>last-name</code>, - <code>gender</code>, and <code>age</code> elements as well as - the <code>id</code> attribute belong to this cardinality class. - The following code fragment shows type definitions as well as the - accessor and modifier functions that are generated for the - <code>gender</code> element in the <code>person_t</code> class:</p> - - <pre class="c++"> -class person_t -{ - // gender - // - typedef gender_t gender_type; - - const gender_type& - gender () const; - - gender_type& - gender (); - - void - gender (const gender_type&); -}; - </pre> - - <p>The <code>gender_type</code> type is an alias for the element's type. - The first two accessor functions return read-only (constant) and - read-write references to the element's value, respectively. The - modifier function sets the new value for the element.</p> - - <p>The <em>optional</em> cardinality class covers all elements that - can occur zero or one time as well as optional attributes. In our - example, the <code>middle-name</code> element belongs to this - cardinality class. The following code fragment shows the type - definitions as well as the accessor and modifier functions that - are generated for this element in the <code>person_t</code> class:</p> - - <pre class="c++"> -class person_t -{ - // middle-name - // - typedef xml_schema::string middle_name_type; - typedef xsd::optional<middle_name_type> middle_name_optional; - - const middle_name_optional& - middle_name () const; - - middle_name_optional& - middle_name (); - - void - middle_name (const middle_name_type&); - - void - middle_name (const middle_name_optional&); -}; - </pre> - - <p>As with the <code>gender</code> element, <code>middle_name_type</code> - is an alias for the element's type. The <code>middle_name_optional</code> - type is a container for the element's optional value. It can be queried - for the presence of the value using the <code>present()</code> function. - The value itself can be retrieved using the <code>get()</code> - accessor and set using the <code>set()</code> modifier. The container - can be reverted to the value not present state with the call to the - <code>reset()</code> function. The following example shows how we - can use this container:</p> - - <pre class="c++"> -person_t::middle_name_optional n ("John"); - -if (n.present ()) -{ - cout << n.get () << endl; -} - -n.set ("Jane"); -n.reset (); - </pre> - - - <p>Unlike the <em>one</em> cardinality class, the accessor functions - for the <em>optional</em> class return read-only (constant) and - read-write references to the container instead of the element's - value directly. The modifier functions set the new value for the - element.</p> - - <p>Finally, the <em>sequence</em> cardinality class covers all elements - that can occur more than once. In our example, the - <code>person</code> element in the <code>people_t</code> type - belongs to this cardinality class. The following code fragment shows - the type definitions as well as the accessor and modifier functions - that are generated for this element in the <code>people_t</code> - class:</p> - - <pre class="c++"> -class people_t -{ - // person - // - typedef person_t person_type; - typedef xsd::sequence<person_type> person_sequence; - typedef person_sequence::iterator person_iterator; - typedef person_sequence::const_iterator person_const_iterator; - - const person_sequence& - person () const; - - person_sequence& - person (); - - void - person (const person_sequence&); -}; - </pre> - - <p>Identical to the other cardinality classes, <code>person_type</code> - is an alias for the element's type. The <code>person_sequence</code> - type is a sequence container for the element's values. It is based - on and has the same interface as <code>std::vector</code> and - therefore can be used in similar ways. The <code>person_iterator</code> - and <code>person_const_iterator</code> types are read-only - (constant) and read-write iterators for the <code>person_sequence</code> - container.</p> - - <p>Similar to the <em>optional</em> cardinality class, the - accessor functions for the <em>sequence</em> class return - read-only (constant) and read-write references to the sequence - container. The modifier functions copies the entries from - the passed sequence.</p> - - <p>C++/Tree is a "flattening" mapping in a sense that many levels of - nested compositors (<code>choice</code> and <code>sequence</code>), - all potentially with their own cardinalities, are in the end mapped - to a flat set of elements with one of the three cardinality classes - discussed above. While this results in a simple and easy to use API - for most types, in certain cases, the order of elements in the actual - XML documents is not preserved once parsed into the object model. To - overcome this limitation we can mark certain schema types, for which - content order is not sufficiently preserved, as ordered. For more - information on this functionality refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.8.4">Section - 2.8.4, "Element Order"</a> in the C++/Tree Mapping User Manual.</p> - - <p>For complex schemas with many levels of nested compositors - (<code>choice</code> and <code>sequence</code>) it can also - be hard to deduce the cardinality class of a particular element. - The generated Doxygen documentation can greatly help with - this task. For each element and attribute the documentation - clearly identifies its cardinality class. Alternatively, you - can study the generated header files to find out the cardinality - class of a particular attribute or element.</p> - - <p>In the next sections we will examine how to access and modify - information stored in an object model using accessor and modifier - functions described in this section.</p> - - <h2><a name="4.2">4.2 Accessing the Object Model</a></h2> - - <p>In this section we will learn how to get to the information - stored in the object model for our person records vocabulary. - The following application accesses and prints the contents - of the <code>people.xml</code> file:</p> - - <pre class="c++"> -#include <iostream> -#include "people.hxx" - -using namespace std; - -int -main () -{ - unique_ptr<people_t> ppl (people ("people.xml")); - - // Iterate over individual person records. - // - people_t::person_sequence& ps (ppl->person ()); - - for (people_t::person_iterator i (ps.begin ()); i != ps.end (); ++i) - { - person_t& p (*i); - - // Print names: first-name and last-name are required elements, - // middle-name is optional. - // - cout << "name: " << p.first_name () << " "; - - if (p.middle_name ().present ()) - cout << p.middle_name ().get () << " "; - - cout << p.last_name () << endl; - - // Print gender, age, and id which are all required. - // - cout << "gender: " << p.gender () << endl - << "age: " << p.age () << endl - << "id: " << p.id () << endl - << endl; - } -} - </pre> - - <p>This code shows common patterns of accessing elements and attributes - with different cardinality classes. For the sequence element - (<code>person</code> in <code>people_t</code>) we first obtain a - reference to the container and then iterate over individual - records. The values of elements and attributes with the - <em>one</em> cardinality class (<code>first-name</code>, - <code>last-name</code>, <code>gender</code>, <code>age</code>, - and <code>id</code>) can be obtained directly by calling the - corresponding accessor functions. For the optional element - <code>middle-name</code> we first check if the value is present - and only then call <code>get()</code> to retrieve it.</p> - - <p>Note that when we want to reduce typing by creating a variable - representing a fragment of the object model that we are currently - working with (<code>ps</code> and <code>p</code> above), we obtain - a reference to that fragment instead of making a potentially - expensive copy. This is generally a good rule to follow when - creating high-performance applications.</p> - - <p>If we run the above application on our sample - <code>people.xml</code>, the output looks as follows:</p> - - <pre class="terminal"> -name: John Doe -gender: male -age: 32 -id: 1 - -name: Jane Mary Doe -gender: female -age: 28 -id: 2 - </pre> - - - <h2><a name="4.3">4.3 Modifying the Object Model</a></h2> - - <p>In this section we will learn how to modify the information - stored in the object model for our person records vocabulary. - The following application changes the contents of the - <code>people.xml</code> file:</p> - - <pre class="c++"> -#include <iostream> -#include "people.hxx" - -using namespace std; - -int -main () -{ - unique_ptr<people_t> ppl (people ("people.xml")); - - // Iterate over individual person records and increment - // the age. - // - people_t::person_sequence& ps (ppl->person ()); - - for (people_t::person_iterator i (ps.begin ()); i != ps.end (); ++i) - { - // Alternative way: i->age ()++; - // - i->age (i->age () + 1); - } - - // Add middle-name to the first record and remove it from - // the second. - // - person_t& john (ps[0]); - person_t& jane (ps[1]); - - john.middle_name ("Mary"); - jane.middle_name ().reset (); - - // Add another John record. - // - ps.push_back (john); - - // Serialize the modified object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "people.xsd"; - - people (cout, *ppl, map); -} - </pre> - - <p>The first modification the above application performs is iterating - over person records and incrementing the age value. This code - fragment shows how to modify the value of a required attribute - or element. The next modification shows how to set a new value - for the optional <code>middle-name</code> element as well - as clear its value. Finally the example adds a copy of the - John Doe record to the <code>person</code> element sequence.</p> - - <p>Note that in this case using references for the <code>ps</code>, - <code>john</code>, and <code>jane</code> variables is no longer - a performance improvement but a requirement for the application - to function correctly. If we hadn't used references, all our changes - would have been made on copies without affecting the object model.</p> - - <p>If we run the above application on our sample <code>people.xml</code>, - the output looks as follows:</p> - - <pre class="xml"> -<?xml version="1.0"?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd"> - - <person id="1"> - <first-name>John</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>33</age> - </person> - - <person id="2"> - <first-name>Jane</first-name> - <last-name>Doe</last-name> - <gender>female</gender> - <age>29</age> - </person> - - <person id="1"> - <first-name>John</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>33</age> - </person> - -</people> - </pre> - - - <h2><a name="4.4">4.4 Creating the Object Model from Scratch</a></h2> - - <p>In this section we will learn how to create a new object model - for our person records vocabulary. The following application - recreates the content of the original <code>people.xml</code> - file:</p> - - <pre class="c++"> -#include <iostream> -#include "people.hxx" - -using namespace std; - -int -main () -{ - people_t ppl; - people_t::person_sequence& ps (ppl.person ()); - - // Add the John Doe record. - // - ps.push_back ( - person_t ("John", // first-name - "Doe", // last-name - gender_t::male, // gender - 32, // age - 1)); - - // Add the Jane Doe record. - // - ps.push_back ( - person_t ("Jane", // first-name - "Doe", // last-name - gender_t::female, // gender - 28, // age - 2)); // id - - // Add middle name to the Jane Doe record. - // - person_t& jane (ps.back ()); - jane.middle_name ("Mary"); - - // Serialize the object model to XML. - // - xml_schema::namespace_infomap map; - map[""].name = ""; - map[""].schema = "people.xsd"; - - people (cout, ppl, map); -} - </pre> - - <p>The only new part in the above application is the calls - to the <code>people_t</code> and <code>person_t</code> - constructors. As a general rule, for each C++ class - XSD generates a constructor with initializers - for each element and attribute belonging to the <em>one</em> - cardinality class. For our vocabulary, the following - constructors are generated:</p> - - <pre class="c++"> -class person_t -{ - person_t (const first_name_type&, - const last_name_type&, - const gender_type&, - const age_type&, - const id_type&); -}; - -class people_t -{ - people_t (); -}; - </pre> - - <p>Note also that we set the <code>middle-name</code> element - on the Jane Doe record by obtaining a reference to that record - in the object model and setting the <code>middle-name</code> - value on it. This is a general rule that should be followed - in order to obtain the best performance: if possible, - direct modifications to the object model should be preferred - to modifications on temporaries with subsequent copying. The - following code fragment shows a semantically equivalent but - slightly slower version:</p> - - <pre class="c++"> -// Add the Jane Doe record. -// -person_t jane ("Jane", // first-name - "Doe", // last-name - gender_t::female, // gender - 28, // age - 2); // id - -jane.middle_name ("Mary"); - -ps.push_back (jane); - </pre> - - <p>We can also go one step further to reduce copying and improve - the performance of our application by using the non-copying - <code>push_back()</code> function which assumes ownership - of the passed objects:</p> - - <pre class="c++"> -// Add the Jane Doe record. C++11 version -// -unique_ptr<person_t> jane_p ( - new person_t ("Jane", // first-name - "Doe", // last-name - gender_t::female, // gender - 28, // age - 2)); // id -ps.push_back (std::move (jane_p)); // assumes ownership - -// Add the John Doe record. C++98 version. -// -auto_ptr<person_t> john_p ( - new person_t ("John", // first-name - "Doe", // last-name - gender_t::male, // gender - 32, // age - 1)); -ps.push_back (john_p); // assumes ownership - </pre> - - <p>For more information on the non-copying modifier functions refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.8">Section - 2.8, "Mapping for Local Elements and Attributes"</a> in the C++/Tree Mapping - User Manual. The above application produces the following output:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd"> - - <person id="1"> - <first-name>John</first-name> - <last-name>Doe</last-name> - <gender>male</gender> - <age>32</age> - </person> - - <person id="2"> - <first-name>Jane</first-name> - <middle-name>Mary</middle-name> - <last-name>Doe</last-name> - <gender>female</gender> - <age>28</age> - </person> - -</people> - </pre> - - <h2><a name="4.5">4.5 Mapping for the Built-in XML Schema Types</a></h2> - - <p>Our person record vocabulary uses several built-in XML Schema - types: <code>string</code>, <code>short</code>, and - <code>unsignedInt</code>. Until now we haven't talked about - the mapping of built-in XML Schema types to C++ types and how - to work with them. This section provides an overview - of the built-in types. For more detailed information refer - to <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.5">Section - 2.5, "Mapping for Built-in Data Types"</a> in the C++/Tree Mapping - User Manual.</p> - - <p>In XML Schema, built-in types are defined in the XML Schema namespace. - By default, the C++/Tree mapping maps this namespace to C++ - namespace <code>xml_schema</code> (this mapping can be altered - with the <code>--namespace-map</code> option). The following table - summarizes the mapping of XML Schema built-in types to C++ types:</p> - - <!-- border="1" is necessary for html2ps --> - <table id="builtin" border="1"> - <tr> - <th>XML Schema type</th> - <th>Alias in the <code>xml_schema</code> namespace</th> - <th>C++ type</th> - </tr> - - <tr> - <th colspan="3">fixed-length integral types</th> - </tr> - <!-- 8-bit --> - <tr> - <td><code>byte</code></td> - <td><code>byte</code></td> - <td><code>signed char</code></td> - </tr> - <tr> - <td><code>unsignedByte</code></td> - <td><code>unsigned_byte</code></td> - <td><code>unsigned char</code></td> - </tr> - - <!-- 16-bit --> - <tr> - <td><code>short</code></td> - <td><code>short_</code></td> - <td><code>short</code></td> - </tr> - <tr> - <td><code>unsignedShort</code></td> - <td><code>unsigned_short</code></td> - <td><code>unsigned short</code></td> - </tr> - - <!-- 32-bit --> - <tr> - <td><code>int</code></td> - <td><code>int_</code></td> - <td><code>int</code></td> - </tr> - <tr> - <td><code>unsignedInt</code></td> - <td><code>unsigned_int</code></td> - <td><code>unsigned int</code></td> - </tr> - - <!-- 64-bit --> - <tr> - <td><code>long</code></td> - <td><code>long_</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>unsignedLong</code></td> - <td><code>unsigned_long</code></td> - <td><code>unsigned long long</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-length integral types</th> - </tr> - <tr> - <td><code>integer</code></td> - <td><code>integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonPositiveInteger</code></td> - <td><code>non_positive_integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonNegativeInteger</code></td> - <td><code>non_negative_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>positiveInteger</code></td> - <td><code>positive_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>negativeInteger</code></td> - <td><code>negative_integer</code></td> - <td><code>long long</code></td> - </tr> - - <tr> - <th colspan="3">boolean types</th> - </tr> - <tr> - <td><code>boolean</code></td> - <td><code>boolean</code></td> - <td><code>bool</code></td> - </tr> - - <tr> - <th colspan="3">fixed-precision floating-point types</th> - </tr> - <tr> - <td><code>float</code></td> - <td><code>float_</code></td> - <td><code>float</code></td> - </tr> - <tr> - <td><code>double</code></td> - <td><code>double_</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-precision floating-point types</th> - </tr> - <tr> - <td><code>decimal</code></td> - <td><code>decimal</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">string types</th> - </tr> - <tr> - <td><code>string</code></td> - <td><code>string</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - <tr> - <td><code>normalizedString</code></td> - <td><code>normalized_string</code></td> - <td>type derived from <code>string</code></td> - </tr> - <tr> - <td><code>token</code></td> - <td><code>token</code></td> - <td>type derived from <code>normalized_string</code></td> - </tr> - <tr> - <td><code>Name</code></td> - <td><code>name</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKEN</code></td> - <td><code>nmtoken</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKENS</code></td> - <td><code>nmtokens</code></td> - <td>type derived from <code>sequence<nmtoken></code></td> - </tr> - <tr> - <td><code>NCName</code></td> - <td><code>ncname</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>language</code></td> - <td><code>language</code></td> - <td>type derived from <code>token</code></td> - </tr> - - <tr> - <th colspan="3">qualified name</th> - </tr> - <tr> - <td><code>QName</code></td> - <td><code>qname</code></td> - <td><code>xml_schema::qname</code></td> - </tr> - - <tr> - <th colspan="3">ID/IDREF types</th> - </tr> - <tr> - <td><code>ID</code></td> - <td><code>id</code></td> - <td>type derived from <code>ncname</code></td> - </tr> - <tr> - <td><code>IDREF</code></td> - <td><code>idref</code></td> - <td>type derived from <code>ncname</code></td> - </tr> - <tr> - <td><code>IDREFS</code></td> - <td><code>idrefs</code></td> - <td>type derived from <code>sequence<idref></code></td> - </tr> - - <tr> - <th colspan="3">URI types</th> - </tr> - <tr> - <td><code>anyURI</code></td> - <td><code>uri</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - - <tr> - <th colspan="3">binary types</th> - </tr> - <tr> - <td><code>base64Binary</code></td> - <td><code>base64_binary</code></td> - <td><code>xml_schema::base64_binary</code></td> - </tr> - <tr> - <td><code>hexBinary</code></td> - <td><code>hex_binary</code></td> - <td><code>xml_schema::hex_binary</code></td> - </tr> - - <tr> - <th colspan="3">date/time types</th> - </tr> - <tr> - <td><code>date</code></td> - <td><code>date</code></td> - <td><code>xml_schema::date</code></td> - </tr> - <tr> - <td><code>dateTime</code></td> - <td><code>date_time</code></td> - <td><code>xml_schema::date_time</code></td> - </tr> - <tr> - <td><code>duration</code></td> - <td><code>duration</code></td> - <td><code>xml_schema::duration</code></td> - </tr> - <tr> - <td><code>gDay</code></td> - <td><code>gday</code></td> - <td><code>xml_schema::gday</code></td> - </tr> - <tr> - <td><code>gMonth</code></td> - <td><code>gmonth</code></td> - <td><code>xml_schema::gmonth</code></td> - </tr> - <tr> - <td><code>gMonthDay</code></td> - <td><code>gmonth_day</code></td> - <td><code>xml_schema::gmonth_day</code></td> - </tr> - <tr> - <td><code>gYear</code></td> - <td><code>gyear</code></td> - <td><code>xml_schema::gyear</code></td> - </tr> - <tr> - <td><code>gYearMonth</code></td> - <td><code>gyear_month</code></td> - <td><code>xml_schema::gyear_month</code></td> - </tr> - <tr> - <td><code>time</code></td> - <td><code>time</code></td> - <td><code>xml_schema::time</code></td> - </tr> - - <tr> - <th colspan="3">entity types</th> - </tr> - <tr> - <td><code>ENTITY</code></td> - <td><code>entity</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>ENTITIES</code></td> - <td><code>entities</code></td> - <td>type derived from <code>sequence<entity></code></td> - </tr> - </table> - - <p>As you can see from the table above a number of built-in - XML Schema types are mapped to fundamental C++ types such - as <code>int</code> or <code>bool</code>. All string-based - XML Schema types are mapped to C++ types that are derived - from either <code>std::string</code> or - <code>std::wstring</code>, depending on the character - type selected. For access and modification purposes these - types can be treated as <code>std::string</code>. A number - of built-in types, such as <code>qname</code>, the binary - types, and the date/time types do not have suitable - fundamental or standard C++ types to map to. As a result, - these types are implemented from scratch in the XSD runtime. - For more information on their interfaces refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.5">Section - 2.5, "Mapping for Built-in Data Types"</a> in the C++/Tree Mapping - User Manual.</p> - - - <!-- Chapater 5 --> - - - <h1><a name="5">5 Parsing</a></h1> - - <p>We have already seen how to parse XML to an object model in this guide - before. In this chapter we will discuss the parsing topic in more - detail.</p> - - <p>By default, the C++/Tree mapping provides a total of 14 overloaded - parsing functions. They differ in the input methods used to - read XML as well as the error reporting mechanisms. It is also possible - to generate types for root elements instead of parsing and serialization - functions. This may be useful if your XML vocabulary has multiple - root elements. For more information on element types refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.9">Section - 2.9, "Mapping for Global Elements"</a> in the C++/Tree Mapping User - Manual.</p> - - - <p>In this section we will discuss the most commonly used versions of - the parsing functions. For a comprehensive description of parsing - refer to <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#3">Chapter - 3, "Parsing"</a> in the C++/Tree Mapping User Manual. For the <code>people</code> - global element from our person record vocabulary, we will concentrate - on the following three parsing functions:</p> - - <pre class="c++"> -std::[unique|auto]_ptr<people_t> -people (const std::string& uri, - xml_schema::flags f = 0, - const xml_schema::properties& p = xml_schema::properties ()); - -std::[unique|auto]_ptr<people_t> -people (std::istream& is, - xml_schema::flags f = 0, - const xml_schema::properties& p = xml_schema::properties ()); - -std::[unique|auto]_ptr<people_t> -people (std::istream& is, - const std::string& resource_id, - xml_schema::flags f = 0, - const xml_schema::properties& p = ::xml_schema::properties ()); - </pre> - - <p>The first function parses a local file or a URI. We have already - used this parsing function in the previous chapters. The second - and third functions read XML from a standard input stream. The - last function also requires a resource id. This id is used to - identify the XML document being parser in diagnostics messages - as well as to resolve relative paths to other documents (for example, - schemas) that might be referenced from the XML document.</p> - - <p>The last two arguments to all three parsing functions are parsing - flags and properties. The flags argument provides a number of ways - to fine-tune the parsing process. The properties argument allows - to pass additional information to the parsing functions. We will - use these two arguments in <a href="#5.1">Section 5.1, "XML Schema - Validation and Searching"</a> below. All three functions return - the object model as either <code>std::unique_ptr</code> (C++11) or - <code>std::auto_ptr</code> (C++98), depending on the C++ standard - selected (<code>--std</code> XSD compiler option). The following - example shows how we can use the above parsing functions:</p> - - <pre class="c++"> -using std::unique_ptr; - -// Parse a local file or URI. -// -unique_ptr<people_t> p1 (people ("people.xml")); -unique_ptr<people_t> p2 (people ("http://example.com/people.xml")); - -// Parse a local file via ifstream. -// -std::ifstream ifs ("people.xml"); -unique_ptr<people_t> p3 (people (ifs, "people.xml")); - -// Parse an XML string. -// -std::string str ("..."); // XML in a string. -std::istringstream iss (str); -unique_ptr<people_t> p4 (people (iss)); - </pre> - - - <h2><a name="5.1">5.1 XML Schema Validation and Searching</a></h2> - - <p>The C++/Tree mapping relies on the underlying Xerces-C++ XML - parser for full XML document validation. The XML Schema - validation is enabled by default and can be disabled by - passing the <code>xml_schema::flags::dont_validate</code> - flag to the parsing functions, for example:</p> - - <pre class="c++"> -unique_ptr<people_t> p ( - people ("people.xml", xml_schema::flags::dont_validate)); - </pre> - - <p>Even when XML Schema validation is disabled, the generated - code still performs a number of checks to prevent - construction of an inconsistent object model (for example, an - object model with missing required attributes or elements).</p> - - <p>When XML Schema validation is enabled, the XML parser needs - to locate a schema to validate against. There are several - methods to provide the schema location information to the - parser. The easiest and most commonly used method is to - specify schema locations in the XML document itself - with the <code>schemaLocation</code> or - <code>noNamespaceSchemaLocation</code> attributes, for example:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<people xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="people.xsd" - xsi:schemaLocation="http://www.w3.org/XML/1998/namespace xml.xsd"> - </pre> - - <p>As you might have noticed, we used this method in all the sample XML - documents presented in this guide up until now. Note that the - schema locations specified with these two attributes are relative - to the document's path unless they are absolute URIs (that is - start with <code>http://</code>, <code>file://</code>, etc.). - In particular, if you specify just file names as your schema - locations, as we did above, then the schemas should reside in - the same directory as the XML document itself.</p> - - <p>Another method of providing the schema location information - is via the <code>xml_schema::properties</code> argument, as - shown in the following example:</p> - - <pre class="c++"> -xml_schema::properties props; -props.no_namespace_schema_location ("people.xsd"); -props.schema_location ("http://www.w3.org/XML/1998/namespace", "xml.xsd"); - -unique_ptr<people_t> p (people ("people.xml", 0, props)); - </pre> - - <p>The schema locations provided with this method overrides - those specified in the XML document. As with the previous - method, the schema locations specified this way are - relative to the document's path unless they are absolute URIs. - In particular, if you want to use local schemas that are - not related to the document being parsed, then you will - need to use the <code>file://</code> URI. The following - example shows how to use schemas that reside in the current - working directory:</p> - - <pre class="c++"> -#include <unistd.h> // getcwd -#include <limits.h> // PATH_MAX - -char cwd[PATH_MAX]; -if (getcwd (cwd, PATH_MAX) == 0) -{ - // Buffer too small? -} - -xml_schema::properties props; - -props.no_namespace_schema_location ( - "file:///" + std::string (cwd) + "/people.xsd"); - -props.schema_location ( - "http://www.w3.org/XML/1998/namespace", - "file:///" + std::string (cwd) + "/xml.xsd"); - -unique_ptr<people_t> p (people ("people.xml", 0, props)); - </pre> - - <p>A third method is the most useful if you are planning to parse - several XML documents of the same vocabulary. In that case - it may be beneficial to pre-parse and cache the schemas in - the XML parser which can then be used to parse all documents - without re-parsing the schemas. For more information on - this method refer to the <code>caching</code> example in the - <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package. - It is also possible to convert the schemas into a pre-compiled - binary representation and embed this representation directly into - the application executable. With this approach your application can - perform XML Schema validation without depending on any external - schema files. For more information on how to achieve this refer to - the <code>embedded</code> example in the <code>cxx/tree/</code> - directory in the <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <p>When the XML parser cannot locate a schema for the - XML document, the validation fails and XML document - elements and attributes for which schema definitions could - not be located are reported in the diagnostics. For - example, if we remove the <code>noNamespaceSchemaLocation</code> - attribute in <code>people.xml</code> from the previous chapter, - then we will get the following diagnostics if we try to parse - this file with validation enabled:</p> - - <pre class="terminal"> -people.xml:2:63 error: no declaration found for element 'people' -people.xml:4:18 error: no declaration found for element 'person' -people.xml:4:18 error: attribute 'id' is not declared for element 'person' -people.xml:5:17 error: no declaration found for element 'first-name' -people.xml:6:18 error: no declaration found for element 'middle-name' -people.xml:7:16 error: no declaration found for element 'last-name' -people.xml:8:13 error: no declaration found for element 'gender' -people.xml:9:10 error: no declaration found for element 'age' - </pre> - - <h2><a name="5.2">5.2 Error Handling</a></h2> - - <p>The parsing functions offer a number of ways to handle error conditions - with the C++ exceptions being the most commonly used mechanism. All - C++/Tree exceptions derive from common base <code>xml_schema::exception</code> - which in turn derives from <code>std::exception</code>. The easiest - way to uniformly handle all possible C++/Tree exceptions and print - detailed information about the error is to catch and print - <code>xml_schema::exception</code>, as shown in the following - example:</p> - - <pre class="c++"> -try -{ - unique_ptr<people_t> p (people ("people.xml")); -} -catch (const xml_schema::exception& e) -{ - cerr << e << endl; -} - </pre> - - <p>Each individual C++/Tree exception also allows you to obtain - error details programmatically. For example, the - <code>xml_schema::parsing</code> exception is thrown when - the XML parsing and validation in the underlying XML parser - fails. It encapsulates various diagnostics information - such as the file name, line and column numbers, as well as the - error or warning message for each entry. For more information - about this and other exceptions that can be thrown during - parsing, refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#3.3">Section - 3.3, "Error Handling"</a> in the C++/Tree Mapping - User Manual.</p> - - <p>Note that if you are parsing <code>std::istream</code> on which - exceptions are not enabled, then you will need to check the - stream state after the call to the parsing function in order - to detect any possible stream failures, for example:</p> - - <pre class="c++"> -std::ifstream ifs ("people.xml"); - -if (ifs.fail ()) -{ - cerr << "people.xml: unable to open" << endl; - return 1; -} - -unique_ptr<people_t> p (people (ifs, "people.xml")); - -if (ifs.fail ()) -{ - cerr << "people.xml: read error" << endl; - return 1; -} - </pre> - - <p>The above example can be rewritten to use exceptions as - shown below:</p> - - <pre class="c++"> -try -{ - std::ifstream ifs; - ifs.exceptions (std::ifstream::badbit | std::ifstream::failbit); - ifs.open ("people.xml"); - - unique_ptr<people_t> p (people (ifs, "people.xml")); -} -catch (const std::ifstream::failure&) -{ - cerr << "people.xml: unable to open or read error" << endl; - return 1; -} - </pre> - - - <!-- Chapater 6 --> - - - <h1><a name="6">6 Serialization</a></h1> - - <p>We have already seen how to serialize an object model back to XML - in this guide before. In this chapter we will discuss the - serialization topic in more detail.</p> - - <p>By default, the C++/Tree mapping provides a total of 8 overloaded - serialization functions. They differ in the output methods used to write - XML as well as the error reporting mechanisms. It is also possible to - generate types for root elements instead of parsing and serialization - functions. This may be useful if your XML vocabulary has multiple - root elements. For more information on element types refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#2.9">Section - 2.9, "Mapping for Global Elements"</a> in the C++/Tree Mapping User - Manual.</p> - - - <p>In this section we will discuss the most commonly - used version of serialization functions. For a comprehensive description - of serialization refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#4">Chapter - 4, "Serialization"</a> in the C++/Tree Mapping User Manual. For the - <code>people</code> global element from our person record vocabulary, - we will concentrate on the following serialization function:</p> - - <pre class="c++"> -void -people (std::ostream& os, - const people_t& x, - const xml_schema::namespace_infomap& map = - xml_schema::namespace_infomap (), - const std::string& encoding = "UTF-8", - xml_schema::flags f = 0); - </pre> - - <p>This function serializes the object model passed as the second - argument to the standard output stream passed as the first - argument. The third argument is a namespace information map - which we will discuss in more detail in the next section. - The fourth argument is a character encoding that the resulting - XML document should be in. Possible valid values for this - argument are "US-ASCII", "ISO8859-1", "UTF-8", "UTF-16BE", - "UTF-16LE", "UCS-4BE", and "UCS-4LE". Finally, the flags - argument allows fine-tuning of the serialization process. - The following example shows how we can use the above serialization - function:</p> - - <pre class="c++"> -people_t& p = ... - -xml_schema::namespace_infomap map; -map[""].schema = "people.xsd"; - -// Serialize to stdout. -// -people (std::cout, p, map); - -// Serialize to a file. -// -std::ofstream ofs ("people.xml"); -people (ofs, p, map); - -// Serialize to a string. -// -std::ostringstream oss; -people (oss, p, map); -std::string xml (oss.str ()); - </pre> - - - <h2><a name="6.1">6.1 Namespace and Schema Information</a></h2> - - <p>While XML serialization can be done just from the object - model alone, it is often desirable to assign meaningful - prefixes to XML namespaces used in the vocabulary as - well as to provide the schema location information. - This is accomplished by passing the namespace information - map to the serialization function. The key in this map is - a namespace prefix that should be assigned to an XML namespace - specified in the <code>name</code> variable of the - map value. You can also assign an optional schema location for - this namespace in the <code>schema</code> variable. Based - on each key-value entry in this map, the serialization - function adds two attributes to the resulting XML document: - the namespace-prefix mapping attribute and schema location - attribute. The empty prefix indicates that the namespace - should be mapped without a prefix. For example, the following - map:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].name = "http://www.example.com/example"; -map[""].schema = "example.xsd"; - -map["x"].name = "http://www.w3.org/XML/1998/namespace"; -map["x"].schema = "xml.xsd"; - </pre> - - <p>Results in the following XML document:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<example - xmlns="http://www.example.com/example" - xmlns:x="http://www.w3.org/XML/1998/namespace" - xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:schemaLocation="http://www.example.com/example example.xsd - http://www.w3.org/XML/1998/namespace xml.xsd"> - </pre> - - <p>The empty namespace indicates that the vocabulary has no target - namespace. For example, the following map results in only the - <code>noNamespaceSchemaLocation</code> attribute being added:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].name = ""; -map[""].schema = "example.xsd"; - </pre> - - <h2><a name="6.2">6.2 Error Handling</a></h2> - - <p>Similar to the parsing functions, the serialization functions offer a - number of ways to handle error conditions with the C++ exceptions being - the most commonly used mechanisms. As with parsing, the easiest way to - uniformly handle all possible serialization exceptions and print - detailed information about the error is to catch and print - <code>xml_schema::exception</code>:</p> - - <pre class="c++"> -try -{ - people_t& p = ... - - xml_schema::namespace_infomap map; - map[""].schema = "people.xsd"; - - people (std::cout, p, map)); -} -catch (const xml_schema::exception& e) -{ - cerr << e << endl; -} - </pre> - - <p>The most commonly encountered serialization exception is - <code>xml_schema::serialization</code>. It is thrown - when the XML serialization in the underlying XML writer - fails. It encapsulates various diagnostics information - such as the file name, line and column numbers, as well as the - error or warning message for each entry. For more information - about this and other exceptions that can be thrown during - serialization, refer to - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/#4.4">Section - 4.4, "Error Handling"</a> in the C++/Tree Mapping - User Manual.</p> - - <p>Note that if you are serializing to <code>std::ostream</code> on - which exceptions are not enabled, then you will need to check the - stream state after the call to the serialization function in order - to detect any possible stream failures, for example:</p> - - <pre class="c++"> -std::ofstream ofs ("people.xml"); - -if (ofs.fail ()) -{ - cerr << "people.xml: unable to open" << endl; - return 1; -} - -people (ofs, p, map)); - -if (ofs.fail ()) -{ - cerr << "people.xml: write error" << endl; - return 1; -} - </pre> - - <p>The above example can be rewritten to use exceptions as - shown below:</p> - - <pre class="c++"> -try -{ - std::ofstream ofs; - ofs.exceptions (std::ofstream::badbit | std::ofstream::failbit); - ofs.open ("people.xml"); - - people (ofs, p, map)); -} -catch (const std::ofstream::failure&) -{ - cerr << "people.xml: unable to open or write error" << endl; - return 1; -} - </pre> - - </div> -</div> - -</body> -</html> diff --git a/doc/cxx/tree/manual/index.xhtml b/doc/cxx/tree/manual/index.xhtml deleted file mode 100644 index f455bff..0000000 --- a/doc/cxx/tree/manual/index.xhtml +++ /dev/null @@ -1,6826 +0,0 @@ -<?xml version="1.0" encoding="iso-8859-1"?> -<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> -<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"> - -<head> - <title>C++/Tree Mapping User Manual</title> - - <meta name="copyright" content="© 2005-2023 Code Synthesis"/> - <meta name="keywords" content="xsd,xml,schema,c++,mapping,data,binding,tree,serialization,guide,manual,examples"/> - <meta name="description" content="C++/Tree Mapping User Manual"/> - <meta name="revision" content="4.1.0"/> - - <link rel="stylesheet" type="text/css" href="../../../default.css" /> - -<style type="text/css"> - pre { - padding : 0 0 0 0em; - margin : 0em 0em 0em 0; - - font-size : 102% - } - - body { - min-width: 48em; - } - - h1 { - font-weight: bold; - font-size: 200%; - } - - h2 { - font-weight : bold; - font-size : 150%; - - padding-top : 0.8em; - } - - h3 { - font-size : 130%; - padding-top : 0.8em; - } - - /* Adjust indentation for three levels. */ - #container { - max-width: 48em; - } - - #content { - padding: 0 0.1em 0 4em; - /*background-color: red;*/ - } - - #content h1 { - margin-left: -2.06em; - } - - #content h2 { - margin-left: -1.33em; - } - - /* Title page */ - - #titlepage { - padding: 2em 0 1em 0; - border-bottom: 1px solid black; - } - - #titlepage #title { - font-weight: bold; - font-size: 200%; - text-align: center; - padding: 1em 0 2em 0; - } - - /* Lists */ - ul.list li { - padding-top : 0.3em; - padding-bottom : 0.3em; - } - - - /* Built-in table */ - #builtin { - margin: 2em 0 2em 0; - - border-collapse : collapse; - border : 1px solid; - border-color : #000000; - - font-size : 11px; - line-height : 14px; - } - - #builtin th, #builtin td { - border: 1px solid; - padding : 0.9em 0.9em 0.7em 0.9em; - } - - #builtin th { - background : #cde8f6; - } - - #builtin td { - text-align: left; - } - - - /* default-fixed */ - #default-fixed { - margin: 2em 0 2em 0; - - border-collapse : collapse; - border : 1px solid; - border-color : #000000; - - font-size : 11px; - line-height : 14px; - } - - #default-fixed th, #default-fixed td { - border: 1px solid; - padding : 0.9em 0.9em 0.7em 0.9em; - } - - #default-fixed th { - background : #cde8f6; - } - - #default-fixed td { - text-align: center; - } - - - /* */ - dl dt { - padding : 0.8em 0 0 0; - } - - - /* TOC */ - table.toc { - border-style : none; - border-collapse : separate; - border-spacing : 0; - - margin : 0.2em 0 0.2em 0; - padding : 0 0 0 0; - } - - table.toc tr { - padding : 0 0 0 0; - margin : 0 0 0 0; - } - - table.toc * td, table.toc * th { - border-style : none; - margin : 0 0 0 0; - vertical-align : top; - } - - table.toc * th { - font-weight : normal; - padding : 0em 0.1em 0em 0; - text-align : left; - white-space : nowrap; - } - - table.toc * table.toc th { - padding-left : 1em; - } - - table.toc * td { - padding : 0em 0 0em 0.7em; - text-align : left; - } -</style> - - -</head> - -<body> -<div id="container"> - <div id="content"> - - <div class="noprint"> - - <div id="titlepage"> - <div id="title">C++/Tree Mapping User Manual</div> - - <p>Copyright © 2005-2023 Code Synthesis.</p> - - <p>Permission is granted to copy, distribute and/or modify this - document under the terms of the - <a href="https://www.codesynthesis.com/licenses/fdl-1.2.txt">GNU Free - Documentation License, version 1.2</a>; with no Invariant Sections, - no Front-Cover Texts and no Back-Cover Texts. - </p> - - <p>This document is available in the following formats: - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/index.xhtml">XHTML</a>, - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.pdf">PDF</a>, and - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.ps">PostScript</a>.</p> - </div> - - <h1>Table of Contents</h1> - - <table class="toc"> - <tr> - <th></th><td><a href="#0">Preface</a> - <table class="toc"> - <tr><th></th><td><a href="#0.1">About This Document</a></td></tr> - <tr><th></th><td><a href="#0.2">More Information</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>1</th><td><a href="#1">Introduction</a></td> - </tr> - - <tr> - <th>2</th><td><a href="#2">C++/Tree Mapping</a> - <table class="toc"> - <tr> - <th>2.1</th><td><a href="#2.1">Preliminary Information</a> - <table class="toc"> - <tr><th>2.1.1</th><td><a href="#2.1.1">C++ Standard</a></td></tr> - <tr><th>2.1.2</th><td><a href="#2.1.2">Identifiers</a></td></tr> - <tr><th>2.1.3</th><td><a href="#2.1.3">Character Type and Encoding</a></td></tr> - <tr><th>2.1.4</th><td><a href="#2.1.4">XML Schema Namespace</a></td></tr> - <tr><th>2.1.5</th><td><a href="#2.1.5">Anonymous Types</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.2</th><td><a href="#2.2">Error Handling</a> - <table class="toc"> - <tr><th>2.2.1</th><td><a href="#2.2.1"><code>xml_schema::duplicate_id</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.3</th><td><a href="#2.3">Mapping for <code>import</code> and <code>include</code></a> - <table class="toc"> - <tr><th>2.3.1</th><td><a href="#2.3.1">Import</a></td></tr> - <tr><th>2.3.2</th><td><a href="#2.3.2">Inclusion with Target Namespace</a></td></tr> - <tr><th>2.3.3</th><td><a href="#2.3.3">Inclusion without Target Namespace</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.4</th><td><a href="#2.4">Mapping for Namespaces</a></td> - </tr> - <tr> - <th>2.5</th><td><a href="#2.5">Mapping for Built-in Data Types</a> - <table class="toc"> - <tr><th>2.5.1</th><td><a href="#2.5.1">Inheritance from Built-in Data Types</a></td></tr> - <tr><th>2.5.2</th><td><a href="#2.5.2">Mapping for <code>anyType</code></a></td></tr> - <tr><th>2.5.3</th><td><a href="#2.5.3">Mapping for <code>anySimpleType</code></a></td></tr> - <tr><th>2.5.4</th><td><a href="#2.5.4">Mapping for <code>QName</code></a></td></tr> - <tr><th>2.5.5</th><td><a href="#2.5.5">Mapping for <code>IDREF</code></a></td></tr> - <tr><th>2.5.6</th><td><a href="#2.5.6">Mapping for <code>base64Binary</code> and <code>hexBinary</code></a></td></tr> - <tr><th>2.5.7</th><td><a href="#2.5.7">Time Zone Representation</a></td></tr> - <tr><th>2.5.8</th><td><a href="#2.5.8">Mapping for <code>date</code></a></td></tr> - <tr><th>2.5.9</th><td><a href="#2.5.9">Mapping for <code>dateTime</code></a></td></tr> - <tr><th>2.5.10</th><td><a href="#2.5.10">Mapping for <code>duration</code></a></td></tr> - <tr><th>2.5.11</th><td><a href="#2.5.11">Mapping for <code>gDay</code></a></td></tr> - <tr><th>2.5.12</th><td><a href="#2.5.12">Mapping for <code>gMonth</code></a></td></tr> - <tr><th>2.5.13</th><td><a href="#2.5.13">Mapping for <code>gMonthDay</code></a></td></tr> - <tr><th>2.5.14</th><td><a href="#2.5.14">Mapping for <code>gYear</code></a></td></tr> - <tr><th>2.5.15</th><td><a href="#2.5.15">Mapping for <code>gYearMonth</code></a></td></tr> - <tr><th>2.5.16</th><td><a href="#2.5.16">Mapping for <code>time</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.6</th><td><a href="#2.6">Mapping for Simple Types</a> - <table class="toc"> - <tr><th>2.6.1</th><td><a href="#2.6.1">Mapping for Derivation by Restriction</a></td></tr> - <tr><th>2.6.2</th><td><a href="#2.6.2">Mapping for Enumerations</a></td></tr> - <tr><th>2.6.3</th><td><a href="#2.6.3">Mapping for Derivation by List</a></td></tr> - <tr><th>2.6.4</th><td><a href="#2.6.4">Mapping for Derivation by Union</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.7</th><td><a href="#2.7">Mapping for Complex Types</a> - <table class="toc"> - <tr><th>2.7.1</th><td><a href="#2.7.1">Mapping for Derivation by Extension</a></td></tr> - <tr><th>2.7.2</th><td><a href="#2.7.2">Mapping for Derivation by Restriction</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.8</th><td><a href="#2.8">Mapping for Local Elements and Attributes</a> - <table class="toc"> - <tr><th>2.8.1</th><td><a href="#2.8.1">Mapping for Members with the One Cardinality Class</a></td></tr> - <tr><th>2.8.2</th><td><a href="#2.8.2">Mapping for Members with the Optional Cardinality Class</a></td></tr> - <tr><th>2.8.3</th><td><a href="#2.8.3">Mapping for Members with the Sequence Cardinality Class</a></td></tr> - <tr><th>2.8.4</th><td><a href="#2.8.4">Element Order</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.9</th><td><a href="#2.9">Mapping for Global Elements</a> - <table class="toc"> - <tr><th>2.9.1</th><td><a href="#2.9.1">Element Types</a></td></tr> - <tr><th>2.9.2</th><td><a href="#2.9.2">Element Map</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.10</th><td><a href="#2.10">Mapping for Global Attributes</a></td> - </tr> - <tr> - <th>2.11</th><td><a href="#2.11">Mapping for <code>xsi:type</code> and Substitution Groups</a></td> - </tr> - <tr> - <th>2.12</th><td><a href="#2.12">Mapping for <code>any</code> and <code>anyAttribute</code></a> - <table class="toc"> - <tr><th>2.12.1</th><td><a href="#2.12.1">Mapping for <code>any</code> with the One Cardinality Class</a></td></tr> - <tr><th>2.12.2</th><td><a href="#2.12.2">Mapping for <code>any</code> with the Optional Cardinality Class</a></td></tr> - <tr><th>2.12.3</th><td><a href="#2.12.3">Mapping for <code>any</code> with the Sequence Cardinality Class</a></td></tr> - <tr><th>2.12.4</th><td><a href="#2.12.4">Element Wildcard Order</a></td></tr> - <tr><th>2.12.5</th><td><a href="#2.12.5">Mapping for <code>anyAttribute</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.13</th><td><a href="#2.13">Mapping for Mixed Content Models</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th>3</th><td><a href="#3">Parsing</a> - <table class="toc"> - <tr> - <th>3.1</th><td><a href="#3.1">Initializing the Xerces-C++ Runtime</a></td> - </tr> - <tr> - <th>3.2</th><td><a href="#3.2">Flags and Properties</a></td> - </tr> - <tr> - <th>3.3</th><td><a href="#3.3">Error Handling</a> - <table class="toc"> - <tr><th>3.3.1</th><td><a href="#3.3.1"><code>xml_schema::parsing</code></a></td></tr> - <tr><th>3.3.2</th><td><a href="#3.3.2"><code>xml_schema::expected_element</code></a></td></tr> - <tr><th>3.3.3</th><td><a href="#3.3.3"><code>xml_schema::unexpected_element</code></a></td></tr> - <tr><th>3.3.4</th><td><a href="#3.3.4"><code>xml_schema::expected_attribute</code></a></td></tr> - <tr><th>3.3.5</th><td><a href="#3.3.5"><code>xml_schema::unexpected_enumerator</code></a></td></tr> - <tr><th>3.3.6</th><td><a href="#3.3.6"><code>xml_schema::expected_text_content</code></a></td></tr> - <tr><th>3.3.7</th><td><a href="#3.3.7"><code>xml_schema::no_type_info</code></a></td></tr> - <tr><th>3.3.8</th><td><a href="#3.3.8"><code>xml_schema::not_derived</code></a></td></tr> - <tr><th>3.3.9</th><td><a href="#3.3.9"><code>xml_schema::not_prefix_mapping</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>3.4</th><td><a href="#3.4">Reading from a Local File or URI</a></td> - </tr> - <tr> - <th>3.5</th><td><a href="#3.5">Reading from <code>std::istream</code></a></td> - </tr> - <tr> - <th>3.6</th><td><a href="#3.6">Reading from <code>xercesc::InputSource</code></a></td> - </tr> - <tr> - <th>3.7</th><td><a href="#3.7">Reading from DOM</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th>4</th><td><a href="#4">Serialization</a> - <table class="toc"> - <tr> - <th>4.1</th><td><a href="#4.1">Initializing the Xerces-C++ Runtime</a></td> - </tr> - <tr> - <th>4.2</th><td><a href="#4.2">Namespace Infomap and Character Encoding</a></td> - </tr> - <tr> - <th>4.3</th><td><a href="#4.3">Flags</a></td> - </tr> - <tr> - <th>4.4</th><td><a href="#4.4">Error Handling</a> - <table class="toc"> - <tr><th>4.4.1</th><td><a href="#4.4.1"><code>xml_schema::serialization</code></a></td></tr> - <tr><th>4.4.2</th><td><a href="#4.4.2"><code>xml_schema::unexpected_element</code></a></td></tr> - <tr><th>4.4.3</th><td><a href="#4.4.3"><code>xml_schema::no_type_info</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>4.5</th><td><a href="#4.5">Serializing to <code>std::ostream</code></a></td> - </tr> - <tr> - <th>4.6</th><td><a href="#4.6">Serializing to <code>xercesc::XMLFormatTarget</code></a></td> - </tr> - <tr> - <th>4.7</th><td><a href="#4.7">Serializing to DOM</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th>5</th><td><a href="#5">Additional Functionality</a> - <table class="toc"> - <tr> - <th>5.1</th><td><a href="#5.1">DOM Association</a></td> - </tr> - <tr> - <th>5.2</th><td><a href="#5.2">Binary Serialization</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th></th><td><a href="#A">Appendix A — Default and Fixed Values</a></td> - </tr> - - </table> - </div> - - <h1><a name="0">Preface</a></h1> - - <h2><a name="0.1">About This Document</a></h2> - - <p>This document describes the mapping of W3C XML Schema - to the C++ programming language as implemented by - <a href="https://www.codesynthesis.com/products/xsd">CodeSynthesis - XSD</a> - an XML Schema to C++ data binding compiler. The mapping - represents information stored in XML instance documents as a - statically-typed, tree-like in-memory data structure and is - called C++/Tree. - </p> - - <p>Revision 4.1.0<br/> <!-- Remember to change revision in other places --> - This revision of the manual describes the C++/Tree - mapping as implemented by CodeSynthesis XSD version 4.1.0. - </p> - - <p>This document is available in the following formats: - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/index.xhtml">XHTML</a>, - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.pdf">PDF</a>, and - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.ps">PostScript</a>.</p> - - <h2><a name="0.2">More Information</a></h2> - - <p>Beyond this manual, you may also find the following sources of - information useful:</p> - - <ul class="list"> - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/">C++/Tree - Mapping Getting Started Guide</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/Customization_guide">C++/Tree - Mapping Customization Guide</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree - Mapping Frequently Asked Questions (FAQ)</a></li> - - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a></li> - - <li>The <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - contains a collection of examples and a README file with an overview - of each example.</li> - - <li>The <code>README</code> file in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - explains how to build the examples.</li> - - <li>The <a href="https://www.codesynthesis.com/mailman/listinfo/xsd-users">xsd-users</a> - mailing list is a place to ask questions. Furthermore the - <a href="https://www.codesynthesis.com/pipermail/xsd-users/">archives</a> - may already have answers to some of your questions.</li> - </ul> - - - <h1><a name="1">1 Introduction</a></h1> - - <p>C++/Tree is a W3C XML Schema to C++ mapping that represents the - data stored in XML as a statically-typed, vocabulary-specific - object model. Based on a formal description of an XML vocabulary - (schema), the C++/Tree mapping produces a tree-like data structure - suitable for in-memory processing as well as XML parsing and - serialization code.</p> - - <p>A typical application that processes XML documents usually - performs the following three steps: it first reads (parses) an XML - instance document to an object model, it then performs - some useful computations on that model which may involve - modification of the model, and finally it may write (serialize) - the modified object model back to XML. - </p> - - <p>The C++/Tree mapping consists of C++ types that represent the - given vocabulary (<a href="#2">Chapter 2, "C++/Tree Mapping"</a>), - a set of parsing functions that convert XML documents to - a tree-like in-memory data structure (<a href="#3">Chapter 3, - "Parsing"</a>), and a set of serialization functions that convert - the object model back to XML (<a href="#4">Chapter 4, - "Serialization"</a>). Furthermore, the mapping provides a number - of additional features, such as DOM association and binary - serialization, that can be useful in some applications - (<a href="#5">Chapter 5, "Additional Functionality"</a>). - </p> - - - <!-- Chapter 2 --> - - - <h1><a name="2">2 C++/Tree Mapping</a></h1> - - <h2><a name="2.1">2.1 Preliminary Information</a></h2> - - <h3><a name="2.1.1">2.1.1 C++ Standard</a></h3> - - <p>The C++/Tree mapping provides support for ISO/IEC C++ 2011 (C++11) - and ISO/IEC C++ 1998/2003 (C++98). To select the C++ standard for the - generated code we use the <code>--std</code> XSD compiler command - line option. While the majority of the examples in this guide use - C++11, the document explains the C++11/98 usage difference and so - they can easily be converted to C++98.</p> - - <h3><a name="2.1.2">2.1.2 Identifiers</a></h3> - - <p>XML Schema names may happen to be reserved C++ keywords or contain - characters that are illegal in C++ identifiers. To avoid C++ compilation - problems, such names are changed (escaped) when mapped to C++. If an - XML Schema name is a C++ keyword, the "_" suffix is added to it. All - character of an XML Schema name that are not allowed in C++ identifiers - are replaced with "_". - </p> - - <p>For example, XML Schema name <code>try</code> will be mapped to - C++ identifier <code>try_</code>. Similarly, XML Schema name - <code>strange.na-me</code> will be mapped to C++ identifier - <code>strange_na_me</code>. - </p> - - <p>Furthermore, conflicts between type names and function names in the - same scope are resolved using name escaping. Such conflicts include - both a global element (which is mapped to a set of parsing and/or - serialization functions or element types, see <a href="#2.9">Section - 2.9, "Mapping for Global Elements"</a>) and a global type sharing the - same name as well as a local element or attribute inside a type having - the same name as the type itself.</p> - - <p>For example, if we had a global type <code>catalog</code> - and a global element with the same name then the type would be - mapped to a C++ class with name <code>catalog</code> while the - parsing functions corresponding to the global element would have - their names escaped as <code>catalog_</code>. - </p> - - <p>By default the mapping uses the so-called K&R (Kernighan and - Ritchie) identifier naming convention which is also used throughout - this manual. In this convention both type and function names are in - lower case and words are separated by underscores. If your application - code or schemas use a different notation, you may want to change the - naming convention used by the mapping for consistency. - The compiler supports a set of widely-used naming conventions - that you can select with the <code>--type-naming</code> and - <code>--function-naming</code> options. You can also further - refine one of the predefined conventions or create a completely - custom naming scheme by using the <code>--*-regex</code> options. - For more detailed information on these options refer to the NAMING - CONVENTION section in the <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>.</p> - - <h3><a name="2.1.3">2.1.3 Character Type and Encoding</a></h3> - - <p>The code that implements the mapping, depending on the - <code>--char-type</code> option, is generated using either - <code>char</code> or <code>wchar_t</code> as the character - type. In this document code samples use symbol <code>C</code> - to refer to the character type you have selected when translating - your schemas, for example <code>std::basic_string<C></code>. - </p> - - <p>Another aspect of the mapping that depends on the character type - is character encoding. For the <code>char</code> character type - the default encoding is UTF-8. Other supported encodings are - ISO-8859-1, Xerces-C++ Local Code Page (LPC), as well as - custom encodings and can be selected with the - <code>--char-encoding</code> command line option.</p> - - <p>For the <code>wchar_t</code> character type the encoding is - automatically selected between UTF-16 and UTF-32/UCS-4 depending - on the size of the <code>wchar_t</code> type. On some platforms - (for example, Windows with Visual C++ and AIX with IBM XL C++) - <code>wchar_t</code> is 2 bytes long. For these platforms the - encoding is UTF-16. On other platforms <code>wchar_t</code> is 4 bytes - long and UTF-32/UCS-4 is used.</p> - - <h3><a name="2.1.4">2.1.4 XML Schema Namespace</a></h3> - - <p>The mapping relies on some predefined types, classes, and functions - that are logically defined in the XML Schema namespace reserved for - the XML Schema language (<code>http://www.w3.org/2001/XMLSchema</code>). - By default, this namespace is mapped to C++ namespace - <code>xml_schema</code>. It is automatically accessible - from a C++ compilation unit that includes a header file generated - from an XML Schema definition. - </p> - - <p>Note that, if desired, the default mapping of this namespace can be - changed as described in <a href="#2.4">Section 2.4, "Mapping for - Namespaces"</a>. - </p> - - - <h3><a name="2.1.5">2.1.5 Anonymous Types</a></h3> - - <p>For the purpose of code generation, anonymous types defined in - XML Schema are automatically assigned names that are derived - from enclosing attributes and elements. Otherwise, such types - follows standard mapping rules for simple and complex type - definitions (see <a href="#2.6">Section 2.6, "Mapping for Simple Types"</a> - and <a href="#2.7">Section 2.7, "Mapping for Complex Types"</a>). - For example, in the following schema fragment: - </p> - - <pre class="xml"> -<element name="object"> - <complexType> - ... - </complexType> -</element> - </pre> - - <p>The anonymous type defined inside element <code>object</code> will - be given name <code>object</code>. The compiler has a number of - options that control the process of anonymous type naming. For more - information refer to the <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>.</p> - - - <h2><a name="2.2">2.2 Error Handling</a></h2> - - <p>The mapping uses the C++ exception handling mechanism as a primary way - of reporting error conditions. All exceptions that are specified in - this mapping derive from <code>xml_schema::exception</code> which - itself is derived from <code>std::exception</code>: - </p> - - <pre class="c++"> -struct exception: virtual std::exception -{ - friend - std::basic_ostream<C>& - operator<< (std::basic_ostream<C>& os, const exception& e) - { - e.print (os); - return os; - } - -protected: - virtual void - print (std::basic_ostream<C>&) const = 0; -}; - </pre> - - <p>The exception hierarchy supports "virtual" <code>operator<<</code> - which allows you to obtain diagnostics corresponding to the thrown - exception using the base exception interface. For example:</p> - - <pre class="c++"> -try -{ - ... -} -catch (const xml_schema::exception& e) -{ - cerr << e << endl; -} - </pre> - - <p>The following sub-sections describe exceptions thrown by the - types that constitute the object model. - <a href="#3.3">Section 3.3, "Error Handling"</a> of - <a href="#3">Chapter 3, "Parsing"</a> describes exceptions - and error handling mechanisms specific to the parsing functions. - <a href="#4.4">Section 4.4, "Error Handling"</a> of - <a href="#4">Chapter 4, "Serialization"</a> describes exceptions - and error handling mechanisms specific to the serialization functions. - </p> - - - <h3><a name="2.2.1">2.2.1 <code>xml_schema::duplicate_id</code></a></h3> - - <pre class="c++"> -struct duplicate_id: virtual exception -{ - duplicate_id (const std::basic_string<C>& id); - - const std::basic_string<C>& - id () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::duplicate_id</code> is thrown when - a conflicting instance of <code>xml_schema::id</code> (see - <a href="#2.5">Section 2.5, "Mapping for Built-in Data Types"</a>) - is added to a tree. The offending ID value can be obtained using - the <code>id</code> function. - </p> - - <h2><a name="2.3">2.3 Mapping for <code>import</code> and <code>include</code></a></h2> - - <h3><a name="2.3.1">2.3.1 Import</a></h3> - - <p>The XML Schema <code>import</code> element is mapped to the C++ - Preprocessor <code>#include</code> directive. The value of - the <code>schemaLocation</code> attribute is used to derive - the name of the header file that appears in the <code>#include</code> - directive. For instance: - </p> - - <pre class="xml"> -<import namespace="https://www.codesynthesis.com/test" - schemaLocation="test.xsd"/> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -#include "test.hxx" - </pre> - - <p>Note that you will need to compile imported schemas separately - in order to produce corresponding header files.</p> - - <h3><a name="2.3.2">2.3.2 Inclusion with Target Namespace</a></h3> - - <p>The XML Schema <code>include</code> element which refers to a schema - with a target namespace or appears in a schema without a target namespace - follows the same mapping rules as the <code>import</code> element, - see <a href="#2.3.1">Section 2.3.1, "Import"</a>. - </p> - - <h3><a name="2.3.3">2.3.3 Inclusion without Target Namespace</a></h3> - - <p>For the XML Schema <code>include</code> element which refers to a schema - without a target namespace and appears in a schema with a target - namespace (such inclusion sometimes called "chameleon inclusion"), - declarations and definitions from the included schema are generated - in-line in the namespace of the including schema as if they were - declared and defined there verbatim. For example, consider the - following two schemas: - </p> - - <pre class="xml"> -<-- common.xsd --> -<schema> - <complexType name="type"> - ... - </complexType> -</schema> - -<-- test.xsd --> -<schema targetNamespace="https://www.codesynthesis.com/test"> - <include schemaLocation="common.xsd"/> -</schema> - </pre> - - <p>The fragment of interest from the generated header file for - <code>text.xsd</code> would look like this:</p> - - <pre class="c++"> -// test.hxx -namespace test -{ - class type - { - ... - }; -} - </pre> - - <h2><a name="2.4">2.4 Mapping for Namespaces</a></h2> - - <p>An XML Schema namespace is mapped to one or more nested C++ - namespaces. XML Schema namespaces are identified by URIs. - By default, a namespace URI is mapped to a sequence of - C++ namespace names by removing the protocol and host parts - and splitting the rest into a sequence of names with '<code>/</code>' - as the name separator. For instance: - </p> - - <pre class="xml"> -<schema targetNamespace="https://www.codesynthesis.com/system/test"> - ... -</schema> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -namespace system -{ - namespace test - { - ... - } -} - </pre> - - <p>The default mapping of namespace URIs to C++ namespace names can be - altered using the <code>--namespace-map</code> and - <code>--namespace-regex</code> options. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information. - </p> - - <h2><a name="2.5">2.5 Mapping for Built-in Data Types</a></h2> - - <p>The mapping of XML Schema built-in data types to C++ types is - summarized in the table below.</p> - - <!-- border="1" is necessary for html2ps --> - <table id="builtin" border="1"> - <tr> - <th>XML Schema type</th> - <th>Alias in the <code>xml_schema</code> namespace</th> - <th>C++ type</th> - </tr> - - <tr> - <th colspan="3">anyType and anySimpleType types</th> - </tr> - <tr> - <td><code>anyType</code></td> - <td><code>type</code></td> - <td><a href="#2.5.2">Section 2.5.2, "Mapping for <code>anyType</code>"</a></td> - </tr> - <tr> - <td><code>anySimpleType</code></td> - <td><code>simple_type</code></td> - <td><a href="#2.5.3">Section 2.5.3, "Mapping for <code>anySimpleType</code>"</a></td> - </tr> - - <tr> - <th colspan="3">fixed-length integral types</th> - </tr> - <!-- 8-bit --> - <tr> - <td><code>byte</code></td> - <td><code>byte</code></td> - <td><code>signed char</code></td> - </tr> - <tr> - <td><code>unsignedByte</code></td> - <td><code>unsigned_byte</code></td> - <td><code>unsigned char</code></td> - </tr> - - <!-- 16-bit --> - <tr> - <td><code>short</code></td> - <td><code>short_</code></td> - <td><code>short</code></td> - </tr> - <tr> - <td><code>unsignedShort</code></td> - <td><code>unsigned_short</code></td> - <td><code>unsigned short</code></td> - </tr> - - <!-- 32-bit --> - <tr> - <td><code>int</code></td> - <td><code>int_</code></td> - <td><code>int</code></td> - </tr> - <tr> - <td><code>unsignedInt</code></td> - <td><code>unsigned_int</code></td> - <td><code>unsigned int</code></td> - </tr> - - <!-- 64-bit --> - <tr> - <td><code>long</code></td> - <td><code>long_</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>unsignedLong</code></td> - <td><code>unsigned_long</code></td> - <td><code>unsigned long long</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-length integral types</th> - </tr> - <tr> - <td><code>integer</code></td> - <td><code>integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonPositiveInteger</code></td> - <td><code>non_positive_integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonNegativeInteger</code></td> - <td><code>non_negative_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>positiveInteger</code></td> - <td><code>positive_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>negativeInteger</code></td> - <td><code>negative_integer</code></td> - <td><code>long long</code></td> - </tr> - - <tr> - <th colspan="3">boolean types</th> - </tr> - <tr> - <td><code>boolean</code></td> - <td><code>boolean</code></td> - <td><code>bool</code></td> - </tr> - - <tr> - <th colspan="3">fixed-precision floating-point types</th> - </tr> - <tr> - <td><code>float</code></td> - <td><code>float_</code></td> - <td><code>float</code></td> - </tr> - <tr> - <td><code>double</code></td> - <td><code>double_</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-precision floating-point types</th> - </tr> - <tr> - <td><code>decimal</code></td> - <td><code>decimal</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">string types</th> - </tr> - <tr> - <td><code>string</code></td> - <td><code>string</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - <tr> - <td><code>normalizedString</code></td> - <td><code>normalized_string</code></td> - <td>type derived from <code>string</code></td> - </tr> - <tr> - <td><code>token</code></td> - <td><code>token</code></td> - <td>type derived from <code>normalized_string</code></td> - </tr> - <tr> - <td><code>Name</code></td> - <td><code>name</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKEN</code></td> - <td><code>nmtoken</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKENS</code></td> - <td><code>nmtokens</code></td> - <td>type derived from <code>sequence<nmtoken></code></td> - </tr> - <tr> - <td><code>NCName</code></td> - <td><code>ncname</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>language</code></td> - <td><code>language</code></td> - <td>type derived from <code>token</code></td> - </tr> - - <tr> - <th colspan="3">qualified name</th> - </tr> - <tr> - <td><code>QName</code></td> - <td><code>qname</code></td> - <td><a href="#2.5.4">Section 2.5.4, "Mapping for <code>QName</code>"</a></td> - </tr> - - <tr> - <th colspan="3">ID/IDREF types</th> - </tr> - <tr> - <td><code>ID</code></td> - <td><code>id</code></td> - <td>type derived from <code>ncname</code></td> - </tr> - <tr> - <td><code>IDREF</code></td> - <td><code>idref</code></td> - <td><a href="#2.5.5">Section 2.5.5, "Mapping for <code>IDREF</code>"</a></td> - </tr> - <tr> - <td><code>IDREFS</code></td> - <td><code>idrefs</code></td> - <td>type derived from <code>sequence<idref></code></td> - </tr> - - <tr> - <th colspan="3">URI types</th> - </tr> - <tr> - <td><code>anyURI</code></td> - <td><code>uri</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - - <tr> - <th colspan="3">binary types</th> - </tr> - <tr> - <td><code>base64Binary</code></td> - <td><code>base64_binary</code></td> - <td rowspan="2"><a href="#2.5.6">Section 2.5.6, "Mapping for - <code>base64Binary</code> and <code>hexBinary</code>"</a></td> - </tr> - <tr> - <td><code>hexBinary</code></td> - <td><code>hex_binary</code></td> - </tr> - - <tr> - <th colspan="3">date/time types</th> - </tr> - <tr> - <td><code>date</code></td> - <td><code>date</code></td> - <td><a href="#2.5.8">Section 2.5.8, "Mapping for - <code>date</code>"</a></td> - </tr> - <tr> - <td><code>dateTime</code></td> - <td><code>date_time</code></td> - <td><a href="#2.5.9">Section 2.5.9, "Mapping for - <code>dateTime</code>"</a></td> - </tr> - <tr> - <td><code>duration</code></td> - <td><code>duration</code></td> - <td><a href="#2.5.10">Section 2.5.10, "Mapping for - <code>duration</code>"</a></td> - </tr> - <tr> - <td><code>gDay</code></td> - <td><code>gday</code></td> - <td><a href="#2.5.11">Section 2.5.11, "Mapping for - <code>gDay</code>"</a></td> - </tr> - <tr> - <td><code>gMonth</code></td> - <td><code>gmonth</code></td> - <td><a href="#2.5.12">Section 2.5.12, "Mapping for - <code>gMonth</code>"</a></td> - </tr> - <tr> - <td><code>gMonthDay</code></td> - <td><code>gmonth_day</code></td> - <td><a href="#2.5.13">Section 2.5.13, "Mapping for - <code>gMonthDay</code>"</a></td> - </tr> - <tr> - <td><code>gYear</code></td> - <td><code>gyear</code></td> - <td><a href="#2.5.14">Section 2.5.14, "Mapping for - <code>gYear</code>"</a></td> - </tr> - <tr> - <td><code>gYearMonth</code></td> - <td><code>gyear_month</code></td> - <td><a href="#2.5.15">Section 2.5.15, "Mapping for - <code>gYearMonth</code>"</a></td> - </tr> - <tr> - <td><code>time</code></td> - <td><code>time</code></td> - <td><a href="#2.5.16">Section 2.5.16, "Mapping for - <code>time</code>"</a></td> - </tr> - - <tr> - <th colspan="3">entity types</th> - </tr> - <tr> - <td><code>ENTITY</code></td> - <td><code>entity</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>ENTITIES</code></td> - <td><code>entities</code></td> - <td>type derived from <code>sequence<entity></code></td> - </tr> - </table> - - <p>All XML Schema built-in types are mapped to C++ classes that are - derived from the <code>xml_schema::simple_type</code> class except - where the mapping is to a fundamental C++ type.</p> - - <p>The <code>sequence</code> class template is defined in an - implementation-specific namespace. It conforms to the - sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences"). - Practically, this means that you can treat such a sequence - as if it was <code>std::vector</code>. One notable extension - to the standard interface that is available only for - sequences of non-fundamental C++ types is the addition of - the overloaded <code>push_back</code> and <code>insert</code> - member functions which instead of the constant reference - to the element type accept automatic pointer (<code>std::unique_ptr</code> - or <code>std::auto_ptr</code>, depending on the C++ standard - selected) to the element type. These functions assume ownership - of the pointed to object and reset the passed automatic pointer. - </p> - - <h3><a name="2.5.1">2.5.1 Inheritance from Built-in Data Types</a></h3> - - <p>In cases where the mapping calls for an inheritance from a built-in - type which is mapped to a fundamental C++ type, a proxy type is - used instead of the fundamental C++ type (C++ does not allow - inheritance from fundamental types). For instance:</p> - - <pre class="xml"> -<simpleType name="my_int"> - <restriction base="int"/> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class my_int: public fundamental_base<int> -{ - ... -}; - </pre> - - <p>The <code>fundamental_base</code> class template provides a close - emulation (though not exact) of a fundamental C++ type. - It is defined in an implementation-specific namespace and has the - following interface:</p> - - <pre class="c++"> -template <typename X> -class fundamental_base: public simple_type -{ -public: - fundamental_base (); - fundamental_base (X) - fundamental_base (const fundamental_base&) - -public: - fundamental_base& - operator= (const X&); - -public: - operator const X & () const; - operator X& (); - - template <typename Y> - operator Y () const; - - template <typename Y> - operator Y (); -}; - </pre> - - <h3><a name="2.5.2">2.5.2 Mapping for <code>anyType</code></a></h3> - - <p>The XML Schema <code>anyType</code> built-in data type is mapped to the - <code>xml_schema::type</code> C++ class:</p> - - <pre class="c++"> -class type -{ -public: - virtual - ~type (); - - type (); - type (const type&); - - type& - operator= (const type&); - - virtual type* - _clone () const; - - // anyType DOM content. - // -public: - typedef element_optional dom_content_optional; - - const dom_content_optional& - dom_content () const; - - dom_content_optional& - dom_content (); - - void - dom_content (const xercesc::DOMElement&); - - void - dom_content (xercesc::DOMElement*); - - void - dom_content (const dom_content_optional&); - - const xercesc::DOMDocument& - dom_content_document () const; - - xercesc::DOMDocument& - dom_content_document (); - - bool - null_content () const; - - // DOM association. - // -public: - const xercesc::DOMNode* - _node () const; - - xercesc::DOMNode* - _node (); -}; - </pre> - - <p>When <code>xml_schema::type</code> is used to create an instance - (as opposed to being a base of a derived type), it represents - the XML Schema <code>anyType</code> type. <code>anyType</code> - allows any attributes and any content in any order. In the - C++/Tree mapping this content can be represented as a DOM - fragment, similar to XML Schema wildcards (<a href="#2.12">Section - 2.12, "Mapping for <code>any</code> and - <code>anyAttribute</code>"</a>).</p> - - <p>To enable automatic extraction of <code>anyType</code> content - during parsing, the <code>--generate-any-type</code> option must be - specified. Because the DOM API is used to access such content, the - Xerces-C++ runtime should be initialized by the application prior to - parsing and should remain initialized for the lifetime of objects - with the DOM content. For more information on the Xerces-C++ runtime - initialization see <a href="#3.1">Section 3.1, "Initializing the - Xerces-C++ Runtime"</a>.</p> - - <p>The DOM content is stored as the optional DOM element container - and the DOM content accessors and modifiers presented above are - identical to those generated for an optional element wildcard. - Refer to <a href="#2.12.2">Section 2.12.2, "Mapping for <code>any</code> - with the Optional Cardinality Class"</a> for details on their - semantics.</p> - - <p>The <code>dom_content_document()</code> function returns the - DOM document used to store the raw XML content corresponding - to the <code>anyType</code> instance. It is equivalent to the - <code>dom_document()</code> function generated for types - with wildcards.</p> - - <p>The <code>null_content()</code> accessor is an optimization function - that allows us to check for the lack of content without actually - creating its empty representation, that is, empty DOM document for - <code>anyType</code> or empty string for <code>anySimpleType</code> - (see the following section for details on <code>anySimpleType</code>).</p> - - <p>For more information on DOM association refer to - <a href="#5.1">Section 5.1, "DOM Association"</a>.</p> - - <h3><a name="2.5.3">2.5.3 Mapping for <code>anySimpleType</code></a></h3> - - <p>The XML Schema <code>anySimpleType</code> built-in data type is mapped - to the <code>xml_schema::simple_type</code> C++ class:</p> - - <pre class="c++"> -class simple_type: public type -{ -public: - simple_type (); - simple_type (const C*); - simple_type (const std::basic_string<C>&); - - simple_type (const simple_type&); - - simple_type& - operator= (const simple_type&); - - virtual simple_type* - _clone () const; - - // anySimpleType text content. - // -public: - const std::basic_string<C>& - text_content () const; - - std::basic_string<C>& - text_content (); - - void - text_content (const std::basic_string<C>&); -}; - </pre> - - <p>When <code>xml_schema::simple_type</code> is used to create an instance - (as opposed to being a base of a derived type), it represents - the XML Schema <code>anySimpleType</code> type. <code>anySimpleType</code> - allows any simple content. In the C++/Tree mapping this content can - be represented as a string and accessed or modified with the - <code>text_content()</code> functions shown above.</p> - - <h3><a name="2.5.4">2.5.4 Mapping for <code>QName</code></a></h3> - - <p>The XML Schema <code>QName</code> built-in data type is mapped to the - <code>xml_schema::qname</code> C++ class:</p> - - <pre class="c++"> -class qname: public simple_type -{ -public: - qname (const ncname&); - qname (const uri&, const ncname&); - qname (const qname&); - -public: - qname& - operator= (const qname&); - -public: - virtual qname* - _clone () const; - -public: - bool - qualified () const; - - const uri& - namespace_ () const; - - const ncname& - name () const; -}; - </pre> - - <p>The <code>qualified</code> accessor function can be used to determine - if the name is qualified.</p> - - <h3><a name="2.5.5">2.5.5 Mapping for <code>IDREF</code></a></h3> - - <p>The XML Schema <code>IDREF</code> built-in data type is mapped to the - <code>xml_schema::idref</code> C++ class. This class implements the - smart pointer C++ idiom:</p> - - <pre class="c++"> -class idref: public ncname -{ -public: - idref (const C* s); - idref (const C* s, std::size_t n); - idref (std::size_t n, C c); - idref (const std::basic_string<C>&); - idref (const std::basic_string<C>&, - std::size_t pos, - std::size_t n = npos); - -public: - idref (const idref&); - -public: - virtual idref* - _clone () const; - -public: - idref& - operator= (C c); - - idref& - operator= (const C* s); - - idref& - operator= (const std::basic_string<C>&) - - idref& - operator= (const idref&); - -public: - const type* - operator-> () const; - - type* - operator-> (); - - const type& - operator* () const; - - type& - operator* (); - - const type* - get () const; - - type* - get (); - - // Conversion to bool. - // -public: - typedef void (idref::*bool_convertible)(); - operator bool_convertible () const; -}; - </pre> - - <p>The object, <code>idref</code> instance refers to, is the immediate - container of the matching <code>id</code> instance. For example, - with the following instance document and schema: - </p> - - - <pre class="xml"> -<!-- test.xml --> -<root> - <object id="obj-1" text="hello"/> - <reference>obj-1</reference> -</root> - -<!-- test.xsd --> -<schema> - <complexType name="object_type"> - <attribute name="id" type="ID"/> - <attribute name="text" type="string"/> - </complexType> - - <complexType name="root_type"> - <sequence> - <element name="object" type="object_type"/> - <element name="reference" type="IDREF"/> - </sequence> - </complexType> - - <element name="root" type="root_type"/> -</schema> - </pre> - - <p>The <code>ref</code> instance in the code below will refer to - an object of type <code>object_type</code>:</p> - - <pre class="c++"> -root_type& root = ...; -xml_schema::idref& ref (root.reference ()); -object_type& obj (dynamic_cast<object_type&> (*ref)); -cout << obj.text () << endl; - </pre> - - <p>The smart pointer interface of the <code>idref</code> class always - returns a pointer or reference to <code>xml_schema::type</code>. - This means that you will need to manually cast such pointer or - reference to its real (dynamic) type before you can use it (unless - all you need is the base interface provided by - <code>xml_schema::type</code>). As a special extension to the XML - Schema language, the mapping supports static typing of <code>idref</code> - references by employing the <code>refType</code> extension attribute. - The following example illustrates this mechanism: - </p> - - <pre class="xml"> -<!-- test.xsd --> -<schema - xmlns:xse="https://www.codesynthesis.com/xmlns/xml-schema-extension"> - - ... - - <element name="reference" type="IDREF" xse:refType="object_type"/> - - ... - -</schema> - </pre> - - <p>With this modification we do not need to do manual casting anymore: - </p> - - <pre class="c++"> -root_type& root = ...; -root_type::reference_type& ref (root.reference ()); -object_type& obj (*ref); -cout << ref->text () << endl; - </pre> - - - <h3><a name="2.5.6">2.5.6 Mapping for <code>base64Binary</code> and - <code>hexBinary</code></a></h3> - - <p>The XML Schema <code>base64Binary</code> and <code>hexBinary</code> - built-in data types are mapped to the - <code>xml_schema::base64_binary</code> and - <code>xml_schema::hex_binary</code> C++ classes, respectively. The - <code>base64_binary</code> and <code>hex_binary</code> classes - support a simple buffer abstraction by inheriting from the - <code>xml_schema::buffer</code> class: - </p> - - <pre class="c++"> -class bounds: public virtual exception -{ -public: - virtual const char* - what () const throw (); -}; - -class buffer -{ -public: - typedef std::size_t size_t; - -public: - buffer (size_t size = 0); - buffer (size_t size, size_t capacity); - buffer (const void* data, size_t size); - buffer (const void* data, size_t size, size_t capacity); - buffer (void* data, - size_t size, - size_t capacity, - bool assume_ownership); - -public: - buffer (const buffer&); - - buffer& - operator= (const buffer&); - - void - swap (buffer&); - -public: - size_t - capacity () const; - - bool - capacity (size_t); - -public: - size_t - size () const; - - bool - size (size_t); - -public: - const char* - data () const; - - char* - data (); - - const char* - begin () const; - - char* - begin (); - - const char* - end () const; - - char* - end (); -}; - </pre> - - <p>The last overloaded constructor reuses an existing data buffer instead - of making a copy. If the <code>assume_ownership</code> argument is - <code>true</code>, the instance assumes ownership of the - memory block pointed to by the <code>data</code> argument and will - eventually release it by calling <code>operator delete</code>. The - <code>capacity</code> and <code>size</code> modifier functions return - <code>true</code> if the underlying buffer has moved. - </p> - - <p>The <code>bounds</code> exception is thrown if the constructor - arguments violate the <code>(size <= capacity)</code> - constraint.</p> - - <p>The <code>base64_binary</code> and <code>hex_binary</code> classes - support the <code>buffer</code> interface and perform automatic - decoding/encoding from/to the Base64 and Hex formats, respectively: - </p> - - <pre class="c++"> -class base64_binary: public simple_type, public buffer -{ -public: - base64_binary (size_t size = 0); - base64_binary (size_t size, size_t capacity); - base64_binary (const void* data, size_t size); - base64_binary (const void* data, size_t size, size_t capacity); - base64_binary (void* data, - size_t size, - size_t capacity, - bool assume_ownership); - -public: - base64_binary (const base64_binary&); - - base64_binary& - operator= (const base64_binary&); - - virtual base64_binary* - _clone () const; - -public: - std::basic_string<C> - encode () const; -}; - </pre> - - <pre class="c++"> -class hex_binary: public simple_type, public buffer -{ -public: - hex_binary (size_t size = 0); - hex_binary (size_t size, size_t capacity); - hex_binary (const void* data, size_t size); - hex_binary (const void* data, size_t size, size_t capacity); - hex_binary (void* data, - size_t size, - size_t capacity, - bool assume_ownership); - -public: - hex_binary (const hex_binary&); - - hex_binary& - operator= (const hex_binary&); - - virtual hex_binary* - _clone () const; - -public: - std::basic_string<C> - encode () const; -}; - </pre> - - - <h2><a name="2.5.7">2.5.7 Time Zone Representation</a></h2> - - <p>The <code>date</code>, <code>dateTime</code>, <code>gDay</code>, - <code>gMonth</code>, <code>gMonthDay</code>, <code>gYear</code>, - <code>gYearMonth</code>, and <code>time</code> XML Schema built-in - types all include an optional time zone component. The following - <code>xml_schema::time_zone</code> base class is used to represent - this information:</p> - - <pre class="c++"> -class time_zone -{ -public: - time_zone (); - time_zone (short hours, short minutes); - - bool - zone_present () const; - - void - zone_reset (); - - short - zone_hours () const; - - void - zone_hours (short); - - short - zone_minutes () const; - - void - zone_minutes (short); -}; - -bool -operator== (const time_zone&, const time_zone&); - -bool -operator!= (const time_zone&, const time_zone&); - </pre> - - <p>The <code>zone_present()</code> accessor function returns <code>true</code> - if the time zone is specified. The <code>zone_reset()</code> modifier - function resets the time zone object to the <em>not specified</em> - state. If the time zone offset is negative then both hours and - minutes components are represented as negative integers.</p> - - - <h2><a name="2.5.8">2.5.8 Mapping for <code>date</code></a></h2> - - <p>The XML Schema <code>date</code> built-in data type is mapped to the - <code>xml_schema::date</code> C++ class which represents a year, a day, - and a month with an optional time zone. Its interface is presented - below. For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class date: public simple_type, public time_zone -{ -public: - date (int year, unsigned short month, unsigned short day); - date (int year, unsigned short month, unsigned short day, - short zone_hours, short zone_minutes); - -public: - date (const date&); - - date& - operator= (const date&); - - virtual date* - _clone () const; - -public: - int - year () const; - - void - year (int); - - unsigned short - month () const; - - void - month (unsigned short); - - unsigned short - day () const; - - void - day (unsigned short); -}; - -bool -operator== (const date&, const date&); - -bool -operator!= (const date&, const date&); - </pre> - - <h2><a name="2.5.9">2.5.9 Mapping for <code>dateTime</code></a></h2> - - <p>The XML Schema <code>dateTime</code> built-in data type is mapped to the - <code>xml_schema::date_time</code> C++ class which represents a year, a month, - a day, hours, minutes, and seconds with an optional time zone. Its interface - is presented below. For more information on the base - <code>xml_schema::time_zone</code> class refer to <a href="#2.5.7">Section - 2.5.7, "Time Zone Representation"</a>.</p> - - <pre class="c++"> -class date_time: public simple_type, public time_zone -{ -public: - date_time (int year, unsigned short month, unsigned short day, - unsigned short hours, unsigned short minutes, - double seconds); - - date_time (int year, unsigned short month, unsigned short day, - unsigned short hours, unsigned short minutes, - double seconds, short zone_hours, short zone_minutes); -public: - date_time (const date_time&); - - date_time& - operator= (const date_time&); - - virtual date_time* - _clone () const; - -public: - int - year () const; - - void - year (int); - - unsigned short - month () const; - - void - month (unsigned short); - - unsigned short - day () const; - - void - day (unsigned short); - - unsigned short - hours () const; - - void - hours (unsigned short); - - unsigned short - minutes () const; - - void - minutes (unsigned short); - - double - seconds () const; - - void - seconds (double); -}; - -bool -operator== (const date_time&, const date_time&); - -bool -operator!= (const date_time&, const date_time&); - </pre> - - - <h2><a name="2.5.10">2.5.10 Mapping for <code>duration</code></a></h2> - - <p>The XML Schema <code>duration</code> built-in data type is mapped to the - <code>xml_schema::duration</code> C++ class which represents a potentially - negative duration in the form of years, months, days, hours, minutes, - and seconds. Its interface is presented below.</p> - - <pre class="c++"> -class duration: public simple_type -{ -public: - duration (bool negative, - unsigned int years, unsigned int months, unsigned int days, - unsigned int hours, unsigned int minutes, double seconds); -public: - duration (const duration&); - - duration& - operator= (const duration&); - - virtual duration* - _clone () const; - -public: - bool - negative () const; - - void - negative (bool); - - unsigned int - years () const; - - void - years (unsigned int); - - unsigned int - months () const; - - void - months (unsigned int); - - unsigned int - days () const; - - void - days (unsigned int); - - unsigned int - hours () const; - - void - hours (unsigned int); - - unsigned int - minutes () const; - - void - minutes (unsigned int); - - double - seconds () const; - - void - seconds (double); -}; - -bool -operator== (const duration&, const duration&); - -bool -operator!= (const duration&, const duration&); - </pre> - - - <h2><a name="2.5.11">2.5.11 Mapping for <code>gDay</code></a></h2> - - <p>The XML Schema <code>gDay</code> built-in data type is mapped to the - <code>xml_schema::gday</code> C++ class which represents a day of the - month with an optional time zone. Its interface is presented below. - For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gday: public simple_type, public time_zone -{ -public: - explicit - gday (unsigned short day); - gday (unsigned short day, short zone_hours, short zone_minutes); - -public: - gday (const gday&); - - gday& - operator= (const gday&); - - virtual gday* - _clone () const; - -public: - unsigned short - day () const; - - void - day (unsigned short); -}; - -bool -operator== (const gday&, const gday&); - -bool -operator!= (const gday&, const gday&); - </pre> - - - <h2><a name="2.5.12">2.5.12 Mapping for <code>gMonth</code></a></h2> - - <p>The XML Schema <code>gMonth</code> built-in data type is mapped to the - <code>xml_schema::gmonth</code> C++ class which represents a month of the - year with an optional time zone. Its interface is presented below. - For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gmonth: public simple_type, public time_zone -{ -public: - explicit - gmonth (unsigned short month); - gmonth (unsigned short month, - short zone_hours, short zone_minutes); - -public: - gmonth (const gmonth&); - - gmonth& - operator= (const gmonth&); - - virtual gmonth* - _clone () const; - -public: - unsigned short - month () const; - - void - month (unsigned short); -}; - -bool -operator== (const gmonth&, const gmonth&); - -bool -operator!= (const gmonth&, const gmonth&); - </pre> - - - <h2><a name="2.5.13">2.5.13 Mapping for <code>gMonthDay</code></a></h2> - - <p>The XML Schema <code>gMonthDay</code> built-in data type is mapped to the - <code>xml_schema::gmonth_day</code> C++ class which represents a day and - a month of the year with an optional time zone. Its interface is presented - below. For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gmonth_day: public simple_type, public time_zone -{ -public: - gmonth_day (unsigned short month, unsigned short day); - gmonth_day (unsigned short month, unsigned short day, - short zone_hours, short zone_minutes); - -public: - gmonth_day (const gmonth_day&); - - gmonth_day& - operator= (const gmonth_day&); - - virtual gmonth_day* - _clone () const; - -public: - unsigned short - month () const; - - void - month (unsigned short); - - unsigned short - day () const; - - void - day (unsigned short); -}; - -bool -operator== (const gmonth_day&, const gmonth_day&); - -bool -operator!= (const gmonth_day&, const gmonth_day&); - </pre> - - - <h2><a name="2.5.14">2.5.14 Mapping for <code>gYear</code></a></h2> - - <p>The XML Schema <code>gYear</code> built-in data type is mapped to the - <code>xml_schema::gyear</code> C++ class which represents a year with - an optional time zone. Its interface is presented below. For more - information on the base <code>xml_schema::time_zone</code> class refer - to <a href="#2.5.7">Section 2.5.7, "Time Zone Representation"</a>.</p> - - <pre class="c++"> -class gyear: public simple_type, public time_zone -{ -public: - explicit - gyear (int year); - gyear (int year, short zone_hours, short zone_minutes); - -public: - gyear (const gyear&); - - gyear& - operator= (const gyear&); - - virtual gyear* - _clone () const; - -public: - int - year () const; - - void - year (int); -}; - -bool -operator== (const gyear&, const gyear&); - -bool -operator!= (const gyear&, const gyear&); - </pre> - - - <h2><a name="2.5.15">2.5.15 Mapping for <code>gYearMonth</code></a></h2> - - <p>The XML Schema <code>gYearMonth</code> built-in data type is mapped to - the <code>xml_schema::gyear_month</code> C++ class which represents - a year and a month with an optional time zone. Its interface is presented - below. For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gyear_month: public simple_type, public time_zone -{ -public: - gyear_month (int year, unsigned short month); - gyear_month (int year, unsigned short month, - short zone_hours, short zone_minutes); -public: - gyear_month (const gyear_month&); - - gyear_month& - operator= (const gyear_month&); - - virtual gyear_month* - _clone () const; - -public: - int - year () const; - - void - year (int); - - unsigned short - month () const; - - void - month (unsigned short); -}; - -bool -operator== (const gyear_month&, const gyear_month&); - -bool -operator!= (const gyear_month&, const gyear_month&); - </pre> - - - <h2><a name="2.5.16">2.5.16 Mapping for <code>time</code></a></h2> - - <p>The XML Schema <code>time</code> built-in data type is mapped to - the <code>xml_schema::time</code> C++ class which represents hours, - minutes, and seconds with an optional time zone. Its interface is - presented below. For more information on the base - <code>xml_schema::time_zone</code> class refer to - <a href="#2.5.7">Section 2.5.7, "Time Zone Representation"</a>.</p> - - <pre class="c++"> -class time: public simple_type, public time_zone -{ -public: - time (unsigned short hours, unsigned short minutes, double seconds); - time (unsigned short hours, unsigned short minutes, double seconds, - short zone_hours, short zone_minutes); - -public: - time (const time&); - - time& - operator= (const time&); - - virtual time* - _clone () const; - -public: - unsigned short - hours () const; - - void - hours (unsigned short); - - unsigned short - minutes () const; - - void - minutes (unsigned short); - - double - seconds () const; - - void - seconds (double); -}; - -bool -operator== (const time&, const time&); - -bool -operator!= (const time&, const time&); - </pre> - - - <!-- Mapping for Simple Types --> - - <h2><a name="2.6">2.6 Mapping for Simple Types</a></h2> - - <p>An XML Schema simple type is mapped to a C++ class with the same - name as the simple type. The class defines a public copy constructor, - a public copy assignment operator, and a public virtual - <code>_clone</code> function. The <code>_clone</code> function is - declared <code>const</code>, does not take any arguments, and returns - a pointer to a complete copy of the instance allocated in the free - store. The <code>_clone</code> function shall be used to make copies - when static type and dynamic type of the instance may differ (see - <a href="#2.11">Section 2.11, "Mapping for <code>xsi:type</code> - and Substitution Groups"</a>). For instance:</p> - - <pre class="xml"> -<simpleType name="object"> - ... -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: ... -{ -public: - object (const object&); - -public: - object& - operator= (const object&); - -public: - virtual object* - _clone () const; - - ... - -}; - </pre> - - <p>The base class specification and the rest of the class definition - depend on the type of derivation used to define the simple type. </p> - - - <h3><a name="2.6.1">2.6.1 Mapping for Derivation by Restriction</a></h3> - - <p>XML Schema derivation by restriction is mapped to C++ public - inheritance. The base type of the restriction becomes the base - type for the resulting C++ class. In addition to the members described - in <a href="#2.6">Section 2.6, "Mapping for Simple Types"</a>, the - resulting C++ class defines a public constructor with the base type - as its single argument. For instance:</p> - - <pre class="xml"> -<simpleType name="object"> - <restriction base="base"> - ... - </restriction> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public base -{ -public: - object (const base&); - object (const object&); - -public: - object& - operator= (const object&); - -public: - virtual object* - _clone () const; -}; - </pre> - - - <h3><a name="2.6.2">2.6.2 Mapping for Enumerations</a></h3> - -<p>XML Schema restriction by enumeration is mapped to a C++ class - with semantics similar to C++ <code>enum</code>. Each XML Schema - enumeration element is mapped to a C++ enumerator with the - name derived from the <code>value</code> attribute and defined - in the class scope. In addition to the members - described in <a href="#2.6">Section 2.6, "Mapping for Simple Types"</a>, - the resulting C++ class defines a public constructor that can be called - with one of the enumerators as its single argument, a public constructor - that can be called with enumeration's base value as its single - argument, a public assignment operator that can be used to assign the - value of one of the enumerators, and a public implicit conversion - operator to the underlying C++ enum type.</p> - -<p>Furthermore, for string-based enumeration types, the resulting C++ - class defines a public constructor with a single argument of type - <code>const C*</code> and a public constructor with a single - argument of type <code>const std::basic_string<C>&</code>. - For instance:</p> - - <pre class="xml"> -<simpleType name="color"> - <restriction base="string"> - <enumeration value="red"/> - <enumeration value="green"/> - <enumeration value="blue"/> - </restriction> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class color: public xml_schema::string -{ -public: - enum value - { - red, - green, - blue - }; - -public: - color (value); - color (const C*); - color (const std::basic_string<C>&); - color (const xml_schema::string&); - color (const color&); - -public: - color& - operator= (value); - - color& - operator= (const color&); - -public: - virtual color* - _clone () const; - -public: - operator value () const; -}; - </pre> - - <h3><a name="2.6.3">2.6.3 Mapping for Derivation by List</a></h3> - - <p>XML Schema derivation by list is mapped to C++ public - inheritance from <code>xml_schema::simple_type</code> - (<a href="#2.5.3">Section 2.5.3, "Mapping for - <code>anySimpleType</code>"</a>) and a suitable sequence type. - The list item type becomes the element type of the sequence. - In addition to the members described in <a href="#2.6">Section 2.6, - "Mapping for Simple Types"</a>, the resulting C++ class defines - a public default constructor, a public constructor - with the first argument of type <code>size_type</code> and - the second argument of list item type that creates - a list object with the specified number of copies of the specified - element value, and a public constructor with the two arguments - of an input iterator type that creates a list object from an - iterator range. For instance: - </p> - - <pre class="xml"> -<simpleType name="int_list"> - <list itemType="int"/> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class int_list: public simple_type, - public sequence<int> -{ -public: - int_list (); - int_list (size_type n, int x); - - template <typename I> - int_list (const I& begin, const I& end); - int_list (const int_list&); - -public: - int_list& - operator= (const int_list&); - -public: - virtual int_list* - _clone () const; -}; - </pre> - - <p>The <code>sequence</code> class template is defined in an - implementation-specific namespace. It conforms to the - sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences"). - Practically, this means that you can treat such a sequence - as if it was <code>std::vector</code>. One notable extension - to the standard interface that is available only for - sequences of non-fundamental C++ types is the addition of - the overloaded <code>push_back</code> and <code>insert</code> - member functions which instead of the constant reference - to the element type accept automatic pointer (<code>std::unique_ptr</code> - or <code>std::auto_ptr</code>, depending on the C++ standard - selected) to the element type. These functions assume ownership - of the pointed to object and reset the passed automatic pointer. - </p> - - <h3><a name="2.6.4">2.6.4 Mapping for Derivation by Union</a></h3> - - <p>XML Schema derivation by union is mapped to C++ public - inheritance from <code>xml_schema::simple_type</code> - (<a href="#2.5.3">Section 2.5.3, "Mapping for - <code>anySimpleType</code>"</a>) and <code>std::basic_string<C></code>. - In addition to the members described in <a href="#2.6">Section 2.6, - "Mapping for Simple Types"</a>, the resulting C++ class defines a - public constructor with a single argument of type <code>const C*</code> - and a public constructor with a single argument of type - <code>const std::basic_string<C>&</code>. For instance: - </p> - - <pre class="xml"> -<simpleType name="int_string_union"> - <xsd:union memberTypes="xsd:int xsd:string"/> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class int_string_union: public simple_type, - public std::basic_string<C> -{ -public: - int_string_union (const C*); - int_string_union (const std::basic_string<C>&); - int_string_union (const int_string_union&); - -public: - int_string_union& - operator= (const int_string_union&); - -public: - virtual int_string_union* - _clone () const; -}; - </pre> - - <h2><a name="2.7">2.7 Mapping for Complex Types</a></h2> - - <p>An XML Schema complex type is mapped to a C++ class with the same - name as the complex type. The class defines a public copy constructor, - a public copy assignment operator, and a public virtual - <code>_clone</code> function. The <code>_clone</code> function is - declared <code>const</code>, does not take any arguments, and returns - a pointer to a complete copy of the instance allocated in the free - store. The <code>_clone</code> function shall be used to make copies - when static type and dynamic type of the instance may differ (see - <a href="#2.11">Section 2.11, "Mapping for <code>xsi:type</code> - and Substitution Groups"</a>).</p> - - <p>Additionally, the resulting C++ class - defines two public constructors that take an initializer for each - member of the complex type and all its base types that belongs to - the One cardinality class (see <a href="#2.8">Section 2.8, "Mapping - for Local Elements and Attributes"</a>). In the first constructor, - the arguments are passed as constant references and the newly created - instance is initialized with copies of the passed objects. In the - second constructor, arguments that are complex types (that is, - they themselves contain elements or attributes) are passed as - either <code>std::unique_ptr</code> (C++11) or <code>std::auto_ptr</code> - (C++98), depending on the C++ standard selected. In this case the newly - created instance is directly initialized with and assumes ownership - of the pointed to objects and the <code>std::[unique|auto]_ptr</code> - arguments are reset to <code>0</code>. For instance:</p> - - <pre class="xml"> -<complexType name="complex"> - <sequence> - <element name="a" type="int"/> - <element name="b" type="string"/> - </sequence> -</complexType> - -<complexType name="object"> - <sequence> - <element name="s-one" type="boolean"/> - <element name="c-one" type="complex"/> - <element name="optional" type="int" minOccurs="0"/> - <element name="sequence" type="string" maxOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class complex: public xml_schema::type -{ -public: - object (const int& a, const xml_schema::string& b); - object (const complex&); - -public: - object& - operator= (const complex&); - -public: - virtual complex* - _clone () const; - - ... - -}; - -class object: public xml_schema::type -{ -public: - object (const bool& s_one, const complex& c_one); - object (const bool& s_one, std::[unique|auto]_ptr<complex> c_one); - object (const object&); - -public: - object& - operator= (const object&); - -public: - virtual object* - _clone () const; - - ... - -}; - </pre> - - <p>Notice that the generated <code>complex</code> class does not - have the second (<code>std::[unique|auto]_ptr</code>) version of the - constructor since all its required members are of simple types.</p> - - <p>If an XML Schema complex type has an ultimate base which is an XML - Schema simple type then the resulting C++ class also defines a public - constructor that takes an initializer for the base type as well as - for each member of the complex type and all its base types that - belongs to the One cardinality class. For instance:</p> - - <pre class="xml"> -<complexType name="object"> - <simpleContent> - <extension base="date"> - <attribute name="lang" type="language" use="required"/> - </extension> - </simpleContent> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::string -{ -public: - object (const xml_schema::language& lang); - - object (const xml_schema::date& base, - const xml_schema::language& lang); - - ... - -}; - </pre> - - <p>Furthermore, for string-based XML Schema complex types, the resulting C++ - class also defines two public constructors with the first arguments - of type <code>const C*</code> and <code>std::basic_string<C>&</code>, - respectively, followed by arguments for each member of the complex - type and all its base types that belongs to the One cardinality - class. For enumeration-based complex types the resulting C++ - class also defines a public constructor with the first arguments - of the underlying enum type followed by arguments for each member - of the complex type and all its base types that belongs to the One - cardinality class. For instance:</p> - - <pre class="xml"> -<simpleType name="color"> - <restriction base="string"> - <enumeration value="red"/> - <enumeration value="green"/> - <enumeration value="blue"/> - </restriction> -</simpleType> - -<complexType name="object"> - <simpleContent> - <extension base="color"> - <attribute name="lang" type="language" use="required"/> - </extension> - </simpleContent> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class color: public xml_schema::string -{ -public: - enum value - { - red, - green, - blue - }; - -public: - color (value); - color (const C*); - color (const std::basic_string<C>&); - - ... - -}; - -class object: color -{ -public: - object (const color& base, - const xml_schema::language& lang); - - object (const color::value& base, - const xml_schema::language& lang); - - object (const C* base, - const xml_schema::language& lang); - - object (const std::basic_string<C>& base, - const xml_schema::language& lang); - - ... - -}; - </pre> - - <p>Additional constructors can be requested with the - <code>--generate-default-ctor</code> and - <code>--generate-from-base-ctor</code> options. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for details.</p> - - <p>If an XML Schema complex type is not explicitly derived from any type, - the resulting C++ class is derived from <code>xml_schema::type</code>. - In cases where an XML Schema complex type is defined using derivation - by extension or restriction, the resulting C++ base class specification - depends on the type of derivation and is described in the subsequent - sections. - </p> - - <p>The mapping for elements and attributes that are defined in a complex - type is described in <a href="#2.8">Section 2.8, "Mapping for Local - Elements and Attributes"</a>. - </p> - - <h3><a name="2.7.1">2.7.1 Mapping for Derivation by Extension</a></h3> - - <p>XML Schema derivation by extension is mapped to C++ public - inheritance. The base type of the extension becomes the base - type for the resulting C++ class. - </p> - - <h3><a name="2.7.2">2.7.2 Mapping for Derivation by Restriction</a></h3> - - <p>XML Schema derivation by restriction is mapped to C++ public - inheritance. The base type of the restriction becomes the base - type for the resulting C++ class. XML Schema elements and - attributes defined within restriction do not result in any - definitions in the resulting C++ class. Instead, corresponding - (unrestricted) definitions are inherited from the base class. - In the future versions of this mapping, such elements and - attributes may result in redefinitions of accessors and - modifiers to reflect their restricted semantics. - </p> - - <!-- 2.8 Mapping for Local Elements and Attributes --> - - <h2><a name="2.8">2.8 Mapping for Local Elements and Attributes</a></h2> - - <p>XML Schema element and attribute definitions are called local - if they appear within a complex type definition, an element group - definition, or an attribute group definitions. - </p> - - <p>Local XML Schema element and attribute definitions have the same - C++ mapping. Therefore, in this section, local elements and - attributes are collectively called members. - </p> - - <p>While there are many different member cardinality combinations - (determined by the <code>use</code> attribute for attributes and - the <code>minOccurs</code> and <code>maxOccurs</code> attributes - for elements), the mapping divides all possible cardinality - combinations into three cardinality classes: - </p> - - <dl> - <dt><i>one</i></dt> - <dd>attributes: <code>use == "required"</code></dd> - <dd>attributes: <code>use == "optional"</code> and has default or fixed value</dd> - <dd>elements: <code>minOccurs == "1"</code> and <code>maxOccurs == "1"</code></dd> - - <dt><i>optional</i></dt> - <dd>attributes: <code>use == "optional"</code> and doesn't have default or fixed value</dd> - <dd>elements: <code>minOccurs == "0"</code> and <code>maxOccurs == "1"</code></dd> - - <dt><i>sequence</i></dt> - <dd>elements: <code>maxOccurs > "1"</code></dd> - </dl> - - <p>An optional attribute with a default or fixed value acquires this value - if the attribute hasn't been specified in an instance document (see - <a href="#A">Appendix A, "Default and Fixed Values"</a>). This - mapping places such optional attributes to the One cardinality - class.</p> - - <p>A member is mapped to a set of public type definitions - (<code>typedef</code>s) and a set of public accessor and modifier - functions. Type definitions have names derived from the member's - name. The accessor and modifier functions have the same name as the - member. For example: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - typedef xml_schema::string member_type; - - const member_type& - member () const; - - ... - -}; - </pre> - - <p>In addition, if a member has a default or fixed value, a static - accessor function is generated that returns this value. For - example:</p> - -<pre class="xml"> -<complexType name="object"> - <attribute name="data" type="string" default="test"/> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - typedef xml_schema::string data_type; - - const data_type& - data () const; - - static const data_type& - data_default_value (); - - ... - -}; - </pre> - - <p>Names and semantics of type definitions for the member as well - as signatures of the accessor and modifier functions depend on - the member's cardinality class and are described in the following - sub-sections. - </p> - - - <h3><a name="2.8.1">2.8.1 Mapping for Members with the One Cardinality Class</a></h3> - - <p>For the One cardinality class, the type definitions consist of - an alias for the member's type with the name created by appending - the <code>_type</code> suffix to the member's name. - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - member and can be used for read-only access. The non-constant - version returns an unrestricted reference to the member and can - be used for read-write access. - </p> - - <p>The first modifier function expects an argument of type reference to - constant of the member's type. It makes a deep copy of its argument. - Except for member's types that are mapped to fundamental C++ types, - the second modifier function is provided that expects an argument - of type automatic pointer (<code>std::unique_ptr</code> or - <code>std::auto_ptr</code>, depending on the C++ standard selected) - to the member's type. It assumes ownership of the pointed to object - and resets the passed automatic pointer. For instance:</p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef xml_schema::string member_type; - - // Accessors. - // - const member_type& - member () const; - - member_type& - member (); - - // Modifiers. - // - void - member (const member_type&); - - void - member (std::[unique|auto]_ptr<member_type>); - ... - -}; - </pre> - - <p>In addition, if requested by specifying the <code>--generate-detach</code> - option and only for members of non-fundamental C++ types, the mapping - provides a detach function that returns an automatic pointer to the - member's type, for example:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - ... - - std::[unique|auto]_ptr<member_type> - detach_member (); - ... - -}; - </pre> - - <p>This function detaches the value from the tree leaving the member - value uninitialized. Accessing such an uninitialized value prior to - re-initializing it results in undefined behavior.</p> - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o) -{ - using xml_schema::string; - - string s (o.member ()); // get - object::member_type& sr (o.member ()); // get - - o.member ("hello"); // set, deep copy - o.member () = "hello"; // set, deep copy - - // C++11 version. - // - std::unique_ptr<string> p (new string ("hello")); - o.member (std::move (p)); // set, assumes ownership - p = o.detach_member (); // detach, member is uninitialized - o.member (std::move (p)); // re-attach - - // C++98 version. - // - std::auto_ptr<string> p (new string ("hello")); - o.member (p); // set, assumes ownership - p = o.detach_member (); // detach, member is uninitialized - o.member (p); // re-attach -} - </pre> - - -<h3><a name="2.8.2">2.8.2 Mapping for Members with the Optional Cardinality Class</a></h3> - - <p>For the Optional cardinality class, the type definitions consist of - an alias for the member's type with the name created by appending - the <code>_type</code> suffix to the member's name and an alias for - the container type with the name created by appending the - <code>_optional</code> suffix to the member's name. - </p> - - <p>Unlike accessor functions for the One cardinality class, accessor - functions for the Optional cardinality class return references to - corresponding containers rather than directly to members. The - accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to - the container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container - and can be used for read-write access. - </p> - - <p>The modifier functions are overloaded for the member's - type and the container type. The first modifier function - expects an argument of type reference to constant of the - member's type. It makes a deep copy of its argument. - Except for member's types that are mapped to fundamental C++ types, - the second modifier function is provided that expects an argument - of type automatic pointer (<code>std::unique_ptr</code> or - <code>std::auto_ptr</code>, depending on the C++ standard selected) - to the member's type. It assumes ownership of the pointed to object - and resets the passed automatic pointer. The last modifier function - expects an argument of type reference to constant of the container - type. It makes a deep copy of its argument. For instance: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string" minOccurs="0"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef xml_schema::string member_type; - typedef optional<member_type> member_optional; - - // Accessors. - // - const member_optional& - member () const; - - member_optional& - member (); - - // Modifiers. - // - void - member (const member_type&); - - void - member (std::[unique|auto]_ptr<member_type>); - - void - member (const member_optional&); - - ... - -}; - </pre> - - - <p>The <code>optional</code> class template is defined in an - implementation-specific namespace and has the following - interface. The <code>[unique|auto]_ptr</code>-based constructor - and modifier function are only available if the template - argument is not a fundamental C++ type. - </p> - - <pre class="c++"> -template <typename X> -class optional -{ -public: - optional (); - - // Makes a deep copy. - // - explicit - optional (const X&); - - // Assumes ownership. - // - explicit - optional (std::[unique|auto]_ptr<X>); - - optional (const optional&); - -public: - optional& - operator= (const X&); - - optional& - operator= (const optional&); - - // Pointer-like interface. - // -public: - const X* - operator-> () const; - - X* - operator-> (); - - const X& - operator* () const; - - X& - operator* (); - - typedef void (optional::*bool_convertible) (); - operator bool_convertible () const; - - // Get/set interface. - // -public: - bool - present () const; - - const X& - get () const; - - X& - get (); - - // Makes a deep copy. - // - void - set (const X&); - - // Assumes ownership. - // - void - set (std::[unique|auto]_ptr<X>); - - // Detach and return the contained value. - // - std::[unique|auto]_ptr<X> - detach (); - - void - reset (); -}; - -template <typename X> -bool -operator== (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator!= (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator< (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator> (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator<= (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator>= (const optional<X>&, const optional<X>&); - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o) -{ - using xml_schema::string; - - if (o.member ().present ()) // test - { - string& s (o.member ().get ()); // get - o.member ("hello"); // set, deep copy - o.member ().set ("hello"); // set, deep copy - o.member ().reset (); // reset - } - - // Same as above but using pointer notation: - // - if (o.member ()) // test - { - string& s (*o.member ()); // get - o.member ("hello"); // set, deep copy - *o.member () = "hello"; // set, deep copy - o.member ().reset (); // reset - } - - // C++11 version. - // - std::unique_ptr<string> p (new string ("hello")); - o.member (std::move (p)); // set, assumes ownership - - p.reset (new string ("hello")); - o.member ().set (std::move (p)); // set, assumes ownership - - p = o.member ().detach (); // detach, member is reset - o.member ().set (std::move (p)); // re-attach - - // C++98 version. - // - std::auto_ptr<string> p (new string ("hello")); - o.member (p); // set, assumes ownership - - p = new string ("hello"); - o.member ().set (p); // set, assumes ownership - - p = o.member ().detach (); // detach, member is reset - o.member ().set (p); // re-attach -} - </pre> - - - <h3><a name="2.8.3">2.8.3 Mapping for Members with the Sequence Cardinality Class</a></h3> - - <p>For the Sequence cardinality class, the type definitions consist of an - alias for the member's type with the name created by appending - the <code>_type</code> suffix to the member's name, an alias of - the container type with the name created by appending the - <code>_sequence</code> suffix to the member's name, an alias of - the iterator type with the name created by appending the - <code>_iterator</code> suffix to the member's name, and an alias - of the constant iterator type with the name created by appending the - <code>_const_iterator</code> suffix to the member's name. - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container and can - be used for read-write access. - </p> - - <p>The modifier function expects an argument of type reference to - constant of the container type. The modifier function - makes a deep copy of its argument. For instance: - </p> - - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string" minOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef xml_schema::string member_type; - typedef sequence<member_type> member_sequence; - typedef member_sequence::iterator member_iterator; - typedef member_sequence::const_iterator member_const_iterator; - - // Accessors. - // - const member_sequence& - member () const; - - member_sequence& - member (); - - // Modifier. - // - void - member (const member_sequence&); - - ... - -}; - </pre> - - <p>The <code>sequence</code> class template is defined in an - implementation-specific namespace. It conforms to the - sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences"). - Practically, this means that you can treat such a sequence - as if it was <code>std::vector</code>. Two notable extensions - to the standard interface that are available only for - sequences of non-fundamental C++ types are the addition of - the overloaded <code>push_back</code> and <code>insert</code> - as well as the <code>detach_back</code> and <code>detach</code> - member functions. The additional <code>push_back</code> and - <code>insert</code> functions accept an automatic pointer - (<code>std::unique_ptr</code> or <code>std::auto_ptr</code>, - depending on the C++ standard selected) to the - element type instead of the constant reference. They assume - ownership of the pointed to object and reset the passed - automatic pointer. The <code>detach_back</code> and - <code>detach</code> functions detach the element - value from the sequence container and, by default, remove - the element from the sequence. These additional functions - have the following signatures:</p> - - <pre class="c++"> -template <typename X> -class sequence -{ -public: - ... - - void - push_back (std::[unique|auto]_ptr<X>) - - iterator - insert (iterator position, std::[unique|auto]_ptr<X>) - - std::[unique|auto]_ptr<X> - detach_back (bool pop = true); - - iterator - detach (iterator position, - std::[unique|auto]_ptr<X>& result, - bool erase = true) - - ... -} - </pre> - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o) -{ - using xml_schema::string; - - object::member_sequence& s (o.member ()); - - // Iteration. - // - for (object::member_iterator i (s.begin ()); i != s.end (); ++i) - { - string& value (*i); - } - - // Modification. - // - s.push_back ("hello"); // deep copy - - // C++11 version. - // - std::unique_ptr<string> p (new string ("hello")); - s.push_back (std::move (p)); // assumes ownership - p = s.detach_back (); // detach and pop - s.push_back (std::move (p)); // re-append - - // C++98 version. - // - std::auto_ptr<string> p (new string ("hello")); - s.push_back (p); // assumes ownership - p = s.detach_back (); // detach and pop - s.push_back (p); // re-append - - // Setting a new container. - // - object::member_sequence n; - n.push_back ("one"); - n.push_back ("two"); - o.member (n); // deep copy -} - </pre> - - <h3><a name="2.8.4">2.8.4 Element Order</a></h3> - - <p>C++/Tree is a "flattening" mapping in a sense that many levels of - nested compositors (<code>choice</code> and <code>sequence</code>), - all potentially with their own cardinalities, are in the end mapped - to a flat set of elements with one of the three cardinality classes - discussed in the previous sections. While this results in a simple - and easy to use API for most types, in certain cases, the order of - elements in the actual XML documents is not preserved once parsed - into the object model. And sometimes such order has - application-specific significance. As an example, consider a schema - that defines a batch of bank transactions:</p> - - <pre class="xml"> -<complexType name="withdraw"> - <sequence> - <element name="account" type="unsignedInt"/> - <element name="amount" type="unsignedInt"/> - </sequence> -</complexType> - -<complexType name="deposit"> - <sequence> - <element name="account" type="unsignedInt"/> - <element name="amount" type="unsignedInt"/> - </sequence> -</complexType> - -<complexType name="batch"> - <choice minOccurs="0" maxOccurs="unbounded"> - <element name="withdraw" type="withdraw"/> - <element name="deposit" type="deposit"/> - </choice> -</complexType> - </pre> - - <p>The batch can contain any number of transactions in any order - but the order of transactions in each actual batch is significant. - For instance, consider what could happen if we reorder the - transactions and apply all the withdrawals before deposits.</p> - - <p>For the <code>batch</code> schema type defined above the default - C++/Tree mapping will produce a C++ class that contains a pair of - sequence containers, one for each of the two elements. While this - will capture the content (transactions), the order of this content - as it appears in XML will be lost. Also, if we try to serialize the - batch we just loaded back to XML, all the withdrawal transactions - will appear before deposits.</p> - - <p>To overcome this limitation of a flattening mapping, C++/Tree - allows us to mark certain XML Schema types, for which content - order is important, as ordered.</p> - - <p>There are several command line options that control which - schema types are treated as ordered. To make an individual - type ordered, we use the <code>--ordered-type</code> option, - for example:</p> - - <pre class="term"> ---ordered-type batch - </pre> - - <p>To automatically treat all the types that are derived from an ordered - type also ordered, we use the <code>--ordered-type-derived</code> - option. This is primarily useful if you would like to iterate - over the complete hierarchy's content using the content order - sequence (discussed below).</p> - - <p>Ordered types are also useful for handling mixed content. To - automatically mark all the types with mixed content as ordered - we use the <code>--ordered-type-mixed</code> option. For more - information on handling mixed content see <a href="#2.13">Section - 2.13, "Mapping for Mixed Content Models"</a>.</p> - - <p>Finally, we can mark all the types in the schema we are - compiling with the <code>--ordered-type-all</code> option. - You should only resort to this option if all the types in - your schema truly suffer from the loss of content - order since, as we will discuss shortly, ordered types - require extra effort to access and, especially, modify. - See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information on - these options.</p> - - <p>Once a type is marked ordered, C++/Tree alters its mapping - in several ways. Firstly, for each local element, element - wildcard (<a href="#2.12.4">Section 2.12.4, "Element Wildcard - Order"</a>), and mixed content text (<a href="#2.13">Section - 2.13, "Mapping for Mixed Content Models"</a>) in this type, a - content id constant is generated. Secondly, an addition sequence - is added to the class that captures the content order. Here - is how the mapping of our <code>batch</code> class changes - once we make it ordered:</p> - - <pre class="c++"> -class batch: public xml_schema::type -{ -public: - // withdraw - // - typedef withdraw withdraw_type; - typedef sequence<withdraw_type> withdraw_sequence; - typedef withdraw_sequence::iterator withdraw_iterator; - typedef withdraw_sequence::const_iterator withdraw_const_iterator; - - static const std::size_t withdraw_id = 1; - - const withdraw_sequence& - withdraw () const; - - withdraw_sequence& - withdraw (); - - void - withdraw (const withdraw_sequence&); - - // deposit - // - typedef deposit deposit_type; - typedef sequence<deposit_type> deposit_sequence; - typedef deposit_sequence::iterator deposit_iterator; - typedef deposit_sequence::const_iterator deposit_const_iterator; - - static const std::size_t deposit_id = 2; - - const deposit_sequence& - deposit () const; - - deposit_sequence& - deposit (); - - void - deposit (const deposit_sequence&); - - // content_order - // - typedef xml_schema::content_order content_order_type; - typedef std::vector<content_order_type> content_order_sequence; - typedef content_order_sequence::iterator content_order_iterator; - typedef content_order_sequence::const_iterator content_order_const_iterator; - - const content_order_sequence& - content_order () const; - - content_order_sequence& - content_order (); - - void - content_order (const content_order_sequence&); - - ... -}; - </pre> - - <p>Notice the <code>withdraw_id</code> and <code>deposit_id</code> - content ids as well as the extra <code>content_order</code> - sequence that does not correspond to any element in the - schema definition. The other changes to the mapping for ordered - types has to do with XML parsing and serialization code. During - parsing the content order is captured in the <code>content_order</code> - sequence while during serialization this sequence is used to - determine the order in which content is serialized. The - <code>content_order</code> sequence is also copied during - copy construction and assigned during copy assignment. It is also - taken into account during comparison.</p> - - <p>The entry type of the <code>content_order</code> sequence is the - <code>xml_schema::content_order</code> type that has the following - interface:</p> - - <pre class="c++"> -namespace xml_schema -{ - struct content_order - { - content_order (std::size_t id, std::size_t index = 0); - - std::size_t id; - std::size_t index; - }; - - bool - operator== (const content_order&, const content_order&); - - bool - operator!= (const content_order&, const content_order&); - - bool - operator< (const content_order&, const content_order&); -} - </pre> - - <p>The <code>content_order</code> sequence describes the order of - content (elements, including wildcards, as well as mixed content - text). Each entry in this sequence consists of the content id - (for example, <code>withdraw_id</code> or <code>deposit_id</code> - in our case) as well as, for elements of the sequence cardinality - class, an index into the corresponding sequence container (the - index is unused for the one and optional cardinality classes). - For example, in our case, if the content id is <code>withdraw_id</code>, - then the index will point into the <code>withdraw</code> element - sequence.</p> - - <p>With all this information we can now examine how to iterate over - transaction in the batch in content order:</p> - - <pre class="c++"> -batch& b = ... - -for (batch::content_order_const_iterator i (b.content_order ().begin ()); - i != b.content_order ().end (); - ++i) -{ - switch (i->id) - { - case batch::withdraw_id: - { - const withdraw& t (b.withdraw ()[i->index]); - cerr << t.account () << " withdraw " << t.amount () << endl; - break; - } - case batch::deposit_id: - { - const deposit& t (b.deposit ()[i->index]); - cerr << t.account () << " deposit " << t.amount () << endl; - break; - } - default: - { - assert (false); // Unknown content id. - } - } -} - </pre> - - <p>If we serialized our batch back to XML, we would also see that the - order of transactions in the output is exactly the same as in the - input rather than all the withdrawals first followed by all the - deposits.</p> - - <p>The most complex aspect of working with ordered types is - modifications. Now we not only need to change the content, - but also remember to update the order information corresponding - to this change. As a first example, we add a deposit transaction - to the batch:</p> - - <pre class="c++"> -using xml_schema::content_order; - -batch::deposit_sequence& d (b.deposit ()); -batch::withdraw_sequence& w (b.withdraw ()); -batch::content_order_sequence& co (b.content_order ()); - -d.push_back (deposit (123456789, 100000)); -co.push_back (content_order (batch::deposit_id, d.size () - 1)); - </pre> - - <p>In the above example we first added the content (deposit - transaction) and then updated the content order information - by adding an entry with <code>deposit_id</code> content - id and the index of the just added deposit transaction.</p> - - <p>Removing the last transaction can be easy if we know which - transaction (deposit or withdrawal) is last:</p> - - <pre class="c++"> -d.pop_back (); -co.pop_back (); - </pre> - - <p>If, however, we do not know which transaction is last, then - things get a bit more complicated:</p> - - <pre class="c++"> -switch (co.back ().id) -{ -case batch::withdraw_id: - { - d.pop_back (); - break; - } -case batch::deposit_id: - { - w.pop_back (); - break; - } -} - -co.pop_back (); - </pre> - - <p>The following example shows how to add a transaction at the - beginning of the batch:</p> - - <pre class="c++"> -w.push_back (withdraw (123456789, 100000)); -co.insert (co.begin (), - content_order (batch::withdraw_id, w.size () - 1)); - </pre> - - <p>Note also that when we merely modify the content of one - of the elements in place, we do not need to update its - order since it doesn't change. For example, here is how - we can change the amount in the first withdrawal:</p> - - <pre class="c++"> -w[0].amount (10000); - </pre> - - <p>For the complete working code shown in this section refer to the - <code>order/element</code> example in the - <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <p>If both the base and derived types are ordered, then the - content order sequence is only added to the base and the content - ids are unique within the whole hierarchy. In this case - the content order sequence for the derived type contains - ordering information for both base and derived content.</p> - - <p>In some applications we may need to perform more complex - content processing. For example, in our case, we may need - to remove all the withdrawal transactions. The default - container, <code>std::vector</code>, is not particularly - suitable for such operations. What may be required by - some applications is a multi-index container that not - only allows us to iterate in content order similar to - <code>std::vector</code> but also search by the content - id as well as the content id and index pair.</p> - - <p>While C++/Tree does not provide this functionality by - default, it allows us to specify a custom container - type for content order with the <code>--order-container</code> - command line option. The only requirement from the - generated code side for such a container is to provide - the <code>vector</code>-like <code>push_back()</code>, - <code>size()</code>, and const iteration interfaces.</p> - - <p>As an example, here is how we can use the Boost Multi-Index - container for content order. First we create the - <code>content-order-container.hxx</code> header with the - following definition:</p> - - <pre class="c++"> -#ifndef CONTENT_ORDER_CONTAINER -#define CONTENT_ORDER_CONTAINER - -#include <cstddef> // std::size_t - -#include <boost/multi_index_container.hpp> -#include <boost/multi_index/member.hpp> -#include <boost/multi_index/identity.hpp> -#include <boost/multi_index/ordered_index.hpp> -#include <boost/multi_index/random_access_index.hpp> - -struct by_id {}; -struct by_id_index {}; - -template <typename T> -using content_order_container = - boost::multi_index::multi_index_container< - T, - boost::multi_index::indexed_by< - boost::multi_index::random_access<>, - boost::multi_index::ordered_unique< - boost::multi_index::tag<by_id_index>, - boost::multi_index::identity<T> - >, - boost::multi_index::ordered_non_unique< - boost::multi_index::tag<by_id>, - boost::multi_index::member<T, std::size_t, &T::id> - > - > - >; - -#endif - </pre> - - <p>Next we add the following two XSD compiler options to include - this header into every generated header file and to use the - custom container type (see the XSD compiler command line manual - for more information on shell quoting for the first option):</p> - - <pre class="term"> ---hxx-prologue '#include "content-order-container.hxx"' ---order-container content_order_container - </pre> - - <p>With these changes we can now use the multi-index functionality, - for example, to search for a specific content id:</p> - - <pre class="c++"> -typedef batch::content_order_sequence::index<by_id>::type id_set; -typedef id_set::iterator id_iterator; - -const id_set& ids (b.content_order ().get<by_id> ()); - -std::pair<id_iterator, id_iterator> r ( - ids.equal_range (std::size_t (batch::deposit_id)); - -for (id_iterator i (r.first); i != r.second; ++i) -{ - const deposit& t (b.deposit ()[i->index]); - cerr << t.account () << " deposit " << t.amount () << endl; -} - </pre> - - <h2><a name="2.9">2.9 Mapping for Global Elements</a></h2> - - <p>An XML Schema element definition is called global if it appears - directly under the <code>schema</code> element. - A global element is a valid root of an instance document. By - default, a global element is mapped to a set of overloaded - parsing and, optionally, serialization functions with the - same name as the element. It is also possible to generate types - for root elements instead of parsing and serialization functions. - This is primarily useful to distinguish object models with the - same root type but with different root elements. See - <a href="#2.9.1">Section 2.9.1, "Element Types"</a> for details. - It is also possible to request the generation of an element map - which allows uniform parsing and serialization of multiple root - elements. See <a href="#2.9.2">Section 2.9.2, "Element Map"</a> - for details. - </p> - - <p>The parsing functions read XML instance documents and return - corresponding object models as an automatic pointer - (<code>std::unique_ptr</code> or <code>std::auto_ptr</code>, - depending on the C++ standard selected). Their signatures - have the following pattern (<code>type</code> denotes - element's type and <code>name</code> denotes element's - name): - </p> - - <pre class="c++"> -std::[unique|auto]_ptr<type> -name (....); - </pre> - - <p>The process of parsing, including the exact signatures of the parsing - functions, is the subject of <a href="#3">Chapter 3, "Parsing"</a>. - </p> - - <p>The serialization functions write object models back to XML instance - documents. Their signatures have the following pattern: - </p> - - <pre class="c++"> -void -name (<stream type>&, const type&, ....); - </pre> - - <p>The process of serialization, including the exact signatures of the - serialization functions, is the subject of <a href="#4">Chapter 4, - "Serialization"</a>. - </p> - - - <h3><a name="2.9.1">2.9.1 Element Types</a></h3> - - <p>The generation of element types is requested with the - <code>--generate-element-type</code> option. With this option - each global element is mapped to a C++ class with the - same name as the element. Such a class is derived from - <code>xml_schema::element_type</code> and contains the same set - of type definitions, constructors, and member function as would a - type containing a single element with the One cardinality class - named <code>"value"</code>. In addition, the element type also - contains a set of member functions for accessing the element - name and namespace as well as its value in a uniform manner. - For example:</p> - - <pre class="xml"> -<complexType name="type"> - <sequence> - ... - </sequence> -</complexType> - -<element name="root" type="type"/> - </pre> - -<p>is mapped to:</p> - - <pre class="c++"> -class type -{ - ... -}; - -class root: public xml_schema::element_type -{ -public: - // Element value. - // - typedef type value_type; - - const value_type& - value () const; - - value_type& - value (); - - void - value (const value_type&); - - void - value (std::[unique|auto]_ptr<value_type>); - - // Constructors. - // - root (const value_type&); - - root (std::[unique|auto]_ptr<value_type>); - - root (const xercesc::DOMElement&, xml_schema::flags = 0); - - root (const root&, xml_schema::flags = 0); - - virtual root* - _clone (xml_schema::flags = 0) const; - - // Element name and namespace. - // - static const std::string& - name (); - - static const std::string& - namespace_ (); - - virtual const std::string& - _name () const; - - virtual const std::string& - _namespace () const; - - // Element value as xml_schema::type. - // - virtual const xml_schema::type* - _value () const; - - virtual xml_schema::type* - _value (); -}; - -void -operator<< (xercesc::DOMElement&, const root&); - </pre> - - <p>The <code>xml_schema::element_type</code> class is a common - base type for all element types and is defined as follows:</p> - - <pre class="c++"> -namespace xml_schema -{ - class element_type - { - public: - virtual - ~element_type (); - - virtual element_type* - _clone (flags f = 0) const = 0; - - virtual const std::basic_string<C>& - _name () const = 0; - - virtual const std::basic_string<C>& - _namespace () const = 0; - - virtual xml_schema::type* - _value () = 0; - - virtual const xml_schema::type* - _value () const = 0; - }; -} - </pre> - - <p>The <code>_value()</code> member function returns a pointer to - the element value or 0 if the element is of a fundamental C++ - type and therefore is not derived from <code>xml_schema::type</code>. - </p> - - <p>Unlike parsing and serialization functions, element types - are only capable of parsing and serializing from/to a - <code>DOMElement</code> object. This means that the application - will need to perform its own XML-to-DOM parsing and DOM-to-XML - serialization. The following section describes a mechanism - provided by the mapping to uniformly parse and serialize - multiple root elements.</p> - - - <h3><a name="2.9.2">2.9.2 Element Map</a></h3> - - <p>When element types are generated for root elements it is also - possible to request the generation of an element map with the - <code>--generate-element-map</code> option. The element map - allows uniform parsing and serialization of multiple root - elements via the common <code>xml_schema::element_type</code> - base type. The <code>xml_schema::element_map</code> class is - defined as follows:</p> - - <pre class="c++"> -namespace xml_schema -{ - class element_map - { - public: - static std::[unique|auto]_ptr<xml_schema::element_type> - parse (const xercesc::DOMElement&, flags = 0); - - static void - serialize (xercesc::DOMElement&, const element_type&); - }; -} - </pre> - - <p>The <code>parse()</code> function creates the corresponding - element type object based on the element name and namespace - and returns it as an automatic pointer (<code>std::unique_ptr</code> - or <code>std::auto_ptr</code>, depending on the C++ standard - selected) to <code>xml_schema::element_type</code>. - The <code>serialize()</code> function serializes the passed element - object to <code>DOMElement</code>. Note that in case of - <code>serialize()</code>, the <code>DOMElement</code> object - should have the correct name and namespace. If no element type is - available for an element, both functions throw the - <code>xml_schema::no_element_info</code> exception:</p> - - <pre class="c++"> -struct no_element_info: virtual exception -{ - no_element_info (const std::basic_string<C>& element_name, - const std::basic_string<C>& element_namespace); - - const std::basic_string<C>& - element_name () const; - - const std::basic_string<C>& - element_namespace () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The application can discover the actual type of the element - object returned by <code>parse()</code> either using - <code>dynamic_cast</code> or by comparing element names and - namespaces. The following code fragments illustrate how the - element map can be used:</p> - - <pre class="c++"> -// Parsing. -// -DOMElement& e = ... // Parse XML to DOM. - -unique_ptr<xml_schema::element_type> r ( - xml_schema::element_map::parse (e)); - -if (root1 r1 = dynamic_cast<root1*> (r.get ())) -{ - ... -} -else if (r->_name == root2::name () && - r->_namespace () == root2::namespace_ ()) -{ - root2& r2 (static_cast<root2&> (*r)); - - ... -} - </pre> - - <pre class="c++"> -// Serialization. -// -xml_schema::element_type& r = ... - -string name (r._name ()); -string ns (r._namespace ()); - -DOMDocument& doc = ... // Create a new DOMDocument with name and ns. -DOMElement& e (*doc->getDocumentElement ()); - -xml_schema::element_map::serialize (e, r); - -// Serialize DOMDocument to XML. - </pre> - - <!-- --> - - <h2><a name="2.10">2.10 Mapping for Global Attributes</a></h2> - - <p>An XML Schema attribute definition is called global if it appears - directly under the <code>schema</code> element. A global - attribute does not have any mapping. - </p> - - <!-- - When it is referenced from - a local attribute definition (using the <code>ref</code> attribute) - it is treated as a local attribute (see Section 2.8, "Mapping for - Local Elements and Attributes"). - --> - - <h2><a name="2.11">2.11 Mapping for <code>xsi:type</code> and Substitution - Groups</a></h2> - - <p>The mapping provides optional support for the XML Schema polymorphism - features (<code>xsi:type</code> and substitution groups) which can - be requested with the <code>--generate-polymorphic</code> option. - When used, the dynamic type of a member may be different from - its static type. Consider the following schema definition and - instance document: - </p> - - <pre class="xml"> -<!-- test.xsd --> -<schema> - <complexType name="base"> - <attribute name="text" type="string"/> - </complexType> - - <complexType name="derived"> - <complexContent> - <extension base="base"> - <attribute name="extra-text" type="string"/> - </extension> - </complexContent> - </complexType> - - <complexType name="root_type"> - <sequence> - <element name="item" type="base" maxOccurs="unbounded"/> - </sequence> - </complexType> - - <element name="root" type="root_type"/> -</schema> - -<!-- test.xml --> -<root xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"> - <item text="hello"/> - <item text="hello" extra-text="world" xsi:type="derived"/> -</root> - </pre> - - <p>In the resulting object model, the container for - the <code>root::item</code> member will have two elements: - the first element's type will be <code>base</code> while - the second element's (dynamic) type will be - <code>derived</code>. This can be discovered using the - <code>dynamic_cast</code> operator as shown in the following - example: - </p> - - <pre class="c++"> -void -f (root& r) -{ - for (root::item_const_iterator i (r.item ().begin ()); - i != r.item ().end () - ++i) - { - if (derived* d = dynamic_cast<derived*> (&(*i))) - { - // derived - } - else - { - // base - } - } -} - </pre> - - <p>The <code>_clone</code> virtual function should be used instead of - copy constructors to make copies of members that might use - polymorphism: - </p> - - <pre class="c++"> -void -f (root& r) -{ - for (root::item_const_iterator i (r.item ().begin ()); - i != r.item ().end () - ++i) - { - std::unique_ptr<base> c (i->_clone ()); - } -} - </pre> - - <p>The mapping can often automatically determine which types are - polymorphic based on the substitution group declarations. However, - if your XML vocabulary is not using substitution groups or if - substitution groups are defined in a separate schema, then you will - need to use the <code>--polymorphic-type</code> option to specify - which types are polymorphic. When using this option you only need - to specify the root of a polymorphic type hierarchy and the mapping - will assume that all the derived types are also polymorphic. - Also note that you need to specify this option when compiling every - schema file that references the polymorphic type. Consider the following - two schemas as an example:</p> - - <pre class="xml"> -<!-- base.xsd --> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:complexType name="base"> - <xs:sequence> - <xs:element name="b" type="xs:int"/> - </xs:sequence> - </xs:complexType> - - <!-- substitution group root --> - <xs:element name="base" type="base"/> - -</xs:schema> - </pre> - - <pre class="xml"> -<!-- derived.xsd --> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <include schemaLocation="base.xsd"/> - - <xs:complexType name="derived"> - <xs:complexContent> - <xs:extension base="base"> - <xs:sequence> - <xs:element name="d" type="xs:string"/> - </xs:sequence> - </xs:extension> - </xs:complexContent> - </xs:complexType> - - <xs:element name="derived" type="derived" substitutionGroup="base"/> - -</xs:schema> - </pre> - - <p>In this example we need to specify "<code>--polymorphic-type base</code>" - when compiling both schemas because the substitution group is declared - in a schema other than the one defining type <code>base</code>.</p> - - <p>You can also indicate that all types should be treated as polymorphic - with the <code>--polymorphic-type-all</code>. However, this may result - in slower generated code with a greater footprint.</p> - - - <!-- Mapping for any and anyAttribute --> - - - <h2><a name="2.12">2.12 Mapping for <code>any</code> and <code>anyAttribute</code></a></h2> - - <p>For the XML Schema <code>any</code> and <code>anyAttribute</code> - wildcards an optional mapping can be requested with the - <code>--generate-wildcard</code> option. The mapping represents - the content matched by wildcards as DOM fragments. Because the - DOM API is used to access such content, the Xerces-C++ runtime - should be initialized by the application prior to parsing and - should remain initialized for the lifetime of objects with - the wildcard content. For more information on the Xerces-C++ - runtime initialization see <a href="#3.1">Section 3.1, - "Initializing the Xerces-C++ Runtime"</a>. - </p> - - <p>The mapping for <code>any</code> is similar to the mapping for - local elements (see <a href="#2.8">Section 2.8, "Mapping for Local - Elements and Attributes"</a>) except that the type used in the - wildcard mapping is <code>xercesc::DOMElement</code>. As with local - elements, the mapping divides all possible cardinality combinations - into three cardinality classes: <i>one</i>, <i>optional</i>, and - <i>sequence</i>. - </p> - - <p>The mapping for <code>anyAttribute</code> represents the attributes - matched by this wildcard as a set of <code>xercesc::DOMAttr</code> - objects with a key being the attribute's name and namespace.</p> - - <p>Similar to local elements and attributes, the <code>any</code> and - <code>anyAttribute</code> wildcards are mapped to a set of public type - definitions (typedefs) and a set of public accessor and modifier - functions. Type definitions have names derived from <code>"any"</code> - for the <code>any</code> wildcard and <code>"any_attribute"</code> - for the <code>anyAttribute</code> wildcard. The accessor and modifier - functions are named <code>"any"</code> for the <code>any</code> wildcard - and <code>"any_attribute"</code> for the <code>anyAttribute</code> - wildcard. Subsequent wildcards in the same type have escaped names - such as <code>"any1"</code> or <code>"any_attribute1"</code>. - </p> - - <p>Because Xerces-C++ DOM nodes always belong to a <code>DOMDocument</code>, - each type with a wildcard has an associated <code>DOMDocument</code> - object. The reference to this object can be obtained using the accessor - function called <code>dom_document</code>. The access to the document - object from the application code may be necessary to create or modify - the wildcard content. For example: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other"/> - </sequence> - <anyAttribute namespace="##other"/> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // any - // - const xercesc::DOMElement& - any () const; - - void - any (const xercesc::DOMElement&); - - ... - - // any_attribute - // - typedef attribute_set any_attribute_set; - typedef any_attribute_set::iterator any_attribute_iterator; - typedef any_attribute_set::const_iterator any_attribute_const_iterator; - - const any_attribute_set& - any_attribute () const; - - any_attribute_set& - any_attribute (); - - ... - - // DOMDocument object for wildcard content. - // - const xercesc::DOMDocument& - dom_document () const; - - xercesc::DOMDocument& - dom_document (); - - ... -}; - </pre> - - - <p>Names and semantics of type definitions for the wildcards as well - as signatures of the accessor and modifier functions depend on the - wildcard type as well as the cardinality class for the <code>any</code> - wildcard. They are described in the following sub-sections. - </p> - - - <h3><a name="2.12.1">2.12.1 Mapping for <code>any</code> with the One Cardinality Class</a></h3> - - <p>For <code>any</code> with the One cardinality class, - there are no type definitions. The accessor functions come in - constant and non-constant versions. The constant accessor function - returns a constant reference to <code>xercesc::DOMElement</code> and - can be used for read-only access. The non-constant version returns - an unrestricted reference to <code>xercesc::DOMElement</code> and can - be used for read-write access. - </p> - - <p>The first modifier function expects an argument of type reference - to constant <code>xercesc::DOMElement</code> and makes a deep copy - of its argument. The second modifier function expects an argument of - type pointer to <code>xercesc::DOMElement</code>. This modifier - function assumes ownership of its argument and expects the element - object to be created using the DOM document associated with this - instance. For example: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Accessors. - // - const xercesc::DOMElement& - any () const; - - xercesc::DOMElement& - any (); - - // Modifiers. - // - void - any (const xercesc::DOMElement&); - - void - any (xercesc::DOMElement*); - - ... - -}; - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMElement& e) -{ - using namespace xercesc; - - DOMElement& e1 (o.any ()); // get - o.any (e) // set, deep copy - DOMDocument& doc (o.dom_document ()); - o.any (doc.createElement (...)); // set, assumes ownership -} - </pre> - - <h3><a name="2.12.2">2.12.2 Mapping for <code>any</code> with the Optional Cardinality Class</a></h3> - - <p>For <code>any</code> with the Optional cardinality class, the type - definitions consist of an alias for the container type with name - <code>any_optional</code> (or <code>any1_optional</code>, etc., for - subsequent wildcards in the type definition). - </p> - - <p>Unlike accessor functions for the One cardinality class, accessor - functions for the Optional cardinality class return references to - corresponding containers rather than directly to <code>DOMElement</code>. - The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to - the container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container - and can be used for read-write access. - </p> - - <p>The modifier functions are overloaded for <code>xercesc::DOMElement</code> - and the container type. The first modifier function expects an argument of - type reference to constant <code>xercesc::DOMElement</code> and - makes a deep copy of its argument. The second modifier function - expects an argument of type pointer to <code>xercesc::DOMElement</code>. - This modifier function assumes ownership of its argument and expects - the element object to be created using the DOM document associated - with this instance. The third modifier function expects an argument - of type reference to constant of the container type and makes a - deep copy of its argument. For instance: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other" minOccurs="0"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef element_optional any_optional; - - // Accessors. - // - const any_optional& - any () const; - - any_optional& - any (); - - // Modifiers. - // - void - any (const xercesc::DOMElement&); - - void - any (xercesc::DOMElement*); - - void - any (const any_optional&); - - ... - -}; - </pre> - - - <p>The <code>element_optional</code> container is a - specialization of the <code>optional</code> class template described - in <a href="#2.8.2">Section 2.8.2, "Mapping for Members with the Optional - Cardinality Class"</a>. Its interface is presented below: - </p> - - <pre class="c++"> -class element_optional -{ -public: - explicit - element_optional (xercesc::DOMDocument&); - - // Makes a deep copy. - // - element_optional (const xercesc::DOMElement&, xercesc::DOMDocument&); - - // Assumes ownership. - // - element_optional (xercesc::DOMElement*, xercesc::DOMDocument&); - - element_optional (const element_optional&, xercesc::DOMDocument&); - -public: - element_optional& - operator= (const xercesc::DOMElement&); - - element_optional& - operator= (const element_optional&); - - // Pointer-like interface. - // -public: - const xercesc::DOMElement* - operator-> () const; - - xercesc::DOMElement* - operator-> (); - - const xercesc::DOMElement& - operator* () const; - - xercesc::DOMElement& - operator* (); - - typedef void (element_optional::*bool_convertible) (); - operator bool_convertible () const; - - // Get/set interface. - // -public: - bool - present () const; - - const xercesc::DOMElement& - get () const; - - xercesc::DOMElement& - get (); - - // Makes a deep copy. - // - void - set (const xercesc::DOMElement&); - - // Assumes ownership. - // - void - set (xercesc::DOMElement*); - - void - reset (); -}; - -bool -operator== (const element_optional&, const element_optional&); - -bool -operator!= (const element_optional&, const element_optional&); - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMElement& e) -{ - using namespace xercesc; - - DOMDocument& doc (o.dom_document ()); - - if (o.any ().present ()) // test - { - DOMElement& e1 (o.any ().get ()); // get - o.any ().set (e); // set, deep copy - o.any ().set (doc.createElement (...)); // set, assumes ownership - o.any ().reset (); // reset - } - - // Same as above but using pointer notation: - // - if (o.member ()) // test - { - DOMElement& e1 (*o.any ()); // get - o.any (e); // set, deep copy - o.any (doc.createElement (...)); // set, assumes ownership - o.any ().reset (); // reset - } -} - </pre> - - - - <h3><a name="2.12.3">2.12.3 Mapping for <code>any</code> with the Sequence Cardinality Class</a></h3> - - <p>For <code>any</code> with the Sequence cardinality class, the type - definitions consist of an alias of the container type with name - <code>any_sequence</code> (or <code>any1_sequence</code>, etc., for - subsequent wildcards in the type definition), an alias of the iterator - type with name <code>any_iterator</code> (or <code>any1_iterator</code>, - etc., for subsequent wildcards in the type definition), and an alias - of the constant iterator type with name <code>any_const_iterator</code> - (or <code>any1_const_iterator</code>, etc., for subsequent wildcards - in the type definition). - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container and can - be used for read-write access. - </p> - - <p>The modifier function expects an argument of type reference to - constant of the container type. The modifier function makes - a deep copy of its argument. For instance: - </p> - - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other" minOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef element_sequence any_sequence; - typedef any_sequence::iterator any_iterator; - typedef any_sequence::const_iterator any_const_iterator; - - // Accessors. - // - const any_sequence& - any () const; - - any_sequence& - any (); - - // Modifier. - // - void - any (const any_sequence&); - - ... - -}; - </pre> - - <p>The <code>element_sequence</code> container is a - specialization of the <code>sequence</code> class template described - in <a href="#2.8.3">Section 2.8.3, "Mapping for Members with the - Sequence Cardinality Class"</a>. Its interface is similar to - the sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences") and is - presented below: - </p> - - <pre class="c++"> -class element_sequence -{ -public: - typedef xercesc::DOMElement value_type; - typedef xercesc::DOMElement* pointer; - typedef const xercesc::DOMElement* const_pointer; - typedef xercesc::DOMElement& reference; - typedef const xercesc::DOMElement& const_reference; - - typedef <implementation-defined> iterator; - typedef <implementation-defined> const_iterator; - typedef <implementation-defined> reverse_iterator; - typedef <implementation-defined> const_reverse_iterator; - - typedef <implementation-defined> size_type; - typedef <implementation-defined> difference_type; - typedef <implementation-defined> allocator_type; - -public: - explicit - element_sequence (xercesc::DOMDocument&); - - // DOMElement cannot be default-constructed. - // - // explicit - // element_sequence (size_type n); - - element_sequence (size_type n, - const xercesc::DOMElement&, - xercesc::DOMDocument&); - - template <typename I> - element_sequence (const I& begin, - const I& end, - xercesc::DOMDocument&); - - element_sequence (const element_sequence&, xercesc::DOMDocument&); - - element_sequence& - operator= (const element_sequence&); - -public: - void - assign (size_type n, const xercesc::DOMElement&); - - template <typename I> - void - assign (const I& begin, const I& end); - -public: - // This version of resize can only be used to shrink the - // sequence because DOMElement cannot be default-constructed. - // - void - resize (size_type); - - void - resize (size_type, const xercesc::DOMElement&); - -public: - size_type - size () const; - - size_type - max_size () const; - - size_type - capacity () const; - - bool - empty () const; - - void - reserve (size_type); - - void - clear (); - -public: - const_iterator - begin () const; - - const_iterator - end () const; - - iterator - begin (); - - iterator - end (); - - const_reverse_iterator - rbegin () const; - - const_reverse_iterator - rend () const - - reverse_iterator - rbegin (); - - reverse_iterator - rend (); - -public: - xercesc::DOMElement& - operator[] (size_type); - - const xercesc::DOMElement& - operator[] (size_type) const; - - xercesc::DOMElement& - at (size_type); - - const xercesc::DOMElement& - at (size_type) const; - - xercesc::DOMElement& - front (); - - const xercesc::DOMElement& - front () const; - - xercesc::DOMElement& - back (); - - const xercesc::DOMElement& - back () const; - -public: - // Makes a deep copy. - // - void - push_back (const xercesc::DOMElement&); - - // Assumes ownership. - // - void - push_back (xercesc::DOMElement*); - - void - pop_back (); - - // Makes a deep copy. - // - iterator - insert (iterator position, const xercesc::DOMElement&); - - // Assumes ownership. - // - iterator - insert (iterator position, xercesc::DOMElement*); - - void - insert (iterator position, size_type n, const xercesc::DOMElement&); - - template <typename I> - void - insert (iterator position, const I& begin, const I& end); - - iterator - erase (iterator position); - - iterator - erase (iterator begin, iterator end); - -public: - // Note that the DOMDocument object of the two sequences being - // swapped should be the same. - // - void - swap (sequence& x); -}; - -inline bool -operator== (const element_sequence&, const element_sequence&); - -inline bool -operator!= (const element_sequence&, const element_sequence&); - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMElement& e) -{ - using namespace xercesc; - - object::any_sequence& s (o.any ()); - - // Iteration. - // - for (object::any_iterator i (s.begin ()); i != s.end (); ++i) - { - DOMElement& e (*i); - } - - // Modification. - // - s.push_back (e); // deep copy - DOMDocument& doc (o.dom_document ()); - s.push_back (doc.createElement (...)); // assumes ownership -} - </pre> - - <h3><a name="2.12.4">2.12.4 Element Wildcard Order</a></h3> - - <p>Similar to elements, element wildcards in ordered types - (<a href="#2.8.4">Section 2.8.4, "Element Order"</a>) are assigned - content ids and are included in the content order sequence. - Continuing with the bank transactions example started in Section - 2.8.4, we can extend the batch by allowing custom transactions:</p> - - <pre class="xml"> -<complexType name="batch"> - <choice minOccurs="0" maxOccurs="unbounded"> - <element name="withdraw" type="withdraw"/> - <element name="deposit" type="deposit"/> - <any namespace="##other" processContents="lax"/> - </choice> -</complexType> - </pre> - - <p>This will lead to the following changes in the generated - <code>batch</code> C++ class:</p> - - <pre class="c++"> -class batch: public xml_schema::type -{ -public: - ... - - // any - // - typedef element_sequence any_sequence; - typedef any_sequence::iterator any_iterator; - typedef any_sequence::const_iterator any_const_iterator; - - static const std::size_t any_id = 3UL; - - const any_sequence& - any () const; - - any_sequence& - any (); - - void - any (const any_sequence&); - - ... -}; - </pre> - - <p>With this change we also need to update the iteration code to handle - the new content id:</p> - - <pre class="c++"> -for (batch::content_order_const_iterator i (b.content_order ().begin ()); - i != b.content_order ().end (); - ++i) -{ - switch (i->id) - { - ... - - case batch::any_id: - { - const DOMElement& e (b.any ()[i->index]); - ... - break; - } - - ... - } -} - </pre> - - <p>For the complete working code that shows the use of wildcards in - ordered types refer to the <code>order/element</code> example in - the <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <h3><a name="2.12.5">2.12.5 Mapping for <code>anyAttribute</code></a></h3> - - <p>For <code>anyAttribute</code> the type definitions consist of an alias - of the container type with name <code>any_attribute_set</code> - (or <code>any1_attribute_set</code>, etc., for subsequent wildcards - in the type definition), an alias of the iterator type with name - <code>any_attribute_iterator</code> (or <code>any1_attribute_iterator</code>, - etc., for subsequent wildcards in the type definition), and an alias - of the constant iterator type with name <code>any_attribute_const_iterator</code> - (or <code>any1_attribute_const_iterator</code>, etc., for subsequent - wildcards in the type definition). - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container and can - be used for read-write access. - </p> - - <p>The modifier function expects an argument of type reference to - constant of the container type. The modifier function makes - a deep copy of its argument. For instance: - </p> - - - <pre class="xml"> -<complexType name="object"> - <sequence> - ... - </sequence> - <anyAttribute namespace="##other"/> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef attribute_set any_attribute_set; - typedef any_attribute_set::iterator any_attribute_iterator; - typedef any_attribute_set::const_iterator any_attribute_const_iterator; - - // Accessors. - // - const any_attribute_set& - any_attribute () const; - - any_attribute_set& - any_attribute (); - - // Modifier. - // - void - any_attribute (const any_attribute_set&); - - ... - -}; - </pre> - - <p>The <code>attribute_set</code> class is an associative container - similar to the <code>std::set</code> class template as defined by - the ISO/ANSI Standard for C++ (ISO/IEC 14882:1998, Section 23.3.3, - "Class template set") with the key being the attribute's name - and namespace. Unlike <code>std::set</code>, <code>attribute_set</code> - allows searching using names and namespaces instead of - <code>xercesc::DOMAttr</code> objects. It is defined in an - implementation-specific namespace and its interface is presented - below: - </p> - - <pre class="c++"> -class attribute_set -{ -public: - typedef xercesc::DOMAttr key_type; - typedef xercesc::DOMAttr value_type; - typedef xercesc::DOMAttr* pointer; - typedef const xercesc::DOMAttr* const_pointer; - typedef xercesc::DOMAttr& reference; - typedef const xercesc::DOMAttr& const_reference; - - typedef <implementation-defined> iterator; - typedef <implementation-defined> const_iterator; - typedef <implementation-defined> reverse_iterator; - typedef <implementation-defined> const_reverse_iterator; - - typedef <implementation-defined> size_type; - typedef <implementation-defined> difference_type; - typedef <implementation-defined> allocator_type; - -public: - attribute_set (xercesc::DOMDocument&); - - template <typename I> - attribute_set (const I& begin, const I& end, xercesc::DOMDocument&); - - attribute_set (const attribute_set&, xercesc::DOMDocument&); - - attribute_set& - operator= (const attribute_set&); - -public: - const_iterator - begin () const; - - const_iterator - end () const; - - iterator - begin (); - - iterator - end (); - - const_reverse_iterator - rbegin () const; - - const_reverse_iterator - rend () const; - - reverse_iterator - rbegin (); - - reverse_iterator - rend (); - -public: - size_type - size () const; - - size_type - max_size () const; - - bool - empty () const; - - void - clear (); - -public: - // Makes a deep copy. - // - std::pair<iterator, bool> - insert (const xercesc::DOMAttr&); - - // Assumes ownership. - // - std::pair<iterator, bool> - insert (xercesc::DOMAttr*); - - // Makes a deep copy. - // - iterator - insert (iterator position, const xercesc::DOMAttr&); - - // Assumes ownership. - // - iterator - insert (iterator position, xercesc::DOMAttr*); - - template <typename I> - void - insert (const I& begin, const I& end); - -public: - void - erase (iterator position); - - size_type - erase (const std::basic_string<C>& name); - - size_type - erase (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name); - - size_type - erase (const XMLCh* name); - - size_type - erase (const XMLCh* namespace_, const XMLCh* name); - - void - erase (iterator begin, iterator end); - -public: - size_type - count (const std::basic_string<C>& name) const; - - size_type - count (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name) const; - - size_type - count (const XMLCh* name) const; - - size_type - count (const XMLCh* namespace_, const XMLCh* name) const; - - iterator - find (const std::basic_string<C>& name); - - iterator - find (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name); - - iterator - find (const XMLCh* name); - - iterator - find (const XMLCh* namespace_, const XMLCh* name); - - const_iterator - find (const std::basic_string<C>& name) const; - - const_iterator - find (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name) const; - - const_iterator - find (const XMLCh* name) const; - - const_iterator - find (const XMLCh* namespace_, const XMLCh* name) const; - -public: - // Note that the DOMDocument object of the two sets being - // swapped should be the same. - // - void - swap (attribute_set&); -}; - -bool -operator== (const attribute_set&, const attribute_set&); - -bool -operator!= (const attribute_set&, const attribute_set&); - </pre> - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMAttr& a) -{ - using namespace xercesc; - - object::any_attribute_set& s (o.any_attribute ()); - - // Iteration. - // - for (object::any_attribute_iterator i (s.begin ()); i != s.end (); ++i) - { - DOMAttr& a (*i); - } - - // Modification. - // - s.insert (a); // deep copy - DOMDocument& doc (o.dom_document ()); - s.insert (doc.createAttribute (...)); // assumes ownership - - // Searching. - // - object::any_attribute_iterator i (s.find ("name")); - i = s.find ("http://www.w3.org/XML/1998/namespace", "lang"); -} - </pre> - - <!-- Mapping for Mixed Content Models --> - - <h2><a name="2.13">2.13 Mapping for Mixed Content Models</a></h2> - - <p>For XML Schema types with mixed content models C++/Tree provides - mapping support only if the type is marked as ordered - (<a href="#2.8.4">Section 2.8.4, "Element Order"</a>). Use the - <code>--ordered-type-mixed</code> XSD compiler option to - automatically mark all types with mixed content as ordered.</p> - - <p>For an ordered type with mixed content, C++/Tree adds an extra - text content sequence that is used to store the text fragments. - This text content sequence is also assigned the content id and - its entries are included in the content order sequence, just - like elements. As a result, it is possible to capture the order - between elements and text fragments.</p> - - <p>As an example, consider the following schema that describes text - with embedded links:</p> - - <pre class="xml"> -<complexType name="anchor"> - <simpleContent> - <extension base="string"> - <attribute name="href" type="anyURI" use="required"/> - </extension> - </simpleContent> -</complexType> - -<complexType name="text" mixed="true"> - <sequence> - <element name="a" type="anchor" minOccurs="0" maxOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>The generated <code>text</code> C++ class will provide the following - API (assuming it is marked as ordered):</p> - - <pre class="c++"> -class text: public xml_schema::type -{ -public: - // a - // - typedef anchor a_type; - typedef sequence<a_type> a_sequence; - typedef a_sequence::iterator a_iterator; - typedef a_sequence::const_iterator a_const_iterator; - - static const std::size_t a_id = 1UL; - - const a_sequence& - a () const; - - a_sequence& - a (); - - void - a (const a_sequence&); - - // text_content - // - typedef xml_schema::string text_content_type; - typedef sequence<text_content_type> text_content_sequence; - typedef text_content_sequence::iterator text_content_iterator; - typedef text_content_sequence::const_iterator text_content_const_iterator; - - static const std::size_t text_content_id = 2UL; - - const text_content_sequence& - text_content () const; - - text_content_sequence& - text_content (); - - void - text_content (const text_content_sequence&); - - // content_order - // - typedef xml_schema::content_order content_order_type; - typedef std::vector<content_order_type> content_order_sequence; - typedef content_order_sequence::iterator content_order_iterator; - typedef content_order_sequence::const_iterator content_order_const_iterator; - - const content_order_sequence& - content_order () const; - - content_order_sequence& - content_order (); - - void - content_order (const content_order_sequence&); - - ... -}; - </pre> - - <p>Given this interface we can iterate over both link elements - and text in content order. The following code fragment converts - our format to plain text with references.</p> - - <pre class="c++"> -const text& t = ... - -for (text::content_order_const_iterator i (t.content_order ().begin ()); - i != t.content_order ().end (); - ++i) -{ - switch (i->id) - { - case text::a_id: - { - const anchor& a (t.a ()[i->index]); - cerr << a << "[" << a.href () << "]"; - break; - } - case text::text_content_id: - { - const xml_schema::string& s (t.text_content ()[i->index]); - cerr << s; - break; - } - default: - { - assert (false); // Unknown content id. - } - } -} - </pre> - - <p>For the complete working code that shows the use of mixed content - in ordered types refer to the <code>order/mixed</code> example in - the <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <!-- Parsing --> - - - <h1><a name="3">3 Parsing</a></h1> - - <p>This chapter covers various aspects of parsing XML instance - documents in order to obtain corresponding tree-like object - model. - </p> - - <p>Each global XML Schema element in the form:</p> - - <pre class="xml"> -<element name="name" type="type"/> - </pre> - - <p>is mapped to 14 overloaded C++ functions in the form:</p> - - <pre class="c++"> -// Read from a URI or a local file. -// - -std::[unique|auto]_ptr<type> -name (const std::basic_string<C>& uri, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (const std::basic_string<C>& uri, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (const std::basic_string<C>& uri, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -// Read from std::istream. -// - -std::[unique|auto]_ptr<type> -name (std::istream&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -std::[unique|auto]_ptr<type> -name (std::istream&, - const std::basic_string<C>& id, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - const std::basic_string<C>& id, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - const std::basic_string<C>& id, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -// Read from InputSource. -// - -std::[unique|auto]_ptr<type> -name (xercesc::InputSource&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (xercesc::InputSource&, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (xercesc::InputSource&, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -// Read from DOM. -// - -std::[unique|auto]_ptr<type> -name (const xercesc::DOMDocument&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (xml_schema::dom::[unique|auto]_ptr<xercesc::DOMDocument>, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - </pre> - - <p>You can choose between reading an XML instance from a local file, - URI, <code>std::istream</code>, <code>xercesc::InputSource</code>, - or a pre-parsed DOM instance in the form of - <code>xercesc::DOMDocument</code>. All the parsing functions - return a dynamically allocated object model as either - <code>std::unique_ptr</code> or <code>std::auto_ptr</code>, - depending on the C++ standard selected. Each of these parsing - functions is discussed in more detail in the following sections. - </p> - - <h2><a name="3.1">3.1 Initializing the Xerces-C++ Runtime</a></h2> - - <p>Some parsing functions expect you to initialize the Xerces-C++ - runtime while others initialize and terminate it as part of their - work. The general rule is as follows: if a function has any arguments - or return a value that is an instance of a Xerces-C++ type, then - this function expects you to initialize the Xerces-C++ runtime. - Otherwise, the function initializes and terminates the runtime for - you. Note that it is legal to have nested calls to the Xerces-C++ - initialize and terminate functions as long as the calls are balanced. - </p> - - <p>You can instruct parsing functions that initialize and terminate - the runtime not to do so by passing the - <code>xml_schema::flags::dont_initialize</code> flag (see - <a href="#3.2">Section 3.2, "Flags and Properties"</a>). - </p> - - - <h2><a name="3.2">3.2 Flags and Properties</a></h2> - - <p>Parsing flags and properties are the last two arguments of every - parsing function. They allow you to fine-tune the process of - instance validation and parsing. Both arguments are optional. - </p> - - - <p>The following flags are recognized by the parsing functions:</p> - - <dl> - <dt><code>xml_schema::flags::keep_dom</code></dt> - <dd>Keep association between DOM nodes and the resulting - object model nodes. For more information about DOM association - refer to <a href="#5.1">Section 5.1, "DOM Association"</a>.</dd> - - <dt><code>xml_schema::flags::own_dom</code></dt> - <dd>Assume ownership of the DOM document passed. This flag only - makes sense together with the <code>keep_dom</code> flag in - the call to the parsing function with the - <code>xml_schema::dom::[unique|auto]_ptr<DOMDocument></code> - argument.</dd> - - <dt><code>xml_schema::flags::dont_validate</code></dt> - <dd>Do not validate instance documents against schemas.</dd> - - <dt><code>xml_schema::flags::dont_initialize</code></dt> - <dd>Do not initialize the Xerces-C++ runtime.</dd> - </dl> - - <p>You can pass several flags by combining them using the bit-wise OR - operator. For example:</p> - - <pre class="c++"> -using xml_schema::flags; - -std::unique_ptr<type> r ( - name ("test.xml", flags::keep_dom | flags::dont_validate)); - </pre> - - <p>By default, validation of instance documents is turned on even - though parsers generated by XSD do not assume instance - documents are valid. They include a number of checks that prevent - construction of inconsistent object models. This, - however, does not mean that an instance document that was - successfully parsed by the XSD-generated parsers is - valid per the corresponding schema. If an instance document is not - "valid enough" for the generated parsers to construct consistent - object model, one of the exceptions defined in - <code>xml_schema</code> namespace is thrown (see - <a href="#3.3">Section 3.3, "Error Handling"</a>). - </p> - - <p>For more information on the Xerces-C++ runtime initialization - refer to <a href="#3.1">Section 3.1, "Initializing the Xerces-C++ - Runtime"</a>. - </p> - - <p>The <code>xml_schema::properties</code> class allows you to - programmatically specify schema locations to be used instead - of those specified with the <code>xsi::schemaLocation</code> - and <code>xsi::noNamespaceSchemaLocation</code> attributes - in instance documents. The interface of the <code>properties</code> - class is presented below: - </p> - - <pre class="c++"> -class properties -{ -public: - void - schema_location (const std::basic_string<C>& namespace_, - const std::basic_string<C>& location); - void - no_namespace_schema_location (const std::basic_string<C>& location); -}; - </pre> - - <p>Note that all locations are relative to an instance document unless - they are URIs. For example, if you want to use a local file as your - schema, then you will need to pass - <code>file:///absolute/path/to/your/schema</code> as the location - argument. - </p> - - <h2><a name="3.3">3.3 Error Handling</a></h2> - - <p>As discussed in <a href="#2.2">Section 2.2, "Error Handling"</a>, - the mapping uses the C++ exception handling mechanism as its primary - way of reporting error conditions. However, to handle recoverable - parsing and validation errors and warnings, a callback interface maybe - preferred by the application.</p> - - <p>To better understand error handling and reporting strategies employed - by the parsing functions, it is useful to know that the - transformation of an XML instance document to a statically-typed - tree happens in two stages. The first stage, performed by Xerces-C++, - consists of parsing an XML document into a DOM instance. For short, - we will call this stage the XML-DOM stage. Validation, if not disabled, - happens during this stage. The second stage, - performed by the generated parsers, consist of parsing the DOM - instance into the statically-typed tree. We will call this stage - the DOM-Tree stage. Additional checks are performed during this - stage in order to prevent construction of inconsistent tree which - could otherwise happen when validation is disabled, for example.</p> - - <p>All parsing functions except the one that operates on a DOM instance - come in overloaded triples. The first function in such a triple - reports error conditions exclusively by throwing exceptions. It - accumulates all the parsing and validation errors of the XML-DOM - stage and throws them in a single instance of the - <code>xml_schema::parsing</code> exception (described below). - The second and the third functions in the triple use callback - interfaces to report parsing and validation errors and warnings. - The two callback interfaces are <code>xml_schema::error_handler</code> - and <code>xercesc::DOMErrorHandler</code>. For more information - on the <code>xercesc::DOMErrorHandler</code> interface refer to - the Xerces-C++ documentation. The <code>xml_schema::error_handler</code> - interface is presented below: - </p> - - <pre class="c++"> -class error_handler -{ -public: - struct severity - { - enum value - { - warning, - error, - fatal - }; - }; - - virtual bool - handle (const std::basic_string<C>& id, - unsigned long line, - unsigned long column, - severity, - const std::basic_string<C>& message) = 0; - - virtual - ~error_handler (); -}; - </pre> - - <p>The <code>id</code> argument of the <code>error_handler::handle</code> - function identifies the resource being parsed (e.g., a file name or - URI). - </p> - - <p>By returning <code>true</code> from the <code>handle</code> function - you instruct the parser to recover and continue parsing. Returning - <code>false</code> results in termination of the parsing process. - An error with the <code>fatal</code> severity level results in - termination of the parsing process no matter what is returned from - the <code>handle</code> function. It is safe to throw an exception - from the <code>handle</code> function. - </p> - - <p>The DOM-Tree stage reports error conditions exclusively by throwing - exceptions. Individual exceptions thrown by the parsing functions - are described in the following sub-sections. - </p> - - - <h3><a name="3.3.1">3.3.1 <code>xml_schema::parsing</code></a></h3> - - <pre class="c++"> -struct severity -{ - enum value - { - warning, - error - }; - - severity (value); - operator value () const; -}; - -struct error -{ - error (severity, - const std::basic_string<C>& id, - unsigned long line, - unsigned long column, - const std::basic_string<C>& message); - - severity - severity () const; - - const std::basic_string<C>& - id () const; - - unsigned long - line () const; - - unsigned long - column () const; - - const std::basic_string<C>& - message () const; -}; - -std::basic_ostream<C>& -operator<< (std::basic_ostream<C>&, const error&); - -struct diagnostics: std::vector<error> -{ -}; - -std::basic_ostream<C>& -operator<< (std::basic_ostream<C>&, const diagnostics&); - -struct parsing: virtual exception -{ - parsing (); - parsing (const diagnostics&); - - const diagnostics& - diagnostics () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::parsing</code> exception is thrown if there - were parsing or validation errors reported during the XML-DOM stage. - If no callback interface was provided to the parsing function, the - exception contains a list of errors and warnings accessible using - the <code>diagnostics</code> function. The usual conditions when - this exception is thrown include malformed XML instances and, if - validation is turned on, invalid instance documents. - </p> - - <h3><a name="3.3.2">3.3.2 <code>xml_schema::expected_element</code></a></h3> - - <pre class="c++"> -struct expected_element: virtual exception -{ - expected_element (const std::basic_string<C>& name, - const std::basic_string<C>& namespace_); - - - const std::basic_string<C>& - name () const; - - const std::basic_string<C>& - namespace_ () const; - - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::expected_element</code> exception is thrown - when an expected element is not encountered by the DOM-Tree stage. - The name and namespace of the expected element can be obtained using - the <code>name</code> and <code>namespace_</code> functions respectively. - </p> - - - <h3><a name="3.3.3">3.3.3 <code>xml_schema::unexpected_element</code></a></h3> - - <pre class="c++"> -struct unexpected_element: virtual exception -{ - unexpected_element (const std::basic_string<C>& encountered_name, - const std::basic_string<C>& encountered_namespace, - const std::basic_string<C>& expected_name, - const std::basic_string<C>& expected_namespace) - - - const std::basic_string<C>& - encountered_name () const; - - const std::basic_string<C>& - encountered_namespace () const; - - - const std::basic_string<C>& - expected_name () const; - - const std::basic_string<C>& - expected_namespace () const; - - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::unexpected_element</code> exception is thrown - when an unexpected element is encountered by the DOM-Tree stage. - The name and namespace of the encountered element can be obtained - using the <code>encountered_name</code> and - <code>encountered_namespace</code> functions respectively. If an - element was expected instead of the encountered one, its name - and namespace can be obtained using the <code>expected_name</code> and - <code>expected_namespace</code> functions respectively. Otherwise - these functions return empty strings. - </p> - - <h3><a name="3.3.4">3.3.4 <code>xml_schema::expected_attribute</code></a></h3> - - <pre class="c++"> -struct expected_attribute: virtual exception -{ - expected_attribute (const std::basic_string<C>& name, - const std::basic_string<C>& namespace_); - - - const std::basic_string<C>& - name () const; - - const std::basic_string<C>& - namespace_ () const; - - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::expected_attribute</code> exception is thrown - when an expected attribute is not encountered by the DOM-Tree stage. - The name and namespace of the expected attribute can be obtained using - the <code>name</code> and <code>namespace_</code> functions respectively. - </p> - - - <h3><a name="3.3.5">3.3.5 <code>xml_schema::unexpected_enumerator</code></a></h3> - - <pre class="c++"> -struct unexpected_enumerator: virtual exception -{ - unexpected_enumerator (const std::basic_string<C>& enumerator); - - const std::basic_string<C>& - enumerator () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::unexpected_enumerator</code> exception is thrown - when an unexpected enumerator is encountered by the DOM-Tree stage. - The enumerator can be obtained using the <code>enumerator</code> - functions. - </p> - - <h3><a name="3.3.6">3.3.6 <code>xml_schema::expected_text_content</code></a></h3> - - <pre class="c++"> -struct expected_text_content: virtual exception -{ - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::expected_text_content</code> exception is thrown - when a content other than text is encountered and the text content was - expected by the DOM-Tree stage. - </p> - - <h3><a name="3.3.7">3.3.7 <code>xml_schema::no_type_info</code></a></h3> - - <pre class="c++"> -struct no_type_info: virtual exception -{ - no_type_info (const std::basic_string<C>& type_name, - const std::basic_string<C>& type_namespace); - - const std::basic_string<C>& - type_name () const; - - const std::basic_string<C>& - type_namespace () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::no_type_info</code> exception is thrown - when there is no type information associated with a type specified - by the <code>xsi:type</code> attribute. This exception is thrown - by the DOM-Tree stage. The name and namespace of the type in question - can be obtained using the <code>type_name</code> and - <code>type_namespace</code> functions respectively. Usually, catching - this exception means that you haven't linked the code generated - from the schema defining the type in question with your application - or this schema has been compiled without the - <code>--generate-polymorphic</code> option. - </p> - - - <h3><a name="3.3.8">3.3.8 <code>xml_schema::not_derived</code></a></h3> - - <pre class="c++"> -struct not_derived: virtual exception -{ - not_derived (const std::basic_string<C>& base_type_name, - const std::basic_string<C>& base_type_namespace, - const std::basic_string<C>& derived_type_name, - const std::basic_string<C>& derived_type_namespace); - - const std::basic_string<C>& - base_type_name () const; - - const std::basic_string<C>& - base_type_namespace () const; - - - const std::basic_string<C>& - derived_type_name () const; - - const std::basic_string<C>& - derived_type_namespace () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::not_derived</code> exception is thrown - when a type specified by the <code>xsi:type</code> attribute is - not derived from the expected base type. This exception is thrown - by the DOM-Tree stage. The name and namespace of the expected - base type can be obtained using the <code>base_type_name</code> and - <code>base_type_namespace</code> functions respectively. The name - and namespace of the offending type can be obtained using the - <code>derived_type_name</code> and - <code>derived_type_namespace</code> functions respectively. - </p> - - <h3><a name="3.3.9">3.3.9 <code>xml_schema::no_prefix_mapping</code></a></h3> - - <pre class="c++"> -struct no_prefix_mapping: virtual exception -{ - no_prefix_mapping (const std::basic_string<C>& prefix); - - const std::basic_string<C>& - prefix () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::no_prefix_mapping</code> exception is thrown - during the DOM-Tree stage if a namespace prefix is encountered for - which a prefix-namespace mapping hasn't been provided. The namespace - prefix in question can be obtained using the <code>prefix</code> - function. - </p> - - <h2><a name="3.4">3.4 Reading from a Local File or URI</a></h2> - - <p>Using a local file or URI is the simplest way to parse an XML instance. - For example:</p> - - <pre class="c++"> -using std::unique_ptr; - -unique_ptr<type> r1 (name ("test.xml")); -unique_ptr<type> r2 (name ("https://www.codesynthesis.com/test.xml")); - </pre> - - <p>Or, in the C++98 mode:</p> - - <pre class="c++"> -using std::auto_ptr; - -auto_ptr<type> r1 (name ("test.xml")); -auto_ptr<type> r2 (name ("https://www.codesynthesis.com/test.xml")); - </pre> - - <h2><a name="3.5">3.5 Reading from <code>std::istream</code></a></h2> - - <p>When using an <code>std::istream</code> instance, you may also - pass an optional resource id. This id is used to identify the - resource (for example in error messages) as well as to resolve - relative paths. For instance:</p> - - <pre class="c++"> -using std::unique_ptr; - -{ - std::ifstream ifs ("test.xml"); - unique_ptr<type> r (name (ifs, "test.xml")); -} - -{ - std::string str ("..."); // Some XML fragment. - std::istringstream iss (str); - unique_ptr<type> r (name (iss)); -} - </pre> - - <h2><a name="3.6">3.6 Reading from <code>xercesc::InputSource</code></a></h2> - - <p>Reading from a <code>xercesc::InputSource</code> instance - is similar to the <code>std::istream</code> case except - the resource id is maintained by the <code>InputSource</code> - object. For instance:</p> - - <pre class="c++"> -xercesc::StdInInputSource is; -std::unique_ptr<type> r (name (is)); - </pre> - - <h2><a name="3.7">3.7 Reading from DOM</a></h2> - - <p>Reading from a <code>xercesc::DOMDocument</code> instance allows - you to setup a custom XML-DOM stage. Things like DOM - parser reuse, schema pre-parsing, and schema caching can be achieved - with this approach. For more information on how to obtain DOM - representation from an XML instance refer to the Xerces-C++ - documentation. In addition, the - <a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree Mapping - FAQ</a> shows how to parse an XML instance to a Xerces-C++ - DOM document using the XSD runtime utilities. - </p> - - <p>The last parsing function is useful when you would like to perform - your own XML-to-DOM parsing and associate the resulting DOM document - with the object model nodes. The automatic <code>DOMDocument</code> - pointer is reset and the resulting object model assumes ownership - of the DOM document passed. For example:</p> - - <pre class="c++"> -// C++11 version. -// -xml_schema::dom::unique_ptr<xercesc::DOMDocument> doc = ... - -std::unique_ptr<type> r ( - name (std::move (doc), - xml_schema::flags::keep_dom | xml_schema::flags::own_dom)); - -// At this point doc is reset to 0. - -// C++98 version. -// -xml_schema::dom::auto_ptr<xercesc::DOMDocument> doc = ... - -std::auto_ptr<type> r ( - name (doc, xml_schema::flags::keep_dom | xml_schema::flags::own_dom)); - -// At this point doc is reset to 0. - </pre> - - <h1><a name="4">4 Serialization</a></h1> - - <p>This chapter covers various aspects of serializing a - tree-like object model to DOM or XML. - In this regard, serialization is complimentary to the reverse - process of parsing a DOM or XML instance into an object model - which is discussed in <a href="#3">Chapter 3, - "Parsing"</a>. Note that the generation of the serialization code - is optional and should be explicitly requested with the - <code>--generate-serialization</code> option. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information. - </p> - - <p>Each global XML Schema element in the form: - </p> - - - <pre class="xml"> -<xsd:element name="name" type="type"/> - </pre> - - <p>is mapped to 8 overloaded C++ functions in the form:</p> - - <pre class="c++"> -// Serialize to std::ostream. -// -void -name (std::ostream&, - const type&, - const xml_schema::namespace_fomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (std::ostream&, - const type&, - xml_schema::error_handler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (std::ostream&, - const type&, - xercesc::DOMErrorHandler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - - -// Serialize to XMLFormatTarget. -// -void -name (xercesc::XMLFormatTarget&, - const type&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (xercesc::XMLFormatTarget&, - const type&, - xml_schema::error_handler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (xercesc::XMLFormatTarget&, - const type&, - xercesc::DOMErrorHandler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - - -// Serialize to DOM. -// -xml_schema::dom::[unique|auto]_ptr<xercesc::DOMDocument> -name (const type&, - const xml_schema::namespace_infomap& - xml_schema::namespace_infomap (), - xml_schema::flags = 0); - -void -name (xercesc::DOMDocument&, - const type&, - xml_schema::flags = 0); - </pre> - - <p>You can choose between writing XML to <code>std::ostream</code> or - <code>xercesc::XMLFormatTarget</code> and creating a DOM instance - in the form of <code>xercesc::DOMDocument</code>. Serialization - to <code>ostream</code> or <code>XMLFormatTarget</code> requires a - considerably less work while serialization to DOM provides - for greater flexibility. Each of these serialization functions - is discussed in more detail in the following sections. - </p> - - - <h2><a name="4.1">4.1 Initializing the Xerces-C++ Runtime</a></h2> - - <p>Some serialization functions expect you to initialize the Xerces-C++ - runtime while others initialize and terminate it as part of their - work. The general rule is as follows: if a function has any arguments - or return a value that is an instance of a Xerces-C++ type, then - this function expects you to initialize the Xerces-C++ runtime. - Otherwise, the function initializes and terminates the runtime for - you. Note that it is legal to have nested calls to the Xerces-C++ - initialize and terminate functions as long as the calls are balanced. - </p> - - <p>You can instruct serialization functions that initialize and terminate - the runtime not to do so by passing the - <code>xml_schema::flags::dont_initialize</code> flag (see - <a href="#4.3">Section 4.3, "Flags"</a>). - </p> - - <h2><a name="4.2">4.2 Namespace Infomap and Character Encoding</a></h2> - - <p>When a document being serialized uses XML namespaces, custom - prefix-namespace associations can to be established. If custom - prefix-namespace mapping is not provided then generic prefixes - (<code>p1</code>, <code>p2</code>, etc) are automatically assigned - to namespaces as needed. Also, if - you would like the resulting instance document to contain the - <code>schemaLocation</code> or <code>noNamespaceSchemaLocation</code> - attributes, you will need to provide namespace-schema associations. - The <code>xml_schema::namespace_infomap</code> class is used - to capture this information:</p> - - <pre class="c++"> -struct namespace_info -{ - namespace_info (); - namespace_info (const std::basic_string<C>& name, - const std::basic_string<C>& schema); - - std::basic_string<C> name; - std::basic_string<C> schema; -}; - -// Map of namespace prefix to namespace_info. -// -struct namespace_infomap: public std::map<std::basic_string<C>, - namespace_info> -{ -}; - </pre> - - <p>Consider the following associations as an example:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - </pre> - - <p>This map, if passed to one of the serialization functions, - could result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<t:name xmlns:t="https://www.codesynthesis.com/test" - xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:schemaLocation="https://www.codesynthesis.com/test test.xsd"> - </pre> - - <p>As you can see, the serialization function automatically added namespace - mapping for the <code>xsi</code> prefix. You can change this by - providing your own prefix:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map["xsn"].name = "http://www.w3.org/2001/XMLSchema-instance"; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - </pre> - - <p>This could result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<t:name xmlns:t="https://www.codesynthesis.com/test" - xmlns:xsn="http://www.w3.org/2001/XMLSchema-instance" - xsn:schemaLocation="https://www.codesynthesis.com/test test.xsd"> - </pre> - - <p>To specify the location of a schema without a namespace you can use - an empty prefix as in the example below: </p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].schema = "test.xsd"; - </pre> - - <p>This would result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<name xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="test.xsd"> - </pre> - - <p>To make a particular namespace default you can use an empty - prefix, for example:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].name = "https://www.codesynthesis.com/test"; -map[""].schema = "test.xsd"; - </pre> - - <p>This could result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<name xmlns="https://www.codesynthesis.com/test" - xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:schemaLocation="https://www.codesynthesis.com/test test.xsd"> - </pre> - - - <p>Another bit of information that you can pass to the serialization - functions is the character encoding method that you would like to use. - Common values for this argument are <code>"US-ASCII"</code>, - <code>"ISO8859-1"</code>, <code>"UTF-8"</code>, - <code>"UTF-16BE"</code>, <code>"UTF-16LE"</code>, - <code>"UCS-4BE"</code>, and <code>"UCS-4LE"</code>. The default - encoding is <code>"UTF-8"</code>. For more information on - encoding methods see the - "<a href="http://en.wikipedia.org/wiki/Character_code">Character - Encoding</a>" article from Wikipedia. - </p> - - <h2><a name="4.3">4.3 Flags</a></h2> - - <p>Serialization flags are the last argument of every serialization - function. They allow you to fine-tune the process of serialization. - The flags argument is optional. - </p> - - - <p>The following flags are recognized by the serialization - functions:</p> - - <dl> - <dt><code>xml_schema::flags::dont_initialize</code></dt> - <dd>Do not initialize the Xerces-C++ runtime.</dd> - - <dt><code>xml_schema::flags::dont_pretty_print</code></dt> - <dd>Do not add extra spaces or new lines that make the resulting XML - slightly bigger but easier to read.</dd> - - <dt><code>xml_schema::flags::no_xml_declaration</code></dt> - <dd>Do not write XML declaration (<?xml ... ?>).</dd> - </dl> - - <p>You can pass several flags by combining them using the bit-wise OR - operator. For example:</p> - - <pre class="c++"> -std::unique_ptr<type> r = ... -std::ofstream ofs ("test.xml"); -xml_schema::namespace_infomap map; -name (ofs, - *r, - map, - "UTF-8", - xml_schema::flags::no_xml_declaration | - xml_schema::flags::dont_pretty_print); - </pre> - - <p>For more information on the Xerces-C++ runtime initialization - refer to <a href="#4.1">Section 4.1, "Initializing the Xerces-C++ - Runtime"</a>. - </p> - - <h2><a name="4.4">4.4 Error Handling</a></h2> - - <p>As with the parsing functions (see <a href="#3.3">Section 3.3, - "Error Handling"</a>), to better understand error handling and - reporting strategies employed by the serialization functions, it - is useful to know that the transformation of a statically-typed - tree to an XML instance document happens in two stages. The first - stage, performed by the generated code, consist of building a DOM - instance from the statically-typed tree . For short, we will call - this stage the Tree-DOM stage. The second stage, performed by - Xerces-C++, consists of serializing the DOM instance into the XML - document. We will call this stage the DOM-XML stage. - </p> - - <p>All serialization functions except the two that serialize into - a DOM instance come in overloaded triples. The first function - in such a triple reports error conditions exclusively by throwing - exceptions. It accumulates all the serialization errors of the - DOM-XML stage and throws them in a single instance of the - <code>xml_schema::serialization</code> exception (described below). - The second and the third functions in the triple use callback - interfaces to report serialization errors and warnings. The two - callback interfaces are <code>xml_schema::error_handler</code> and - <code>xercesc::DOMErrorHandler</code>. The - <code>xml_schema::error_handler</code> interface is described in - <a href="#3.3">Section 3.3, "Error Handling"</a>. For more information - on the <code>xercesc::DOMErrorHandler</code> interface refer to the - Xerces-C++ documentation. - </p> - - <p>The Tree-DOM stage reports error conditions exclusively by throwing - exceptions. Individual exceptions thrown by the serialization functions - are described in the following sub-sections. - </p> - - <h3><a name="4.4.1">4.4.1 <code>xml_schema::serialization</code></a></h3> - - <pre class="c++"> -struct serialization: virtual exception -{ - serialization (); - serialization (const diagnostics&); - - const diagnostics& - diagnostics () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::diagnostics</code> class is described in - <a href="#3.3.1">Section 3.3.1, "<code>xml_schema::parsing</code>"</a>. - The <code>xml_schema::serialization</code> exception is thrown if - there were serialization errors reported during the DOM-XML stage. - If no callback interface was provided to the serialization function, - the exception contains a list of errors and warnings accessible using - the <code>diagnostics</code> function. - </p> - - - <h3><a name="4.4.2">4.4.2 <code>xml_schema::unexpected_element</code></a></h3> - - <p>The <code>xml_schema::unexpected_element</code> exception is - described in <a href="#3.3.3">Section 3.3.3, - "<code>xml_schema::unexpected_element</code>"</a>. It is thrown - by the serialization functions during the Tree-DOM stage if the - root element name of the provided DOM instance does not match with - the name of the element this serialization function is for. - </p> - - <h3><a name="4.4.3">4.4.3 <code>xml_schema::no_type_info</code></a></h3> - - <p>The <code>xml_schema::no_type_info</code> exception is - described in <a href="#3.3.7">Section 3.3.7, - "<code>xml_schema::no_type_info</code>"</a>. It is thrown - by the serialization functions during the Tree-DOM stage when there - is no type information associated with a dynamic type of an - element. Usually, catching this exception means that you haven't - linked the code generated from the schema defining the type in - question with your application or this schema has been compiled - without the <code>--generate-polymorphic</code> option. - </p> - - <h2><a name="4.5">4.5 Serializing to <code>std::ostream</code></a></h2> - - <p>In order to serialize to <code>std::ostream</code> you will need - an object model, an output stream and, optionally, a namespace - infomap. For instance:</p> - - <pre class="c++"> -// Obtain the object model. -// -std::unique_ptr<type> r = ... - -// Prepare namespace mapping and schema location information. -// -xml_schema::namespace_infomap map; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - -// Write it out. -// -name (std::cout, *r, map); - </pre> - - <p>Note that the output stream is treated as a binary stream. This - becomes important when you use a character encoding that is wider - than 8-bit <code>char</code>, for instance UTF-16 or UCS-4. For - example, things will most likely break if you try to serialize - to <code>std::ostringstream</code> with UTF-16 or UCS-4 as an - encoding. This is due to the special value, - <code>'\0'</code>, that will most likely occur as part of such - serialization and it won't have the special meaning assumed by - <code>std::ostringstream</code>. - </p> - - - <h2><a name="4.6">4.6 Serializing to <code>xercesc::XMLFormatTarget</code></a></h2> - - <p>Serializing to an <code>xercesc::XMLFormatTarget</code> instance - is similar the <code>std::ostream</code> case. For instance: - </p> - - <pre class="c++"> -using std::unique_ptr; - -// Obtain the object model. -// -unique_ptr<type> r = ... - -// Prepare namespace mapping and schema location information. -// -xml_schema::namespace_infomap map; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Choose a target. - // - unique_ptr<XMLFormatTarget> ft; - - if (argc != 2) - { - ft = unique_ptr<XMLFormatTarget> (new StdOutFormatTarget ()); - } - else - { - ft = unique_ptr<XMLFormatTarget> ( - new LocalFileFormatTarget (argv[1])); - } - - // Write it out. - // - name (*ft, *r, map); -} - -XMLPlatformUtils::Terminate (); - </pre> - - <p>Note that we had to initialize the Xerces-C++ runtime before we - could call this serialization function.</p> - - <h2><a name="4.7">4.7 Serializing to DOM</a></h2> - - <p>The mapping provides two overloaded functions that implement - serialization to a DOM instance. The first creates a DOM instance - for you and the second serializes to an existing DOM instance. - While serializing to a new DOM instance is similar to serializing - to <code>std::ostream</code> or <code>xercesc::XMLFormatTarget</code>, - serializing to an existing DOM instance requires quite a bit of work - from your side. You will need to set all the custom namespace mapping - attributes as well as the <code>schemaLocation</code> and/or - <code>noNamespaceSchemaLocation</code> attributes. The following - listing should give you an idea about what needs to be done: - </p> - - <pre class="c++"> -// Obtain the object model. -// -std::unique_ptr<type> r = ... - -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Create a DOM instance. Set custom namespace mapping and schema - // location attributes. - // - DOMDocument& doc = ... - - // Serialize to DOM. - // - name (doc, *r); - - // Serialize the DOM document to XML. - // - ... -} - -XMLPlatformUtils::Terminate (); - </pre> - - <p>For more information on how to create and serialize a DOM instance - refer to the Xerces-C++ documentation. In addition, the - <a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree Mapping - FAQ</a> shows how to implement these operations using the XSD - runtime utilities. - </p> - - <h1><a name="5">5 Additional Functionality</a></h1> - - <p>The C++/Tree mapping provides a number of optional features - that can be useful in certain situations. They are described - in the following sections.</p> - - <h2><a name="5.1">5.1 DOM Association</a></h2> - - <p>Normally, after parsing is complete, the DOM document which - was used to extract the data is discarded. However, the parsing - functions can be instructed to preserve the DOM document - and create an association between the DOM nodes and object model - nodes. When there is an association between the DOM and - object model nodes, you can obtain the corresponding DOM element - or attribute node from an object model node as well as perform - the reverse transition: obtain the corresponding object model - from a DOM element or attribute node.</p> - - <p>Maintaining DOM association is normally useful when the application - needs access to XML constructs that are not preserved in the - object model, for example, XML comments. - Another useful aspect of DOM association is the ability of the - application to navigate the document tree using the generic DOM - interface (for example, with the help of an XPath processor) - and then move back to the statically-typed object model. Note - also that while you can change the underlying DOM document, - these changes are not reflected in the object model and will - be ignored during serialization. If you need to not only access - but also modify some aspects of XML that are not preserved in - the object model, then type customization with custom parsing - constructors and serialization operators should be used instead.</p> - - <p>To request DOM association you will need to pass the - <code>xml_schema::flags::keep_dom</code> flag to one of the - parsing functions (see <a href="#3.2">Section 3.2, - "Flags and Properties"</a> for more information). In this case the - DOM document is retained and will be released when the object model - is deleted. Note that since DOM nodes "out-live" the parsing function - call, you need to initialize the Xerces-C++ runtime before calling - one of the parsing functions with the <code>keep_dom</code> flag and - terminate it after the object model is destroyed (see - <a href="#3.1">Section 3.1, "Initializing the Xerces-C++ Runtime"</a>).</p> - - <p>If the <code>keep_dom</code> flag is passed - as the second argument to the copy constructor and the copy - being made is of a complete tree, then the DOM association - is also maintained in the copy by cloning the underlying - DOM document and reestablishing the associations. For example:</p> - - <pre class="c++"> -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Parse XML to object model. - // - std::unique_ptr<type> r (root ( - "root.xml", - xml_schema::flags::keep_dom | - xml_schema::flags::dont_initialize)); - - // Copy without DOM association. - // - type copy1 (*r); - - // Copy with DOM association. - // - type copy2 (*r, xml_schema::flags::keep_dom); -} - -XMLPlatformUtils::Terminate (); - </pre> - - - <p>To obtain the corresponding DOM node from an object model node - you will need to call the <code>_node</code> accessor function - which returns a pointer to <code>DOMNode</code>. You can then query - this DOM node's type and cast it to either <code>DOMAttr*</code> - or <code>DOMElement*</code>. To obtain the corresponding object - model node from a DOM node, the DOM user data API is used. The - <code>xml_schema::dom::tree_node_key</code> variable contains - the key for object model nodes. The following schema and code - fragment show how to navigate from DOM to object model nodes - and in the opposite direction:</p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="a" type="string"/> - </sequence> -</complexType> - -<element name="root" type="object"/> - </pre> - - <pre class="c++"> -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Parse XML to object model. - // - std::unique_ptr<type> r (root ( - "root.xml", - xml_schema::flags::keep_dom | - xml_schema::flags::dont_initialize)); - - DOMNode* n = r->_node (); - assert (n->getNodeType () == DOMNode::ELEMENT_NODE); - DOMElement* re = static_cast<DOMElement*> (n); - - // Get the 'a' element. Note that it is not necessarily the - // first child node of 'root' since there could be whitespace - // nodes before it. - // - DOMElement* ae; - - for (n = re->getFirstChild (); n != 0; n = n->getNextSibling ()) - { - if (n->getNodeType () == DOMNode::ELEMENT_NODE) - { - ae = static_cast<DOMElement*> (n); - break; - } - } - - // Get from the 'a' DOM element to xml_schema::string object model - // node. - // - xml_schema::type& t ( - *reinterpret_cast<xml_schema::type*> ( - ae->getUserData (xml_schema::dom::tree_node_key))); - - xml_schema::string& a (dynamic_cast<xml_schema::string&> (t)); -} - -XMLPlatformUtils::Terminate (); - </pre> - - <p>The 'mixed' example which can be found in the XSD distribution - shows how to handle the mixed content using DOM association.</p> - - <h2><a name="5.2">5.2 Binary Serialization</a></h2> - - <p>Besides reading from and writing to XML, the C++/Tree mapping - also allows you to save the object model to and load it from a - number of predefined as well as custom data representation - formats. The predefined binary formats are CDR (Common Data - Representation) and XDR (eXternal Data Representation). A - custom format can easily be supported by providing - insertion and extraction operators for basic types.</p> - - <p>Binary serialization saves only the data without any meta - information or markup. As a result, saving to and loading - from a binary representation can be an order of magnitude - faster than parsing and serializing the same data in XML. - Furthermore, the resulting representation is normally several - times smaller than the equivalent XML representation. These - properties make binary serialization ideal for internal data - exchange and storage. A typical application that uses this - facility stores the data and communicates within the - system using a binary format and reads/writes the data - in XML when communicating with the outside world.</p> - - <p>In order to request the generation of insertion operators and - extraction constructors for a specific predefined or custom - data representation stream, you will need to use the - <code>--generate-insertion</code> and <code>--generate-extraction</code> - compiler options. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information.</p> - - <p>Once the insertion operators and extraction constructors are - generated, you can use the <code>xml_schema::istream</code> - and <code>xml_schema::ostream</code> wrapper stream templates - to save the object model to and load it from a specific format. - The following code fragment shows how to do this using ACE - (Adaptive Communication Environment) CDR streams as an example:</p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="a" type="string"/> - <element name="b" type="int"/> - </sequence> -</complexType> - -<element name="root" type="object"/> - </pre> - - <pre class="c++"> -// Parse XML to object model. -// -std::unique_ptr<type> r (root ("root.xml")); - -// Save to a CDR stream. -// -ACE_OutputCDR ace_ocdr; -xml_schema::ostream<ACE_OutputCDR> ocdr (ace_ocdr); - -ocdr << *r; - -// Load from a CDR stream. -// -ACE_InputCDR ace_icdr (buf, size); -xml_schema::istream<ACE_InputCDR> icdr (ace_icdr); - -std::unique_ptr<object> copy (new object (icdr)); - -// Serialize to XML. -// -root (std::cout, *copy); - </pre> - - <p>The XSD distribution contains a number of examples that - show how to save the object model to and load it from - CDR, XDR, and a custom format.</p> - - <!-- Appendix A --> - - - <h1><a name="A">Appendix A — Default and Fixed Values</a></h1> - - <p>The following table summarizes the effect of default and fixed - values (specified with the <code>default</code> and <code>fixed</code> - attributes, respectively) on attribute and element values. The - <code>default</code> and <code>fixed</code> attributes are mutually - exclusive. It is also worthwhile to note that the fixed value semantics - is a superset of the default value semantics. - </p> - - <!-- border="1" is necessary for html2ps --> - <table id="default-fixed" border="1"> - <tr> - <th></th> - <th></th> - <th colspan="2">default</th> - <th colspan="2">fixed</th> - </tr> - - <!-- element --> - - <tr> - <th rowspan="4">element</th> - <th rowspan="2">not present</th> - <th>optional</th> - <th>required</th> - <th>optional</th> - <th>required</th> - </tr> - <tr> - <td>not present</td> - <td>invalid instance</td> - <td>not present</td> - <td>invalid instance</td> - </tr> - - - <tr> - <th>empty</th> - <td colspan="2">default value is used</td> - <td colspan="2">fixed value is used</td> - </tr> - - <tr> - <th>value</th> - <td colspan="2">value is used</td> - <td colspan="2">value is used provided it's the same as fixed</td> - </tr> - - <!-- attribute --> - - <!-- element --> - - <tr> - <th rowspan="4">attribute</th> - <th rowspan="2">not present</th> - <th>optional</th> - <th>required</th> - <th>optional</th> - <th>required</th> - </tr> - <tr> - <td>default value is used</td> - <td>invalid schema</td> - <td>fixed value is used</td> - <td>invalid instance</td> - </tr> - - - <tr> - <th>empty</th> - <td colspan="2">empty value is used</td> - <td colspan="2">empty value is used provided it's the same as fixed</td> - </tr> - - <tr> - <th>value</th> - <td colspan="2">value is used</td> - <td colspan="2">value is used provided it's the same as fixed</td> - </tr> - - </table> - - </div> -</div> - - -</body> -</html> diff --git a/doc/cxx/tree/manual/index.xhtml.in b/doc/cxx/tree/manual/index.xhtml.in deleted file mode 100644 index 5a7240a..0000000 --- a/doc/cxx/tree/manual/index.xhtml.in +++ /dev/null @@ -1,6826 +0,0 @@ -<?xml version="1.0" encoding="iso-8859-1"?> -<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> -<html xmlns="http://www.w3.org/1999/xhtml" lang="en" xml:lang="en"> - -<head> - <title>C++/Tree Mapping User Manual</title> - - <meta name="copyright" content="© @copyright@"/> - <meta name="keywords" content="xsd,xml,schema,c++,mapping,data,binding,tree,serialization,guide,manual,examples"/> - <meta name="description" content="C++/Tree Mapping User Manual"/> - <meta name="revision" content="4.1.0"/> - - <link rel="stylesheet" type="text/css" href="../../../default.css" /> - -<style type="text/css"> - pre { - padding : 0 0 0 0em; - margin : 0em 0em 0em 0; - - font-size : 102% - } - - body { - min-width: 48em; - } - - h1 { - font-weight: bold; - font-size: 200%; - } - - h2 { - font-weight : bold; - font-size : 150%; - - padding-top : 0.8em; - } - - h3 { - font-size : 130%; - padding-top : 0.8em; - } - - /* Adjust indentation for three levels. */ - #container { - max-width: 48em; - } - - #content { - padding: 0 0.1em 0 4em; - /*background-color: red;*/ - } - - #content h1 { - margin-left: -2.06em; - } - - #content h2 { - margin-left: -1.33em; - } - - /* Title page */ - - #titlepage { - padding: 2em 0 1em 0; - border-bottom: 1px solid black; - } - - #titlepage #title { - font-weight: bold; - font-size: 200%; - text-align: center; - padding: 1em 0 2em 0; - } - - /* Lists */ - ul.list li { - padding-top : 0.3em; - padding-bottom : 0.3em; - } - - - /* Built-in table */ - #builtin { - margin: 2em 0 2em 0; - - border-collapse : collapse; - border : 1px solid; - border-color : #000000; - - font-size : 11px; - line-height : 14px; - } - - #builtin th, #builtin td { - border: 1px solid; - padding : 0.9em 0.9em 0.7em 0.9em; - } - - #builtin th { - background : #cde8f6; - } - - #builtin td { - text-align: left; - } - - - /* default-fixed */ - #default-fixed { - margin: 2em 0 2em 0; - - border-collapse : collapse; - border : 1px solid; - border-color : #000000; - - font-size : 11px; - line-height : 14px; - } - - #default-fixed th, #default-fixed td { - border: 1px solid; - padding : 0.9em 0.9em 0.7em 0.9em; - } - - #default-fixed th { - background : #cde8f6; - } - - #default-fixed td { - text-align: center; - } - - - /* */ - dl dt { - padding : 0.8em 0 0 0; - } - - - /* TOC */ - table.toc { - border-style : none; - border-collapse : separate; - border-spacing : 0; - - margin : 0.2em 0 0.2em 0; - padding : 0 0 0 0; - } - - table.toc tr { - padding : 0 0 0 0; - margin : 0 0 0 0; - } - - table.toc * td, table.toc * th { - border-style : none; - margin : 0 0 0 0; - vertical-align : top; - } - - table.toc * th { - font-weight : normal; - padding : 0em 0.1em 0em 0; - text-align : left; - white-space : nowrap; - } - - table.toc * table.toc th { - padding-left : 1em; - } - - table.toc * td { - padding : 0em 0 0em 0.7em; - text-align : left; - } -</style> - - -</head> - -<body> -<div id="container"> - <div id="content"> - - <div class="noprint"> - - <div id="titlepage"> - <div id="title">C++/Tree Mapping User Manual</div> - - <p>Copyright © @copyright@.</p> - - <p>Permission is granted to copy, distribute and/or modify this - document under the terms of the - <a href="https://www.codesynthesis.com/licenses/fdl-1.2.txt">GNU Free - Documentation License, version 1.2</a>; with no Invariant Sections, - no Front-Cover Texts and no Back-Cover Texts. - </p> - - <p>This document is available in the following formats: - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/index.xhtml">XHTML</a>, - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.pdf">PDF</a>, and - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.ps">PostScript</a>.</p> - </div> - - <h1>Table of Contents</h1> - - <table class="toc"> - <tr> - <th></th><td><a href="#0">Preface</a> - <table class="toc"> - <tr><th></th><td><a href="#0.1">About This Document</a></td></tr> - <tr><th></th><td><a href="#0.2">More Information</a></td></tr> - </table> - </td> - </tr> - - <tr> - <th>1</th><td><a href="#1">Introduction</a></td> - </tr> - - <tr> - <th>2</th><td><a href="#2">C++/Tree Mapping</a> - <table class="toc"> - <tr> - <th>2.1</th><td><a href="#2.1">Preliminary Information</a> - <table class="toc"> - <tr><th>2.1.1</th><td><a href="#2.1.1">C++ Standard</a></td></tr> - <tr><th>2.1.2</th><td><a href="#2.1.2">Identifiers</a></td></tr> - <tr><th>2.1.3</th><td><a href="#2.1.3">Character Type and Encoding</a></td></tr> - <tr><th>2.1.4</th><td><a href="#2.1.4">XML Schema Namespace</a></td></tr> - <tr><th>2.1.5</th><td><a href="#2.1.5">Anonymous Types</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.2</th><td><a href="#2.2">Error Handling</a> - <table class="toc"> - <tr><th>2.2.1</th><td><a href="#2.2.1"><code>xml_schema::duplicate_id</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.3</th><td><a href="#2.3">Mapping for <code>import</code> and <code>include</code></a> - <table class="toc"> - <tr><th>2.3.1</th><td><a href="#2.3.1">Import</a></td></tr> - <tr><th>2.3.2</th><td><a href="#2.3.2">Inclusion with Target Namespace</a></td></tr> - <tr><th>2.3.3</th><td><a href="#2.3.3">Inclusion without Target Namespace</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.4</th><td><a href="#2.4">Mapping for Namespaces</a></td> - </tr> - <tr> - <th>2.5</th><td><a href="#2.5">Mapping for Built-in Data Types</a> - <table class="toc"> - <tr><th>2.5.1</th><td><a href="#2.5.1">Inheritance from Built-in Data Types</a></td></tr> - <tr><th>2.5.2</th><td><a href="#2.5.2">Mapping for <code>anyType</code></a></td></tr> - <tr><th>2.5.3</th><td><a href="#2.5.3">Mapping for <code>anySimpleType</code></a></td></tr> - <tr><th>2.5.4</th><td><a href="#2.5.4">Mapping for <code>QName</code></a></td></tr> - <tr><th>2.5.5</th><td><a href="#2.5.5">Mapping for <code>IDREF</code></a></td></tr> - <tr><th>2.5.6</th><td><a href="#2.5.6">Mapping for <code>base64Binary</code> and <code>hexBinary</code></a></td></tr> - <tr><th>2.5.7</th><td><a href="#2.5.7">Time Zone Representation</a></td></tr> - <tr><th>2.5.8</th><td><a href="#2.5.8">Mapping for <code>date</code></a></td></tr> - <tr><th>2.5.9</th><td><a href="#2.5.9">Mapping for <code>dateTime</code></a></td></tr> - <tr><th>2.5.10</th><td><a href="#2.5.10">Mapping for <code>duration</code></a></td></tr> - <tr><th>2.5.11</th><td><a href="#2.5.11">Mapping for <code>gDay</code></a></td></tr> - <tr><th>2.5.12</th><td><a href="#2.5.12">Mapping for <code>gMonth</code></a></td></tr> - <tr><th>2.5.13</th><td><a href="#2.5.13">Mapping for <code>gMonthDay</code></a></td></tr> - <tr><th>2.5.14</th><td><a href="#2.5.14">Mapping for <code>gYear</code></a></td></tr> - <tr><th>2.5.15</th><td><a href="#2.5.15">Mapping for <code>gYearMonth</code></a></td></tr> - <tr><th>2.5.16</th><td><a href="#2.5.16">Mapping for <code>time</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.6</th><td><a href="#2.6">Mapping for Simple Types</a> - <table class="toc"> - <tr><th>2.6.1</th><td><a href="#2.6.1">Mapping for Derivation by Restriction</a></td></tr> - <tr><th>2.6.2</th><td><a href="#2.6.2">Mapping for Enumerations</a></td></tr> - <tr><th>2.6.3</th><td><a href="#2.6.3">Mapping for Derivation by List</a></td></tr> - <tr><th>2.6.4</th><td><a href="#2.6.4">Mapping for Derivation by Union</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.7</th><td><a href="#2.7">Mapping for Complex Types</a> - <table class="toc"> - <tr><th>2.7.1</th><td><a href="#2.7.1">Mapping for Derivation by Extension</a></td></tr> - <tr><th>2.7.2</th><td><a href="#2.7.2">Mapping for Derivation by Restriction</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.8</th><td><a href="#2.8">Mapping for Local Elements and Attributes</a> - <table class="toc"> - <tr><th>2.8.1</th><td><a href="#2.8.1">Mapping for Members with the One Cardinality Class</a></td></tr> - <tr><th>2.8.2</th><td><a href="#2.8.2">Mapping for Members with the Optional Cardinality Class</a></td></tr> - <tr><th>2.8.3</th><td><a href="#2.8.3">Mapping for Members with the Sequence Cardinality Class</a></td></tr> - <tr><th>2.8.4</th><td><a href="#2.8.4">Element Order</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.9</th><td><a href="#2.9">Mapping for Global Elements</a> - <table class="toc"> - <tr><th>2.9.1</th><td><a href="#2.9.1">Element Types</a></td></tr> - <tr><th>2.9.2</th><td><a href="#2.9.2">Element Map</a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.10</th><td><a href="#2.10">Mapping for Global Attributes</a></td> - </tr> - <tr> - <th>2.11</th><td><a href="#2.11">Mapping for <code>xsi:type</code> and Substitution Groups</a></td> - </tr> - <tr> - <th>2.12</th><td><a href="#2.12">Mapping for <code>any</code> and <code>anyAttribute</code></a> - <table class="toc"> - <tr><th>2.12.1</th><td><a href="#2.12.1">Mapping for <code>any</code> with the One Cardinality Class</a></td></tr> - <tr><th>2.12.2</th><td><a href="#2.12.2">Mapping for <code>any</code> with the Optional Cardinality Class</a></td></tr> - <tr><th>2.12.3</th><td><a href="#2.12.3">Mapping for <code>any</code> with the Sequence Cardinality Class</a></td></tr> - <tr><th>2.12.4</th><td><a href="#2.12.4">Element Wildcard Order</a></td></tr> - <tr><th>2.12.5</th><td><a href="#2.12.5">Mapping for <code>anyAttribute</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>2.13</th><td><a href="#2.13">Mapping for Mixed Content Models</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th>3</th><td><a href="#3">Parsing</a> - <table class="toc"> - <tr> - <th>3.1</th><td><a href="#3.1">Initializing the Xerces-C++ Runtime</a></td> - </tr> - <tr> - <th>3.2</th><td><a href="#3.2">Flags and Properties</a></td> - </tr> - <tr> - <th>3.3</th><td><a href="#3.3">Error Handling</a> - <table class="toc"> - <tr><th>3.3.1</th><td><a href="#3.3.1"><code>xml_schema::parsing</code></a></td></tr> - <tr><th>3.3.2</th><td><a href="#3.3.2"><code>xml_schema::expected_element</code></a></td></tr> - <tr><th>3.3.3</th><td><a href="#3.3.3"><code>xml_schema::unexpected_element</code></a></td></tr> - <tr><th>3.3.4</th><td><a href="#3.3.4"><code>xml_schema::expected_attribute</code></a></td></tr> - <tr><th>3.3.5</th><td><a href="#3.3.5"><code>xml_schema::unexpected_enumerator</code></a></td></tr> - <tr><th>3.3.6</th><td><a href="#3.3.6"><code>xml_schema::expected_text_content</code></a></td></tr> - <tr><th>3.3.7</th><td><a href="#3.3.7"><code>xml_schema::no_type_info</code></a></td></tr> - <tr><th>3.3.8</th><td><a href="#3.3.8"><code>xml_schema::not_derived</code></a></td></tr> - <tr><th>3.3.9</th><td><a href="#3.3.9"><code>xml_schema::not_prefix_mapping</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>3.4</th><td><a href="#3.4">Reading from a Local File or URI</a></td> - </tr> - <tr> - <th>3.5</th><td><a href="#3.5">Reading from <code>std::istream</code></a></td> - </tr> - <tr> - <th>3.6</th><td><a href="#3.6">Reading from <code>xercesc::InputSource</code></a></td> - </tr> - <tr> - <th>3.7</th><td><a href="#3.7">Reading from DOM</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th>4</th><td><a href="#4">Serialization</a> - <table class="toc"> - <tr> - <th>4.1</th><td><a href="#4.1">Initializing the Xerces-C++ Runtime</a></td> - </tr> - <tr> - <th>4.2</th><td><a href="#4.2">Namespace Infomap and Character Encoding</a></td> - </tr> - <tr> - <th>4.3</th><td><a href="#4.3">Flags</a></td> - </tr> - <tr> - <th>4.4</th><td><a href="#4.4">Error Handling</a> - <table class="toc"> - <tr><th>4.4.1</th><td><a href="#4.4.1"><code>xml_schema::serialization</code></a></td></tr> - <tr><th>4.4.2</th><td><a href="#4.4.2"><code>xml_schema::unexpected_element</code></a></td></tr> - <tr><th>4.4.3</th><td><a href="#4.4.3"><code>xml_schema::no_type_info</code></a></td></tr> - </table> - </td> - </tr> - <tr> - <th>4.5</th><td><a href="#4.5">Serializing to <code>std::ostream</code></a></td> - </tr> - <tr> - <th>4.6</th><td><a href="#4.6">Serializing to <code>xercesc::XMLFormatTarget</code></a></td> - </tr> - <tr> - <th>4.7</th><td><a href="#4.7">Serializing to DOM</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th>5</th><td><a href="#5">Additional Functionality</a> - <table class="toc"> - <tr> - <th>5.1</th><td><a href="#5.1">DOM Association</a></td> - </tr> - <tr> - <th>5.2</th><td><a href="#5.2">Binary Serialization</a></td> - </tr> - </table> - </td> - </tr> - - <tr> - <th></th><td><a href="#A">Appendix A — Default and Fixed Values</a></td> - </tr> - - </table> - </div> - - <h1><a name="0">Preface</a></h1> - - <h2><a name="0.1">About This Document</a></h2> - - <p>This document describes the mapping of W3C XML Schema - to the C++ programming language as implemented by - <a href="https://www.codesynthesis.com/products/xsd">CodeSynthesis - XSD</a> - an XML Schema to C++ data binding compiler. The mapping - represents information stored in XML instance documents as a - statically-typed, tree-like in-memory data structure and is - called C++/Tree. - </p> - - <p>Revision 4.1.0<br/> <!-- Remember to change revision in other places --> - This revision of the manual describes the C++/Tree - mapping as implemented by CodeSynthesis XSD version 4.1.0. - </p> - - <p>This document is available in the following formats: - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/index.xhtml">XHTML</a>, - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.pdf">PDF</a>, and - <a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.ps">PostScript</a>.</p> - - <h2><a name="0.2">More Information</a></h2> - - <p>Beyond this manual, you may also find the following sources of - information useful:</p> - - <ul class="list"> - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/guide/">C++/Tree - Mapping Getting Started Guide</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/Customization_guide">C++/Tree - Mapping Customization Guide</a></li> - - <li><a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree - Mapping Frequently Asked Questions (FAQ)</a></li> - - <li><a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a></li> - - <li>The <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - contains a collection of examples and a README file with an overview - of each example.</li> - - <li>The <code>README</code> file in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> package - explains how to build the examples.</li> - - <li>The <a href="https://www.codesynthesis.com/mailman/listinfo/xsd-users">xsd-users</a> - mailing list is a place to ask questions. Furthermore the - <a href="https://www.codesynthesis.com/pipermail/xsd-users/">archives</a> - may already have answers to some of your questions.</li> - </ul> - - - <h1><a name="1">1 Introduction</a></h1> - - <p>C++/Tree is a W3C XML Schema to C++ mapping that represents the - data stored in XML as a statically-typed, vocabulary-specific - object model. Based on a formal description of an XML vocabulary - (schema), the C++/Tree mapping produces a tree-like data structure - suitable for in-memory processing as well as XML parsing and - serialization code.</p> - - <p>A typical application that processes XML documents usually - performs the following three steps: it first reads (parses) an XML - instance document to an object model, it then performs - some useful computations on that model which may involve - modification of the model, and finally it may write (serialize) - the modified object model back to XML. - </p> - - <p>The C++/Tree mapping consists of C++ types that represent the - given vocabulary (<a href="#2">Chapter 2, "C++/Tree Mapping"</a>), - a set of parsing functions that convert XML documents to - a tree-like in-memory data structure (<a href="#3">Chapter 3, - "Parsing"</a>), and a set of serialization functions that convert - the object model back to XML (<a href="#4">Chapter 4, - "Serialization"</a>). Furthermore, the mapping provides a number - of additional features, such as DOM association and binary - serialization, that can be useful in some applications - (<a href="#5">Chapter 5, "Additional Functionality"</a>). - </p> - - - <!-- Chapter 2 --> - - - <h1><a name="2">2 C++/Tree Mapping</a></h1> - - <h2><a name="2.1">2.1 Preliminary Information</a></h2> - - <h3><a name="2.1.1">2.1.1 C++ Standard</a></h3> - - <p>The C++/Tree mapping provides support for ISO/IEC C++ 2011 (C++11) - and ISO/IEC C++ 1998/2003 (C++98). To select the C++ standard for the - generated code we use the <code>--std</code> XSD compiler command - line option. While the majority of the examples in this guide use - C++11, the document explains the C++11/98 usage difference and so - they can easily be converted to C++98.</p> - - <h3><a name="2.1.2">2.1.2 Identifiers</a></h3> - - <p>XML Schema names may happen to be reserved C++ keywords or contain - characters that are illegal in C++ identifiers. To avoid C++ compilation - problems, such names are changed (escaped) when mapped to C++. If an - XML Schema name is a C++ keyword, the "_" suffix is added to it. All - character of an XML Schema name that are not allowed in C++ identifiers - are replaced with "_". - </p> - - <p>For example, XML Schema name <code>try</code> will be mapped to - C++ identifier <code>try_</code>. Similarly, XML Schema name - <code>strange.na-me</code> will be mapped to C++ identifier - <code>strange_na_me</code>. - </p> - - <p>Furthermore, conflicts between type names and function names in the - same scope are resolved using name escaping. Such conflicts include - both a global element (which is mapped to a set of parsing and/or - serialization functions or element types, see <a href="#2.9">Section - 2.9, "Mapping for Global Elements"</a>) and a global type sharing the - same name as well as a local element or attribute inside a type having - the same name as the type itself.</p> - - <p>For example, if we had a global type <code>catalog</code> - and a global element with the same name then the type would be - mapped to a C++ class with name <code>catalog</code> while the - parsing functions corresponding to the global element would have - their names escaped as <code>catalog_</code>. - </p> - - <p>By default the mapping uses the so-called K&R (Kernighan and - Ritchie) identifier naming convention which is also used throughout - this manual. In this convention both type and function names are in - lower case and words are separated by underscores. If your application - code or schemas use a different notation, you may want to change the - naming convention used by the mapping for consistency. - The compiler supports a set of widely-used naming conventions - that you can select with the <code>--type-naming</code> and - <code>--function-naming</code> options. You can also further - refine one of the predefined conventions or create a completely - custom naming scheme by using the <code>--*-regex</code> options. - For more detailed information on these options refer to the NAMING - CONVENTION section in the <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>.</p> - - <h3><a name="2.1.3">2.1.3 Character Type and Encoding</a></h3> - - <p>The code that implements the mapping, depending on the - <code>--char-type</code> option, is generated using either - <code>char</code> or <code>wchar_t</code> as the character - type. In this document code samples use symbol <code>C</code> - to refer to the character type you have selected when translating - your schemas, for example <code>std::basic_string<C></code>. - </p> - - <p>Another aspect of the mapping that depends on the character type - is character encoding. For the <code>char</code> character type - the default encoding is UTF-8. Other supported encodings are - ISO-8859-1, Xerces-C++ Local Code Page (LPC), as well as - custom encodings and can be selected with the - <code>--char-encoding</code> command line option.</p> - - <p>For the <code>wchar_t</code> character type the encoding is - automatically selected between UTF-16 and UTF-32/UCS-4 depending - on the size of the <code>wchar_t</code> type. On some platforms - (for example, Windows with Visual C++ and AIX with IBM XL C++) - <code>wchar_t</code> is 2 bytes long. For these platforms the - encoding is UTF-16. On other platforms <code>wchar_t</code> is 4 bytes - long and UTF-32/UCS-4 is used.</p> - - <h3><a name="2.1.4">2.1.4 XML Schema Namespace</a></h3> - - <p>The mapping relies on some predefined types, classes, and functions - that are logically defined in the XML Schema namespace reserved for - the XML Schema language (<code>http://www.w3.org/2001/XMLSchema</code>). - By default, this namespace is mapped to C++ namespace - <code>xml_schema</code>. It is automatically accessible - from a C++ compilation unit that includes a header file generated - from an XML Schema definition. - </p> - - <p>Note that, if desired, the default mapping of this namespace can be - changed as described in <a href="#2.4">Section 2.4, "Mapping for - Namespaces"</a>. - </p> - - - <h3><a name="2.1.5">2.1.5 Anonymous Types</a></h3> - - <p>For the purpose of code generation, anonymous types defined in - XML Schema are automatically assigned names that are derived - from enclosing attributes and elements. Otherwise, such types - follows standard mapping rules for simple and complex type - definitions (see <a href="#2.6">Section 2.6, "Mapping for Simple Types"</a> - and <a href="#2.7">Section 2.7, "Mapping for Complex Types"</a>). - For example, in the following schema fragment: - </p> - - <pre class="xml"> -<element name="object"> - <complexType> - ... - </complexType> -</element> - </pre> - - <p>The anonymous type defined inside element <code>object</code> will - be given name <code>object</code>. The compiler has a number of - options that control the process of anonymous type naming. For more - information refer to the <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a>.</p> - - - <h2><a name="2.2">2.2 Error Handling</a></h2> - - <p>The mapping uses the C++ exception handling mechanism as a primary way - of reporting error conditions. All exceptions that are specified in - this mapping derive from <code>xml_schema::exception</code> which - itself is derived from <code>std::exception</code>: - </p> - - <pre class="c++"> -struct exception: virtual std::exception -{ - friend - std::basic_ostream<C>& - operator<< (std::basic_ostream<C>& os, const exception& e) - { - e.print (os); - return os; - } - -protected: - virtual void - print (std::basic_ostream<C>&) const = 0; -}; - </pre> - - <p>The exception hierarchy supports "virtual" <code>operator<<</code> - which allows you to obtain diagnostics corresponding to the thrown - exception using the base exception interface. For example:</p> - - <pre class="c++"> -try -{ - ... -} -catch (const xml_schema::exception& e) -{ - cerr << e << endl; -} - </pre> - - <p>The following sub-sections describe exceptions thrown by the - types that constitute the object model. - <a href="#3.3">Section 3.3, "Error Handling"</a> of - <a href="#3">Chapter 3, "Parsing"</a> describes exceptions - and error handling mechanisms specific to the parsing functions. - <a href="#4.4">Section 4.4, "Error Handling"</a> of - <a href="#4">Chapter 4, "Serialization"</a> describes exceptions - and error handling mechanisms specific to the serialization functions. - </p> - - - <h3><a name="2.2.1">2.2.1 <code>xml_schema::duplicate_id</code></a></h3> - - <pre class="c++"> -struct duplicate_id: virtual exception -{ - duplicate_id (const std::basic_string<C>& id); - - const std::basic_string<C>& - id () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::duplicate_id</code> is thrown when - a conflicting instance of <code>xml_schema::id</code> (see - <a href="#2.5">Section 2.5, "Mapping for Built-in Data Types"</a>) - is added to a tree. The offending ID value can be obtained using - the <code>id</code> function. - </p> - - <h2><a name="2.3">2.3 Mapping for <code>import</code> and <code>include</code></a></h2> - - <h3><a name="2.3.1">2.3.1 Import</a></h3> - - <p>The XML Schema <code>import</code> element is mapped to the C++ - Preprocessor <code>#include</code> directive. The value of - the <code>schemaLocation</code> attribute is used to derive - the name of the header file that appears in the <code>#include</code> - directive. For instance: - </p> - - <pre class="xml"> -<import namespace="https://www.codesynthesis.com/test" - schemaLocation="test.xsd"/> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -#include "test.hxx" - </pre> - - <p>Note that you will need to compile imported schemas separately - in order to produce corresponding header files.</p> - - <h3><a name="2.3.2">2.3.2 Inclusion with Target Namespace</a></h3> - - <p>The XML Schema <code>include</code> element which refers to a schema - with a target namespace or appears in a schema without a target namespace - follows the same mapping rules as the <code>import</code> element, - see <a href="#2.3.1">Section 2.3.1, "Import"</a>. - </p> - - <h3><a name="2.3.3">2.3.3 Inclusion without Target Namespace</a></h3> - - <p>For the XML Schema <code>include</code> element which refers to a schema - without a target namespace and appears in a schema with a target - namespace (such inclusion sometimes called "chameleon inclusion"), - declarations and definitions from the included schema are generated - in-line in the namespace of the including schema as if they were - declared and defined there verbatim. For example, consider the - following two schemas: - </p> - - <pre class="xml"> -<-- common.xsd --> -<schema> - <complexType name="type"> - ... - </complexType> -</schema> - -<-- test.xsd --> -<schema targetNamespace="https://www.codesynthesis.com/test"> - <include schemaLocation="common.xsd"/> -</schema> - </pre> - - <p>The fragment of interest from the generated header file for - <code>text.xsd</code> would look like this:</p> - - <pre class="c++"> -// test.hxx -namespace test -{ - class type - { - ... - }; -} - </pre> - - <h2><a name="2.4">2.4 Mapping for Namespaces</a></h2> - - <p>An XML Schema namespace is mapped to one or more nested C++ - namespaces. XML Schema namespaces are identified by URIs. - By default, a namespace URI is mapped to a sequence of - C++ namespace names by removing the protocol and host parts - and splitting the rest into a sequence of names with '<code>/</code>' - as the name separator. For instance: - </p> - - <pre class="xml"> -<schema targetNamespace="https://www.codesynthesis.com/system/test"> - ... -</schema> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -namespace system -{ - namespace test - { - ... - } -} - </pre> - - <p>The default mapping of namespace URIs to C++ namespace names can be - altered using the <code>--namespace-map</code> and - <code>--namespace-regex</code> options. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information. - </p> - - <h2><a name="2.5">2.5 Mapping for Built-in Data Types</a></h2> - - <p>The mapping of XML Schema built-in data types to C++ types is - summarized in the table below.</p> - - <!-- border="1" is necessary for html2ps --> - <table id="builtin" border="1"> - <tr> - <th>XML Schema type</th> - <th>Alias in the <code>xml_schema</code> namespace</th> - <th>C++ type</th> - </tr> - - <tr> - <th colspan="3">anyType and anySimpleType types</th> - </tr> - <tr> - <td><code>anyType</code></td> - <td><code>type</code></td> - <td><a href="#2.5.2">Section 2.5.2, "Mapping for <code>anyType</code>"</a></td> - </tr> - <tr> - <td><code>anySimpleType</code></td> - <td><code>simple_type</code></td> - <td><a href="#2.5.3">Section 2.5.3, "Mapping for <code>anySimpleType</code>"</a></td> - </tr> - - <tr> - <th colspan="3">fixed-length integral types</th> - </tr> - <!-- 8-bit --> - <tr> - <td><code>byte</code></td> - <td><code>byte</code></td> - <td><code>signed char</code></td> - </tr> - <tr> - <td><code>unsignedByte</code></td> - <td><code>unsigned_byte</code></td> - <td><code>unsigned char</code></td> - </tr> - - <!-- 16-bit --> - <tr> - <td><code>short</code></td> - <td><code>short_</code></td> - <td><code>short</code></td> - </tr> - <tr> - <td><code>unsignedShort</code></td> - <td><code>unsigned_short</code></td> - <td><code>unsigned short</code></td> - </tr> - - <!-- 32-bit --> - <tr> - <td><code>int</code></td> - <td><code>int_</code></td> - <td><code>int</code></td> - </tr> - <tr> - <td><code>unsignedInt</code></td> - <td><code>unsigned_int</code></td> - <td><code>unsigned int</code></td> - </tr> - - <!-- 64-bit --> - <tr> - <td><code>long</code></td> - <td><code>long_</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>unsignedLong</code></td> - <td><code>unsigned_long</code></td> - <td><code>unsigned long long</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-length integral types</th> - </tr> - <tr> - <td><code>integer</code></td> - <td><code>integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonPositiveInteger</code></td> - <td><code>non_positive_integer</code></td> - <td><code>long long</code></td> - </tr> - <tr> - <td><code>nonNegativeInteger</code></td> - <td><code>non_negative_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>positiveInteger</code></td> - <td><code>positive_integer</code></td> - <td><code>unsigned long long</code></td> - </tr> - <tr> - <td><code>negativeInteger</code></td> - <td><code>negative_integer</code></td> - <td><code>long long</code></td> - </tr> - - <tr> - <th colspan="3">boolean types</th> - </tr> - <tr> - <td><code>boolean</code></td> - <td><code>boolean</code></td> - <td><code>bool</code></td> - </tr> - - <tr> - <th colspan="3">fixed-precision floating-point types</th> - </tr> - <tr> - <td><code>float</code></td> - <td><code>float_</code></td> - <td><code>float</code></td> - </tr> - <tr> - <td><code>double</code></td> - <td><code>double_</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">arbitrary-precision floating-point types</th> - </tr> - <tr> - <td><code>decimal</code></td> - <td><code>decimal</code></td> - <td><code>double</code></td> - </tr> - - <tr> - <th colspan="3">string types</th> - </tr> - <tr> - <td><code>string</code></td> - <td><code>string</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - <tr> - <td><code>normalizedString</code></td> - <td><code>normalized_string</code></td> - <td>type derived from <code>string</code></td> - </tr> - <tr> - <td><code>token</code></td> - <td><code>token</code></td> - <td>type derived from <code>normalized_string</code></td> - </tr> - <tr> - <td><code>Name</code></td> - <td><code>name</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKEN</code></td> - <td><code>nmtoken</code></td> - <td>type derived from <code>token</code></td> - </tr> - <tr> - <td><code>NMTOKENS</code></td> - <td><code>nmtokens</code></td> - <td>type derived from <code>sequence<nmtoken></code></td> - </tr> - <tr> - <td><code>NCName</code></td> - <td><code>ncname</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>language</code></td> - <td><code>language</code></td> - <td>type derived from <code>token</code></td> - </tr> - - <tr> - <th colspan="3">qualified name</th> - </tr> - <tr> - <td><code>QName</code></td> - <td><code>qname</code></td> - <td><a href="#2.5.4">Section 2.5.4, "Mapping for <code>QName</code>"</a></td> - </tr> - - <tr> - <th colspan="3">ID/IDREF types</th> - </tr> - <tr> - <td><code>ID</code></td> - <td><code>id</code></td> - <td>type derived from <code>ncname</code></td> - </tr> - <tr> - <td><code>IDREF</code></td> - <td><code>idref</code></td> - <td><a href="#2.5.5">Section 2.5.5, "Mapping for <code>IDREF</code>"</a></td> - </tr> - <tr> - <td><code>IDREFS</code></td> - <td><code>idrefs</code></td> - <td>type derived from <code>sequence<idref></code></td> - </tr> - - <tr> - <th colspan="3">URI types</th> - </tr> - <tr> - <td><code>anyURI</code></td> - <td><code>uri</code></td> - <td>type derived from <code>std::basic_string</code></td> - </tr> - - <tr> - <th colspan="3">binary types</th> - </tr> - <tr> - <td><code>base64Binary</code></td> - <td><code>base64_binary</code></td> - <td rowspan="2"><a href="#2.5.6">Section 2.5.6, "Mapping for - <code>base64Binary</code> and <code>hexBinary</code>"</a></td> - </tr> - <tr> - <td><code>hexBinary</code></td> - <td><code>hex_binary</code></td> - </tr> - - <tr> - <th colspan="3">date/time types</th> - </tr> - <tr> - <td><code>date</code></td> - <td><code>date</code></td> - <td><a href="#2.5.8">Section 2.5.8, "Mapping for - <code>date</code>"</a></td> - </tr> - <tr> - <td><code>dateTime</code></td> - <td><code>date_time</code></td> - <td><a href="#2.5.9">Section 2.5.9, "Mapping for - <code>dateTime</code>"</a></td> - </tr> - <tr> - <td><code>duration</code></td> - <td><code>duration</code></td> - <td><a href="#2.5.10">Section 2.5.10, "Mapping for - <code>duration</code>"</a></td> - </tr> - <tr> - <td><code>gDay</code></td> - <td><code>gday</code></td> - <td><a href="#2.5.11">Section 2.5.11, "Mapping for - <code>gDay</code>"</a></td> - </tr> - <tr> - <td><code>gMonth</code></td> - <td><code>gmonth</code></td> - <td><a href="#2.5.12">Section 2.5.12, "Mapping for - <code>gMonth</code>"</a></td> - </tr> - <tr> - <td><code>gMonthDay</code></td> - <td><code>gmonth_day</code></td> - <td><a href="#2.5.13">Section 2.5.13, "Mapping for - <code>gMonthDay</code>"</a></td> - </tr> - <tr> - <td><code>gYear</code></td> - <td><code>gyear</code></td> - <td><a href="#2.5.14">Section 2.5.14, "Mapping for - <code>gYear</code>"</a></td> - </tr> - <tr> - <td><code>gYearMonth</code></td> - <td><code>gyear_month</code></td> - <td><a href="#2.5.15">Section 2.5.15, "Mapping for - <code>gYearMonth</code>"</a></td> - </tr> - <tr> - <td><code>time</code></td> - <td><code>time</code></td> - <td><a href="#2.5.16">Section 2.5.16, "Mapping for - <code>time</code>"</a></td> - </tr> - - <tr> - <th colspan="3">entity types</th> - </tr> - <tr> - <td><code>ENTITY</code></td> - <td><code>entity</code></td> - <td>type derived from <code>name</code></td> - </tr> - <tr> - <td><code>ENTITIES</code></td> - <td><code>entities</code></td> - <td>type derived from <code>sequence<entity></code></td> - </tr> - </table> - - <p>All XML Schema built-in types are mapped to C++ classes that are - derived from the <code>xml_schema::simple_type</code> class except - where the mapping is to a fundamental C++ type.</p> - - <p>The <code>sequence</code> class template is defined in an - implementation-specific namespace. It conforms to the - sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences"). - Practically, this means that you can treat such a sequence - as if it was <code>std::vector</code>. One notable extension - to the standard interface that is available only for - sequences of non-fundamental C++ types is the addition of - the overloaded <code>push_back</code> and <code>insert</code> - member functions which instead of the constant reference - to the element type accept automatic pointer (<code>std::unique_ptr</code> - or <code>std::auto_ptr</code>, depending on the C++ standard - selected) to the element type. These functions assume ownership - of the pointed to object and reset the passed automatic pointer. - </p> - - <h3><a name="2.5.1">2.5.1 Inheritance from Built-in Data Types</a></h3> - - <p>In cases where the mapping calls for an inheritance from a built-in - type which is mapped to a fundamental C++ type, a proxy type is - used instead of the fundamental C++ type (C++ does not allow - inheritance from fundamental types). For instance:</p> - - <pre class="xml"> -<simpleType name="my_int"> - <restriction base="int"/> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class my_int: public fundamental_base<int> -{ - ... -}; - </pre> - - <p>The <code>fundamental_base</code> class template provides a close - emulation (though not exact) of a fundamental C++ type. - It is defined in an implementation-specific namespace and has the - following interface:</p> - - <pre class="c++"> -template <typename X> -class fundamental_base: public simple_type -{ -public: - fundamental_base (); - fundamental_base (X) - fundamental_base (const fundamental_base&) - -public: - fundamental_base& - operator= (const X&); - -public: - operator const X & () const; - operator X& (); - - template <typename Y> - operator Y () const; - - template <typename Y> - operator Y (); -}; - </pre> - - <h3><a name="2.5.2">2.5.2 Mapping for <code>anyType</code></a></h3> - - <p>The XML Schema <code>anyType</code> built-in data type is mapped to the - <code>xml_schema::type</code> C++ class:</p> - - <pre class="c++"> -class type -{ -public: - virtual - ~type (); - - type (); - type (const type&); - - type& - operator= (const type&); - - virtual type* - _clone () const; - - // anyType DOM content. - // -public: - typedef element_optional dom_content_optional; - - const dom_content_optional& - dom_content () const; - - dom_content_optional& - dom_content (); - - void - dom_content (const xercesc::DOMElement&); - - void - dom_content (xercesc::DOMElement*); - - void - dom_content (const dom_content_optional&); - - const xercesc::DOMDocument& - dom_content_document () const; - - xercesc::DOMDocument& - dom_content_document (); - - bool - null_content () const; - - // DOM association. - // -public: - const xercesc::DOMNode* - _node () const; - - xercesc::DOMNode* - _node (); -}; - </pre> - - <p>When <code>xml_schema::type</code> is used to create an instance - (as opposed to being a base of a derived type), it represents - the XML Schema <code>anyType</code> type. <code>anyType</code> - allows any attributes and any content in any order. In the - C++/Tree mapping this content can be represented as a DOM - fragment, similar to XML Schema wildcards (<a href="#2.12">Section - 2.12, "Mapping for <code>any</code> and - <code>anyAttribute</code>"</a>).</p> - - <p>To enable automatic extraction of <code>anyType</code> content - during parsing, the <code>--generate-any-type</code> option must be - specified. Because the DOM API is used to access such content, the - Xerces-C++ runtime should be initialized by the application prior to - parsing and should remain initialized for the lifetime of objects - with the DOM content. For more information on the Xerces-C++ runtime - initialization see <a href="#3.1">Section 3.1, "Initializing the - Xerces-C++ Runtime"</a>.</p> - - <p>The DOM content is stored as the optional DOM element container - and the DOM content accessors and modifiers presented above are - identical to those generated for an optional element wildcard. - Refer to <a href="#2.12.2">Section 2.12.2, "Mapping for <code>any</code> - with the Optional Cardinality Class"</a> for details on their - semantics.</p> - - <p>The <code>dom_content_document()</code> function returns the - DOM document used to store the raw XML content corresponding - to the <code>anyType</code> instance. It is equivalent to the - <code>dom_document()</code> function generated for types - with wildcards.</p> - - <p>The <code>null_content()</code> accessor is an optimization function - that allows us to check for the lack of content without actually - creating its empty representation, that is, empty DOM document for - <code>anyType</code> or empty string for <code>anySimpleType</code> - (see the following section for details on <code>anySimpleType</code>).</p> - - <p>For more information on DOM association refer to - <a href="#5.1">Section 5.1, "DOM Association"</a>.</p> - - <h3><a name="2.5.3">2.5.3 Mapping for <code>anySimpleType</code></a></h3> - - <p>The XML Schema <code>anySimpleType</code> built-in data type is mapped - to the <code>xml_schema::simple_type</code> C++ class:</p> - - <pre class="c++"> -class simple_type: public type -{ -public: - simple_type (); - simple_type (const C*); - simple_type (const std::basic_string<C>&); - - simple_type (const simple_type&); - - simple_type& - operator= (const simple_type&); - - virtual simple_type* - _clone () const; - - // anySimpleType text content. - // -public: - const std::basic_string<C>& - text_content () const; - - std::basic_string<C>& - text_content (); - - void - text_content (const std::basic_string<C>&); -}; - </pre> - - <p>When <code>xml_schema::simple_type</code> is used to create an instance - (as opposed to being a base of a derived type), it represents - the XML Schema <code>anySimpleType</code> type. <code>anySimpleType</code> - allows any simple content. In the C++/Tree mapping this content can - be represented as a string and accessed or modified with the - <code>text_content()</code> functions shown above.</p> - - <h3><a name="2.5.4">2.5.4 Mapping for <code>QName</code></a></h3> - - <p>The XML Schema <code>QName</code> built-in data type is mapped to the - <code>xml_schema::qname</code> C++ class:</p> - - <pre class="c++"> -class qname: public simple_type -{ -public: - qname (const ncname&); - qname (const uri&, const ncname&); - qname (const qname&); - -public: - qname& - operator= (const qname&); - -public: - virtual qname* - _clone () const; - -public: - bool - qualified () const; - - const uri& - namespace_ () const; - - const ncname& - name () const; -}; - </pre> - - <p>The <code>qualified</code> accessor function can be used to determine - if the name is qualified.</p> - - <h3><a name="2.5.5">2.5.5 Mapping for <code>IDREF</code></a></h3> - - <p>The XML Schema <code>IDREF</code> built-in data type is mapped to the - <code>xml_schema::idref</code> C++ class. This class implements the - smart pointer C++ idiom:</p> - - <pre class="c++"> -class idref: public ncname -{ -public: - idref (const C* s); - idref (const C* s, std::size_t n); - idref (std::size_t n, C c); - idref (const std::basic_string<C>&); - idref (const std::basic_string<C>&, - std::size_t pos, - std::size_t n = npos); - -public: - idref (const idref&); - -public: - virtual idref* - _clone () const; - -public: - idref& - operator= (C c); - - idref& - operator= (const C* s); - - idref& - operator= (const std::basic_string<C>&) - - idref& - operator= (const idref&); - -public: - const type* - operator-> () const; - - type* - operator-> (); - - const type& - operator* () const; - - type& - operator* (); - - const type* - get () const; - - type* - get (); - - // Conversion to bool. - // -public: - typedef void (idref::*bool_convertible)(); - operator bool_convertible () const; -}; - </pre> - - <p>The object, <code>idref</code> instance refers to, is the immediate - container of the matching <code>id</code> instance. For example, - with the following instance document and schema: - </p> - - - <pre class="xml"> -<!-- test.xml --> -<root> - <object id="obj-1" text="hello"/> - <reference>obj-1</reference> -</root> - -<!-- test.xsd --> -<schema> - <complexType name="object_type"> - <attribute name="id" type="ID"/> - <attribute name="text" type="string"/> - </complexType> - - <complexType name="root_type"> - <sequence> - <element name="object" type="object_type"/> - <element name="reference" type="IDREF"/> - </sequence> - </complexType> - - <element name="root" type="root_type"/> -</schema> - </pre> - - <p>The <code>ref</code> instance in the code below will refer to - an object of type <code>object_type</code>:</p> - - <pre class="c++"> -root_type& root = ...; -xml_schema::idref& ref (root.reference ()); -object_type& obj (dynamic_cast<object_type&> (*ref)); -cout << obj.text () << endl; - </pre> - - <p>The smart pointer interface of the <code>idref</code> class always - returns a pointer or reference to <code>xml_schema::type</code>. - This means that you will need to manually cast such pointer or - reference to its real (dynamic) type before you can use it (unless - all you need is the base interface provided by - <code>xml_schema::type</code>). As a special extension to the XML - Schema language, the mapping supports static typing of <code>idref</code> - references by employing the <code>refType</code> extension attribute. - The following example illustrates this mechanism: - </p> - - <pre class="xml"> -<!-- test.xsd --> -<schema - xmlns:xse="https://www.codesynthesis.com/xmlns/xml-schema-extension"> - - ... - - <element name="reference" type="IDREF" xse:refType="object_type"/> - - ... - -</schema> - </pre> - - <p>With this modification we do not need to do manual casting anymore: - </p> - - <pre class="c++"> -root_type& root = ...; -root_type::reference_type& ref (root.reference ()); -object_type& obj (*ref); -cout << ref->text () << endl; - </pre> - - - <h3><a name="2.5.6">2.5.6 Mapping for <code>base64Binary</code> and - <code>hexBinary</code></a></h3> - - <p>The XML Schema <code>base64Binary</code> and <code>hexBinary</code> - built-in data types are mapped to the - <code>xml_schema::base64_binary</code> and - <code>xml_schema::hex_binary</code> C++ classes, respectively. The - <code>base64_binary</code> and <code>hex_binary</code> classes - support a simple buffer abstraction by inheriting from the - <code>xml_schema::buffer</code> class: - </p> - - <pre class="c++"> -class bounds: public virtual exception -{ -public: - virtual const char* - what () const throw (); -}; - -class buffer -{ -public: - typedef std::size_t size_t; - -public: - buffer (size_t size = 0); - buffer (size_t size, size_t capacity); - buffer (const void* data, size_t size); - buffer (const void* data, size_t size, size_t capacity); - buffer (void* data, - size_t size, - size_t capacity, - bool assume_ownership); - -public: - buffer (const buffer&); - - buffer& - operator= (const buffer&); - - void - swap (buffer&); - -public: - size_t - capacity () const; - - bool - capacity (size_t); - -public: - size_t - size () const; - - bool - size (size_t); - -public: - const char* - data () const; - - char* - data (); - - const char* - begin () const; - - char* - begin (); - - const char* - end () const; - - char* - end (); -}; - </pre> - - <p>The last overloaded constructor reuses an existing data buffer instead - of making a copy. If the <code>assume_ownership</code> argument is - <code>true</code>, the instance assumes ownership of the - memory block pointed to by the <code>data</code> argument and will - eventually release it by calling <code>operator delete</code>. The - <code>capacity</code> and <code>size</code> modifier functions return - <code>true</code> if the underlying buffer has moved. - </p> - - <p>The <code>bounds</code> exception is thrown if the constructor - arguments violate the <code>(size <= capacity)</code> - constraint.</p> - - <p>The <code>base64_binary</code> and <code>hex_binary</code> classes - support the <code>buffer</code> interface and perform automatic - decoding/encoding from/to the Base64 and Hex formats, respectively: - </p> - - <pre class="c++"> -class base64_binary: public simple_type, public buffer -{ -public: - base64_binary (size_t size = 0); - base64_binary (size_t size, size_t capacity); - base64_binary (const void* data, size_t size); - base64_binary (const void* data, size_t size, size_t capacity); - base64_binary (void* data, - size_t size, - size_t capacity, - bool assume_ownership); - -public: - base64_binary (const base64_binary&); - - base64_binary& - operator= (const base64_binary&); - - virtual base64_binary* - _clone () const; - -public: - std::basic_string<C> - encode () const; -}; - </pre> - - <pre class="c++"> -class hex_binary: public simple_type, public buffer -{ -public: - hex_binary (size_t size = 0); - hex_binary (size_t size, size_t capacity); - hex_binary (const void* data, size_t size); - hex_binary (const void* data, size_t size, size_t capacity); - hex_binary (void* data, - size_t size, - size_t capacity, - bool assume_ownership); - -public: - hex_binary (const hex_binary&); - - hex_binary& - operator= (const hex_binary&); - - virtual hex_binary* - _clone () const; - -public: - std::basic_string<C> - encode () const; -}; - </pre> - - - <h2><a name="2.5.7">2.5.7 Time Zone Representation</a></h2> - - <p>The <code>date</code>, <code>dateTime</code>, <code>gDay</code>, - <code>gMonth</code>, <code>gMonthDay</code>, <code>gYear</code>, - <code>gYearMonth</code>, and <code>time</code> XML Schema built-in - types all include an optional time zone component. The following - <code>xml_schema::time_zone</code> base class is used to represent - this information:</p> - - <pre class="c++"> -class time_zone -{ -public: - time_zone (); - time_zone (short hours, short minutes); - - bool - zone_present () const; - - void - zone_reset (); - - short - zone_hours () const; - - void - zone_hours (short); - - short - zone_minutes () const; - - void - zone_minutes (short); -}; - -bool -operator== (const time_zone&, const time_zone&); - -bool -operator!= (const time_zone&, const time_zone&); - </pre> - - <p>The <code>zone_present()</code> accessor function returns <code>true</code> - if the time zone is specified. The <code>zone_reset()</code> modifier - function resets the time zone object to the <em>not specified</em> - state. If the time zone offset is negative then both hours and - minutes components are represented as negative integers.</p> - - - <h2><a name="2.5.8">2.5.8 Mapping for <code>date</code></a></h2> - - <p>The XML Schema <code>date</code> built-in data type is mapped to the - <code>xml_schema::date</code> C++ class which represents a year, a day, - and a month with an optional time zone. Its interface is presented - below. For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class date: public simple_type, public time_zone -{ -public: - date (int year, unsigned short month, unsigned short day); - date (int year, unsigned short month, unsigned short day, - short zone_hours, short zone_minutes); - -public: - date (const date&); - - date& - operator= (const date&); - - virtual date* - _clone () const; - -public: - int - year () const; - - void - year (int); - - unsigned short - month () const; - - void - month (unsigned short); - - unsigned short - day () const; - - void - day (unsigned short); -}; - -bool -operator== (const date&, const date&); - -bool -operator!= (const date&, const date&); - </pre> - - <h2><a name="2.5.9">2.5.9 Mapping for <code>dateTime</code></a></h2> - - <p>The XML Schema <code>dateTime</code> built-in data type is mapped to the - <code>xml_schema::date_time</code> C++ class which represents a year, a month, - a day, hours, minutes, and seconds with an optional time zone. Its interface - is presented below. For more information on the base - <code>xml_schema::time_zone</code> class refer to <a href="#2.5.7">Section - 2.5.7, "Time Zone Representation"</a>.</p> - - <pre class="c++"> -class date_time: public simple_type, public time_zone -{ -public: - date_time (int year, unsigned short month, unsigned short day, - unsigned short hours, unsigned short minutes, - double seconds); - - date_time (int year, unsigned short month, unsigned short day, - unsigned short hours, unsigned short minutes, - double seconds, short zone_hours, short zone_minutes); -public: - date_time (const date_time&); - - date_time& - operator= (const date_time&); - - virtual date_time* - _clone () const; - -public: - int - year () const; - - void - year (int); - - unsigned short - month () const; - - void - month (unsigned short); - - unsigned short - day () const; - - void - day (unsigned short); - - unsigned short - hours () const; - - void - hours (unsigned short); - - unsigned short - minutes () const; - - void - minutes (unsigned short); - - double - seconds () const; - - void - seconds (double); -}; - -bool -operator== (const date_time&, const date_time&); - -bool -operator!= (const date_time&, const date_time&); - </pre> - - - <h2><a name="2.5.10">2.5.10 Mapping for <code>duration</code></a></h2> - - <p>The XML Schema <code>duration</code> built-in data type is mapped to the - <code>xml_schema::duration</code> C++ class which represents a potentially - negative duration in the form of years, months, days, hours, minutes, - and seconds. Its interface is presented below.</p> - - <pre class="c++"> -class duration: public simple_type -{ -public: - duration (bool negative, - unsigned int years, unsigned int months, unsigned int days, - unsigned int hours, unsigned int minutes, double seconds); -public: - duration (const duration&); - - duration& - operator= (const duration&); - - virtual duration* - _clone () const; - -public: - bool - negative () const; - - void - negative (bool); - - unsigned int - years () const; - - void - years (unsigned int); - - unsigned int - months () const; - - void - months (unsigned int); - - unsigned int - days () const; - - void - days (unsigned int); - - unsigned int - hours () const; - - void - hours (unsigned int); - - unsigned int - minutes () const; - - void - minutes (unsigned int); - - double - seconds () const; - - void - seconds (double); -}; - -bool -operator== (const duration&, const duration&); - -bool -operator!= (const duration&, const duration&); - </pre> - - - <h2><a name="2.5.11">2.5.11 Mapping for <code>gDay</code></a></h2> - - <p>The XML Schema <code>gDay</code> built-in data type is mapped to the - <code>xml_schema::gday</code> C++ class which represents a day of the - month with an optional time zone. Its interface is presented below. - For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gday: public simple_type, public time_zone -{ -public: - explicit - gday (unsigned short day); - gday (unsigned short day, short zone_hours, short zone_minutes); - -public: - gday (const gday&); - - gday& - operator= (const gday&); - - virtual gday* - _clone () const; - -public: - unsigned short - day () const; - - void - day (unsigned short); -}; - -bool -operator== (const gday&, const gday&); - -bool -operator!= (const gday&, const gday&); - </pre> - - - <h2><a name="2.5.12">2.5.12 Mapping for <code>gMonth</code></a></h2> - - <p>The XML Schema <code>gMonth</code> built-in data type is mapped to the - <code>xml_schema::gmonth</code> C++ class which represents a month of the - year with an optional time zone. Its interface is presented below. - For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gmonth: public simple_type, public time_zone -{ -public: - explicit - gmonth (unsigned short month); - gmonth (unsigned short month, - short zone_hours, short zone_minutes); - -public: - gmonth (const gmonth&); - - gmonth& - operator= (const gmonth&); - - virtual gmonth* - _clone () const; - -public: - unsigned short - month () const; - - void - month (unsigned short); -}; - -bool -operator== (const gmonth&, const gmonth&); - -bool -operator!= (const gmonth&, const gmonth&); - </pre> - - - <h2><a name="2.5.13">2.5.13 Mapping for <code>gMonthDay</code></a></h2> - - <p>The XML Schema <code>gMonthDay</code> built-in data type is mapped to the - <code>xml_schema::gmonth_day</code> C++ class which represents a day and - a month of the year with an optional time zone. Its interface is presented - below. For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gmonth_day: public simple_type, public time_zone -{ -public: - gmonth_day (unsigned short month, unsigned short day); - gmonth_day (unsigned short month, unsigned short day, - short zone_hours, short zone_minutes); - -public: - gmonth_day (const gmonth_day&); - - gmonth_day& - operator= (const gmonth_day&); - - virtual gmonth_day* - _clone () const; - -public: - unsigned short - month () const; - - void - month (unsigned short); - - unsigned short - day () const; - - void - day (unsigned short); -}; - -bool -operator== (const gmonth_day&, const gmonth_day&); - -bool -operator!= (const gmonth_day&, const gmonth_day&); - </pre> - - - <h2><a name="2.5.14">2.5.14 Mapping for <code>gYear</code></a></h2> - - <p>The XML Schema <code>gYear</code> built-in data type is mapped to the - <code>xml_schema::gyear</code> C++ class which represents a year with - an optional time zone. Its interface is presented below. For more - information on the base <code>xml_schema::time_zone</code> class refer - to <a href="#2.5.7">Section 2.5.7, "Time Zone Representation"</a>.</p> - - <pre class="c++"> -class gyear: public simple_type, public time_zone -{ -public: - explicit - gyear (int year); - gyear (int year, short zone_hours, short zone_minutes); - -public: - gyear (const gyear&); - - gyear& - operator= (const gyear&); - - virtual gyear* - _clone () const; - -public: - int - year () const; - - void - year (int); -}; - -bool -operator== (const gyear&, const gyear&); - -bool -operator!= (const gyear&, const gyear&); - </pre> - - - <h2><a name="2.5.15">2.5.15 Mapping for <code>gYearMonth</code></a></h2> - - <p>The XML Schema <code>gYearMonth</code> built-in data type is mapped to - the <code>xml_schema::gyear_month</code> C++ class which represents - a year and a month with an optional time zone. Its interface is presented - below. For more information on the base <code>xml_schema::time_zone</code> - class refer to <a href="#2.5.7">Section 2.5.7, "Time Zone - Representation"</a>.</p> - - <pre class="c++"> -class gyear_month: public simple_type, public time_zone -{ -public: - gyear_month (int year, unsigned short month); - gyear_month (int year, unsigned short month, - short zone_hours, short zone_minutes); -public: - gyear_month (const gyear_month&); - - gyear_month& - operator= (const gyear_month&); - - virtual gyear_month* - _clone () const; - -public: - int - year () const; - - void - year (int); - - unsigned short - month () const; - - void - month (unsigned short); -}; - -bool -operator== (const gyear_month&, const gyear_month&); - -bool -operator!= (const gyear_month&, const gyear_month&); - </pre> - - - <h2><a name="2.5.16">2.5.16 Mapping for <code>time</code></a></h2> - - <p>The XML Schema <code>time</code> built-in data type is mapped to - the <code>xml_schema::time</code> C++ class which represents hours, - minutes, and seconds with an optional time zone. Its interface is - presented below. For more information on the base - <code>xml_schema::time_zone</code> class refer to - <a href="#2.5.7">Section 2.5.7, "Time Zone Representation"</a>.</p> - - <pre class="c++"> -class time: public simple_type, public time_zone -{ -public: - time (unsigned short hours, unsigned short minutes, double seconds); - time (unsigned short hours, unsigned short minutes, double seconds, - short zone_hours, short zone_minutes); - -public: - time (const time&); - - time& - operator= (const time&); - - virtual time* - _clone () const; - -public: - unsigned short - hours () const; - - void - hours (unsigned short); - - unsigned short - minutes () const; - - void - minutes (unsigned short); - - double - seconds () const; - - void - seconds (double); -}; - -bool -operator== (const time&, const time&); - -bool -operator!= (const time&, const time&); - </pre> - - - <!-- Mapping for Simple Types --> - - <h2><a name="2.6">2.6 Mapping for Simple Types</a></h2> - - <p>An XML Schema simple type is mapped to a C++ class with the same - name as the simple type. The class defines a public copy constructor, - a public copy assignment operator, and a public virtual - <code>_clone</code> function. The <code>_clone</code> function is - declared <code>const</code>, does not take any arguments, and returns - a pointer to a complete copy of the instance allocated in the free - store. The <code>_clone</code> function shall be used to make copies - when static type and dynamic type of the instance may differ (see - <a href="#2.11">Section 2.11, "Mapping for <code>xsi:type</code> - and Substitution Groups"</a>). For instance:</p> - - <pre class="xml"> -<simpleType name="object"> - ... -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: ... -{ -public: - object (const object&); - -public: - object& - operator= (const object&); - -public: - virtual object* - _clone () const; - - ... - -}; - </pre> - - <p>The base class specification and the rest of the class definition - depend on the type of derivation used to define the simple type. </p> - - - <h3><a name="2.6.1">2.6.1 Mapping for Derivation by Restriction</a></h3> - - <p>XML Schema derivation by restriction is mapped to C++ public - inheritance. The base type of the restriction becomes the base - type for the resulting C++ class. In addition to the members described - in <a href="#2.6">Section 2.6, "Mapping for Simple Types"</a>, the - resulting C++ class defines a public constructor with the base type - as its single argument. For instance:</p> - - <pre class="xml"> -<simpleType name="object"> - <restriction base="base"> - ... - </restriction> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public base -{ -public: - object (const base&); - object (const object&); - -public: - object& - operator= (const object&); - -public: - virtual object* - _clone () const; -}; - </pre> - - - <h3><a name="2.6.2">2.6.2 Mapping for Enumerations</a></h3> - -<p>XML Schema restriction by enumeration is mapped to a C++ class - with semantics similar to C++ <code>enum</code>. Each XML Schema - enumeration element is mapped to a C++ enumerator with the - name derived from the <code>value</code> attribute and defined - in the class scope. In addition to the members - described in <a href="#2.6">Section 2.6, "Mapping for Simple Types"</a>, - the resulting C++ class defines a public constructor that can be called - with one of the enumerators as its single argument, a public constructor - that can be called with enumeration's base value as its single - argument, a public assignment operator that can be used to assign the - value of one of the enumerators, and a public implicit conversion - operator to the underlying C++ enum type.</p> - -<p>Furthermore, for string-based enumeration types, the resulting C++ - class defines a public constructor with a single argument of type - <code>const C*</code> and a public constructor with a single - argument of type <code>const std::basic_string<C>&</code>. - For instance:</p> - - <pre class="xml"> -<simpleType name="color"> - <restriction base="string"> - <enumeration value="red"/> - <enumeration value="green"/> - <enumeration value="blue"/> - </restriction> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class color: public xml_schema::string -{ -public: - enum value - { - red, - green, - blue - }; - -public: - color (value); - color (const C*); - color (const std::basic_string<C>&); - color (const xml_schema::string&); - color (const color&); - -public: - color& - operator= (value); - - color& - operator= (const color&); - -public: - virtual color* - _clone () const; - -public: - operator value () const; -}; - </pre> - - <h3><a name="2.6.3">2.6.3 Mapping for Derivation by List</a></h3> - - <p>XML Schema derivation by list is mapped to C++ public - inheritance from <code>xml_schema::simple_type</code> - (<a href="#2.5.3">Section 2.5.3, "Mapping for - <code>anySimpleType</code>"</a>) and a suitable sequence type. - The list item type becomes the element type of the sequence. - In addition to the members described in <a href="#2.6">Section 2.6, - "Mapping for Simple Types"</a>, the resulting C++ class defines - a public default constructor, a public constructor - with the first argument of type <code>size_type</code> and - the second argument of list item type that creates - a list object with the specified number of copies of the specified - element value, and a public constructor with the two arguments - of an input iterator type that creates a list object from an - iterator range. For instance: - </p> - - <pre class="xml"> -<simpleType name="int_list"> - <list itemType="int"/> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class int_list: public simple_type, - public sequence<int> -{ -public: - int_list (); - int_list (size_type n, int x); - - template <typename I> - int_list (const I& begin, const I& end); - int_list (const int_list&); - -public: - int_list& - operator= (const int_list&); - -public: - virtual int_list* - _clone () const; -}; - </pre> - - <p>The <code>sequence</code> class template is defined in an - implementation-specific namespace. It conforms to the - sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences"). - Practically, this means that you can treat such a sequence - as if it was <code>std::vector</code>. One notable extension - to the standard interface that is available only for - sequences of non-fundamental C++ types is the addition of - the overloaded <code>push_back</code> and <code>insert</code> - member functions which instead of the constant reference - to the element type accept automatic pointer (<code>std::unique_ptr</code> - or <code>std::auto_ptr</code>, depending on the C++ standard - selected) to the element type. These functions assume ownership - of the pointed to object and reset the passed automatic pointer. - </p> - - <h3><a name="2.6.4">2.6.4 Mapping for Derivation by Union</a></h3> - - <p>XML Schema derivation by union is mapped to C++ public - inheritance from <code>xml_schema::simple_type</code> - (<a href="#2.5.3">Section 2.5.3, "Mapping for - <code>anySimpleType</code>"</a>) and <code>std::basic_string<C></code>. - In addition to the members described in <a href="#2.6">Section 2.6, - "Mapping for Simple Types"</a>, the resulting C++ class defines a - public constructor with a single argument of type <code>const C*</code> - and a public constructor with a single argument of type - <code>const std::basic_string<C>&</code>. For instance: - </p> - - <pre class="xml"> -<simpleType name="int_string_union"> - <xsd:union memberTypes="xsd:int xsd:string"/> -</simpleType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class int_string_union: public simple_type, - public std::basic_string<C> -{ -public: - int_string_union (const C*); - int_string_union (const std::basic_string<C>&); - int_string_union (const int_string_union&); - -public: - int_string_union& - operator= (const int_string_union&); - -public: - virtual int_string_union* - _clone () const; -}; - </pre> - - <h2><a name="2.7">2.7 Mapping for Complex Types</a></h2> - - <p>An XML Schema complex type is mapped to a C++ class with the same - name as the complex type. The class defines a public copy constructor, - a public copy assignment operator, and a public virtual - <code>_clone</code> function. The <code>_clone</code> function is - declared <code>const</code>, does not take any arguments, and returns - a pointer to a complete copy of the instance allocated in the free - store. The <code>_clone</code> function shall be used to make copies - when static type and dynamic type of the instance may differ (see - <a href="#2.11">Section 2.11, "Mapping for <code>xsi:type</code> - and Substitution Groups"</a>).</p> - - <p>Additionally, the resulting C++ class - defines two public constructors that take an initializer for each - member of the complex type and all its base types that belongs to - the One cardinality class (see <a href="#2.8">Section 2.8, "Mapping - for Local Elements and Attributes"</a>). In the first constructor, - the arguments are passed as constant references and the newly created - instance is initialized with copies of the passed objects. In the - second constructor, arguments that are complex types (that is, - they themselves contain elements or attributes) are passed as - either <code>std::unique_ptr</code> (C++11) or <code>std::auto_ptr</code> - (C++98), depending on the C++ standard selected. In this case the newly - created instance is directly initialized with and assumes ownership - of the pointed to objects and the <code>std::[unique|auto]_ptr</code> - arguments are reset to <code>0</code>. For instance:</p> - - <pre class="xml"> -<complexType name="complex"> - <sequence> - <element name="a" type="int"/> - <element name="b" type="string"/> - </sequence> -</complexType> - -<complexType name="object"> - <sequence> - <element name="s-one" type="boolean"/> - <element name="c-one" type="complex"/> - <element name="optional" type="int" minOccurs="0"/> - <element name="sequence" type="string" maxOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class complex: public xml_schema::type -{ -public: - object (const int& a, const xml_schema::string& b); - object (const complex&); - -public: - object& - operator= (const complex&); - -public: - virtual complex* - _clone () const; - - ... - -}; - -class object: public xml_schema::type -{ -public: - object (const bool& s_one, const complex& c_one); - object (const bool& s_one, std::[unique|auto]_ptr<complex> c_one); - object (const object&); - -public: - object& - operator= (const object&); - -public: - virtual object* - _clone () const; - - ... - -}; - </pre> - - <p>Notice that the generated <code>complex</code> class does not - have the second (<code>std::[unique|auto]_ptr</code>) version of the - constructor since all its required members are of simple types.</p> - - <p>If an XML Schema complex type has an ultimate base which is an XML - Schema simple type then the resulting C++ class also defines a public - constructor that takes an initializer for the base type as well as - for each member of the complex type and all its base types that - belongs to the One cardinality class. For instance:</p> - - <pre class="xml"> -<complexType name="object"> - <simpleContent> - <extension base="date"> - <attribute name="lang" type="language" use="required"/> - </extension> - </simpleContent> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::string -{ -public: - object (const xml_schema::language& lang); - - object (const xml_schema::date& base, - const xml_schema::language& lang); - - ... - -}; - </pre> - - <p>Furthermore, for string-based XML Schema complex types, the resulting C++ - class also defines two public constructors with the first arguments - of type <code>const C*</code> and <code>std::basic_string<C>&</code>, - respectively, followed by arguments for each member of the complex - type and all its base types that belongs to the One cardinality - class. For enumeration-based complex types the resulting C++ - class also defines a public constructor with the first arguments - of the underlying enum type followed by arguments for each member - of the complex type and all its base types that belongs to the One - cardinality class. For instance:</p> - - <pre class="xml"> -<simpleType name="color"> - <restriction base="string"> - <enumeration value="red"/> - <enumeration value="green"/> - <enumeration value="blue"/> - </restriction> -</simpleType> - -<complexType name="object"> - <simpleContent> - <extension base="color"> - <attribute name="lang" type="language" use="required"/> - </extension> - </simpleContent> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class color: public xml_schema::string -{ -public: - enum value - { - red, - green, - blue - }; - -public: - color (value); - color (const C*); - color (const std::basic_string<C>&); - - ... - -}; - -class object: color -{ -public: - object (const color& base, - const xml_schema::language& lang); - - object (const color::value& base, - const xml_schema::language& lang); - - object (const C* base, - const xml_schema::language& lang); - - object (const std::basic_string<C>& base, - const xml_schema::language& lang); - - ... - -}; - </pre> - - <p>Additional constructors can be requested with the - <code>--generate-default-ctor</code> and - <code>--generate-from-base-ctor</code> options. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for details.</p> - - <p>If an XML Schema complex type is not explicitly derived from any type, - the resulting C++ class is derived from <code>xml_schema::type</code>. - In cases where an XML Schema complex type is defined using derivation - by extension or restriction, the resulting C++ base class specification - depends on the type of derivation and is described in the subsequent - sections. - </p> - - <p>The mapping for elements and attributes that are defined in a complex - type is described in <a href="#2.8">Section 2.8, "Mapping for Local - Elements and Attributes"</a>. - </p> - - <h3><a name="2.7.1">2.7.1 Mapping for Derivation by Extension</a></h3> - - <p>XML Schema derivation by extension is mapped to C++ public - inheritance. The base type of the extension becomes the base - type for the resulting C++ class. - </p> - - <h3><a name="2.7.2">2.7.2 Mapping for Derivation by Restriction</a></h3> - - <p>XML Schema derivation by restriction is mapped to C++ public - inheritance. The base type of the restriction becomes the base - type for the resulting C++ class. XML Schema elements and - attributes defined within restriction do not result in any - definitions in the resulting C++ class. Instead, corresponding - (unrestricted) definitions are inherited from the base class. - In the future versions of this mapping, such elements and - attributes may result in redefinitions of accessors and - modifiers to reflect their restricted semantics. - </p> - - <!-- 2.8 Mapping for Local Elements and Attributes --> - - <h2><a name="2.8">2.8 Mapping for Local Elements and Attributes</a></h2> - - <p>XML Schema element and attribute definitions are called local - if they appear within a complex type definition, an element group - definition, or an attribute group definitions. - </p> - - <p>Local XML Schema element and attribute definitions have the same - C++ mapping. Therefore, in this section, local elements and - attributes are collectively called members. - </p> - - <p>While there are many different member cardinality combinations - (determined by the <code>use</code> attribute for attributes and - the <code>minOccurs</code> and <code>maxOccurs</code> attributes - for elements), the mapping divides all possible cardinality - combinations into three cardinality classes: - </p> - - <dl> - <dt><i>one</i></dt> - <dd>attributes: <code>use == "required"</code></dd> - <dd>attributes: <code>use == "optional"</code> and has default or fixed value</dd> - <dd>elements: <code>minOccurs == "1"</code> and <code>maxOccurs == "1"</code></dd> - - <dt><i>optional</i></dt> - <dd>attributes: <code>use == "optional"</code> and doesn't have default or fixed value</dd> - <dd>elements: <code>minOccurs == "0"</code> and <code>maxOccurs == "1"</code></dd> - - <dt><i>sequence</i></dt> - <dd>elements: <code>maxOccurs > "1"</code></dd> - </dl> - - <p>An optional attribute with a default or fixed value acquires this value - if the attribute hasn't been specified in an instance document (see - <a href="#A">Appendix A, "Default and Fixed Values"</a>). This - mapping places such optional attributes to the One cardinality - class.</p> - - <p>A member is mapped to a set of public type definitions - (<code>typedef</code>s) and a set of public accessor and modifier - functions. Type definitions have names derived from the member's - name. The accessor and modifier functions have the same name as the - member. For example: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - typedef xml_schema::string member_type; - - const member_type& - member () const; - - ... - -}; - </pre> - - <p>In addition, if a member has a default or fixed value, a static - accessor function is generated that returns this value. For - example:</p> - -<pre class="xml"> -<complexType name="object"> - <attribute name="data" type="string" default="test"/> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - typedef xml_schema::string data_type; - - const data_type& - data () const; - - static const data_type& - data_default_value (); - - ... - -}; - </pre> - - <p>Names and semantics of type definitions for the member as well - as signatures of the accessor and modifier functions depend on - the member's cardinality class and are described in the following - sub-sections. - </p> - - - <h3><a name="2.8.1">2.8.1 Mapping for Members with the One Cardinality Class</a></h3> - - <p>For the One cardinality class, the type definitions consist of - an alias for the member's type with the name created by appending - the <code>_type</code> suffix to the member's name. - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - member and can be used for read-only access. The non-constant - version returns an unrestricted reference to the member and can - be used for read-write access. - </p> - - <p>The first modifier function expects an argument of type reference to - constant of the member's type. It makes a deep copy of its argument. - Except for member's types that are mapped to fundamental C++ types, - the second modifier function is provided that expects an argument - of type automatic pointer (<code>std::unique_ptr</code> or - <code>std::auto_ptr</code>, depending on the C++ standard selected) - to the member's type. It assumes ownership of the pointed to object - and resets the passed automatic pointer. For instance:</p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef xml_schema::string member_type; - - // Accessors. - // - const member_type& - member () const; - - member_type& - member (); - - // Modifiers. - // - void - member (const member_type&); - - void - member (std::[unique|auto]_ptr<member_type>); - ... - -}; - </pre> - - <p>In addition, if requested by specifying the <code>--generate-detach</code> - option and only for members of non-fundamental C++ types, the mapping - provides a detach function that returns an automatic pointer to the - member's type, for example:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - ... - - std::[unique|auto]_ptr<member_type> - detach_member (); - ... - -}; - </pre> - - <p>This function detaches the value from the tree leaving the member - value uninitialized. Accessing such an uninitialized value prior to - re-initializing it results in undefined behavior.</p> - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o) -{ - using xml_schema::string; - - string s (o.member ()); // get - object::member_type& sr (o.member ()); // get - - o.member ("hello"); // set, deep copy - o.member () = "hello"; // set, deep copy - - // C++11 version. - // - std::unique_ptr<string> p (new string ("hello")); - o.member (std::move (p)); // set, assumes ownership - p = o.detach_member (); // detach, member is uninitialized - o.member (std::move (p)); // re-attach - - // C++98 version. - // - std::auto_ptr<string> p (new string ("hello")); - o.member (p); // set, assumes ownership - p = o.detach_member (); // detach, member is uninitialized - o.member (p); // re-attach -} - </pre> - - -<h3><a name="2.8.2">2.8.2 Mapping for Members with the Optional Cardinality Class</a></h3> - - <p>For the Optional cardinality class, the type definitions consist of - an alias for the member's type with the name created by appending - the <code>_type</code> suffix to the member's name and an alias for - the container type with the name created by appending the - <code>_optional</code> suffix to the member's name. - </p> - - <p>Unlike accessor functions for the One cardinality class, accessor - functions for the Optional cardinality class return references to - corresponding containers rather than directly to members. The - accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to - the container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container - and can be used for read-write access. - </p> - - <p>The modifier functions are overloaded for the member's - type and the container type. The first modifier function - expects an argument of type reference to constant of the - member's type. It makes a deep copy of its argument. - Except for member's types that are mapped to fundamental C++ types, - the second modifier function is provided that expects an argument - of type automatic pointer (<code>std::unique_ptr</code> or - <code>std::auto_ptr</code>, depending on the C++ standard selected) - to the member's type. It assumes ownership of the pointed to object - and resets the passed automatic pointer. The last modifier function - expects an argument of type reference to constant of the container - type. It makes a deep copy of its argument. For instance: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string" minOccurs="0"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef xml_schema::string member_type; - typedef optional<member_type> member_optional; - - // Accessors. - // - const member_optional& - member () const; - - member_optional& - member (); - - // Modifiers. - // - void - member (const member_type&); - - void - member (std::[unique|auto]_ptr<member_type>); - - void - member (const member_optional&); - - ... - -}; - </pre> - - - <p>The <code>optional</code> class template is defined in an - implementation-specific namespace and has the following - interface. The <code>[unique|auto]_ptr</code>-based constructor - and modifier function are only available if the template - argument is not a fundamental C++ type. - </p> - - <pre class="c++"> -template <typename X> -class optional -{ -public: - optional (); - - // Makes a deep copy. - // - explicit - optional (const X&); - - // Assumes ownership. - // - explicit - optional (std::[unique|auto]_ptr<X>); - - optional (const optional&); - -public: - optional& - operator= (const X&); - - optional& - operator= (const optional&); - - // Pointer-like interface. - // -public: - const X* - operator-> () const; - - X* - operator-> (); - - const X& - operator* () const; - - X& - operator* (); - - typedef void (optional::*bool_convertible) (); - operator bool_convertible () const; - - // Get/set interface. - // -public: - bool - present () const; - - const X& - get () const; - - X& - get (); - - // Makes a deep copy. - // - void - set (const X&); - - // Assumes ownership. - // - void - set (std::[unique|auto]_ptr<X>); - - // Detach and return the contained value. - // - std::[unique|auto]_ptr<X> - detach (); - - void - reset (); -}; - -template <typename X> -bool -operator== (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator!= (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator< (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator> (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator<= (const optional<X>&, const optional<X>&); - -template <typename X> -bool -operator>= (const optional<X>&, const optional<X>&); - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o) -{ - using xml_schema::string; - - if (o.member ().present ()) // test - { - string& s (o.member ().get ()); // get - o.member ("hello"); // set, deep copy - o.member ().set ("hello"); // set, deep copy - o.member ().reset (); // reset - } - - // Same as above but using pointer notation: - // - if (o.member ()) // test - { - string& s (*o.member ()); // get - o.member ("hello"); // set, deep copy - *o.member () = "hello"; // set, deep copy - o.member ().reset (); // reset - } - - // C++11 version. - // - std::unique_ptr<string> p (new string ("hello")); - o.member (std::move (p)); // set, assumes ownership - - p.reset (new string ("hello")); - o.member ().set (std::move (p)); // set, assumes ownership - - p = o.member ().detach (); // detach, member is reset - o.member ().set (std::move (p)); // re-attach - - // C++98 version. - // - std::auto_ptr<string> p (new string ("hello")); - o.member (p); // set, assumes ownership - - p = new string ("hello"); - o.member ().set (p); // set, assumes ownership - - p = o.member ().detach (); // detach, member is reset - o.member ().set (p); // re-attach -} - </pre> - - - <h3><a name="2.8.3">2.8.3 Mapping for Members with the Sequence Cardinality Class</a></h3> - - <p>For the Sequence cardinality class, the type definitions consist of an - alias for the member's type with the name created by appending - the <code>_type</code> suffix to the member's name, an alias of - the container type with the name created by appending the - <code>_sequence</code> suffix to the member's name, an alias of - the iterator type with the name created by appending the - <code>_iterator</code> suffix to the member's name, and an alias - of the constant iterator type with the name created by appending the - <code>_const_iterator</code> suffix to the member's name. - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container and can - be used for read-write access. - </p> - - <p>The modifier function expects an argument of type reference to - constant of the container type. The modifier function - makes a deep copy of its argument. For instance: - </p> - - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="member" type="string" minOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef xml_schema::string member_type; - typedef sequence<member_type> member_sequence; - typedef member_sequence::iterator member_iterator; - typedef member_sequence::const_iterator member_const_iterator; - - // Accessors. - // - const member_sequence& - member () const; - - member_sequence& - member (); - - // Modifier. - // - void - member (const member_sequence&); - - ... - -}; - </pre> - - <p>The <code>sequence</code> class template is defined in an - implementation-specific namespace. It conforms to the - sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences"). - Practically, this means that you can treat such a sequence - as if it was <code>std::vector</code>. Two notable extensions - to the standard interface that are available only for - sequences of non-fundamental C++ types are the addition of - the overloaded <code>push_back</code> and <code>insert</code> - as well as the <code>detach_back</code> and <code>detach</code> - member functions. The additional <code>push_back</code> and - <code>insert</code> functions accept an automatic pointer - (<code>std::unique_ptr</code> or <code>std::auto_ptr</code>, - depending on the C++ standard selected) to the - element type instead of the constant reference. They assume - ownership of the pointed to object and reset the passed - automatic pointer. The <code>detach_back</code> and - <code>detach</code> functions detach the element - value from the sequence container and, by default, remove - the element from the sequence. These additional functions - have the following signatures:</p> - - <pre class="c++"> -template <typename X> -class sequence -{ -public: - ... - - void - push_back (std::[unique|auto]_ptr<X>) - - iterator - insert (iterator position, std::[unique|auto]_ptr<X>) - - std::[unique|auto]_ptr<X> - detach_back (bool pop = true); - - iterator - detach (iterator position, - std::[unique|auto]_ptr<X>& result, - bool erase = true) - - ... -} - </pre> - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o) -{ - using xml_schema::string; - - object::member_sequence& s (o.member ()); - - // Iteration. - // - for (object::member_iterator i (s.begin ()); i != s.end (); ++i) - { - string& value (*i); - } - - // Modification. - // - s.push_back ("hello"); // deep copy - - // C++11 version. - // - std::unique_ptr<string> p (new string ("hello")); - s.push_back (std::move (p)); // assumes ownership - p = s.detach_back (); // detach and pop - s.push_back (std::move (p)); // re-append - - // C++98 version. - // - std::auto_ptr<string> p (new string ("hello")); - s.push_back (p); // assumes ownership - p = s.detach_back (); // detach and pop - s.push_back (p); // re-append - - // Setting a new container. - // - object::member_sequence n; - n.push_back ("one"); - n.push_back ("two"); - o.member (n); // deep copy -} - </pre> - - <h3><a name="2.8.4">2.8.4 Element Order</a></h3> - - <p>C++/Tree is a "flattening" mapping in a sense that many levels of - nested compositors (<code>choice</code> and <code>sequence</code>), - all potentially with their own cardinalities, are in the end mapped - to a flat set of elements with one of the three cardinality classes - discussed in the previous sections. While this results in a simple - and easy to use API for most types, in certain cases, the order of - elements in the actual XML documents is not preserved once parsed - into the object model. And sometimes such order has - application-specific significance. As an example, consider a schema - that defines a batch of bank transactions:</p> - - <pre class="xml"> -<complexType name="withdraw"> - <sequence> - <element name="account" type="unsignedInt"/> - <element name="amount" type="unsignedInt"/> - </sequence> -</complexType> - -<complexType name="deposit"> - <sequence> - <element name="account" type="unsignedInt"/> - <element name="amount" type="unsignedInt"/> - </sequence> -</complexType> - -<complexType name="batch"> - <choice minOccurs="0" maxOccurs="unbounded"> - <element name="withdraw" type="withdraw"/> - <element name="deposit" type="deposit"/> - </choice> -</complexType> - </pre> - - <p>The batch can contain any number of transactions in any order - but the order of transactions in each actual batch is significant. - For instance, consider what could happen if we reorder the - transactions and apply all the withdrawals before deposits.</p> - - <p>For the <code>batch</code> schema type defined above the default - C++/Tree mapping will produce a C++ class that contains a pair of - sequence containers, one for each of the two elements. While this - will capture the content (transactions), the order of this content - as it appears in XML will be lost. Also, if we try to serialize the - batch we just loaded back to XML, all the withdrawal transactions - will appear before deposits.</p> - - <p>To overcome this limitation of a flattening mapping, C++/Tree - allows us to mark certain XML Schema types, for which content - order is important, as ordered.</p> - - <p>There are several command line options that control which - schema types are treated as ordered. To make an individual - type ordered, we use the <code>--ordered-type</code> option, - for example:</p> - - <pre class="term"> ---ordered-type batch - </pre> - - <p>To automatically treat all the types that are derived from an ordered - type also ordered, we use the <code>--ordered-type-derived</code> - option. This is primarily useful if you would like to iterate - over the complete hierarchy's content using the content order - sequence (discussed below).</p> - - <p>Ordered types are also useful for handling mixed content. To - automatically mark all the types with mixed content as ordered - we use the <code>--ordered-type-mixed</code> option. For more - information on handling mixed content see <a href="#2.13">Section - 2.13, "Mapping for Mixed Content Models"</a>.</p> - - <p>Finally, we can mark all the types in the schema we are - compiling with the <code>--ordered-type-all</code> option. - You should only resort to this option if all the types in - your schema truly suffer from the loss of content - order since, as we will discuss shortly, ordered types - require extra effort to access and, especially, modify. - See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information on - these options.</p> - - <p>Once a type is marked ordered, C++/Tree alters its mapping - in several ways. Firstly, for each local element, element - wildcard (<a href="#2.12.4">Section 2.12.4, "Element Wildcard - Order"</a>), and mixed content text (<a href="#2.13">Section - 2.13, "Mapping for Mixed Content Models"</a>) in this type, a - content id constant is generated. Secondly, an addition sequence - is added to the class that captures the content order. Here - is how the mapping of our <code>batch</code> class changes - once we make it ordered:</p> - - <pre class="c++"> -class batch: public xml_schema::type -{ -public: - // withdraw - // - typedef withdraw withdraw_type; - typedef sequence<withdraw_type> withdraw_sequence; - typedef withdraw_sequence::iterator withdraw_iterator; - typedef withdraw_sequence::const_iterator withdraw_const_iterator; - - static const std::size_t withdraw_id = 1; - - const withdraw_sequence& - withdraw () const; - - withdraw_sequence& - withdraw (); - - void - withdraw (const withdraw_sequence&); - - // deposit - // - typedef deposit deposit_type; - typedef sequence<deposit_type> deposit_sequence; - typedef deposit_sequence::iterator deposit_iterator; - typedef deposit_sequence::const_iterator deposit_const_iterator; - - static const std::size_t deposit_id = 2; - - const deposit_sequence& - deposit () const; - - deposit_sequence& - deposit (); - - void - deposit (const deposit_sequence&); - - // content_order - // - typedef xml_schema::content_order content_order_type; - typedef std::vector<content_order_type> content_order_sequence; - typedef content_order_sequence::iterator content_order_iterator; - typedef content_order_sequence::const_iterator content_order_const_iterator; - - const content_order_sequence& - content_order () const; - - content_order_sequence& - content_order (); - - void - content_order (const content_order_sequence&); - - ... -}; - </pre> - - <p>Notice the <code>withdraw_id</code> and <code>deposit_id</code> - content ids as well as the extra <code>content_order</code> - sequence that does not correspond to any element in the - schema definition. The other changes to the mapping for ordered - types has to do with XML parsing and serialization code. During - parsing the content order is captured in the <code>content_order</code> - sequence while during serialization this sequence is used to - determine the order in which content is serialized. The - <code>content_order</code> sequence is also copied during - copy construction and assigned during copy assignment. It is also - taken into account during comparison.</p> - - <p>The entry type of the <code>content_order</code> sequence is the - <code>xml_schema::content_order</code> type that has the following - interface:</p> - - <pre class="c++"> -namespace xml_schema -{ - struct content_order - { - content_order (std::size_t id, std::size_t index = 0); - - std::size_t id; - std::size_t index; - }; - - bool - operator== (const content_order&, const content_order&); - - bool - operator!= (const content_order&, const content_order&); - - bool - operator< (const content_order&, const content_order&); -} - </pre> - - <p>The <code>content_order</code> sequence describes the order of - content (elements, including wildcards, as well as mixed content - text). Each entry in this sequence consists of the content id - (for example, <code>withdraw_id</code> or <code>deposit_id</code> - in our case) as well as, for elements of the sequence cardinality - class, an index into the corresponding sequence container (the - index is unused for the one and optional cardinality classes). - For example, in our case, if the content id is <code>withdraw_id</code>, - then the index will point into the <code>withdraw</code> element - sequence.</p> - - <p>With all this information we can now examine how to iterate over - transaction in the batch in content order:</p> - - <pre class="c++"> -batch& b = ... - -for (batch::content_order_const_iterator i (b.content_order ().begin ()); - i != b.content_order ().end (); - ++i) -{ - switch (i->id) - { - case batch::withdraw_id: - { - const withdraw& t (b.withdraw ()[i->index]); - cerr << t.account () << " withdraw " << t.amount () << endl; - break; - } - case batch::deposit_id: - { - const deposit& t (b.deposit ()[i->index]); - cerr << t.account () << " deposit " << t.amount () << endl; - break; - } - default: - { - assert (false); // Unknown content id. - } - } -} - </pre> - - <p>If we serialized our batch back to XML, we would also see that the - order of transactions in the output is exactly the same as in the - input rather than all the withdrawals first followed by all the - deposits.</p> - - <p>The most complex aspect of working with ordered types is - modifications. Now we not only need to change the content, - but also remember to update the order information corresponding - to this change. As a first example, we add a deposit transaction - to the batch:</p> - - <pre class="c++"> -using xml_schema::content_order; - -batch::deposit_sequence& d (b.deposit ()); -batch::withdraw_sequence& w (b.withdraw ()); -batch::content_order_sequence& co (b.content_order ()); - -d.push_back (deposit (123456789, 100000)); -co.push_back (content_order (batch::deposit_id, d.size () - 1)); - </pre> - - <p>In the above example we first added the content (deposit - transaction) and then updated the content order information - by adding an entry with <code>deposit_id</code> content - id and the index of the just added deposit transaction.</p> - - <p>Removing the last transaction can be easy if we know which - transaction (deposit or withdrawal) is last:</p> - - <pre class="c++"> -d.pop_back (); -co.pop_back (); - </pre> - - <p>If, however, we do not know which transaction is last, then - things get a bit more complicated:</p> - - <pre class="c++"> -switch (co.back ().id) -{ -case batch::withdraw_id: - { - d.pop_back (); - break; - } -case batch::deposit_id: - { - w.pop_back (); - break; - } -} - -co.pop_back (); - </pre> - - <p>The following example shows how to add a transaction at the - beginning of the batch:</p> - - <pre class="c++"> -w.push_back (withdraw (123456789, 100000)); -co.insert (co.begin (), - content_order (batch::withdraw_id, w.size () - 1)); - </pre> - - <p>Note also that when we merely modify the content of one - of the elements in place, we do not need to update its - order since it doesn't change. For example, here is how - we can change the amount in the first withdrawal:</p> - - <pre class="c++"> -w[0].amount (10000); - </pre> - - <p>For the complete working code shown in this section refer to the - <code>order/element</code> example in the - <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <p>If both the base and derived types are ordered, then the - content order sequence is only added to the base and the content - ids are unique within the whole hierarchy. In this case - the content order sequence for the derived type contains - ordering information for both base and derived content.</p> - - <p>In some applications we may need to perform more complex - content processing. For example, in our case, we may need - to remove all the withdrawal transactions. The default - container, <code>std::vector</code>, is not particularly - suitable for such operations. What may be required by - some applications is a multi-index container that not - only allows us to iterate in content order similar to - <code>std::vector</code> but also search by the content - id as well as the content id and index pair.</p> - - <p>While C++/Tree does not provide this functionality by - default, it allows us to specify a custom container - type for content order with the <code>--order-container</code> - command line option. The only requirement from the - generated code side for such a container is to provide - the <code>vector</code>-like <code>push_back()</code>, - <code>size()</code>, and const iteration interfaces.</p> - - <p>As an example, here is how we can use the Boost Multi-Index - container for content order. First we create the - <code>content-order-container.hxx</code> header with the - following definition:</p> - - <pre class="c++"> -#ifndef CONTENT_ORDER_CONTAINER -#define CONTENT_ORDER_CONTAINER - -#include <cstddef> // std::size_t - -#include <boost/multi_index_container.hpp> -#include <boost/multi_index/member.hpp> -#include <boost/multi_index/identity.hpp> -#include <boost/multi_index/ordered_index.hpp> -#include <boost/multi_index/random_access_index.hpp> - -struct by_id {}; -struct by_id_index {}; - -template <typename T> -using content_order_container = - boost::multi_index::multi_index_container< - T, - boost::multi_index::indexed_by< - boost::multi_index::random_access<>, - boost::multi_index::ordered_unique< - boost::multi_index::tag<by_id_index>, - boost::multi_index::identity<T> - >, - boost::multi_index::ordered_non_unique< - boost::multi_index::tag<by_id>, - boost::multi_index::member<T, std::size_t, &T::id> - > - > - >; - -#endif - </pre> - - <p>Next we add the following two XSD compiler options to include - this header into every generated header file and to use the - custom container type (see the XSD compiler command line manual - for more information on shell quoting for the first option):</p> - - <pre class="term"> ---hxx-prologue '#include "content-order-container.hxx"' ---order-container content_order_container - </pre> - - <p>With these changes we can now use the multi-index functionality, - for example, to search for a specific content id:</p> - - <pre class="c++"> -typedef batch::content_order_sequence::index<by_id>::type id_set; -typedef id_set::iterator id_iterator; - -const id_set& ids (b.content_order ().get<by_id> ()); - -std::pair<id_iterator, id_iterator> r ( - ids.equal_range (std::size_t (batch::deposit_id)); - -for (id_iterator i (r.first); i != r.second; ++i) -{ - const deposit& t (b.deposit ()[i->index]); - cerr << t.account () << " deposit " << t.amount () << endl; -} - </pre> - - <h2><a name="2.9">2.9 Mapping for Global Elements</a></h2> - - <p>An XML Schema element definition is called global if it appears - directly under the <code>schema</code> element. - A global element is a valid root of an instance document. By - default, a global element is mapped to a set of overloaded - parsing and, optionally, serialization functions with the - same name as the element. It is also possible to generate types - for root elements instead of parsing and serialization functions. - This is primarily useful to distinguish object models with the - same root type but with different root elements. See - <a href="#2.9.1">Section 2.9.1, "Element Types"</a> for details. - It is also possible to request the generation of an element map - which allows uniform parsing and serialization of multiple root - elements. See <a href="#2.9.2">Section 2.9.2, "Element Map"</a> - for details. - </p> - - <p>The parsing functions read XML instance documents and return - corresponding object models as an automatic pointer - (<code>std::unique_ptr</code> or <code>std::auto_ptr</code>, - depending on the C++ standard selected). Their signatures - have the following pattern (<code>type</code> denotes - element's type and <code>name</code> denotes element's - name): - </p> - - <pre class="c++"> -std::[unique|auto]_ptr<type> -name (....); - </pre> - - <p>The process of parsing, including the exact signatures of the parsing - functions, is the subject of <a href="#3">Chapter 3, "Parsing"</a>. - </p> - - <p>The serialization functions write object models back to XML instance - documents. Their signatures have the following pattern: - </p> - - <pre class="c++"> -void -name (<stream type>&, const type&, ....); - </pre> - - <p>The process of serialization, including the exact signatures of the - serialization functions, is the subject of <a href="#4">Chapter 4, - "Serialization"</a>. - </p> - - - <h3><a name="2.9.1">2.9.1 Element Types</a></h3> - - <p>The generation of element types is requested with the - <code>--generate-element-type</code> option. With this option - each global element is mapped to a C++ class with the - same name as the element. Such a class is derived from - <code>xml_schema::element_type</code> and contains the same set - of type definitions, constructors, and member function as would a - type containing a single element with the One cardinality class - named <code>"value"</code>. In addition, the element type also - contains a set of member functions for accessing the element - name and namespace as well as its value in a uniform manner. - For example:</p> - - <pre class="xml"> -<complexType name="type"> - <sequence> - ... - </sequence> -</complexType> - -<element name="root" type="type"/> - </pre> - -<p>is mapped to:</p> - - <pre class="c++"> -class type -{ - ... -}; - -class root: public xml_schema::element_type -{ -public: - // Element value. - // - typedef type value_type; - - const value_type& - value () const; - - value_type& - value (); - - void - value (const value_type&); - - void - value (std::[unique|auto]_ptr<value_type>); - - // Constructors. - // - root (const value_type&); - - root (std::[unique|auto]_ptr<value_type>); - - root (const xercesc::DOMElement&, xml_schema::flags = 0); - - root (const root&, xml_schema::flags = 0); - - virtual root* - _clone (xml_schema::flags = 0) const; - - // Element name and namespace. - // - static const std::string& - name (); - - static const std::string& - namespace_ (); - - virtual const std::string& - _name () const; - - virtual const std::string& - _namespace () const; - - // Element value as xml_schema::type. - // - virtual const xml_schema::type* - _value () const; - - virtual xml_schema::type* - _value (); -}; - -void -operator<< (xercesc::DOMElement&, const root&); - </pre> - - <p>The <code>xml_schema::element_type</code> class is a common - base type for all element types and is defined as follows:</p> - - <pre class="c++"> -namespace xml_schema -{ - class element_type - { - public: - virtual - ~element_type (); - - virtual element_type* - _clone (flags f = 0) const = 0; - - virtual const std::basic_string<C>& - _name () const = 0; - - virtual const std::basic_string<C>& - _namespace () const = 0; - - virtual xml_schema::type* - _value () = 0; - - virtual const xml_schema::type* - _value () const = 0; - }; -} - </pre> - - <p>The <code>_value()</code> member function returns a pointer to - the element value or 0 if the element is of a fundamental C++ - type and therefore is not derived from <code>xml_schema::type</code>. - </p> - - <p>Unlike parsing and serialization functions, element types - are only capable of parsing and serializing from/to a - <code>DOMElement</code> object. This means that the application - will need to perform its own XML-to-DOM parsing and DOM-to-XML - serialization. The following section describes a mechanism - provided by the mapping to uniformly parse and serialize - multiple root elements.</p> - - - <h3><a name="2.9.2">2.9.2 Element Map</a></h3> - - <p>When element types are generated for root elements it is also - possible to request the generation of an element map with the - <code>--generate-element-map</code> option. The element map - allows uniform parsing and serialization of multiple root - elements via the common <code>xml_schema::element_type</code> - base type. The <code>xml_schema::element_map</code> class is - defined as follows:</p> - - <pre class="c++"> -namespace xml_schema -{ - class element_map - { - public: - static std::[unique|auto]_ptr<xml_schema::element_type> - parse (const xercesc::DOMElement&, flags = 0); - - static void - serialize (xercesc::DOMElement&, const element_type&); - }; -} - </pre> - - <p>The <code>parse()</code> function creates the corresponding - element type object based on the element name and namespace - and returns it as an automatic pointer (<code>std::unique_ptr</code> - or <code>std::auto_ptr</code>, depending on the C++ standard - selected) to <code>xml_schema::element_type</code>. - The <code>serialize()</code> function serializes the passed element - object to <code>DOMElement</code>. Note that in case of - <code>serialize()</code>, the <code>DOMElement</code> object - should have the correct name and namespace. If no element type is - available for an element, both functions throw the - <code>xml_schema::no_element_info</code> exception:</p> - - <pre class="c++"> -struct no_element_info: virtual exception -{ - no_element_info (const std::basic_string<C>& element_name, - const std::basic_string<C>& element_namespace); - - const std::basic_string<C>& - element_name () const; - - const std::basic_string<C>& - element_namespace () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The application can discover the actual type of the element - object returned by <code>parse()</code> either using - <code>dynamic_cast</code> or by comparing element names and - namespaces. The following code fragments illustrate how the - element map can be used:</p> - - <pre class="c++"> -// Parsing. -// -DOMElement& e = ... // Parse XML to DOM. - -unique_ptr<xml_schema::element_type> r ( - xml_schema::element_map::parse (e)); - -if (root1 r1 = dynamic_cast<root1*> (r.get ())) -{ - ... -} -else if (r->_name == root2::name () && - r->_namespace () == root2::namespace_ ()) -{ - root2& r2 (static_cast<root2&> (*r)); - - ... -} - </pre> - - <pre class="c++"> -// Serialization. -// -xml_schema::element_type& r = ... - -string name (r._name ()); -string ns (r._namespace ()); - -DOMDocument& doc = ... // Create a new DOMDocument with name and ns. -DOMElement& e (*doc->getDocumentElement ()); - -xml_schema::element_map::serialize (e, r); - -// Serialize DOMDocument to XML. - </pre> - - <!-- --> - - <h2><a name="2.10">2.10 Mapping for Global Attributes</a></h2> - - <p>An XML Schema attribute definition is called global if it appears - directly under the <code>schema</code> element. A global - attribute does not have any mapping. - </p> - - <!-- - When it is referenced from - a local attribute definition (using the <code>ref</code> attribute) - it is treated as a local attribute (see Section 2.8, "Mapping for - Local Elements and Attributes"). - --> - - <h2><a name="2.11">2.11 Mapping for <code>xsi:type</code> and Substitution - Groups</a></h2> - - <p>The mapping provides optional support for the XML Schema polymorphism - features (<code>xsi:type</code> and substitution groups) which can - be requested with the <code>--generate-polymorphic</code> option. - When used, the dynamic type of a member may be different from - its static type. Consider the following schema definition and - instance document: - </p> - - <pre class="xml"> -<!-- test.xsd --> -<schema> - <complexType name="base"> - <attribute name="text" type="string"/> - </complexType> - - <complexType name="derived"> - <complexContent> - <extension base="base"> - <attribute name="extra-text" type="string"/> - </extension> - </complexContent> - </complexType> - - <complexType name="root_type"> - <sequence> - <element name="item" type="base" maxOccurs="unbounded"/> - </sequence> - </complexType> - - <element name="root" type="root_type"/> -</schema> - -<!-- test.xml --> -<root xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"> - <item text="hello"/> - <item text="hello" extra-text="world" xsi:type="derived"/> -</root> - </pre> - - <p>In the resulting object model, the container for - the <code>root::item</code> member will have two elements: - the first element's type will be <code>base</code> while - the second element's (dynamic) type will be - <code>derived</code>. This can be discovered using the - <code>dynamic_cast</code> operator as shown in the following - example: - </p> - - <pre class="c++"> -void -f (root& r) -{ - for (root::item_const_iterator i (r.item ().begin ()); - i != r.item ().end () - ++i) - { - if (derived* d = dynamic_cast<derived*> (&(*i))) - { - // derived - } - else - { - // base - } - } -} - </pre> - - <p>The <code>_clone</code> virtual function should be used instead of - copy constructors to make copies of members that might use - polymorphism: - </p> - - <pre class="c++"> -void -f (root& r) -{ - for (root::item_const_iterator i (r.item ().begin ()); - i != r.item ().end () - ++i) - { - std::unique_ptr<base> c (i->_clone ()); - } -} - </pre> - - <p>The mapping can often automatically determine which types are - polymorphic based on the substitution group declarations. However, - if your XML vocabulary is not using substitution groups or if - substitution groups are defined in a separate schema, then you will - need to use the <code>--polymorphic-type</code> option to specify - which types are polymorphic. When using this option you only need - to specify the root of a polymorphic type hierarchy and the mapping - will assume that all the derived types are also polymorphic. - Also note that you need to specify this option when compiling every - schema file that references the polymorphic type. Consider the following - two schemas as an example:</p> - - <pre class="xml"> -<!-- base.xsd --> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <xs:complexType name="base"> - <xs:sequence> - <xs:element name="b" type="xs:int"/> - </xs:sequence> - </xs:complexType> - - <!-- substitution group root --> - <xs:element name="base" type="base"/> - -</xs:schema> - </pre> - - <pre class="xml"> -<!-- derived.xsd --> -<xs:schema xmlns:xs="http://www.w3.org/2001/XMLSchema"> - - <include schemaLocation="base.xsd"/> - - <xs:complexType name="derived"> - <xs:complexContent> - <xs:extension base="base"> - <xs:sequence> - <xs:element name="d" type="xs:string"/> - </xs:sequence> - </xs:extension> - </xs:complexContent> - </xs:complexType> - - <xs:element name="derived" type="derived" substitutionGroup="base"/> - -</xs:schema> - </pre> - - <p>In this example we need to specify "<code>--polymorphic-type base</code>" - when compiling both schemas because the substitution group is declared - in a schema other than the one defining type <code>base</code>.</p> - - <p>You can also indicate that all types should be treated as polymorphic - with the <code>--polymorphic-type-all</code>. However, this may result - in slower generated code with a greater footprint.</p> - - - <!-- Mapping for any and anyAttribute --> - - - <h2><a name="2.12">2.12 Mapping for <code>any</code> and <code>anyAttribute</code></a></h2> - - <p>For the XML Schema <code>any</code> and <code>anyAttribute</code> - wildcards an optional mapping can be requested with the - <code>--generate-wildcard</code> option. The mapping represents - the content matched by wildcards as DOM fragments. Because the - DOM API is used to access such content, the Xerces-C++ runtime - should be initialized by the application prior to parsing and - should remain initialized for the lifetime of objects with - the wildcard content. For more information on the Xerces-C++ - runtime initialization see <a href="#3.1">Section 3.1, - "Initializing the Xerces-C++ Runtime"</a>. - </p> - - <p>The mapping for <code>any</code> is similar to the mapping for - local elements (see <a href="#2.8">Section 2.8, "Mapping for Local - Elements and Attributes"</a>) except that the type used in the - wildcard mapping is <code>xercesc::DOMElement</code>. As with local - elements, the mapping divides all possible cardinality combinations - into three cardinality classes: <i>one</i>, <i>optional</i>, and - <i>sequence</i>. - </p> - - <p>The mapping for <code>anyAttribute</code> represents the attributes - matched by this wildcard as a set of <code>xercesc::DOMAttr</code> - objects with a key being the attribute's name and namespace.</p> - - <p>Similar to local elements and attributes, the <code>any</code> and - <code>anyAttribute</code> wildcards are mapped to a set of public type - definitions (typedefs) and a set of public accessor and modifier - functions. Type definitions have names derived from <code>"any"</code> - for the <code>any</code> wildcard and <code>"any_attribute"</code> - for the <code>anyAttribute</code> wildcard. The accessor and modifier - functions are named <code>"any"</code> for the <code>any</code> wildcard - and <code>"any_attribute"</code> for the <code>anyAttribute</code> - wildcard. Subsequent wildcards in the same type have escaped names - such as <code>"any1"</code> or <code>"any_attribute1"</code>. - </p> - - <p>Because Xerces-C++ DOM nodes always belong to a <code>DOMDocument</code>, - each type with a wildcard has an associated <code>DOMDocument</code> - object. The reference to this object can be obtained using the accessor - function called <code>dom_document</code>. The access to the document - object from the application code may be necessary to create or modify - the wildcard content. For example: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other"/> - </sequence> - <anyAttribute namespace="##other"/> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // any - // - const xercesc::DOMElement& - any () const; - - void - any (const xercesc::DOMElement&); - - ... - - // any_attribute - // - typedef attribute_set any_attribute_set; - typedef any_attribute_set::iterator any_attribute_iterator; - typedef any_attribute_set::const_iterator any_attribute_const_iterator; - - const any_attribute_set& - any_attribute () const; - - any_attribute_set& - any_attribute (); - - ... - - // DOMDocument object for wildcard content. - // - const xercesc::DOMDocument& - dom_document () const; - - xercesc::DOMDocument& - dom_document (); - - ... -}; - </pre> - - - <p>Names and semantics of type definitions for the wildcards as well - as signatures of the accessor and modifier functions depend on the - wildcard type as well as the cardinality class for the <code>any</code> - wildcard. They are described in the following sub-sections. - </p> - - - <h3><a name="2.12.1">2.12.1 Mapping for <code>any</code> with the One Cardinality Class</a></h3> - - <p>For <code>any</code> with the One cardinality class, - there are no type definitions. The accessor functions come in - constant and non-constant versions. The constant accessor function - returns a constant reference to <code>xercesc::DOMElement</code> and - can be used for read-only access. The non-constant version returns - an unrestricted reference to <code>xercesc::DOMElement</code> and can - be used for read-write access. - </p> - - <p>The first modifier function expects an argument of type reference - to constant <code>xercesc::DOMElement</code> and makes a deep copy - of its argument. The second modifier function expects an argument of - type pointer to <code>xercesc::DOMElement</code>. This modifier - function assumes ownership of its argument and expects the element - object to be created using the DOM document associated with this - instance. For example: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Accessors. - // - const xercesc::DOMElement& - any () const; - - xercesc::DOMElement& - any (); - - // Modifiers. - // - void - any (const xercesc::DOMElement&); - - void - any (xercesc::DOMElement*); - - ... - -}; - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMElement& e) -{ - using namespace xercesc; - - DOMElement& e1 (o.any ()); // get - o.any (e) // set, deep copy - DOMDocument& doc (o.dom_document ()); - o.any (doc.createElement (...)); // set, assumes ownership -} - </pre> - - <h3><a name="2.12.2">2.12.2 Mapping for <code>any</code> with the Optional Cardinality Class</a></h3> - - <p>For <code>any</code> with the Optional cardinality class, the type - definitions consist of an alias for the container type with name - <code>any_optional</code> (or <code>any1_optional</code>, etc., for - subsequent wildcards in the type definition). - </p> - - <p>Unlike accessor functions for the One cardinality class, accessor - functions for the Optional cardinality class return references to - corresponding containers rather than directly to <code>DOMElement</code>. - The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to - the container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container - and can be used for read-write access. - </p> - - <p>The modifier functions are overloaded for <code>xercesc::DOMElement</code> - and the container type. The first modifier function expects an argument of - type reference to constant <code>xercesc::DOMElement</code> and - makes a deep copy of its argument. The second modifier function - expects an argument of type pointer to <code>xercesc::DOMElement</code>. - This modifier function assumes ownership of its argument and expects - the element object to be created using the DOM document associated - with this instance. The third modifier function expects an argument - of type reference to constant of the container type and makes a - deep copy of its argument. For instance: - </p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other" minOccurs="0"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef element_optional any_optional; - - // Accessors. - // - const any_optional& - any () const; - - any_optional& - any (); - - // Modifiers. - // - void - any (const xercesc::DOMElement&); - - void - any (xercesc::DOMElement*); - - void - any (const any_optional&); - - ... - -}; - </pre> - - - <p>The <code>element_optional</code> container is a - specialization of the <code>optional</code> class template described - in <a href="#2.8.2">Section 2.8.2, "Mapping for Members with the Optional - Cardinality Class"</a>. Its interface is presented below: - </p> - - <pre class="c++"> -class element_optional -{ -public: - explicit - element_optional (xercesc::DOMDocument&); - - // Makes a deep copy. - // - element_optional (const xercesc::DOMElement&, xercesc::DOMDocument&); - - // Assumes ownership. - // - element_optional (xercesc::DOMElement*, xercesc::DOMDocument&); - - element_optional (const element_optional&, xercesc::DOMDocument&); - -public: - element_optional& - operator= (const xercesc::DOMElement&); - - element_optional& - operator= (const element_optional&); - - // Pointer-like interface. - // -public: - const xercesc::DOMElement* - operator-> () const; - - xercesc::DOMElement* - operator-> (); - - const xercesc::DOMElement& - operator* () const; - - xercesc::DOMElement& - operator* (); - - typedef void (element_optional::*bool_convertible) (); - operator bool_convertible () const; - - // Get/set interface. - // -public: - bool - present () const; - - const xercesc::DOMElement& - get () const; - - xercesc::DOMElement& - get (); - - // Makes a deep copy. - // - void - set (const xercesc::DOMElement&); - - // Assumes ownership. - // - void - set (xercesc::DOMElement*); - - void - reset (); -}; - -bool -operator== (const element_optional&, const element_optional&); - -bool -operator!= (const element_optional&, const element_optional&); - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMElement& e) -{ - using namespace xercesc; - - DOMDocument& doc (o.dom_document ()); - - if (o.any ().present ()) // test - { - DOMElement& e1 (o.any ().get ()); // get - o.any ().set (e); // set, deep copy - o.any ().set (doc.createElement (...)); // set, assumes ownership - o.any ().reset (); // reset - } - - // Same as above but using pointer notation: - // - if (o.member ()) // test - { - DOMElement& e1 (*o.any ()); // get - o.any (e); // set, deep copy - o.any (doc.createElement (...)); // set, assumes ownership - o.any ().reset (); // reset - } -} - </pre> - - - - <h3><a name="2.12.3">2.12.3 Mapping for <code>any</code> with the Sequence Cardinality Class</a></h3> - - <p>For <code>any</code> with the Sequence cardinality class, the type - definitions consist of an alias of the container type with name - <code>any_sequence</code> (or <code>any1_sequence</code>, etc., for - subsequent wildcards in the type definition), an alias of the iterator - type with name <code>any_iterator</code> (or <code>any1_iterator</code>, - etc., for subsequent wildcards in the type definition), and an alias - of the constant iterator type with name <code>any_const_iterator</code> - (or <code>any1_const_iterator</code>, etc., for subsequent wildcards - in the type definition). - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container and can - be used for read-write access. - </p> - - <p>The modifier function expects an argument of type reference to - constant of the container type. The modifier function makes - a deep copy of its argument. For instance: - </p> - - - <pre class="xml"> -<complexType name="object"> - <sequence> - <any namespace="##other" minOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef element_sequence any_sequence; - typedef any_sequence::iterator any_iterator; - typedef any_sequence::const_iterator any_const_iterator; - - // Accessors. - // - const any_sequence& - any () const; - - any_sequence& - any (); - - // Modifier. - // - void - any (const any_sequence&); - - ... - -}; - </pre> - - <p>The <code>element_sequence</code> container is a - specialization of the <code>sequence</code> class template described - in <a href="#2.8.3">Section 2.8.3, "Mapping for Members with the - Sequence Cardinality Class"</a>. Its interface is similar to - the sequence interface as defined by the ISO/ANSI Standard for - C++ (ISO/IEC 14882:1998, Section 23.1.1, "Sequences") and is - presented below: - </p> - - <pre class="c++"> -class element_sequence -{ -public: - typedef xercesc::DOMElement value_type; - typedef xercesc::DOMElement* pointer; - typedef const xercesc::DOMElement* const_pointer; - typedef xercesc::DOMElement& reference; - typedef const xercesc::DOMElement& const_reference; - - typedef <implementation-defined> iterator; - typedef <implementation-defined> const_iterator; - typedef <implementation-defined> reverse_iterator; - typedef <implementation-defined> const_reverse_iterator; - - typedef <implementation-defined> size_type; - typedef <implementation-defined> difference_type; - typedef <implementation-defined> allocator_type; - -public: - explicit - element_sequence (xercesc::DOMDocument&); - - // DOMElement cannot be default-constructed. - // - // explicit - // element_sequence (size_type n); - - element_sequence (size_type n, - const xercesc::DOMElement&, - xercesc::DOMDocument&); - - template <typename I> - element_sequence (const I& begin, - const I& end, - xercesc::DOMDocument&); - - element_sequence (const element_sequence&, xercesc::DOMDocument&); - - element_sequence& - operator= (const element_sequence&); - -public: - void - assign (size_type n, const xercesc::DOMElement&); - - template <typename I> - void - assign (const I& begin, const I& end); - -public: - // This version of resize can only be used to shrink the - // sequence because DOMElement cannot be default-constructed. - // - void - resize (size_type); - - void - resize (size_type, const xercesc::DOMElement&); - -public: - size_type - size () const; - - size_type - max_size () const; - - size_type - capacity () const; - - bool - empty () const; - - void - reserve (size_type); - - void - clear (); - -public: - const_iterator - begin () const; - - const_iterator - end () const; - - iterator - begin (); - - iterator - end (); - - const_reverse_iterator - rbegin () const; - - const_reverse_iterator - rend () const - - reverse_iterator - rbegin (); - - reverse_iterator - rend (); - -public: - xercesc::DOMElement& - operator[] (size_type); - - const xercesc::DOMElement& - operator[] (size_type) const; - - xercesc::DOMElement& - at (size_type); - - const xercesc::DOMElement& - at (size_type) const; - - xercesc::DOMElement& - front (); - - const xercesc::DOMElement& - front () const; - - xercesc::DOMElement& - back (); - - const xercesc::DOMElement& - back () const; - -public: - // Makes a deep copy. - // - void - push_back (const xercesc::DOMElement&); - - // Assumes ownership. - // - void - push_back (xercesc::DOMElement*); - - void - pop_back (); - - // Makes a deep copy. - // - iterator - insert (iterator position, const xercesc::DOMElement&); - - // Assumes ownership. - // - iterator - insert (iterator position, xercesc::DOMElement*); - - void - insert (iterator position, size_type n, const xercesc::DOMElement&); - - template <typename I> - void - insert (iterator position, const I& begin, const I& end); - - iterator - erase (iterator position); - - iterator - erase (iterator begin, iterator end); - -public: - // Note that the DOMDocument object of the two sequences being - // swapped should be the same. - // - void - swap (sequence& x); -}; - -inline bool -operator== (const element_sequence&, const element_sequence&); - -inline bool -operator!= (const element_sequence&, const element_sequence&); - </pre> - - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMElement& e) -{ - using namespace xercesc; - - object::any_sequence& s (o.any ()); - - // Iteration. - // - for (object::any_iterator i (s.begin ()); i != s.end (); ++i) - { - DOMElement& e (*i); - } - - // Modification. - // - s.push_back (e); // deep copy - DOMDocument& doc (o.dom_document ()); - s.push_back (doc.createElement (...)); // assumes ownership -} - </pre> - - <h3><a name="2.12.4">2.12.4 Element Wildcard Order</a></h3> - - <p>Similar to elements, element wildcards in ordered types - (<a href="#2.8.4">Section 2.8.4, "Element Order"</a>) are assigned - content ids and are included in the content order sequence. - Continuing with the bank transactions example started in Section - 2.8.4, we can extend the batch by allowing custom transactions:</p> - - <pre class="xml"> -<complexType name="batch"> - <choice minOccurs="0" maxOccurs="unbounded"> - <element name="withdraw" type="withdraw"/> - <element name="deposit" type="deposit"/> - <any namespace="##other" processContents="lax"/> - </choice> -</complexType> - </pre> - - <p>This will lead to the following changes in the generated - <code>batch</code> C++ class:</p> - - <pre class="c++"> -class batch: public xml_schema::type -{ -public: - ... - - // any - // - typedef element_sequence any_sequence; - typedef any_sequence::iterator any_iterator; - typedef any_sequence::const_iterator any_const_iterator; - - static const std::size_t any_id = 3UL; - - const any_sequence& - any () const; - - any_sequence& - any (); - - void - any (const any_sequence&); - - ... -}; - </pre> - - <p>With this change we also need to update the iteration code to handle - the new content id:</p> - - <pre class="c++"> -for (batch::content_order_const_iterator i (b.content_order ().begin ()); - i != b.content_order ().end (); - ++i) -{ - switch (i->id) - { - ... - - case batch::any_id: - { - const DOMElement& e (b.any ()[i->index]); - ... - break; - } - - ... - } -} - </pre> - - <p>For the complete working code that shows the use of wildcards in - ordered types refer to the <code>order/element</code> example in - the <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <h3><a name="2.12.5">2.12.5 Mapping for <code>anyAttribute</code></a></h3> - - <p>For <code>anyAttribute</code> the type definitions consist of an alias - of the container type with name <code>any_attribute_set</code> - (or <code>any1_attribute_set</code>, etc., for subsequent wildcards - in the type definition), an alias of the iterator type with name - <code>any_attribute_iterator</code> (or <code>any1_attribute_iterator</code>, - etc., for subsequent wildcards in the type definition), and an alias - of the constant iterator type with name <code>any_attribute_const_iterator</code> - (or <code>any1_attribute_const_iterator</code>, etc., for subsequent - wildcards in the type definition). - </p> - - <p>The accessor functions come in constant and non-constant versions. - The constant accessor function returns a constant reference to the - container and can be used for read-only access. The non-constant - version returns an unrestricted reference to the container and can - be used for read-write access. - </p> - - <p>The modifier function expects an argument of type reference to - constant of the container type. The modifier function makes - a deep copy of its argument. For instance: - </p> - - - <pre class="xml"> -<complexType name="object"> - <sequence> - ... - </sequence> - <anyAttribute namespace="##other"/> -</complexType> - </pre> - - <p>is mapped to:</p> - - <pre class="c++"> -class object: public xml_schema::type -{ -public: - // Type definitions. - // - typedef attribute_set any_attribute_set; - typedef any_attribute_set::iterator any_attribute_iterator; - typedef any_attribute_set::const_iterator any_attribute_const_iterator; - - // Accessors. - // - const any_attribute_set& - any_attribute () const; - - any_attribute_set& - any_attribute (); - - // Modifier. - // - void - any_attribute (const any_attribute_set&); - - ... - -}; - </pre> - - <p>The <code>attribute_set</code> class is an associative container - similar to the <code>std::set</code> class template as defined by - the ISO/ANSI Standard for C++ (ISO/IEC 14882:1998, Section 23.3.3, - "Class template set") with the key being the attribute's name - and namespace. Unlike <code>std::set</code>, <code>attribute_set</code> - allows searching using names and namespaces instead of - <code>xercesc::DOMAttr</code> objects. It is defined in an - implementation-specific namespace and its interface is presented - below: - </p> - - <pre class="c++"> -class attribute_set -{ -public: - typedef xercesc::DOMAttr key_type; - typedef xercesc::DOMAttr value_type; - typedef xercesc::DOMAttr* pointer; - typedef const xercesc::DOMAttr* const_pointer; - typedef xercesc::DOMAttr& reference; - typedef const xercesc::DOMAttr& const_reference; - - typedef <implementation-defined> iterator; - typedef <implementation-defined> const_iterator; - typedef <implementation-defined> reverse_iterator; - typedef <implementation-defined> const_reverse_iterator; - - typedef <implementation-defined> size_type; - typedef <implementation-defined> difference_type; - typedef <implementation-defined> allocator_type; - -public: - attribute_set (xercesc::DOMDocument&); - - template <typename I> - attribute_set (const I& begin, const I& end, xercesc::DOMDocument&); - - attribute_set (const attribute_set&, xercesc::DOMDocument&); - - attribute_set& - operator= (const attribute_set&); - -public: - const_iterator - begin () const; - - const_iterator - end () const; - - iterator - begin (); - - iterator - end (); - - const_reverse_iterator - rbegin () const; - - const_reverse_iterator - rend () const; - - reverse_iterator - rbegin (); - - reverse_iterator - rend (); - -public: - size_type - size () const; - - size_type - max_size () const; - - bool - empty () const; - - void - clear (); - -public: - // Makes a deep copy. - // - std::pair<iterator, bool> - insert (const xercesc::DOMAttr&); - - // Assumes ownership. - // - std::pair<iterator, bool> - insert (xercesc::DOMAttr*); - - // Makes a deep copy. - // - iterator - insert (iterator position, const xercesc::DOMAttr&); - - // Assumes ownership. - // - iterator - insert (iterator position, xercesc::DOMAttr*); - - template <typename I> - void - insert (const I& begin, const I& end); - -public: - void - erase (iterator position); - - size_type - erase (const std::basic_string<C>& name); - - size_type - erase (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name); - - size_type - erase (const XMLCh* name); - - size_type - erase (const XMLCh* namespace_, const XMLCh* name); - - void - erase (iterator begin, iterator end); - -public: - size_type - count (const std::basic_string<C>& name) const; - - size_type - count (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name) const; - - size_type - count (const XMLCh* name) const; - - size_type - count (const XMLCh* namespace_, const XMLCh* name) const; - - iterator - find (const std::basic_string<C>& name); - - iterator - find (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name); - - iterator - find (const XMLCh* name); - - iterator - find (const XMLCh* namespace_, const XMLCh* name); - - const_iterator - find (const std::basic_string<C>& name) const; - - const_iterator - find (const std::basic_string<C>& namespace_, - const std::basic_string<C>& name) const; - - const_iterator - find (const XMLCh* name) const; - - const_iterator - find (const XMLCh* namespace_, const XMLCh* name) const; - -public: - // Note that the DOMDocument object of the two sets being - // swapped should be the same. - // - void - swap (attribute_set&); -}; - -bool -operator== (const attribute_set&, const attribute_set&); - -bool -operator!= (const attribute_set&, const attribute_set&); - </pre> - - <p>The following code shows how one could use this mapping:</p> - - <pre class="c++"> -void -f (object& o, const xercesc::DOMAttr& a) -{ - using namespace xercesc; - - object::any_attribute_set& s (o.any_attribute ()); - - // Iteration. - // - for (object::any_attribute_iterator i (s.begin ()); i != s.end (); ++i) - { - DOMAttr& a (*i); - } - - // Modification. - // - s.insert (a); // deep copy - DOMDocument& doc (o.dom_document ()); - s.insert (doc.createAttribute (...)); // assumes ownership - - // Searching. - // - object::any_attribute_iterator i (s.find ("name")); - i = s.find ("http://www.w3.org/XML/1998/namespace", "lang"); -} - </pre> - - <!-- Mapping for Mixed Content Models --> - - <h2><a name="2.13">2.13 Mapping for Mixed Content Models</a></h2> - - <p>For XML Schema types with mixed content models C++/Tree provides - mapping support only if the type is marked as ordered - (<a href="#2.8.4">Section 2.8.4, "Element Order"</a>). Use the - <code>--ordered-type-mixed</code> XSD compiler option to - automatically mark all types with mixed content as ordered.</p> - - <p>For an ordered type with mixed content, C++/Tree adds an extra - text content sequence that is used to store the text fragments. - This text content sequence is also assigned the content id and - its entries are included in the content order sequence, just - like elements. As a result, it is possible to capture the order - between elements and text fragments.</p> - - <p>As an example, consider the following schema that describes text - with embedded links:</p> - - <pre class="xml"> -<complexType name="anchor"> - <simpleContent> - <extension base="string"> - <attribute name="href" type="anyURI" use="required"/> - </extension> - </simpleContent> -</complexType> - -<complexType name="text" mixed="true"> - <sequence> - <element name="a" type="anchor" minOccurs="0" maxOccurs="unbounded"/> - </sequence> -</complexType> - </pre> - - <p>The generated <code>text</code> C++ class will provide the following - API (assuming it is marked as ordered):</p> - - <pre class="c++"> -class text: public xml_schema::type -{ -public: - // a - // - typedef anchor a_type; - typedef sequence<a_type> a_sequence; - typedef a_sequence::iterator a_iterator; - typedef a_sequence::const_iterator a_const_iterator; - - static const std::size_t a_id = 1UL; - - const a_sequence& - a () const; - - a_sequence& - a (); - - void - a (const a_sequence&); - - // text_content - // - typedef xml_schema::string text_content_type; - typedef sequence<text_content_type> text_content_sequence; - typedef text_content_sequence::iterator text_content_iterator; - typedef text_content_sequence::const_iterator text_content_const_iterator; - - static const std::size_t text_content_id = 2UL; - - const text_content_sequence& - text_content () const; - - text_content_sequence& - text_content (); - - void - text_content (const text_content_sequence&); - - // content_order - // - typedef xml_schema::content_order content_order_type; - typedef std::vector<content_order_type> content_order_sequence; - typedef content_order_sequence::iterator content_order_iterator; - typedef content_order_sequence::const_iterator content_order_const_iterator; - - const content_order_sequence& - content_order () const; - - content_order_sequence& - content_order (); - - void - content_order (const content_order_sequence&); - - ... -}; - </pre> - - <p>Given this interface we can iterate over both link elements - and text in content order. The following code fragment converts - our format to plain text with references.</p> - - <pre class="c++"> -const text& t = ... - -for (text::content_order_const_iterator i (t.content_order ().begin ()); - i != t.content_order ().end (); - ++i) -{ - switch (i->id) - { - case text::a_id: - { - const anchor& a (t.a ()[i->index]); - cerr << a << "[" << a.href () << "]"; - break; - } - case text::text_content_id: - { - const xml_schema::string& s (t.text_content ()[i->index]); - cerr << s; - break; - } - default: - { - assert (false); // Unknown content id. - } - } -} - </pre> - - <p>For the complete working code that shows the use of mixed content - in ordered types refer to the <code>order/mixed</code> example in - the <code>cxx/tree/</code> directory in the - <a href="https://cppget.org/xsd-examples">xsd-examples</a> - package.</p> - - <!-- Parsing --> - - - <h1><a name="3">3 Parsing</a></h1> - - <p>This chapter covers various aspects of parsing XML instance - documents in order to obtain corresponding tree-like object - model. - </p> - - <p>Each global XML Schema element in the form:</p> - - <pre class="xml"> -<element name="name" type="type"/> - </pre> - - <p>is mapped to 14 overloaded C++ functions in the form:</p> - - <pre class="c++"> -// Read from a URI or a local file. -// - -std::[unique|auto]_ptr<type> -name (const std::basic_string<C>& uri, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (const std::basic_string<C>& uri, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (const std::basic_string<C>& uri, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -// Read from std::istream. -// - -std::[unique|auto]_ptr<type> -name (std::istream&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -std::[unique|auto]_ptr<type> -name (std::istream&, - const std::basic_string<C>& id, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - const std::basic_string<C>& id, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (std::istream&, - const std::basic_string<C>& id, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -// Read from InputSource. -// - -std::[unique|auto]_ptr<type> -name (xercesc::InputSource&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (xercesc::InputSource&, - xml_schema::error_handler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (xercesc::InputSource&, - xercesc::DOMErrorHandler&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - - -// Read from DOM. -// - -std::[unique|auto]_ptr<type> -name (const xercesc::DOMDocument&, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - -std::[unique|auto]_ptr<type> -name (xml_schema::dom::[unique|auto]_ptr<xercesc::DOMDocument>, - xml_schema::flags = 0, - const xml_schema::properties& = xml_schema::properties ()); - </pre> - - <p>You can choose between reading an XML instance from a local file, - URI, <code>std::istream</code>, <code>xercesc::InputSource</code>, - or a pre-parsed DOM instance in the form of - <code>xercesc::DOMDocument</code>. All the parsing functions - return a dynamically allocated object model as either - <code>std::unique_ptr</code> or <code>std::auto_ptr</code>, - depending on the C++ standard selected. Each of these parsing - functions is discussed in more detail in the following sections. - </p> - - <h2><a name="3.1">3.1 Initializing the Xerces-C++ Runtime</a></h2> - - <p>Some parsing functions expect you to initialize the Xerces-C++ - runtime while others initialize and terminate it as part of their - work. The general rule is as follows: if a function has any arguments - or return a value that is an instance of a Xerces-C++ type, then - this function expects you to initialize the Xerces-C++ runtime. - Otherwise, the function initializes and terminates the runtime for - you. Note that it is legal to have nested calls to the Xerces-C++ - initialize and terminate functions as long as the calls are balanced. - </p> - - <p>You can instruct parsing functions that initialize and terminate - the runtime not to do so by passing the - <code>xml_schema::flags::dont_initialize</code> flag (see - <a href="#3.2">Section 3.2, "Flags and Properties"</a>). - </p> - - - <h2><a name="3.2">3.2 Flags and Properties</a></h2> - - <p>Parsing flags and properties are the last two arguments of every - parsing function. They allow you to fine-tune the process of - instance validation and parsing. Both arguments are optional. - </p> - - - <p>The following flags are recognized by the parsing functions:</p> - - <dl> - <dt><code>xml_schema::flags::keep_dom</code></dt> - <dd>Keep association between DOM nodes and the resulting - object model nodes. For more information about DOM association - refer to <a href="#5.1">Section 5.1, "DOM Association"</a>.</dd> - - <dt><code>xml_schema::flags::own_dom</code></dt> - <dd>Assume ownership of the DOM document passed. This flag only - makes sense together with the <code>keep_dom</code> flag in - the call to the parsing function with the - <code>xml_schema::dom::[unique|auto]_ptr<DOMDocument></code> - argument.</dd> - - <dt><code>xml_schema::flags::dont_validate</code></dt> - <dd>Do not validate instance documents against schemas.</dd> - - <dt><code>xml_schema::flags::dont_initialize</code></dt> - <dd>Do not initialize the Xerces-C++ runtime.</dd> - </dl> - - <p>You can pass several flags by combining them using the bit-wise OR - operator. For example:</p> - - <pre class="c++"> -using xml_schema::flags; - -std::unique_ptr<type> r ( - name ("test.xml", flags::keep_dom | flags::dont_validate)); - </pre> - - <p>By default, validation of instance documents is turned on even - though parsers generated by XSD do not assume instance - documents are valid. They include a number of checks that prevent - construction of inconsistent object models. This, - however, does not mean that an instance document that was - successfully parsed by the XSD-generated parsers is - valid per the corresponding schema. If an instance document is not - "valid enough" for the generated parsers to construct consistent - object model, one of the exceptions defined in - <code>xml_schema</code> namespace is thrown (see - <a href="#3.3">Section 3.3, "Error Handling"</a>). - </p> - - <p>For more information on the Xerces-C++ runtime initialization - refer to <a href="#3.1">Section 3.1, "Initializing the Xerces-C++ - Runtime"</a>. - </p> - - <p>The <code>xml_schema::properties</code> class allows you to - programmatically specify schema locations to be used instead - of those specified with the <code>xsi::schemaLocation</code> - and <code>xsi::noNamespaceSchemaLocation</code> attributes - in instance documents. The interface of the <code>properties</code> - class is presented below: - </p> - - <pre class="c++"> -class properties -{ -public: - void - schema_location (const std::basic_string<C>& namespace_, - const std::basic_string<C>& location); - void - no_namespace_schema_location (const std::basic_string<C>& location); -}; - </pre> - - <p>Note that all locations are relative to an instance document unless - they are URIs. For example, if you want to use a local file as your - schema, then you will need to pass - <code>file:///absolute/path/to/your/schema</code> as the location - argument. - </p> - - <h2><a name="3.3">3.3 Error Handling</a></h2> - - <p>As discussed in <a href="#2.2">Section 2.2, "Error Handling"</a>, - the mapping uses the C++ exception handling mechanism as its primary - way of reporting error conditions. However, to handle recoverable - parsing and validation errors and warnings, a callback interface maybe - preferred by the application.</p> - - <p>To better understand error handling and reporting strategies employed - by the parsing functions, it is useful to know that the - transformation of an XML instance document to a statically-typed - tree happens in two stages. The first stage, performed by Xerces-C++, - consists of parsing an XML document into a DOM instance. For short, - we will call this stage the XML-DOM stage. Validation, if not disabled, - happens during this stage. The second stage, - performed by the generated parsers, consist of parsing the DOM - instance into the statically-typed tree. We will call this stage - the DOM-Tree stage. Additional checks are performed during this - stage in order to prevent construction of inconsistent tree which - could otherwise happen when validation is disabled, for example.</p> - - <p>All parsing functions except the one that operates on a DOM instance - come in overloaded triples. The first function in such a triple - reports error conditions exclusively by throwing exceptions. It - accumulates all the parsing and validation errors of the XML-DOM - stage and throws them in a single instance of the - <code>xml_schema::parsing</code> exception (described below). - The second and the third functions in the triple use callback - interfaces to report parsing and validation errors and warnings. - The two callback interfaces are <code>xml_schema::error_handler</code> - and <code>xercesc::DOMErrorHandler</code>. For more information - on the <code>xercesc::DOMErrorHandler</code> interface refer to - the Xerces-C++ documentation. The <code>xml_schema::error_handler</code> - interface is presented below: - </p> - - <pre class="c++"> -class error_handler -{ -public: - struct severity - { - enum value - { - warning, - error, - fatal - }; - }; - - virtual bool - handle (const std::basic_string<C>& id, - unsigned long line, - unsigned long column, - severity, - const std::basic_string<C>& message) = 0; - - virtual - ~error_handler (); -}; - </pre> - - <p>The <code>id</code> argument of the <code>error_handler::handle</code> - function identifies the resource being parsed (e.g., a file name or - URI). - </p> - - <p>By returning <code>true</code> from the <code>handle</code> function - you instruct the parser to recover and continue parsing. Returning - <code>false</code> results in termination of the parsing process. - An error with the <code>fatal</code> severity level results in - termination of the parsing process no matter what is returned from - the <code>handle</code> function. It is safe to throw an exception - from the <code>handle</code> function. - </p> - - <p>The DOM-Tree stage reports error conditions exclusively by throwing - exceptions. Individual exceptions thrown by the parsing functions - are described in the following sub-sections. - </p> - - - <h3><a name="3.3.1">3.3.1 <code>xml_schema::parsing</code></a></h3> - - <pre class="c++"> -struct severity -{ - enum value - { - warning, - error - }; - - severity (value); - operator value () const; -}; - -struct error -{ - error (severity, - const std::basic_string<C>& id, - unsigned long line, - unsigned long column, - const std::basic_string<C>& message); - - severity - severity () const; - - const std::basic_string<C>& - id () const; - - unsigned long - line () const; - - unsigned long - column () const; - - const std::basic_string<C>& - message () const; -}; - -std::basic_ostream<C>& -operator<< (std::basic_ostream<C>&, const error&); - -struct diagnostics: std::vector<error> -{ -}; - -std::basic_ostream<C>& -operator<< (std::basic_ostream<C>&, const diagnostics&); - -struct parsing: virtual exception -{ - parsing (); - parsing (const diagnostics&); - - const diagnostics& - diagnostics () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::parsing</code> exception is thrown if there - were parsing or validation errors reported during the XML-DOM stage. - If no callback interface was provided to the parsing function, the - exception contains a list of errors and warnings accessible using - the <code>diagnostics</code> function. The usual conditions when - this exception is thrown include malformed XML instances and, if - validation is turned on, invalid instance documents. - </p> - - <h3><a name="3.3.2">3.3.2 <code>xml_schema::expected_element</code></a></h3> - - <pre class="c++"> -struct expected_element: virtual exception -{ - expected_element (const std::basic_string<C>& name, - const std::basic_string<C>& namespace_); - - - const std::basic_string<C>& - name () const; - - const std::basic_string<C>& - namespace_ () const; - - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::expected_element</code> exception is thrown - when an expected element is not encountered by the DOM-Tree stage. - The name and namespace of the expected element can be obtained using - the <code>name</code> and <code>namespace_</code> functions respectively. - </p> - - - <h3><a name="3.3.3">3.3.3 <code>xml_schema::unexpected_element</code></a></h3> - - <pre class="c++"> -struct unexpected_element: virtual exception -{ - unexpected_element (const std::basic_string<C>& encountered_name, - const std::basic_string<C>& encountered_namespace, - const std::basic_string<C>& expected_name, - const std::basic_string<C>& expected_namespace) - - - const std::basic_string<C>& - encountered_name () const; - - const std::basic_string<C>& - encountered_namespace () const; - - - const std::basic_string<C>& - expected_name () const; - - const std::basic_string<C>& - expected_namespace () const; - - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::unexpected_element</code> exception is thrown - when an unexpected element is encountered by the DOM-Tree stage. - The name and namespace of the encountered element can be obtained - using the <code>encountered_name</code> and - <code>encountered_namespace</code> functions respectively. If an - element was expected instead of the encountered one, its name - and namespace can be obtained using the <code>expected_name</code> and - <code>expected_namespace</code> functions respectively. Otherwise - these functions return empty strings. - </p> - - <h3><a name="3.3.4">3.3.4 <code>xml_schema::expected_attribute</code></a></h3> - - <pre class="c++"> -struct expected_attribute: virtual exception -{ - expected_attribute (const std::basic_string<C>& name, - const std::basic_string<C>& namespace_); - - - const std::basic_string<C>& - name () const; - - const std::basic_string<C>& - namespace_ () const; - - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::expected_attribute</code> exception is thrown - when an expected attribute is not encountered by the DOM-Tree stage. - The name and namespace of the expected attribute can be obtained using - the <code>name</code> and <code>namespace_</code> functions respectively. - </p> - - - <h3><a name="3.3.5">3.3.5 <code>xml_schema::unexpected_enumerator</code></a></h3> - - <pre class="c++"> -struct unexpected_enumerator: virtual exception -{ - unexpected_enumerator (const std::basic_string<C>& enumerator); - - const std::basic_string<C>& - enumerator () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::unexpected_enumerator</code> exception is thrown - when an unexpected enumerator is encountered by the DOM-Tree stage. - The enumerator can be obtained using the <code>enumerator</code> - functions. - </p> - - <h3><a name="3.3.6">3.3.6 <code>xml_schema::expected_text_content</code></a></h3> - - <pre class="c++"> -struct expected_text_content: virtual exception -{ - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::expected_text_content</code> exception is thrown - when a content other than text is encountered and the text content was - expected by the DOM-Tree stage. - </p> - - <h3><a name="3.3.7">3.3.7 <code>xml_schema::no_type_info</code></a></h3> - - <pre class="c++"> -struct no_type_info: virtual exception -{ - no_type_info (const std::basic_string<C>& type_name, - const std::basic_string<C>& type_namespace); - - const std::basic_string<C>& - type_name () const; - - const std::basic_string<C>& - type_namespace () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::no_type_info</code> exception is thrown - when there is no type information associated with a type specified - by the <code>xsi:type</code> attribute. This exception is thrown - by the DOM-Tree stage. The name and namespace of the type in question - can be obtained using the <code>type_name</code> and - <code>type_namespace</code> functions respectively. Usually, catching - this exception means that you haven't linked the code generated - from the schema defining the type in question with your application - or this schema has been compiled without the - <code>--generate-polymorphic</code> option. - </p> - - - <h3><a name="3.3.8">3.3.8 <code>xml_schema::not_derived</code></a></h3> - - <pre class="c++"> -struct not_derived: virtual exception -{ - not_derived (const std::basic_string<C>& base_type_name, - const std::basic_string<C>& base_type_namespace, - const std::basic_string<C>& derived_type_name, - const std::basic_string<C>& derived_type_namespace); - - const std::basic_string<C>& - base_type_name () const; - - const std::basic_string<C>& - base_type_namespace () const; - - - const std::basic_string<C>& - derived_type_name () const; - - const std::basic_string<C>& - derived_type_namespace () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::not_derived</code> exception is thrown - when a type specified by the <code>xsi:type</code> attribute is - not derived from the expected base type. This exception is thrown - by the DOM-Tree stage. The name and namespace of the expected - base type can be obtained using the <code>base_type_name</code> and - <code>base_type_namespace</code> functions respectively. The name - and namespace of the offending type can be obtained using the - <code>derived_type_name</code> and - <code>derived_type_namespace</code> functions respectively. - </p> - - <h3><a name="3.3.9">3.3.9 <code>xml_schema::no_prefix_mapping</code></a></h3> - - <pre class="c++"> -struct no_prefix_mapping: virtual exception -{ - no_prefix_mapping (const std::basic_string<C>& prefix); - - const std::basic_string<C>& - prefix () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::no_prefix_mapping</code> exception is thrown - during the DOM-Tree stage if a namespace prefix is encountered for - which a prefix-namespace mapping hasn't been provided. The namespace - prefix in question can be obtained using the <code>prefix</code> - function. - </p> - - <h2><a name="3.4">3.4 Reading from a Local File or URI</a></h2> - - <p>Using a local file or URI is the simplest way to parse an XML instance. - For example:</p> - - <pre class="c++"> -using std::unique_ptr; - -unique_ptr<type> r1 (name ("test.xml")); -unique_ptr<type> r2 (name ("https://www.codesynthesis.com/test.xml")); - </pre> - - <p>Or, in the C++98 mode:</p> - - <pre class="c++"> -using std::auto_ptr; - -auto_ptr<type> r1 (name ("test.xml")); -auto_ptr<type> r2 (name ("https://www.codesynthesis.com/test.xml")); - </pre> - - <h2><a name="3.5">3.5 Reading from <code>std::istream</code></a></h2> - - <p>When using an <code>std::istream</code> instance, you may also - pass an optional resource id. This id is used to identify the - resource (for example in error messages) as well as to resolve - relative paths. For instance:</p> - - <pre class="c++"> -using std::unique_ptr; - -{ - std::ifstream ifs ("test.xml"); - unique_ptr<type> r (name (ifs, "test.xml")); -} - -{ - std::string str ("..."); // Some XML fragment. - std::istringstream iss (str); - unique_ptr<type> r (name (iss)); -} - </pre> - - <h2><a name="3.6">3.6 Reading from <code>xercesc::InputSource</code></a></h2> - - <p>Reading from a <code>xercesc::InputSource</code> instance - is similar to the <code>std::istream</code> case except - the resource id is maintained by the <code>InputSource</code> - object. For instance:</p> - - <pre class="c++"> -xercesc::StdInInputSource is; -std::unique_ptr<type> r (name (is)); - </pre> - - <h2><a name="3.7">3.7 Reading from DOM</a></h2> - - <p>Reading from a <code>xercesc::DOMDocument</code> instance allows - you to setup a custom XML-DOM stage. Things like DOM - parser reuse, schema pre-parsing, and schema caching can be achieved - with this approach. For more information on how to obtain DOM - representation from an XML instance refer to the Xerces-C++ - documentation. In addition, the - <a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree Mapping - FAQ</a> shows how to parse an XML instance to a Xerces-C++ - DOM document using the XSD runtime utilities. - </p> - - <p>The last parsing function is useful when you would like to perform - your own XML-to-DOM parsing and associate the resulting DOM document - with the object model nodes. The automatic <code>DOMDocument</code> - pointer is reset and the resulting object model assumes ownership - of the DOM document passed. For example:</p> - - <pre class="c++"> -// C++11 version. -// -xml_schema::dom::unique_ptr<xercesc::DOMDocument> doc = ... - -std::unique_ptr<type> r ( - name (std::move (doc), - xml_schema::flags::keep_dom | xml_schema::flags::own_dom)); - -// At this point doc is reset to 0. - -// C++98 version. -// -xml_schema::dom::auto_ptr<xercesc::DOMDocument> doc = ... - -std::auto_ptr<type> r ( - name (doc, xml_schema::flags::keep_dom | xml_schema::flags::own_dom)); - -// At this point doc is reset to 0. - </pre> - - <h1><a name="4">4 Serialization</a></h1> - - <p>This chapter covers various aspects of serializing a - tree-like object model to DOM or XML. - In this regard, serialization is complimentary to the reverse - process of parsing a DOM or XML instance into an object model - which is discussed in <a href="#3">Chapter 3, - "Parsing"</a>. Note that the generation of the serialization code - is optional and should be explicitly requested with the - <code>--generate-serialization</code> option. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information. - </p> - - <p>Each global XML Schema element in the form: - </p> - - - <pre class="xml"> -<xsd:element name="name" type="type"/> - </pre> - - <p>is mapped to 8 overloaded C++ functions in the form:</p> - - <pre class="c++"> -// Serialize to std::ostream. -// -void -name (std::ostream&, - const type&, - const xml_schema::namespace_fomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (std::ostream&, - const type&, - xml_schema::error_handler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (std::ostream&, - const type&, - xercesc::DOMErrorHandler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - - -// Serialize to XMLFormatTarget. -// -void -name (xercesc::XMLFormatTarget&, - const type&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (xercesc::XMLFormatTarget&, - const type&, - xml_schema::error_handler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - -void -name (xercesc::XMLFormatTarget&, - const type&, - xercesc::DOMErrorHandler&, - const xml_schema::namespace_infomap& = - xml_schema::namespace_infomap (), - const std::basic_string<C>& encoding = "UTF-8", - xml_schema::flags = 0); - - -// Serialize to DOM. -// -xml_schema::dom::[unique|auto]_ptr<xercesc::DOMDocument> -name (const type&, - const xml_schema::namespace_infomap& - xml_schema::namespace_infomap (), - xml_schema::flags = 0); - -void -name (xercesc::DOMDocument&, - const type&, - xml_schema::flags = 0); - </pre> - - <p>You can choose between writing XML to <code>std::ostream</code> or - <code>xercesc::XMLFormatTarget</code> and creating a DOM instance - in the form of <code>xercesc::DOMDocument</code>. Serialization - to <code>ostream</code> or <code>XMLFormatTarget</code> requires a - considerably less work while serialization to DOM provides - for greater flexibility. Each of these serialization functions - is discussed in more detail in the following sections. - </p> - - - <h2><a name="4.1">4.1 Initializing the Xerces-C++ Runtime</a></h2> - - <p>Some serialization functions expect you to initialize the Xerces-C++ - runtime while others initialize and terminate it as part of their - work. The general rule is as follows: if a function has any arguments - or return a value that is an instance of a Xerces-C++ type, then - this function expects you to initialize the Xerces-C++ runtime. - Otherwise, the function initializes and terminates the runtime for - you. Note that it is legal to have nested calls to the Xerces-C++ - initialize and terminate functions as long as the calls are balanced. - </p> - - <p>You can instruct serialization functions that initialize and terminate - the runtime not to do so by passing the - <code>xml_schema::flags::dont_initialize</code> flag (see - <a href="#4.3">Section 4.3, "Flags"</a>). - </p> - - <h2><a name="4.2">4.2 Namespace Infomap and Character Encoding</a></h2> - - <p>When a document being serialized uses XML namespaces, custom - prefix-namespace associations can to be established. If custom - prefix-namespace mapping is not provided then generic prefixes - (<code>p1</code>, <code>p2</code>, etc) are automatically assigned - to namespaces as needed. Also, if - you would like the resulting instance document to contain the - <code>schemaLocation</code> or <code>noNamespaceSchemaLocation</code> - attributes, you will need to provide namespace-schema associations. - The <code>xml_schema::namespace_infomap</code> class is used - to capture this information:</p> - - <pre class="c++"> -struct namespace_info -{ - namespace_info (); - namespace_info (const std::basic_string<C>& name, - const std::basic_string<C>& schema); - - std::basic_string<C> name; - std::basic_string<C> schema; -}; - -// Map of namespace prefix to namespace_info. -// -struct namespace_infomap: public std::map<std::basic_string<C>, - namespace_info> -{ -}; - </pre> - - <p>Consider the following associations as an example:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - </pre> - - <p>This map, if passed to one of the serialization functions, - could result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<t:name xmlns:t="https://www.codesynthesis.com/test" - xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:schemaLocation="https://www.codesynthesis.com/test test.xsd"> - </pre> - - <p>As you can see, the serialization function automatically added namespace - mapping for the <code>xsi</code> prefix. You can change this by - providing your own prefix:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map["xsn"].name = "http://www.w3.org/2001/XMLSchema-instance"; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - </pre> - - <p>This could result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<t:name xmlns:t="https://www.codesynthesis.com/test" - xmlns:xsn="http://www.w3.org/2001/XMLSchema-instance" - xsn:schemaLocation="https://www.codesynthesis.com/test test.xsd"> - </pre> - - <p>To specify the location of a schema without a namespace you can use - an empty prefix as in the example below: </p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].schema = "test.xsd"; - </pre> - - <p>This would result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<name xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:noNamespaceSchemaLocation="test.xsd"> - </pre> - - <p>To make a particular namespace default you can use an empty - prefix, for example:</p> - - <pre class="c++"> -xml_schema::namespace_infomap map; - -map[""].name = "https://www.codesynthesis.com/test"; -map[""].schema = "test.xsd"; - </pre> - - <p>This could result in the following XML fragment:</p> - - <pre class="xml"> -<?xml version="1.0" ?> -<name xmlns="https://www.codesynthesis.com/test" - xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" - xsi:schemaLocation="https://www.codesynthesis.com/test test.xsd"> - </pre> - - - <p>Another bit of information that you can pass to the serialization - functions is the character encoding method that you would like to use. - Common values for this argument are <code>"US-ASCII"</code>, - <code>"ISO8859-1"</code>, <code>"UTF-8"</code>, - <code>"UTF-16BE"</code>, <code>"UTF-16LE"</code>, - <code>"UCS-4BE"</code>, and <code>"UCS-4LE"</code>. The default - encoding is <code>"UTF-8"</code>. For more information on - encoding methods see the - "<a href="http://en.wikipedia.org/wiki/Character_code">Character - Encoding</a>" article from Wikipedia. - </p> - - <h2><a name="4.3">4.3 Flags</a></h2> - - <p>Serialization flags are the last argument of every serialization - function. They allow you to fine-tune the process of serialization. - The flags argument is optional. - </p> - - - <p>The following flags are recognized by the serialization - functions:</p> - - <dl> - <dt><code>xml_schema::flags::dont_initialize</code></dt> - <dd>Do not initialize the Xerces-C++ runtime.</dd> - - <dt><code>xml_schema::flags::dont_pretty_print</code></dt> - <dd>Do not add extra spaces or new lines that make the resulting XML - slightly bigger but easier to read.</dd> - - <dt><code>xml_schema::flags::no_xml_declaration</code></dt> - <dd>Do not write XML declaration (<?xml ... ?>).</dd> - </dl> - - <p>You can pass several flags by combining them using the bit-wise OR - operator. For example:</p> - - <pre class="c++"> -std::unique_ptr<type> r = ... -std::ofstream ofs ("test.xml"); -xml_schema::namespace_infomap map; -name (ofs, - *r, - map, - "UTF-8", - xml_schema::flags::no_xml_declaration | - xml_schema::flags::dont_pretty_print); - </pre> - - <p>For more information on the Xerces-C++ runtime initialization - refer to <a href="#4.1">Section 4.1, "Initializing the Xerces-C++ - Runtime"</a>. - </p> - - <h2><a name="4.4">4.4 Error Handling</a></h2> - - <p>As with the parsing functions (see <a href="#3.3">Section 3.3, - "Error Handling"</a>), to better understand error handling and - reporting strategies employed by the serialization functions, it - is useful to know that the transformation of a statically-typed - tree to an XML instance document happens in two stages. The first - stage, performed by the generated code, consist of building a DOM - instance from the statically-typed tree . For short, we will call - this stage the Tree-DOM stage. The second stage, performed by - Xerces-C++, consists of serializing the DOM instance into the XML - document. We will call this stage the DOM-XML stage. - </p> - - <p>All serialization functions except the two that serialize into - a DOM instance come in overloaded triples. The first function - in such a triple reports error conditions exclusively by throwing - exceptions. It accumulates all the serialization errors of the - DOM-XML stage and throws them in a single instance of the - <code>xml_schema::serialization</code> exception (described below). - The second and the third functions in the triple use callback - interfaces to report serialization errors and warnings. The two - callback interfaces are <code>xml_schema::error_handler</code> and - <code>xercesc::DOMErrorHandler</code>. The - <code>xml_schema::error_handler</code> interface is described in - <a href="#3.3">Section 3.3, "Error Handling"</a>. For more information - on the <code>xercesc::DOMErrorHandler</code> interface refer to the - Xerces-C++ documentation. - </p> - - <p>The Tree-DOM stage reports error conditions exclusively by throwing - exceptions. Individual exceptions thrown by the serialization functions - are described in the following sub-sections. - </p> - - <h3><a name="4.4.1">4.4.1 <code>xml_schema::serialization</code></a></h3> - - <pre class="c++"> -struct serialization: virtual exception -{ - serialization (); - serialization (const diagnostics&); - - const diagnostics& - diagnostics () const; - - virtual const char* - what () const throw (); -}; - </pre> - - <p>The <code>xml_schema::diagnostics</code> class is described in - <a href="#3.3.1">Section 3.3.1, "<code>xml_schema::parsing</code>"</a>. - The <code>xml_schema::serialization</code> exception is thrown if - there were serialization errors reported during the DOM-XML stage. - If no callback interface was provided to the serialization function, - the exception contains a list of errors and warnings accessible using - the <code>diagnostics</code> function. - </p> - - - <h3><a name="4.4.2">4.4.2 <code>xml_schema::unexpected_element</code></a></h3> - - <p>The <code>xml_schema::unexpected_element</code> exception is - described in <a href="#3.3.3">Section 3.3.3, - "<code>xml_schema::unexpected_element</code>"</a>. It is thrown - by the serialization functions during the Tree-DOM stage if the - root element name of the provided DOM instance does not match with - the name of the element this serialization function is for. - </p> - - <h3><a name="4.4.3">4.4.3 <code>xml_schema::no_type_info</code></a></h3> - - <p>The <code>xml_schema::no_type_info</code> exception is - described in <a href="#3.3.7">Section 3.3.7, - "<code>xml_schema::no_type_info</code>"</a>. It is thrown - by the serialization functions during the Tree-DOM stage when there - is no type information associated with a dynamic type of an - element. Usually, catching this exception means that you haven't - linked the code generated from the schema defining the type in - question with your application or this schema has been compiled - without the <code>--generate-polymorphic</code> option. - </p> - - <h2><a name="4.5">4.5 Serializing to <code>std::ostream</code></a></h2> - - <p>In order to serialize to <code>std::ostream</code> you will need - an object model, an output stream and, optionally, a namespace - infomap. For instance:</p> - - <pre class="c++"> -// Obtain the object model. -// -std::unique_ptr<type> r = ... - -// Prepare namespace mapping and schema location information. -// -xml_schema::namespace_infomap map; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - -// Write it out. -// -name (std::cout, *r, map); - </pre> - - <p>Note that the output stream is treated as a binary stream. This - becomes important when you use a character encoding that is wider - than 8-bit <code>char</code>, for instance UTF-16 or UCS-4. For - example, things will most likely break if you try to serialize - to <code>std::ostringstream</code> with UTF-16 or UCS-4 as an - encoding. This is due to the special value, - <code>'\0'</code>, that will most likely occur as part of such - serialization and it won't have the special meaning assumed by - <code>std::ostringstream</code>. - </p> - - - <h2><a name="4.6">4.6 Serializing to <code>xercesc::XMLFormatTarget</code></a></h2> - - <p>Serializing to an <code>xercesc::XMLFormatTarget</code> instance - is similar the <code>std::ostream</code> case. For instance: - </p> - - <pre class="c++"> -using std::unique_ptr; - -// Obtain the object model. -// -unique_ptr<type> r = ... - -// Prepare namespace mapping and schema location information. -// -xml_schema::namespace_infomap map; - -map["t"].name = "https://www.codesynthesis.com/test"; -map["t"].schema = "test.xsd"; - -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Choose a target. - // - unique_ptr<XMLFormatTarget> ft; - - if (argc != 2) - { - ft = unique_ptr<XMLFormatTarget> (new StdOutFormatTarget ()); - } - else - { - ft = unique_ptr<XMLFormatTarget> ( - new LocalFileFormatTarget (argv[1])); - } - - // Write it out. - // - name (*ft, *r, map); -} - -XMLPlatformUtils::Terminate (); - </pre> - - <p>Note that we had to initialize the Xerces-C++ runtime before we - could call this serialization function.</p> - - <h2><a name="4.7">4.7 Serializing to DOM</a></h2> - - <p>The mapping provides two overloaded functions that implement - serialization to a DOM instance. The first creates a DOM instance - for you and the second serializes to an existing DOM instance. - While serializing to a new DOM instance is similar to serializing - to <code>std::ostream</code> or <code>xercesc::XMLFormatTarget</code>, - serializing to an existing DOM instance requires quite a bit of work - from your side. You will need to set all the custom namespace mapping - attributes as well as the <code>schemaLocation</code> and/or - <code>noNamespaceSchemaLocation</code> attributes. The following - listing should give you an idea about what needs to be done: - </p> - - <pre class="c++"> -// Obtain the object model. -// -std::unique_ptr<type> r = ... - -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Create a DOM instance. Set custom namespace mapping and schema - // location attributes. - // - DOMDocument& doc = ... - - // Serialize to DOM. - // - name (doc, *r); - - // Serialize the DOM document to XML. - // - ... -} - -XMLPlatformUtils::Terminate (); - </pre> - - <p>For more information on how to create and serialize a DOM instance - refer to the Xerces-C++ documentation. In addition, the - <a href="http://wiki.codesynthesis.com/Tree/FAQ">C++/Tree Mapping - FAQ</a> shows how to implement these operations using the XSD - runtime utilities. - </p> - - <h1><a name="5">5 Additional Functionality</a></h1> - - <p>The C++/Tree mapping provides a number of optional features - that can be useful in certain situations. They are described - in the following sections.</p> - - <h2><a name="5.1">5.1 DOM Association</a></h2> - - <p>Normally, after parsing is complete, the DOM document which - was used to extract the data is discarded. However, the parsing - functions can be instructed to preserve the DOM document - and create an association between the DOM nodes and object model - nodes. When there is an association between the DOM and - object model nodes, you can obtain the corresponding DOM element - or attribute node from an object model node as well as perform - the reverse transition: obtain the corresponding object model - from a DOM element or attribute node.</p> - - <p>Maintaining DOM association is normally useful when the application - needs access to XML constructs that are not preserved in the - object model, for example, XML comments. - Another useful aspect of DOM association is the ability of the - application to navigate the document tree using the generic DOM - interface (for example, with the help of an XPath processor) - and then move back to the statically-typed object model. Note - also that while you can change the underlying DOM document, - these changes are not reflected in the object model and will - be ignored during serialization. If you need to not only access - but also modify some aspects of XML that are not preserved in - the object model, then type customization with custom parsing - constructors and serialization operators should be used instead.</p> - - <p>To request DOM association you will need to pass the - <code>xml_schema::flags::keep_dom</code> flag to one of the - parsing functions (see <a href="#3.2">Section 3.2, - "Flags and Properties"</a> for more information). In this case the - DOM document is retained and will be released when the object model - is deleted. Note that since DOM nodes "out-live" the parsing function - call, you need to initialize the Xerces-C++ runtime before calling - one of the parsing functions with the <code>keep_dom</code> flag and - terminate it after the object model is destroyed (see - <a href="#3.1">Section 3.1, "Initializing the Xerces-C++ Runtime"</a>).</p> - - <p>If the <code>keep_dom</code> flag is passed - as the second argument to the copy constructor and the copy - being made is of a complete tree, then the DOM association - is also maintained in the copy by cloning the underlying - DOM document and reestablishing the associations. For example:</p> - - <pre class="c++"> -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Parse XML to object model. - // - std::unique_ptr<type> r (root ( - "root.xml", - xml_schema::flags::keep_dom | - xml_schema::flags::dont_initialize)); - - // Copy without DOM association. - // - type copy1 (*r); - - // Copy with DOM association. - // - type copy2 (*r, xml_schema::flags::keep_dom); -} - -XMLPlatformUtils::Terminate (); - </pre> - - - <p>To obtain the corresponding DOM node from an object model node - you will need to call the <code>_node</code> accessor function - which returns a pointer to <code>DOMNode</code>. You can then query - this DOM node's type and cast it to either <code>DOMAttr*</code> - or <code>DOMElement*</code>. To obtain the corresponding object - model node from a DOM node, the DOM user data API is used. The - <code>xml_schema::dom::tree_node_key</code> variable contains - the key for object model nodes. The following schema and code - fragment show how to navigate from DOM to object model nodes - and in the opposite direction:</p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="a" type="string"/> - </sequence> -</complexType> - -<element name="root" type="object"/> - </pre> - - <pre class="c++"> -using namespace xercesc; - -XMLPlatformUtils::Initialize (); - -{ - // Parse XML to object model. - // - std::unique_ptr<type> r (root ( - "root.xml", - xml_schema::flags::keep_dom | - xml_schema::flags::dont_initialize)); - - DOMNode* n = r->_node (); - assert (n->getNodeType () == DOMNode::ELEMENT_NODE); - DOMElement* re = static_cast<DOMElement*> (n); - - // Get the 'a' element. Note that it is not necessarily the - // first child node of 'root' since there could be whitespace - // nodes before it. - // - DOMElement* ae; - - for (n = re->getFirstChild (); n != 0; n = n->getNextSibling ()) - { - if (n->getNodeType () == DOMNode::ELEMENT_NODE) - { - ae = static_cast<DOMElement*> (n); - break; - } - } - - // Get from the 'a' DOM element to xml_schema::string object model - // node. - // - xml_schema::type& t ( - *reinterpret_cast<xml_schema::type*> ( - ae->getUserData (xml_schema::dom::tree_node_key))); - - xml_schema::string& a (dynamic_cast<xml_schema::string&> (t)); -} - -XMLPlatformUtils::Terminate (); - </pre> - - <p>The 'mixed' example which can be found in the XSD distribution - shows how to handle the mixed content using DOM association.</p> - - <h2><a name="5.2">5.2 Binary Serialization</a></h2> - - <p>Besides reading from and writing to XML, the C++/Tree mapping - also allows you to save the object model to and load it from a - number of predefined as well as custom data representation - formats. The predefined binary formats are CDR (Common Data - Representation) and XDR (eXternal Data Representation). A - custom format can easily be supported by providing - insertion and extraction operators for basic types.</p> - - <p>Binary serialization saves only the data without any meta - information or markup. As a result, saving to and loading - from a binary representation can be an order of magnitude - faster than parsing and serializing the same data in XML. - Furthermore, the resulting representation is normally several - times smaller than the equivalent XML representation. These - properties make binary serialization ideal for internal data - exchange and storage. A typical application that uses this - facility stores the data and communicates within the - system using a binary format and reads/writes the data - in XML when communicating with the outside world.</p> - - <p>In order to request the generation of insertion operators and - extraction constructors for a specific predefined or custom - data representation stream, you will need to use the - <code>--generate-insertion</code> and <code>--generate-extraction</code> - compiler options. See the - <a href="https://www.codesynthesis.com/projects/xsd/documentation/xsd.xhtml">XSD - Compiler Command Line Manual</a> for more information.</p> - - <p>Once the insertion operators and extraction constructors are - generated, you can use the <code>xml_schema::istream</code> - and <code>xml_schema::ostream</code> wrapper stream templates - to save the object model to and load it from a specific format. - The following code fragment shows how to do this using ACE - (Adaptive Communication Environment) CDR streams as an example:</p> - - <pre class="xml"> -<complexType name="object"> - <sequence> - <element name="a" type="string"/> - <element name="b" type="int"/> - </sequence> -</complexType> - -<element name="root" type="object"/> - </pre> - - <pre class="c++"> -// Parse XML to object model. -// -std::unique_ptr<type> r (root ("root.xml")); - -// Save to a CDR stream. -// -ACE_OutputCDR ace_ocdr; -xml_schema::ostream<ACE_OutputCDR> ocdr (ace_ocdr); - -ocdr << *r; - -// Load from a CDR stream. -// -ACE_InputCDR ace_icdr (buf, size); -xml_schema::istream<ACE_InputCDR> icdr (ace_icdr); - -std::unique_ptr<object> copy (new object (icdr)); - -// Serialize to XML. -// -root (std::cout, *copy); - </pre> - - <p>The XSD distribution contains a number of examples that - show how to save the object model to and load it from - CDR, XDR, and a custom format.</p> - - <!-- Appendix A --> - - - <h1><a name="A">Appendix A — Default and Fixed Values</a></h1> - - <p>The following table summarizes the effect of default and fixed - values (specified with the <code>default</code> and <code>fixed</code> - attributes, respectively) on attribute and element values. The - <code>default</code> and <code>fixed</code> attributes are mutually - exclusive. It is also worthwhile to note that the fixed value semantics - is a superset of the default value semantics. - </p> - - <!-- border="1" is necessary for html2ps --> - <table id="default-fixed" border="1"> - <tr> - <th></th> - <th></th> - <th colspan="2">default</th> - <th colspan="2">fixed</th> - </tr> - - <!-- element --> - - <tr> - <th rowspan="4">element</th> - <th rowspan="2">not present</th> - <th>optional</th> - <th>required</th> - <th>optional</th> - <th>required</th> - </tr> - <tr> - <td>not present</td> - <td>invalid instance</td> - <td>not present</td> - <td>invalid instance</td> - </tr> - - - <tr> - <th>empty</th> - <td colspan="2">default value is used</td> - <td colspan="2">fixed value is used</td> - </tr> - - <tr> - <th>value</th> - <td colspan="2">value is used</td> - <td colspan="2">value is used provided it's the same as fixed</td> - </tr> - - <!-- attribute --> - - <!-- element --> - - <tr> - <th rowspan="4">attribute</th> - <th rowspan="2">not present</th> - <th>optional</th> - <th>required</th> - <th>optional</th> - <th>required</th> - </tr> - <tr> - <td>default value is used</td> - <td>invalid schema</td> - <td>fixed value is used</td> - <td>invalid instance</td> - </tr> - - - <tr> - <th>empty</th> - <td colspan="2">empty value is used</td> - <td colspan="2">empty value is used provided it's the same as fixed</td> - </tr> - - <tr> - <th>value</th> - <td colspan="2">value is used</td> - <td colspan="2">value is used provided it's the same as fixed</td> - </tr> - - </table> - - </div> -</div> - - -</body> -</html> diff --git a/doc/cxx/tree/manual/manual.html2ps.in b/doc/cxx/tree/manual/manual.html2ps.in deleted file mode 100644 index 5629122..0000000 --- a/doc/cxx/tree/manual/manual.html2ps.in +++ /dev/null @@ -1,66 +0,0 @@ -@@html2ps { - option { - toc: hb; - colour: 1; - hyphenate: 1; - titlepage: 1; - } - - datefmt: "%B %Y"; - - titlepage { - content: " -<div align=center> - <h1><big>C++/Tree Mapping User Manual</big></h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> - <h1> </h1> -</div> - <p>Revision $[revision] $D</p> - <p>Copyright © @copyright@.</p> - - <p>Permission is granted to copy, distribute and/or modify this - document under the terms of the - <a href='https://www.codesynthesis.com/licenses/fdl-1.2.txt'>GNU Free - Documentation License, version 1.2</a>; with no Invariant Sections, - no Front-Cover Texts and no Back-Cover Texts. - </p> - - <p>This document is available in the following formats: - <a href='https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/index.xhtml'>XHTML</a>, - <a href='https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.pdf'>PDF</a>, and - <a href='https://www.codesynthesis.com/projects/xsd/documentation/cxx/tree/manual/cxx-tree-manual.ps'>PostScript</a>.</p>"; - } - - toc { - indent: 2em; - } - - header { - odd-right: $H; - even-left: $H; - } - - footer { - odd-left: $D; - odd-center: $T, v$[revision]; - odd-right: $N; - - even-left: $N; - even-center: $T, v$[revision]; - even-right: $D; - } -} - -body { - font-size: 12pt; - text-align: justify; -} - -pre { - font-size: 10pt; -} |