Genivia Home Documentation
XML DOM & XPath

updated Fri Nov 4 2016
 
XML DOM & XPath

Table of Contents

Introduction

XML data bindings provide a powerful mechanism to parse, validate, manipulate, and write XML data with ease as if we are working with native C and C++ data. Occasionally we will need to work with raw XML data. Raw XML can be parsed with a DOM parser and stored in a DOM node graph. The DOM (Document Object Model) node graph (actually a node tree) can be inspected to retrieve text and other values. We can iteratively and recursively traverse the XML tree via its DOM node graph at increasingly deeper levels to inspect and retrieve values of XML elements and attributes. We can also use XPath queries to locate data deep in the DOM node hierarchy to retrieve specific values.

The gSOAP XML DOM API is compact in size but offers a wealth of operations to construct, inspect, traverse, and search data in DOM node graphs. The DOM API offers powerful constructors to construct a DOM node graph for XML with smart XML namespace handling to simplify coding. The DOM API also offers a hybrid DOM + data binding approach that allows you to embed serializable C/C++ data types including user-defined structs and classes. You can also search and retrieve values from XML and XML with namespaces using the DOM API's smart XML namespace handling.

We also offer a new code generation tool, domcpp, that greatly simplifies writing DOM API code by generating DOM construction code, DOM inspection code, and XPath code to query a DOM. The generated code uses the DOM API transparently in C and C++ (C++ by default and pure C as an option). So it is easy to get familiar with the DOM API by using the domcpp tool. In the first part of this introduction we show how to effectively use the domcpp tool to get started.

This document describes both the C and C++/C++11 APIs, see table of contents.

XML DOM and gSOAP

To help you quickly develop C/C++ code to construct, inspect, and search DOM node graphs using the XML DOM API, we offer a new code generator domcpp with the gSOAP package (version 2.8.28 and up). You can find the domcpp tool with the DOM examples in gsoap/samples/dom.

The domcpp command-line tool auto-generates C or C++ code from an XML fragment. The generated code creates a DOM node graph for this fragment, which can be further tweaked as necessary. Let's demonstrate this with an example menu.xml XML file, where we show each command executed in a command shell followed by the results displayed in the terminal:

cat menu.xml
<menu id="file" value="File">
<popup>
<menuitem value="New" onclick="CreateNewDoc()" />
<menuitem value="Open" onclick="OpenDoc()" />
<menuitem value="Close" onclick="CloseDoc()" />
</popup>
</menu>
domcpp menu.xml
#include "soapH.h"
{
struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
ctx->double_format = "%lG";
xsd__anyType dom(ctx, "menu");
dom.att("id") = "file";
dom.att("value") = "File";
dom["popup"]["menuitem"][1].att("value") = "New";
dom["popup"]["menuitem"][1].att("onclick") = "CreateNewDoc()";
dom["popup"]["menuitem"][2].att("value") = "Open";
dom["popup"]["menuitem"][2].att("onclick") = "OpenDoc()";
dom["popup"]["menuitem"][3].att("value") = "Close";
dom["popup"]["menuitem"][3].att("onclick") = "CloseDoc()";
std::cout << dom << std::endl;
soap_destroy(ctx); // delete objects
soap_end(ctx); // delete DOM data
soap_free(ctx); // free context
}

The C++ DOM API is designed around the concept of XPath's path location to construct a node graph by aligning the C++ [] overloaded operator to XPath's path steps and position predicates. You can use domcpp option -e to reveal the XPath for each assignment statement. For example:

xsd__anyType dom(ctx, "menu");
// /menu/@id = file
dom.att("id") = "file";
// /menu/@value = File
dom.att("value") = "File";
// /menu/popup/menuitem[1]/@value = New
dom["popup"]["menuitem"][1].att("value") = "New"

Domcpp option -e adds explanatory comments to the generated code, which explains what the code does to help you understand the DOM API by including XPath path locations and other details.

You can use domcpp option -M to narrow the generated code down to the essentials, without the initialization and cleanup parts of the code. This makes the generated code suitable for direct inclusion in your codebase.

Generating code to populate a DOM node graph is one option. Another option is to generate code to inspect a DOM node graph. Use domcpp option -i to generate code to inspect the DOM node graph of XML parsed from input, given that the XML file provided with option -i serves as a generic template:

domcpp -i menu.xml
#include "soapH.h"
{
struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
ctx->double_format = "%lG";
xsd__anyType dom(ctx);
std::cin >> dom;
if (dom.soap->error)
exit(EXIT_FAILURE); // error parsing XML
#define USE_ATT(path, text) std::cout << path << " = " << text << std::endl
#define USE_ELT(path, text) std::cout << path << " = " << text << std::endl
if ((att = dom.att_get("id")))
USE_ATT("/menu/@id", att->get_text());
if ((att = dom.att_get("value")))
USE_ATT("/menu/@value", att->get_text());
if ((elt = dom.elt_get("popup")))
{
xsd__anyType& dom_popup = *elt;
for (xsd__anyType *it = dom_popup.elt_get("menuitem"); it; it = it->get_next())
{
xsd__anyType& dom_popup_menuitem = *it;
if ((att = dom_popup_menuitem.att_get("value")))
USE_ATT("/menu/popup/menuitem/@value", att->get_text());
if ((att = dom_popup_menuitem.att_get("onclick")))
USE_ATT("/menu/popup/menuitem/@onclick", att->get_text());
}
}
std::cout << dom << std::endl;
soap_destroy(ctx); // delete objects
soap_end(ctx); // delete DOM data
soap_free(ctx); // free context
}

You can also use domcpp option -p to generate efficient XPath query code to query and retrieve specific XML values.

For example, let's write an XPath query to display the authors of books in a store. We will read the XML data from std:cin and filter the authors with the query /store/book/@author to collect them in a DOM node graph y containing the query results with domcpp option -y. We generate the code from the command line with domcpp as follows:

domcpp -M -p'/store/book/@author' -y
xsd__anyType dom(ctx);
std::cin >> dom;
// XPath: /store/book/@author
xsd__anyType y(ctx, "results");
#define QUERY_YIELD_ATT(v) y.add(xsd__anyType(ctx, "result").add(v));
#define QUERY_YIELD_ELT(v) y.add(xsd__anyType(ctx, "result").add(v));
if (dom.match("store"))
{
size_t pos = 1;
for (xsd__anyType *it = dom.elt_get("book"); it; it = it->get_next(), ++pos)
{
xsd__anyType& v = *it;
if ((att = v.att_get("author")))
{
xsd__anyAttribute& v = *att;
QUERY_YIELD_ATT(v);
}
}
}

Let's apply this query to the store.xml file that you can find in section XPath by example:

./query < store.xml
<results>
<result author="Nigel Rees"/>
<result author="Evelyn Waugh"/>
<result author="Herman Melville"/>
<result author="J. R. R. Tolkien"/>
</results>

You can compile this example yourself with:

domcpp -o query.cpp -m -p'/store/book/@author' -y
soapcpp2 -CSL import/dom.h
c++ -o query query.cpp soapC.cpp dom.cpp stdsoap2.cpp

You can also embed C/C++ code in XPath queries to filter and select values from XML at runtime.

The domcpp code generator aims to produce clean, high-quality and readable C and C++ code. You can also embed C/C++ code in XPath queries to filter and select values from XML data at runtime.

We will present domcpp in more detail in the next section. The remainder of this document explains how you can use the DOM C and C++ APIs to create XML data, access XML data, send/recv data via REST, and to read/write XML data to files, streams, and string buffers.

XML DOM versus XML data bindings

XML data bindings greatly simplify the development of XML applications. Compared to DOM manipulation and search by DOM tree traversal, all that is needed are declarations of the C/C++ data types that need to be serializable in XML by using the soapcpp2 tool the generate the data binding implementation. Furthermore, XML parsing is much faster with XML data bindings. XML is efficiently pulled and converted by the C/C++ deserializers.

For more details, see XML Data Bindings.

On the other hand, you can use the XML DOM API to develop REST operations for SOAP and XML messaging by composing and decomposing the XML manually. For example in C++:

#include "soapH.h"
struct Namespace namespaces[] = {
{"SOAP-ENV", "http://schemas.xmlsoap.org/soap/envelope/"},
{"SOAP-ENC", "http://schemas.xmlsoap.org/soap/encoding/"},
{"xsi", "http://www.w3.org/2001/XMLSchema-instance"},
{"xsd", "http://www.w3.org/2001/XMLSchema"},
{"t", "http://tempuri.org/t.xsd"},
{NULL, NULL}
};
int main()
{
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
const char *endpoint = "http://www.cs.fsu.edu/~engelen/gmtlitserver.cgi";
xsd__anyType req(ctx, "SOAP-ENV:Envelope"); // DOM with SOAP envelope
xsd__anyType res(ctx); // DOM for SOAP response
req["SOAP-ENV:Body"]["t:gmt"] = ""; // SOAP Body with <t:gmt>
soap_dom_call(ctx, // make a call (POST)
endpoint, // the service endpoint URL
NULL, // no SOAPAction header
req, // request t:gmt
res); // response, if OK
if (ctx->error) ... // handle IO error
xsd__anyType::iterator i = res.find("param-1");
if (i != res.end()) // if found, display time
cout << "Current time = " << i->get_text() << endl;
}

The SOAP_DOM_TREE context flag used here disables the deserialization of embedded objects and data, which will be discussed in the next section.

Other REST methods can be used to manage the state of resources via URL references, allowing for the storage of data (HTTP PUT), retrieval of data (HTTP GET), and removal of data (HTTP DELETE) from a resource. The soap_dom_call function takes NULL as the service input request and service output response parameters to facilitate PUT (NULL for response), GET (NULL for request), and DELETE (both request and response are NULL).

However, XML data bindings greatly simplify the development of XML applications such as client and server runtimes for XML messaging and message validation. Compared to DOM manipulation and search by DOM tree traversal, all that is needed are declarations of server operations as C functions and the C/C++ data types passed to these functions.

For example, with XML data bindings it only takes one line to declare the SOAP/XML service operation:

t__gmt(time_t*);

The soapcpp2 tool generates the data binding implementation such that t__gmt can be invoked as a function in a client-side C or C++ runtime:

#include "soapH.h"
#include "t.nsmap"
int main ()
{
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT);
time_t t;
soap_call_t__gmt(ctx, endpoint, NULL, &t);
if (ctx->error == SOAP_OK)
cout << "Current time = " << ctime(t) << endl;
}

The gSOAP DOM API offers a hybrid DOM + data binding approach that allows you to use DOM nodes a members of serializable structs and classes (these DOM parts are auto-generated by wsdl2h with option -d). This hybrid approach also allows you to embed serializable C/C++ data into a DOM node graph. This way, you get the best of both XML DOM and XML data bindings.

Embedding serializable objects and data in DOM element nodes

To embed serializable data types we first declare these types in a header file for soapcpp2 to generate the data binding code. We also import dom.h from gsoap/import to use the DOM API:

#import "import/dom.h"
//gsoap t schema namespace: urn:demotime
typedef time_t _t__time;

This is a very small example to illustrate the concept. You can add declarations of data types to serialize. If namespaces are used in the XML document, then follow the naming conventions outlined in XML Data Bindings: Colon notation versus name prefixing.

Importing dom.h is automated with wsdl2h option -d when running wsdl2h on WSDL and XSD files to produce a header file with the data binding interface for soapcpp2. Here, we assume you are starting with a header file for soapcpp2 with the declarations of C/C++ types to serialize.

We use soapcpp2 to generate the data binding implementation code together with the DOM API in soapStub.h, soapH.h, soapC.cpp and a namespace table file t.nsmap and t.xsd schema that declares a time type (at line 11) and also an element (at line 15) because by convention the _t__time type name is prefixed with a _:

<schema targetNamespace="urn:demotime"
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:t="urn:demotime"
xmlns="http://www.w3.org/2001/XMLSchema"
elementFormDefault="unqualified"
attributeFormDefault="unqualified">
<import namespace="http://schemas.xmlsoap.org/soap/encoding/"/>
<simpleType name="time">
<restriction base="xsd:dateTime">
</restriction>
</simpleType>
<element name="time" type="xsd:dateTime"/>
</schema>

For details on declaring C/C++ types for XML data bindings and serialization, see XML Data Bindings: Defining document root elements.

You can use these serializable types declared in the header file as embedded serializable objects and data in DOM element nodes. The SOAP_DOM_NODE context flag should be set to enable this feature.

In the following example we use _t__time to construct a DOM for an XML root element <t:time> that contains an embedded serializable time_t value:

#include "soapH.h"
#include "t.nsmap"
int main()
{
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_NODE);
time_t t = time(0);
xsd__anyType dom(ctx, NULL, "t:time", &t, SOAP_TYPE__t__time);
cout << dom << endl; // write XML <t:time>...</t:time>
if (dom.soap->error)
... // handle IO error
}

and to parse and deserialize a <t:time> into a time_t value:

#include "soapH.h"
#include "t.nsmap"
int main()
{
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_NODE);
xsd__anyType dom(ctx);
cin >> dom; // parse XML <t:time>...<t:time>
if (dom.soap->error)
... // handle IO error
const time_t *t = static_cast<const time_t*>(dom.get_node(SOAP_TYPE__t__time));
if (t)
cout << "Time = " << ctime(t) << endl;
}

With the SOAP_DOM_NODE flag set, there are two ways the C/C++ types are deserialized into DOM elements: by element tag name matching or by xsi:type matching:

  1. If an xsi:type attribute is present (which requires preparations on the sending side), then its value is matched against the names of the C/C++ types to select a deserializer to parse and deserialize the data into a DOM element node as embedded objects/data. If no xsi:type attribute is present or if no serializer is found, then:
  2. The element tag is matched against the names of serializable C/C++ types to select a deserializer to deserialize the data into a DOM element node as embedded objects/data. If no deserializer is found then the XML element is parsed and stored as a DOM element node.

While SOAP_DOM_NODE instructs the DOM parser to deserialize the values of C/C++ types that match element tags, the opposite, SOAP_DOM_TREE prevents deserialization, even when an id attribute is present in the XML payload. The default is to deserialze only when an id attribute is present in the XML payload. This is to ensure that the SOAP encoding protocol deos not break when id-ref attibutes are used for multireferenced objects.

The domcpp command-line tool

The domcpp command produces high-quality, readable and reusable source code. The generated code can be readily used in your projects to populate XML data and retrieve XML data, thereby saving you substantial time and effort to write DOM API code in C or C++.

The domcpp command-line tool generates C or C++ source code to populate a DOM node graph with the XML data given in an XML file. The command also has an option -i to generate source code to inspect parsed XML by using an XML file as a generic template for this code. And option -p generates efficient source code for XPath queries. Even stand-alone XPath query filter applications can be auto-generated with option -m (for main).

Compiling the domcpp command

You will find domcpp and the XML DOM examples in the gSOAP package in gsoap/samples/dom.

To build domcpp, install gSOAP and build all sample codes as follows:

./configure --enable-samples
make
make install

This builds the command-line tool domcpp in gsoap/samples/dom from where you can use it and/or copy it for use with your projects.

If you do not have the samples built, you can use soapcpp2 (or soapcpp2.exe in gsoap/bin/win32) from the command line to generate the C++ code required for domcpp:

cd gsoap/samples/dom
soapcpp2 -CSL ../../import/dom.h
c++ -I../.. -o domcpp domcpp.cpp soapC.cpp ../../dom.cpp ../../stdsoap2.cpp

The above builds the domcpp command-line tool.

Command-line options

The domcpp command takes several options and an optional XML input file:

domcpp [-c] [-e] [-f%fmt] [-h] [-i] [-l] [-m] [-M] [-n] [-O] [-ofile] [-ppath] [-rroot] [-xcode] [-y] [infile]

where the domcpp command-line options are:

Option Description
-c generate C code instead of C++
-e add explanatory comments to the generated code
-f%fmtuse %fmt to format double floats, e.g. -f%lG
-h display help message
-i generate code to inspect DOM node graph parsed from XML input
-l generate code for option -i to store values in local variables
-m generate stand-alone code by adding main()
-M generate minimal code unadorned with initialization and cleanup
-n generate XML namespace table
-O optimize code by factoring common indices when applicable
-ofile save source code to file
-ppath generate XPath query code for path
-rroot use root instead of root value dom in the generated code
-xcode generate code that executes code for each XPath query result
-y generate code that yields an array y of XPath query results
infile optional XML file to parse
- read XML from standard input

The domcpp command takes an XML input file infile to generate code to construct a DOM node graph in C/C++ for this XML, or, with option -i, to generate code that reads XML from input and traverses it to inspect its value by using the XML input file infile as a template to match against. For option -i, if you want additional code that uses local variables to store boolean, integer, and floating point values retrieved from the DOM node graph, then also use option -l with option -i.

Use option -c to generate C code instead of C++ and use option -e to add explanatory comments to the generated code.

The domcpp command emits source code to standard output or to the file specified with option -o.

Minimalistic code is generated with option -M, which is useful to automate pasting of the unadorned source code into the source code of your project.

Optimized code is generated with option -O by factoring common array indices and object field names. This produces more elaborate code that is more efficient but may be harder to read and modify. This option has no effect on the code generated with option -i.

The default name of the root value in the generated source code is dom. To change this name use option -r. Do not use the name v, which represents the current value in XPath query C/C++ code. Other variable names to avoid are it, att, elt, and pos, since these are internally used by the generated code.

To include a namespace table in the generated code, use option -n. This option simplifies the use of the DOM API by removing namespace URIs passed to API functions, as the use of qualified tag names will suffice in most cases.

Use option -p to generate code that filters XML data from a source of input with an XPath query path. Option -x specifies XPath query code to execute for each query result. The default action in the generated code is to print each query result value in XML separated by newlines. Option -y yields an XML DOM with root results and each query result stored in a result element. Option -x overrides option -y.

To generate a stand-alone application use option -m. This option is useful for testing XPath query filters given with option -p.

Option -f%fmt sets the floating point double precision format to use in the generated code. By default, domcpp emits floating point numbers with up to 17 digit mantissas to preserve precision. Use -f%lG for the shortest floating point representation.

XPath syntax support for domcpp

XPath (the XML Path Language), is a query language for selecting nodes from an XML document. A XPath query returns the XML elements and attributes of a DOM node structure of an XML document that match the selection criteria.

An XPath expression specifies a data query to select elements and attributes (and their values) typically starting from the root node, and descending deeper into the node graph to match child nodes.

For example, suppose we have a <store> root element with a number of <book> child elements. Each <book> element has a title attribute with the title text and some other attributes which we will ignore for now. The following XPath query returns the titles of all books in the store:

/store/book/@title

This example illustrates the most important kind of expression in XPath, which is a location path. A location path consists of a sequence of location steps. Locations steps are separated by '/'. Each step consists of an axis followed by an optional node test and zero or more predicates.

The axis, node tests, and predicates supported by domcpp XPath expressions are listed in the three tables below.

Axis Nodes matched and returned
/ document root (when used at the start of the location path)
// descendant or self
. self
.. parent node
@ attribute node
tag tag name of attributes or elements in the null namespace
ns:tag tag name of attributes or elements in the ns namespace
* or *:* any tag name in any namespace
*:tag tag name in any namespace
ns:* all tags in the ns namespace

Note that domcpp XPath does not (yet) support the full XPath axis syntax.

Node test Nodes matched and returned
text() text-only child nodes

A text() node test is useful to select text-only child nodes when its siblings are a mix of text and elements. This mixed content of elements and text is more common in XHTML markup, but not in XML Web service messaging for example.

Zero or more predicates may follow to filter nodes by position, by path, or by a C/C++ expression:

Predicate Nodes filtered
[n] select nodes that are at the n-th position
[path] select nodes if XPath path matches
[?expr] select nodes if C/C++ expression expr is true or nonzero

Note that domcpp XPath does not (yet) support functions and operators in predicates. Instead, you can use C/C++ expressions to filter nodes where variable v refers to the axis node (i.e. self, which is an xsd__anyType element node or an xsd__anyAttribute attribute node) and the integer variable pos that refers to the query position.

Path subexpressions are grouped with ( and ). Grouping is useful to apply predicates to a specific part of the location path. Without grouping, a predicate only applies to the axis that immediately preceeds it. For example, x/y[1] selects the first child element y of all x elements. By contrast, (x/y)[1] selects the first y child element of the first x element that has at least one y child element.

XPath by example

Consider the following XML document:

<store>
<book
category="reference"
author="Nigel Rees"
title="Sayings of the Century"
price="8.95" />
<book
category="fiction"
author="Evelyn Waugh"
title="Sword of Honour"
price="12.99" />
<book
category="fiction"
author="Herman Melville"
title="Moby Dick"
isbn="0-553-21311-3"
price="8.99" />
<book
category="fiction"
author="J. R. R. Tolkien"
title="The Lord of the Rings"
isbn="0-395-19395-8"
price="22.99" />
<bicycle
color="red"
price="19.95" />
</store>

To match the title of the first book in a store, starting at the root indicated by /, we use the following XPath query expression:

/store/book[1]/@title

This query returns "Sayings of the Century" when applied to the XML document.

To try the XPath examples yourself, we suggest to create a store.xml file with the above XML and run domcpp from the command line to compile an XPath query as follows:

cd gsoap/samples/dom
soapcpp2 -CSL ../../import/dom.h
domcpp -o test-dom.cpp -m -p'/store/book[1]/@title'
c++ -I../.. -o test-dom test-dom.cpp soapC.cpp ../../dom.cpp ../../stdsoap2.cpp
./test-dom < store.xml

The soapC.cpp file, and the header files soapStub.h and soapH.h, are generated with the command soapcpp2 -CSL dom.h, which is done just once for all C++ DOM applications. Use soapcpp2 option -c to generate soapC.c for C (with the corresponding soapStub.h and soapH.h in C).

The compiled XPath query is applied to the store.xml document and returns the matches found. Use domcpp option -y to collect and return all matches in an XML document with root results:

<results>
  <result title="Sayings of the Century"/>
</results>

To match any tag name of an element or attribute, we use a wildcard *:

/store/*/@price

This matches and returns the prices of all items in the store. If we want to obtain all price attributes at any level of the node graph, we can also use // called "recursive descent" or simply "recurse" to select descendents (and self when used at the root):

//@price

This query selects all price attributes in an XML document at any level, including the price attribute of the document root if any. For our example XML store this query returns the prices of all items in the store.

To select all attributes of a node we use a wildcard @*, for example to obtain all attributes of the first book in the store:

/store/book[1]/@*

We can predicate node selections with criteria based on path location matches. For example, to select items from the store that have an isbn attribute, and then obtain their price, we use a path location predicate [@isbn]:

/store/*[@isbn]/@price

If we only want to find the first result, we specify a position predicate [1] and apply it to all location steps in the XPath expression:

/(store/*[@isbn]/@price)[1]

A predicate applies to the axis and its predicates located to the left of the predicate as part of one location step, so we used parenthesis to group the location steps to apply the [1] position predicate.

There can be several ways an XPath query can be formulated. Here is a different one without parenthesis to obtain the same result:

/store/*[@isbn][@price][1]/@price

This query matches books with both an isbn and price attribute and returns the price of the first match found.

Because all store items are priced in our XML store example, we can also use the following XPath query:

/store/*[@isbn][1]/@price

However, it is generally better to use more robust XPath queries that are not based on assumptions, such as the previous queries for our XML store that cope with price omissions.

More complex queries can be formulated by embedding C/C++ expressions in the query as predicates with [?expr]. This is an extension of domcpp and not standard XPath expression syntax.

For example, the following XPath query with C++ predicate selects store items with prices under 10.0:

/store//@price[?(double)v < 10.0]

and in C:

/store//@price[?soap_att_get_double(v) < 10.0]

We can combine the C/C++ predicate with a path predicate to get books priced under 10.0:

/store/book[@price[?(double)v < 10.0]]

and to obtain the book titles of these books only:

/store/book[@price[?(double)v < 10.0]]/@title

XPath queries do not modify the node graph searched. So you do not need to worry about predicates with position indices that are out of bounds or about tag names that refer to non-existent attributes or elements.

Warning
In this respect we should caution you about using C/C++ expressions that modify DOM nodes, since this may affect the query results in unpredictable ways. For example:
//[?(double)v.att("price") < 10.0]
The att("price") method sets an attribute, not just reads it, so disaster strikes as we visit every node in the DOM node graph to set a price attribute!

As you can see, C/C++ predicates can inspect the current XPath node by accessing variable v. This variable is either an xsd__anyType DOM element node or an xsd__anyAttribute DOM attribute node, depending on the path. Variable v is a reference in C++ and a pointer in C.

Besides the current node v you can also access the XPath root node dom. Instead of the default root name dom, you can change dom to another name with domcpp option -r. You can also access pos which corresponds to the XPath position() value.

You can access variables and functions in C/C++ predicates, but do not modify the variables it and pos which are internally used by the generated XPath query code.

For example, you can access variable pos (but you should not change it):

/(store/*[@isbn]/@price)[?pos > 1]

This XPath query returns the price of books with an isbn, but after the first match was discarded by the predicate.

Other temporary variables used internally are a pointer to an attribute att and a pointer to an element elt. You may set and use these as temporaries in the scope of a C/C++ predicate.

The C/C++ predicates are quite powerful and can be used to filter values as shown earlier but also to select attributes and elements by matching their tag name using v,match(tag) in C++ and soap_att_match(v, NULL, tag) or soap_elt_match(v, NULL, tag) in C:

//book/@*[?v.match(argv[1])]

This assumes that the command-line argument (argv[1]) of the application is a book attribute name. Otherwise, no results are returned.

After compiling the XPath query with

domcpp -o test-dom.cpp -m -p'//book/@*[?v.match(argv[1])]'
c++ -I../.. -o test-dom test-dom.cpp soapC.cpp ../../dom.cpp ../../stdsoap2.cpp

we can obtain the book titles with:

./test-dom title < store.xml

Finally, let's use the value of argv to filter products in the store by a given command-line argument price:

domcpp -m -p'//@price[?(double)v < strtod(argv[1], NULL))]'

and in C:

domcpp -c -m -p'//@price[?soap_att_get_double(v) < strtod(argv[1], NULL))]'

Note that the strtod function returns the double float value of argv[1] and is repeately called. A more efficient implementation would store the value in a temporary variable and use the temporary in the C/C++ predicate.

Understanding XML namespaces

XML namespaces are commonly used with XML documents that are instances of XML schemas. A full introduction is beyond the scope of this document. We instroduce the basics here to help you understand why XML namespaces are important and how they are used in the C/C++ DOM API of gSOAP.

XML namespace are important when XML documents contain instances of multiple XML schemas. XML elements and attributes that are declared in separate schemas should be distinguishable and their content should be verifyable. Tag name conflicts that may be caused by combining multiple XML schemas to define an XML document can be resolved by binding elements and attributes to specific XML namespaces.

XML elements and attributes are bound to an XML namespace by using a namespace prefix in qualified tag names of elements and attributes. A qualified tag is of the form prefix:name, such that the prefix is bound in the scope of the tag to a namespace URI with an xmlns:prefix="URI" declaration. Unqualified tags can be bound to a namesapce with a xmlns="URI" default namespace declaration.

An XML namespace is a Uniform Resource Identifier (URI). There are two general forms of URI: Uniform Resource Locators (URL) and Uniform Resource Names (URN).

Consider for example the XML document with elements and attributes in three distinct XML namespaces "urn:one", "urn:two", and "urn:three":

<x:e1 a1="1" xmlns:x="urn:one" xmlns:y="urn:two" y:a2="2">
<e2 xmlns="urn:three" a3="3">
<e3 />
</e2>
</x:e1>

Element e1 is bound to "urn:one", e2 and e3 are bound to "urn:three". Attribute a1 has no namespace (it is in the null namespace), a2 is bound to "urn:two" and a3 is bound to "urn:three".

A namespace declaration (xmlns) scope extends to all descendants or until redeclared. The x and y prefixes declared in element e1 have a scope of visibility in all elements, since e1 is the root. The default namespace declaration in element e2 has a scope of visibility in element e2 and extends to all of its attributes and all of its child elements.

An XML namespace may refer to the defining XML schema by a URL, but often they do not. The URI string itself should be sufficiently unique, which is important to distinguish XML elements and attributes in different XML namespaces. This is similar to C++ namespaces. A unique c++ namespace name suffices (we do not need to know where the source files are located).

The DOM API aims to simplify XML document construction and XML analysis by offering an API that does not require full namespace URIs to passed to the API functions when the URIs are defined in a namespace table. You can simply use qualified tag names to construct XML documents in a DOM and to analyze a DOM after XML parsing. This approach works on the basis of internal normalization so that two prefixes match when their associated URI matches (i.e. prefix names may differ but still match when their URIs match).

Assume we have the following namespace table of {"prefix", "URI"} bindings that includes the three namespace URIs used in our example XML document, but we pick three different prefixes a, b, and c:

struct Namespace namespaces[] =
{
{"SOAP-ENV", "http://schemas.xmlsoap.org/soap/envelope/", "http://www.w3.org/*/soap-envelope", NULL},
{"SOAP-ENC", "http://schemas.xmlsoap.org/soap/encoding/", "http://www.w3.org/*/soap-encoding", NULL},
{"xsi", "http://www.w3.org/2001/XMLSchema-instance", "http://www.w3.org/*/XMLSchema-instance", NULL},
{"xsd", "http://www.w3.org/2001/XMLSchema", "http://www.w3.org/*/XMLSchema", NULL},
{"a", "urn:one"},
{"b", "urn:two"},
{"c", "urn:three"},
{NULL, NULL, NULL, NULL}
};

With this table defined globally (or by using soap_set_namespaces() see next section), we can conveniently use qualified tag names with the prefixes a, b, and c to construct and/or match XML elements and attributes in these namespaces, because namespace matching is based on the URI of a prefix.

Therefore, the following qualified tag names match the following elements and attributes in the example XML document:

Tag name Matches
"a:e1" element <x:e1> in namespace scope xmlns:x="urn:one"
"a1" attribute a1 in the null namespace
"b:a2" attribute y:a2 in namespace scope xmlns:y="urn:two"
"c:e2" element <e2> in namespace scope xmlns="urn:three"
"c:a3" attribute a3 in namespace scope xmlns="urn:three"
"c:e3" element e3 in namespace scope xmlns="urn:three"

The choice of prefixes a, b, and c is rather arbitrary. We can pick prefix names for the table as long as they are unique and are valid names (we refer to the XML syntax of "colonized names" and "non-colonized names").

The domcpp option -n includes XML namespaces in a namespace table in the auto-generated code by copying namespace bindings from the XML document analyzed by the tool to a namespace table.

Note
You can also use the wsdl2h tool to retrieve namespaces and declare your own prefix names in typemap.dat. The soapcpp2 tool generates the data binding implementation and a .nsmap file with the namespace bindings.

You can still use the DOM API without a namespace table or with an incomplete table, but the namespace URIs that are left out from the table must be specified as additional ns parameters in the DOM API functions.

C++ XML DOM API

XML is represented as a DOM node graph internally consisting of values (text strings of character data in UTF-8), attribute nodes with tag name and optional namespace URI, and element nodes with optional tag names and optional namespace URI. Tag names are stored in strings of UTF-8. Tag names can be qualified of the form q:tag or unqualified. An element tag name that is NULL represents a text-only node, which will only be present in mixed content of elements and text character data (as in XHTML markup).

XML namespace handling is "smart" in the gSOAP DOM API: the DOM engine matches and stores XML with namespace URIs to ensure that a prefix qualifier is always locally bound to a namespace URI in the node graph. Furthermore, you can use qualified tags when constructing a DOM node instead of the full namespace URI, but only if the prefix and URI are defined in the namespace table (global struct Namespace namesapces[]. When working with wsdl2h and soapcpp2 you should include the soapcpp2-generated .nsmap file with namespace bindings that are applicable to your XML tags.

List of C++ files

The following files located in the package under gsoap are required to use the C++ DOM API:

The gSOAP header file dom.h declares the DOM for use in data bindings and should generally be imported with #import in gSOAP header files for the soapcpp2 tool.

You can also run soapcpp2 directly on import/dom.h for DOM-only projects. This generates soapStub.h, soapH.h and soapC.cpp. To auto-generate these files, execute:

soapcpp2 -CSL import/dom.h

Then compile and link the dom.cpp files listed above with the auto-generated soapC.cpp and stdsoap2.cpp:

c++ -I../.. -o myapp myapp.cpp soapC.cpp ../../dom.cpp ../../stdsoap2.cpp

Instead of stdsoap2.cpp you can link -lgsoap++.a when installed by the gSOAP package. Or link -lgsoapssl++.a to get both stdsoap2.cpp and dom.cpp with OpenSSL enabled for HTTPS:

c++ -DWITH_OPENSSL -I../.. -o myapp myapp.cpp soapC.cpp -lgsoapssl++ -lssl -lcrypto

Note that we compile with -DWITH_OPENSSL and link -lssl, and -lcrypto.

Because XML namespaces are required except for the most simple plain XML applications, you should include the generated .nsmap file:

#include "soapH.h"
#include "soap.nsmap"

Or define your own namespace table. For example, at a minimum you want a table with SOAP and XSD bindings:

#include "soapH.h"
struct Namespace namespaces[] =
{
{"SOAP-ENV", "http://schemas.xmlsoap.org/soap/envelope/", "http://www.w3.org/*/soap-envelope", NULL},
{"SOAP-ENC", "http://schemas.xmlsoap.org/soap/encoding/", "http://www.w3.org/*/soap-encoding", NULL},
{"xsi", "http://www.w3.org/2001/XMLSchema-instance", "http://www.w3.org/*/XMLSchema-instance", NULL},
{"xsd", "http://www.w3.org/2001/XMLSchema", "http://www.w3.org/*/XMLSchema", NULL},
{NULL, NULL, NULL, NULL}
};

Then also add entries of the form {"prefix", "URI"} before the last NULL entry to complete the table with xmlns prefix and URI bindings for your XML project.

C++ DOM API functions

This overview of the DOM API functions refers to the following types of parameters and values, which are used throughout this section:

Name Type and value
ctx struct soap *ctx context pointer (manages memory and IO)
ns const char *ns XML namespace URI string or NULL
tag const char *tag or const wchar_t *tag (un)qualified tag name
utag const char *utag or const wchar_t *utag unqualified tag name
qtag const char *qtag or const wchar_t *qtag qualified tag name
patt (wide) string (un)qualified tag name pattern with * wildcard(s)
text (wide) string
node const void *node pointer to serializable C/C++ data object
type SOAP_TYPE_T type identifier of serializable C/C++ data type T
att xsd__anyAttribute att DOM attribute node
elt xsd__anyType elt DOM element node
dom a DOM node (an element or an attribute node)
val bool, integer, float, or (wide) string
pos element position > 0 (XPath position numbering)

The API includes functions that take wide strings and normalizes these internally to UTF-8 encoded strings. However, when retrieving tags and text you will only be able to obtain UTF-8 strings. You can convert UTF-8 strings to wide strings that are managed by the context by using:

Creating a DOM root element node

Creating a DOM node graph starts with creating the root element node using one of the xsd__anyType constructors:

// 8 ways to create a root element node 'dom'
xsd__anyType dom(ctx); // unnamed element node managed by context
xsd__anyType dom(elt); // copy constructor (shares context and data)
xsd__anyType dom(ctx, utag); // element node <utag>
xsd__anyType dom(ctx, qtag); // element node <q:tag xmlns:q="ns">
xsd__anyType dom(ctx, ns, utag); // element node <utag xmlns="ns">
xsd__anyType dom(ctx, ns, qtag); // element node <q:tag xmlns:q="ns">
xsd__anyType dom(ctx, ns, tag, text); // element node as above with a text value
xsd__anyType dom(ctx, ns, tag, node, type); // element node as above with serializable object

The third and fourth constructors omit the ns parameter, which is the same as passing a NULL ns parameter.

The fourth constructor is convenient to use to construct an element node that is bound to an XML namespace by using a qualified tag name qtag without specifying the namespace URI with the ns parameter.

To use this approach, you should define the namespace prefix in qtag in a global namespace table, see Understanding XML namespaces, and as shown below:

// a global namespace table with {"prefix", "URI"} xmlns bindings
struct Namespace namespaces[] =
{
...
{"q", "urn:example"},
{NULL, NULL, NULL, NULL}
};
xsd__anyType dom(ctx, "q:tag"); // OK: "q" is in the namespace table
xsd__anyType dom(ctx, "bad:tag"); // BAD: "bad" is not in the namespace table
xsd__anyType dom(ctx, "urn:ok", "ok:tag"); // OK: "ok" is locally scoped in the DOM

You can also use multiple namespace tables and select the one you want to use with the soap_set_namespaces() function:

// a local (static) namespace table with {"prefix", "URI"} xmlns bindings
struct Namespace my_dom_namespaces[] =
{
...
{"q", "urn:example"},
{NULL, NULL, NULL, NULL}
};
soap_set_namespaces(ctx, my_dom_namespaces); // use namespace table for qualified tags
xsd__anyType dom(ctx, "q:tag"); // OK: "q" is in the namespace table
...
soap_set_namespaces(ctx, namespaces); // restore to global table

To set or change the tag of an element, use:

dom.set(ns, utag); // element node <utag xmlns="ns">
dom.set(ns, qtag); // element node <q:tag xmlns:q="ns">

Assigning child nodes and values to a DOM node

After creating a root element node we can start populating the root with attributes, child elements, and values.

To assign a value to a DOM attribute node or element node dom, we can use the the xsd__anyType::set() and xsd__anyAttribute::set() methods or use the assignment operator as follows:

// 4 ways to assign a value to an element node or to an attribute node
dom.set(val); // set text character data from bool, integer, float, string value
dom = val; // same as above
dom = att; // copy attribute node to attribute dom node (shallow, shares data)
dom = elt; // copy element node to element dom node (shallow, shares data)

In addition, element nodes can be assigned serializable objects and data as values, meaning that the object and data will be serialized in place of the element content:

// two ways to assign serializable data to an element node 'dom'
dom.set(node, type); // assign serializable C/C++ object or data to element node
dom = node; // assign serializable C/C++ object (object class must have soap_type() method)

To learn more about embedding serializable objects and data in DOM node graphs, see Embedding serializable objects and data in DOM element nodes.

To add attributes to the element node xsd__anyType dom we use one of the xsd__anyType::att() methods as follows:

// 4 ways to add an attribute to an element node 'dom' or extend an attribute list 'dom'
dom.att(utag) = val; // add attribute utag="val"
dom.att(qtag) = val; // add attribute q:tag="val" xmlns:q="ns"
dom.att(ns, utag) = val; // add attribute _1:utag="val" xmlns:_1="ns" with temporary _1, _2, _3, ...
dom.att(ns, qtag) = val; // add attribute q:tag="val" xmlns:q="ns"

To add child elements to the element node xsd__anyType dom we use one of the xsd__anyType::elt() methods and the bracket operator as follows:

// 7 ways to add a child element to an element node 'dom'
xsd__anyType& elt = dom.elt(); // add text-only child, unnamed (no tags)
xsd__anyType& elt = dom.elt(utag); // add child element <utag>
xsd__anyType& elt = dom[utag]; // same as above
xsd__anyType& elt = dom.elt(qtag); // add child element <q:tag xmlns:q="ns">
xsd__anyType& elt = dom[qtag]; // same as above
xsd__anyType& elt = dom.elt(ns, utag); // add child <tag xmlns="ns">
xsd__anyType& elt = dom.elt(ns, qtag); // add child <q:tag xmlns:q="ns">

Given the elt reference to a child element of the dom node above, you can use this reference to build more complex node graphs by assigning attributes, child elements, and values at increasingly deeper levels of the tree.

Because element nodes can have multiple children with the same tag name, essentially representing an array, you can use the bracket operator to index the child elements at their relative position among its siblings that have the same tag name. Position indexing starts at one (i.e. [1] XPath style) and child elements are added up to the highest position index used:

// add 'array' of three child elements to an element node 'dom'
xsd__anyType& elt = dom[tag]; // add child (using one of the methods above)
elt[1] = val; // assign val to first child element <tag>
elt[3] = val; // add children <tag> at positions 2 and 3, assign val to the 3rd

Note that elt[1] refers to elt itself, because indexing starts at 1 (as in XPath). Thus, position index [1] is optional to use.

You can combine these operations and methods in various ways to create more complex DOM node graphs. Let's illustrate this with an example:

soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
xsd__anyType dom(ctx, "menu");
dom.att("id") = 7;
dom.att("key") = L"⌘F";
dom["popup"]["menuitem"][1].att("value") = "New";
dom["popup"]["menuitem"][1].att("onclick") = "CreateNewDoc()";
dom["popup"]["menuitem"][2].att("value") = "Open";
dom["popup"]["menuitem"][2].att("onclick") = "OpenDoc()";
dom["popup"]["menuitem"][3].att("value") = "Close";
dom["popup"]["menuitem"][3].att("onclick") = "CloseDoc()";
std::cout << dom;

When executed, this code emits the following XML output on the terminal:

<menu id="7" key="⌘F">
  <popup>
    <menuitem value="New" onclick="CreateNewDoc()"/>
    <menuitem value="Open" onclick="OpenDoc()"/>
    <menuitem value="Close" onclick="CloseDoc()"/>
  </popup>
</menu>

The xsd__anyAttribute DOM attribute node constructors are identical to the xsd__anyType DOM element node constructors and can be used to create an attribute node or to create a list of attributes.

As an alternative to creating attributes and elements in place, you can also construct attribute and element nodes and add them to other element nodes by using the xsd__anyType::add() method. For example:

// add an attribute 'att' and an element node 'elt' to another element node 'dom'
xsd__anyAttribute att(ctx, NULL, "id", "7"); // id attribute
xsd__anyType elt(ctx, "popup"); // popup element
dom.add(att).add(elt); // add id attribute and popup as child element

To create a list of two attributes and add them to an element node dom:

// add a list of attributes 'atts' to an element node 'dom'
atts.att("id") = 7;
atts.att("key") = L"⌘F";
dom.adds(atts);

Finally, you can use domcpp to generate DOM construction code from a sample XML document.

Matching and retrieving DOM node values and properties

Given a DOM node graph parsed from an XML document, we can match its tag names and use methods to retrieve values, attribuets, child elements, and properties.

Given a DOM attribute node xsd__anyAttribute dom or element node xsd__anyType dom, you can use the following methods to match and retrieve the node's namespace and tag name:

bool b = dom.match(patt); // true if tag matches pattern
bool b = dom.match(ns, patt); // true if ns and tag match the patterns
const char *s = dom.ns(); // namespace URI string or NULL if none
const char *t = dom.tag(); // (un)qualified tag name string, NULL if unnamed

Given a DOM attribute node xsd__anyAttribute dom or element node xsd__anyType dom, the following methods can be used to retrieve the value of the node's character data content:

bool b = dom.is_true(); // true if text is "true" or "1"
bool b = dom.is_false(); // true if text is "false" or "0"
LONG64 n = dom.get_int(); // 64 bit integer value of text, or 0
double x = dom.get_double(); // double float value of text, or NaN
const char *t = dom.get_text(); // text string in UTF-8, or NULL if none

You can also cast an element or attribute node to a bool, an int, a double, or a string to obtain its value. For example:

int n = dom; // integer value cast from text in DOM node, or 0
const char *t = dom; // text string in UTF-8 of the DOM node, or NULL if none

Element nodes may contain deserialized objects and data as values, which can be retrieved with one of two methods:

T *node = (T*)dom.get_node(SOAP_TYPE_T); // get deserialized C/C++ object or data of type T, NULL if none
int type = dom.get_type(&node); // get deserialized C/C++ object or data of any type, 0 if none

To learn more about embedding serializable objects and data in DOM node graphs, see Embedding serializable objects and data in DOM element nodes.

Given a DOM element node xsd__anyType dom, one of the following methods can be used to retrieve an attribute by its unqualfied tag name (default namespace), by qualified tag name (assuming the prefix is defined in the namespace table), or by namespace and unqualified tag name:

xsd__anyAttribute *att = dom.att_get(utag); // get attribute utag, NULL if none
xsd__anyAttribute *att = dom.att_get(qtag); // get attribute q:tag
xsd__anyAttribute *att = dom.att_get(ns, utag); // get attribute utag in namespace ns, NULL if none

The second form is convenient to use to retrieve attributes in namespaces by using a qualified tag name qtag without having to specify the ns namespace URI parameter. This requires the namespace prefix used in qtag to be defined in the global namespace table.

Given a DOM element node xsd__anyType dom, one of the following methods can be used to retrieve a child element by its unqualfied tag name (default namespace), by qualified tag name (assuming the prefix is defined in the namespace table), or by namespace and unqualified tag name:

xsd__anyType *elt = dom.elt_get(); // get first text-only child element, NULL if none
xsd__anyType *elt = dom.elt_get(utag); // get first child element <utag>, NULL if none
xsd__anyType *elt = dom.elt_get(qtag); // get first child element <q:tag xmlns:q="ns">, NULL if none
xsd__anyType *elt = dom.elt_get(ns, utag); // get first child element <s:utag xmlsn:s="ns">, NULL if none
elt = elt.get_next(); // get next matching child element, NULL if none
elt = elt.get_nth(pos); // get matching child element at pos counting from 1, NULL if none

The get_next() method can be used in loops to iterate over child elements with the same namespace and tag, as if traversing an array. For example:

// iterate over <menuitem> child elements of element node 'dom'
for (xsd__anyType *it = dom.elt_get("menuitem"); it; it = it->get_next())
...

To obtain various properties of an xsd__anyType dom element node:

xsd__anyType *elt = dom.parent(); // parent node, or NULL when root
size_t n = dom.depth(); // depth from the root node (0 is root)
size_t n = dom.len(); // number of siblings that have the same tag name
size_t pos = dom.nth(); // node is at pos with the same tag name as some siblings, or 0 if alone

The following invariants hold (assuming that pointer dereferences (->) are verified and valid):

dom[tag].match(tag) == dom.elt(tag).match(tag) == true
dom[tag].parent() == dom.elt(tag).parent() == dom
dom[tag].tag() == dom.elt(tag).tag() == tag
dom.elt(ns, tag).match(ns, tag) == dom
dom.elt(ns, tag).parent() == dom
dom.elt(ns, tag).ns() == ns
dom.elt(ns, tag).tag() == tag
dom.att(tag).match(tag) == true
dom.att(ns, tag).match(ns, tag) == true
dom.att(ns, tag).ns() == ns
dom.att(ns, tag).tag() == tag
dom.set(val) == val // by casting to bool, int, float, string
dom.elt_get(tag)->parent() == dom
dom.elt_get(tag)->depth() == dom.depth() + 1
dom.elt_get(tag)->nth() <= dom.elt_get(tag)->len()
dom.elt_get(tag)->nth() == 0 if dom.elt_get(tag)->len() == 1
dom.elt_get(tag)->nth() == 1 if dom.elt_get(tag)->len() > 1
dom.elt_get(tag)->get_next()->nth() == 2
dom.elt_get(tag)->get_nth(pos)->nth() == pos

Traversing a DOM node graph

Given a DOM node graph, we can traverse its attribute list and its child elements by using the iterators xsd__anyAttribute::iterator and xsd__anyType::iterator, respectively:

// print attribute tags
for (xsd__anyAttribute::iterator it = dom.att_begin(); it != dom.att_end(); ++it)
std::cout << "@" << it->tag() << std::endl;
// print child element tags
for (xsd__anyType::iterator it = dom.elt_begin(); it != dom.elt_end(); ++it)
std::cout << "<" << it->tag() << ">" << std::endl;

You can also traverse the entire node graph below a certain DOM element node xsd__anyType dom in depth-first order starting with the current dom node:

// print all element tags in depth-first order
for (xsd__anyType::iterator it = dom.begin(); it != dom.end(); ++it)
std::cout << "<" << it->tag() << "> at depth " << it->depth() << std::endl;

Searching a DOM node graph

Given a DOM node graph, we can search its attribute list and its child elements by using a pattern with wildcards * to match namespace and/or tag names, similar to XPath * pattern matching:

// print attribute tags that match *:id (* matches any namespace, including default)
for (xsd__anyAttribute::iterator it = dom.att_find("*:id"); it != dom.att_end(); ++it)
std::cout << "@" << it->tag() << std::endl;
// print element tags in namespace "urn:one" (assuming "x" is "urn:one" in the namespace table)
for (xsd__anyType::iterator it = dom.elt_find("x:*"); it != dom.elt_end(); ++it)
std::cout << "<" << it->tag() << ">" << std::endl;

The above assumes that a namespace table is used that defines namespace x. Alternatively, you can specify namespace parameters that can be NULL or "*" to match any namespace URI (NULL can be used in place of a "*" pattern):

// print attribute tags that match *:id
for (xsd__anyAttribute::iterator it = dom.att_find("*", "id"); it != dom.att_end(); ++it)
std::cout << "@" << it->tag() << std::endl;
// print element tags in namespace "urn:one"
for (xsd__anyType::iterator it = dom.elt_find("urn:one", "*"); it != dom.elt_end(); ++it)
std::cout << "<" << it->tag() << "<" << std::endl;

Patterns for namespaces and tags are not restricted to wildcards *. You can specify parts of a namespace URI or tag name and use one or more * to match the rest.

You can also search the entire node graph below a certain DOM element node in depth-first order starting with the current node:

// find all element nodes with attribute "id" deep in the node graph
for (xsd__anyType::iterator it = dom.find("@*:id"); it != dom.end(); ++it)
std::cout << "element " << it->tag() << " has an @id=" << it->att_get("*:id")->get_text() << std::endl;
// print all element tags in namespace "urn:one"
for (xsd__anyType::iterator it = dom.find("urn:one", "*"); it != dom.end(); ++it)
std::cout << "<" << it->tag() << ">" << std::endl;

XML DOM parsing and writing from/to streams

To parse XML from a stream into a DOM node graph you can use the >> operator on an input stream and a DOM element node:

// create new context and parse XML from 'cin' into 'dom'
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
xsd__anyType dom(ctx);
std::cin >> dom;
if (dom.soap->error)
... // check for IO end parse errors
...
soap_destroy(ctx);
soap_end(ctx);
soap_free(ctx);

To write a DOM node graph in XML to a stream, you can use the << operator on an output stream and a DOM element node:

// create new context and create a 'dom', write its XML to 'cout'
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
xsd__anyType dom(ctx, "store");
...
std::cout << dom << std::endl;
if (dom.soap->error)
... // check for IO errors
...
soap_destroy(ctx);
soap_end(ctx);
soap_free(ctx);

You can also parse XML from a string into a DOM node graph by using a std::istringstream object:

// create new context and parse XML from a string into 'dom'
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
std::istringstream in;
in.str("<store><book title=\"Sayings of the Century\"/></store>");
xsd__anyType dom(ctx);
in >> dom; // parse XML into DOM
if (dom.soap->error)
... // check for IO and parse errors
...
soap_destroy(ctx);
soap_end(ctx);
soap_free(ctx);

Likewise, you can write a DOM node graph in XML to a string by using a std::ostringstream object:

// create new context and create a 'dom', write its XML to a string
soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
std::ostringstream out;
xsd__anyType dom(ctx, "store");
...
out << dom; // convert DOM to XML string in UTF-8
...
soap_destroy(ctx);
soap_end(ctx);
soap_free(ctx);

It is also possible to send and receive XML over HTTP as REST operations as was illustrated in the introduction. To make REST HTTP POST, GET, OUT, and DELETE calls, use:

This function is overloaded to accept references to the in and out parameters in C++.

C XML DOM API

XML is represented as a DOM node graph internally consisting of values (text strings of character data in UTF-8), attribute nodes with tag name and optional namespace URI, and element nodes with optional tag names and optional namespace URI. Tag names are stored in strings of UTF-8. Tag names can be qualified of the form q:tag or unqualified. An element tag name that is NULL represents a text-only node, which will only be present in mixed content of elements and text character data (as in XHTML markup).

XML namespace handling is "smart" in the gSOAP DOM API: the DOM engine matches and stores XML with namespace URIs to ensure that a prefix qualifier is always locally bound to a namespace URI in the node graph. Furthermore, you can use qualified tags when constructing a DOM node instead of the full namespace URI, but only if the prefix and URI are defined in the namespace table (global struct Namespace namesapces[]. When working with wsdl2h and soapcpp2 you should include the soapcpp2-generated .nsmap file with namespace bindings that are applicable to your XML tags.

List of C files

The following files located in the package under gsoap are required to use the C DOM API:

The gSOAP header file dom.h declares the DOM for use in data bindings and should generally be imported with #import in gSOAP header files for the soapcpp2 tool.

You can also run soapcpp2 directly on import/dom.h for DOM-only projects. This generates soapStub.h, soapH.h and soapC.cpp. To auto-generate these files, execute:

soapcpp2 -c -CSL import/dom.h

Then compile and link the dom.c files listed above with the auto-generated soapC.c and stdsoap2.c:

cc -I../.. -o myapp myapp.c soapC.c ../../dom.c ../../stdsoap2.c

Instead of stdsoap2.c you can link -lgsoap.a when installed by the gSOAP package. Or link -lgsoapssl.a to get both stdsoap2.c and dom.c with OpenSSL enabled for HTTPS:

cc -DWITH_OPENSSL -I../.. -o myapp myapp.c soapC.c -lgsoapssl -lssl -lcrypto

Note that we compile with -DWITH_OPENSSL and link -lssl, and -lcrypto.

Because XML namespaces are required except for the most simple plain XML applications, you should include the generated .nsmap file:

#include "soapH.h"
#include "soap.nsmap"

Or define your own namespace table. For example, at a minimum you want a table with SOAP and XSD bindings:

#include "soapH.h"
struct Namespace namespaces[] =
{
{"SOAP-ENV", "http://schemas.xmlsoap.org/soap/envelope/", "http://www.w3.org/*/soap-envelope", NULL},
{"SOAP-ENC", "http://schemas.xmlsoap.org/soap/encoding/", "http://www.w3.org/*/soap-encoding", NULL},
{"xsi", "http://www.w3.org/2001/XMLSchema-instance", "http://www.w3.org/*/XMLSchema-instance", NULL},
{"xsd", "http://www.w3.org/2001/XMLSchema", "http://www.w3.org/*/XMLSchema", NULL},
{NULL, NULL, NULL, NULL}
};

Then also add entries of the form {"prefix", "URI"} before the last NULL entry to complete the table with xmlns prefix and URI bindings for your XML project.

C DOM API functions

This overview of the DOM API functions refers to the following types of parameters and values, which are used throughout this section:

Name Type and value
ctx struct soap *ctx context pointer (manages memory and IO)
ns const char *ns XML namespace URI string or NULL
tag const char *tag or const wchar_t *tag (un)qualified tag name
utag const char *utag or const wchar_t *utag unqualified tag name
qtag const char *qtag or const wchar_t *qtag qualified tag name
patt (wide) string (un)qualified tag name pattern with * wildcard(s)
text (wide) string
node const void *node pointer to serializable C/C++ data object
type SOAP_TYPE_T type identifier of serializable C/C++ data type T
att xsd__anyAttribute *att DOM attribute node pointer
elt xsd__anyType *elt DOM element node pointer
dom a DOM node (a pointer to an element or an attribute node)
val bool, integer, float, or (wide) string
pos element position > 0 (XPath position numbering)

The API includes functions that take wide strings and normalizes these internally to UTF-8 encoded strings. However, when retrieving tags and text you will only be able to obtain UTF-8 strings. You can convert UTF-8 strings to wide strings that are managed by the context by using:

Creating a DOM root element node

Creating a DOM node graph starts with creating the root element node using one of the xsd__anyType allocators:

// 4 ways to create a root element node pointer 'dom'
xsd__anyType *dom = soap_elt_new(ctx, NULL, utag); // element node <utag>
xsd__anyType *dom = soap_elt_new(ctx, NULL, qtag); // element node <q:tag xmlns:q="ns">
xsd__anyType *dom = soap_elt_new(ctx, ns, utag); // element node <utag xmlns="ns">
xsd__anyType *dom = soap_elt_new(ctx, ns, qtag); // element node <q:tag xmlns:q="ns">

For wide string tags, use soap_elt_new_w(ctx, ns, tag). All C API functions that support wide strings end in _w by convention.

The second function call in the list is convenient to use to construct an element node that is bound to an XML namespace by using a qualified tag name qtag while specifying the namespace URI as NULL.

To use this approach, you should define the namespace prefix in qtag in a global namespace table, see Understanding XML namespaces, and as shown below:

// a global namespace table with {"prefix", "URI"} xmlns bindings
struct Namespace namespaces[] =
{
...
{"q", "urn:example"},
{NULL, NULL, NULL, NULL}
};
xsd__anyType *dom = soap_elt_new(ctx, NULL, "q:tag"); // OK: "q" is in the namespace table
xsd__anyType *dom = soap_elt_new(ctx, NULL, "bad:tag"); // BAD: "bad" is not in the namespace table
xsd__anyType *dom = soap_elt_new(ctx, "urn:ok", "ok:tag"); // OK: "ok" is locally scoped in the DOM

You can also use multiple namespace tables and select the one you want to use with the soap_set_namespaces() function:

// a local (static) namespace table with {"prefix", "URI"} xmlns bindings
struct Namespace my_dom_namespaces[] =
{
...
{"q", "urn:example"},
{NULL, NULL, NULL, NULL}
};
soap_set_namespaces(ctx, my_dom_namespaces); // use namespace table for qualified tags
xsd__anyType *dom = soap_elt_new(ctx, NULL, "q:tag"); // OK: "q" is in the namespace table
...
soap_set_namespaces(ctx, namespaces); // restore to global table

To set or change the tag of an element, use:

soap_elt_set(dom, ns, utag); // element node <utag xmlns="ns">
soap_elt_set(dom, ns, qtag); // element node <q:tag xmlns:q="ns">

Assigning child nodes and values to a DOM node

After creating a root element node we can start populating the root with attributes, child elements, and values.

To assign a value to a DOM element node dom you can use one the the following functions:

// 6 ways to assign a value to an element node 'dom'
soap_elt_bool(dom, val); // set text to boolean "false" or "true"
soap_elt_int(dom, val); // set text to integer (64 bit)
soap_elt_double(dom, val); // set text to (double) float
soap_elt_text(dom, val); // set text with UTF-8 string content
soap_elt_text_w(dom, val); // set text with wide string content
soap_elt_copy(dom, elt); // copy element node to element dom node (shallow, shares data)

You can use similar functions for DOM attribute nodes, altough you do not need these to populate DOM element nodes because it is easier to use the soap_att() function shown further below:

// 6 ways to assign a value to an attribute node 'dom'
soap_att_bool(dom, val); // set text to boolean "false" or "true"
soap_att_int(dom, val); // set text to integer (64 bit)
soap_att_double(dom, val); // set text to (double) float
soap_att_text(dom, val); // set text with UTF-8 string content
soap_att_text_w(dom, val); // set text with wide string content
soap_att_copy(dom, elt); // copy attribute node to attribute dom node (shallow, shares data)

In addition, element nodes can be assigned serializable objects and data as values, meaning that the object and data will be serialized in place of the element content:

// assign serializable data to an element node 'dom'
soap_elt_node(dom, node, type); // assign serializable C data to element node

To learn more about embedding serializable objects and data in DOM node graphs, see Embedding serializable objects and data in DOM element nodes.

To add attributes to an element node dom you can use one of the following functions (use soap_att_w for wide string tags):

// 4 ways to add an attribute to an element node 'dom' or extend an attribute list 'dom'
xsd__anyAttribute *att = soap_att(dom, NULL, utag); // add attribute utag="val"
xsd__anyAttribute *att = soap_att(dom, NULL, qtag); // add attribute q:tag="val" xmlns:q="ns"
xsd__anyAttribute *att = soap_att(dom, ns, utag); // add attribute _1:utag="val" xmlns:_1="ns" with temporary _1, _2, _3, ...
xsd__anyAttribute *att = soap_att(dom, ns, qtag); // add attribute q:tag="val" xmlns:q="ns"

You can then set the value of the attribute att as was shown earlier.

To add child elements to the element node dom we use one of the following functions (use soap_elt_w for wide string tags):

// 5 ways to add a child element to an element node 'dom'
xsd__anyType *elt = soap_elt(dom, NULL, NULL); // add text-only child, unnamed (no tags)
xsd__anyType *elt = soap_elt(dom, NULL, utag); // add child element <utag>
xsd__anyType *elt = soap_elt(dom, NULL, qtag); // add child element <q:tag xmlns:q="ns">
xsd__anyType *elt = soap_elt(dom, ns, utag); // add child <tag xmlns="ns">
xsd__anyType *elt = soap_elt(dom, ns, qtag); // add child <q:tag xmlns:q="ns">

Given the elt pointer to a child element node of the dom node, you can use the pointer to build more complex node graphs by assigning attributes, child elements, and values at increasingly deeper levels of the tree.

Because element nodes can have multiple children with the same tag name, essentially representing an array, you can index the child elements at their relative position among its siblings that have the same tag name. Position indexing starts at one (i.e. XPath style) and child elements are added up to the highest position index used:

// add 'array' of three child elements to an element node 'dom'
xsd__anyType *elt = soap_elt(dom, NULL, tag); // add child (using one of the functions above)
soap_elt_int(soap_nth(elt, 1), val); // assign val to first child element <tag>
soap_elt_int(soap_nth(elt, 3), val); // add children <tag> at positions 2 and 3, assign val to the 3rd

The combination soap_nth(soap_elt(dom, ns, tag), pos) can be shortend to soap_nth_elt(dom, ns, tag, pos).

Note that soap_nth(elt, 1) refers to elt itself, because indexing starts at 1 (as in XPath).

You can combine these operations in various ways to create more complex DOM node graphs. Let's illustrate this with an example:

struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
xsd__anyType *dom = soap_elt_new(ctx, NULL, "menu");
soap_att_int(soap_att(dom, NULL, "id"), 7);
soap_att_text_w(soap_att(dom, NULL, "value"), "⌘F");
{
xsd__anyType *dom_popup = soap_elt(dom, NULL, "popup");
{
xsd__anyType *dom_popup_menuitem = soap_nth_elt(dom_popup, NULL, "menuitem", 1);
soap_att_text(soap_att(dom_popup_menuitem, NULL, "value"), "New");
soap_att_text(soap_att(dom_popup_menuitem, NULL, "onclick"), "CreateNewDoc()");
}
{
xsd__anyType *dom_popup_menuitem = soap_nth_elt(dom_popup, NULL, "menuitem", 2);
soap_att_text(soap_att(dom_popup_menuitem, NULL, "value"), "Open");
soap_att_text(soap_att(dom_popup_menuitem, NULL, "onclick"), "OpenDoc()");
}
{
xsd__anyType *dom_popup_menuitem = soap_nth_elt(dom_popup, NULL, "menuitem", 3);
soap_att_text(soap_att(dom_popup_menuitem, NULL, "value"), "Close");
soap_att_text(soap_att(dom_popup_menuitem, NULL, "onclick"), "CloseDoc()");
}
}

When executed, this code emits the following XML output on the terminal:

<menu id="7" key="⌘F">
  <popup>
    <menuitem value="New" onclick="CreateNewDoc()"/>
    <menuitem value="Open" onclick="OpenDoc()"/>
    <menuitem value="Close" onclick="CloseDoc()"/>
  </popup>
</menu>

The xsd__anyAttribute DOM attribute node allocation functions are identical to the xsd__anyType DOM element node allocation functions and can be used to create an attribute node or to create a list of attributes.

As an alternative to creating attributes and elements in place, you can also construct attribute and element nodes and add them to other element nodes:

// add an attribute 'att' and an element node 'elt' to another element node 'dom'
xsd__anyAttribute *att = soap_att_new(ctx, NULL, "id");
xsd__anyType *elt = soap_elt_new(ctx, NULL, "popup");
soap_att_int(att, 7);
soap_add_att(dom, att);
soap_add_elt(dom, elt);

To create a list of two attributes and add them to an element node dom:

// add a list of attributes 'atts' to an element node 'dom'
xsd__anyAttribute *atts = soap_att_new(ctx, NULL, NULL);
soap_att_int(soap_att_add(atts, NULL, "id"), 7);
soap_att_text_w(soap_att_add(atts, NULL, "key"), L"⌘F");
soap_add_atts(dom, atts);

Finally, you can use domcpp to generate DOM construction code from a sample XML document.

Matching and retrieving DOM node values and properties

Given a DOM node graph parsed from an XML document, we can match its tag names and use functions to retrieve values, attribuets, child elements, and properties.

Given a DOM element node dom, you can use the following functions to match and retrieve the node's namespace and tag name:

int b = soap_elt_match(dom, NULL, patt); // true if tag matches pattern
int b = soap_elt_match(dom, ns, patt); // true if ns and tag match the patterns
const char *s = soap_elt_get_ns(dom)); // namespace URI string or NULL if none
const char *t = soap_elt_get_tag(dom); // (un)qualified tag name string, NULL if unnamed

Similar for attribute node dom:

int b = soap_att_match(dom, NULL, patt); // true if tag matches pattern
int b = soap_att_match(dom, ns, patt); // true if ns and tag match the patterns
const char *s = soap_att_get_ns(dom)); // namespace URI string or NULL if none
const char *t = soap_att_get_tag(dom); // (un)qualified tag name string, NULL if unnamed

Given a DOM element node dom, the following functions can be used to retrieve the value of the node's character data content:

int b = soap_elt_is_true(dom); // true if text is "true" or "1"
int b = soap_elt_is_false(dom); // true if text is "false" or "0"
LONG64 n = soap_elt_get_int(dom)); // 64 bit integer value of text, or 0
double x = soap_elt_get_double(dom); // double float value of text, or NaN
const char *t = soap_elt_get_text(dom)); // text string in UTF-8, or NULL if none

Similar for attribute node dom:

int b = soap_att_is_true(dom); // true if text is "true" or "1"
int b = soap_att_is_false(dom); // true if text is "false" or "0"
LONG64 n = soap_att_get_int(dom)); // 64 bit integer value of text, or 0
double x = soap_att_get_double(dom); // double float value of text, or NaN
const char *t = soap_att_get_text(dom)); // text string in UTF-8, or NULL if none

Element nodes may contain deserialized objects and data as values, which can be retrieved with one of two methods:

T *node = (T*)soap_elt_get_node(dom, SOAP_TYPE_T); // get deserialized C data of type T, NULL if none
int type = soap_elt_get_type(dom, &node); // get deserialized C data of any type, 0 if none

To learn more about embedding serializable objects and data in DOM node graphs, see Embedding serializable objects and data in DOM element nodes.

Given a DOM element node dom, one of the following methods can be used to retrieve an attribute by its unqualfied tag name (has no namespace), by qualified tag name (assuming the prefix is defined in the namespace table), or by namespace and unqualified tag name:

xsd__anyAttribute *att = soap_att_get(dom, NULL, utag); // get attribute utag, NULL if none
xsd__anyAttribute *att = soap_att_get(dom, NULL, qtag); // get attribute q:tag
xsd__anyAttribute *att = soap_att_get(dom, ns, utag); // get attribute utag in namespace ns, NULL if none

The second form is convenient to use to retrieve attributes in namespaces by using a qualified tag name qtag without having to specify the ns namespace URI parameter. This requires the namespace prefix used in qtag to be defined in the global namespace table.

Given a DOM element node dom, one of the following functions can be used to retrieve a child element by its unqualfied tag name (has no namespace), by qualified tag name (assuming the prefix is defined in the namespace table), or by namespace and unqualified tag name:

xsd__anyType *elt = soap_elt_get(dom, NULL, NULL); // get first text-only child element, NULL if none
xsd__anyType *elt = soap_elt_get(dom, NULL, utag); // get first child element <utag>, NULL if none
xsd__anyType *elt = soap_elt_get(dom, NULL, qtag); // get first child element <q:tag xmlns:q="ns">, NULL if none
xsd__anyType *elt = soap_elt_get(dom, ns, utag); // get first child element <s:tag xmlsn:s="ns">, NULL if none
elt = soap_elt_get_next(elt); // get next matching child element, NULL if none
elt = soap_elt_get_nth(elt, pos); // get matching child element at pos counting from 1, NULL if none

The soap_elt_get_next(elt) function can be used in loops to iterate over child elements with the same namespace and tag, as if traversing an array. For example:

// iterate over <menuitem> child elements of element node 'dom'
for (it = soap_elt_get(dom, NULL, "menuitem"); it; it = soap_elt_get_next(it))
...

To obtain various properties of a dom element node:

xsd__anyType *elt = soap_elt_parent(dom); // parent node, or NULL when root
size_t n = soap_elt_depth(dom); // depth from the root node (0 is root)
size_t n = soap_elt_len(dom); // number of siblings that have the same tag name
size_t pos = soap_elt_nth(dom); // node is at pos with the same tag name as some siblings, or 0 if alone

The following invariants hold (assuming that soap_elt_get(dom, ns, tag) is not NULL and thus dom has a child element that matches namespace ns and tag):

soap_elt_match(soap_elt(dom, NULL, tag), NULL, tag) == true (nonzero)
soap_elt_parent(soap_elt(dom, NULL, tag)) == dom
soap_elt_get_tag(soap_elt(dom, NULL, tag)) == tag
soap_elt_match(soap_elt(dom, ns, tag), ns, tag) == true (nonzero)
soap_elt_parent(soap_elt(dom, ns, tag)) == dom
soap_elt_get_ns(soap_elt(dom, ns, tag)) == ns
soap_elt_get_tag(soap_elt(dom, ns, tag)) == tag
soap_att_match(soap_att(dom, ns, tag), ns, tag) == true (nonzero)
soap_att_get_ns(soap_att(dom, ns, tag)) == ns
soap_att_get_tag(soap_att(dom, ns, tag)) == tag
soap_elt_get_int(soap_elt_int(dom, val)) == val // also holds for bool, double, string
soap_att_get_int(soap_att_int(dom, val)) == val // also holds for bool, double, string
soap_elt_parent(soap_elt_get(dom, ns, tag)) == dom
soap_elt_depth(soap_elt_get(dom, ns, tag)) == soap_elt_depth(dom) + 1
soap_elt_nth(soap_elt_get(dom, ns, tag)) <= soap_elt_len(soap_elt_get(dom, ns, tag))
soap_elt_nth(soap_elt_get(dom, ns, tag)) == 0 if soap_elt_len(soap_elt_get(dom, ns, tag)) == 1
soap_elt_nth(soap_elt_get(dom, ns, tag)) == 1 if soap_elt_len(soap_elt_get(dom, ns, tag)) > 1
soap_elt_nth(soap_elt_get_nth(soap_elt_get(dom, ns, tag), pos)) == pos

Traversing a DOM node graph

Given a DOM node graph dom, we can traverse its attribute list xsd__anyType::atts and its child elements xsd__anyType::elts as follows:

// print attribute tags
for (it = dom->atts; it; it = it->next)
printf("@%s\n", soap_att_get_tag(it));
// print child element tags
for (it = dom->elts; it; it = it->next)
printf("<%s>\n", soap_elt_get_tag(it));

You can also traverse the entire node graph below a certain DOM element node dom in depth-first order starting with the current dom and stepping through all element nodes with soap_dom_next_element(elt, end) where end is a pointer to the ending element node (exclusive), usually the starting node of the graph traversal:

// print all element tags in depth-first order
for (it = dom; it; it = soap_dom_next_element(it, dom))
printf("<%s> at depth %zd\n", soap_elt_get_tag(it), soap_elt_depth(it));

Searching a DOM node graph

Given a DOM node graph, we can search its attribute list and its child elements by using a pattern with wildcards * to match namespace and/or tag names, similar to XPath * pattern matching:

// print attribute tags that match *:id (* matches any namespace, including default)
for (it = soap_att_find(dom, NULL, "*:id"); it; it = soap_att_find_next(it, NULL, "*:id"))
printf("@%s\n", soap_att_get_tag(it));
// print element tags in namespace "urn:one" (assuming "x" is "urn:one" in the namespace table)
for (it = soap_elt_find(dom, NULL, "x:*"); it; it = soap_elt_find_next(it, NULL, "x:*"))
printf("<%s>\n", soap_elt_get_tag(it));

The above assumes that a namespace table is used that defines namespace x. Alternatively, you can specify namespace parameters that can be NULL or "*" to match any namespace URI (NULL can be used in place of a "*" pattern):

// print attribute tags that match *:id
for (it = soap_att_find(dom, "*", "id"); it; it = soap_att_find_next(it, "*", "id"))
printf("@%s\n", soap_att_get_tag(it));
// print element tags in namespace "urn:one"
for (it = soap_elt_find(dom, "urn:one", "*"); it; it = soap_elt_find_next(it, "urn:one", "*"))
printf("<%s>\n", soap_elt_get_tag(it));

Patterns for namespaces and tags are not restricted to wildcards *. You can specify parts of a namespace URI or tag name and use one or more * to match the rest.

You can also search the entire node graph below a certain DOM element node in depth-first order starting with the current node:

// find all element nodes with attribute "id" deep in the node graph
for (it = soap_dom_find(dom, NULL, "@*:id"); it; it = soap_dom_find_next(it, dom, NULL, "@*:id")
printf("element <%s> has an @id=%s\n", soap_elt_get_tag(it), soap_att_get_text(soap_att_find(it, NULL, "*:id")));
// print all element tags in namespace "urn:one"
for (it = soap_dom_find(dom, "urn:one", "*"); it; it = soap_dom_find_next(it, dom, "urn:one", "*"))
printf("<%s>\n", soap_elt_get_tag(it));

XML DOM parsing and writing from/to streams

Reading and writing XML from/to files, streams and string buffers is done via the managing context by setting one of the following context members that control IO sources and sinks:

ctx->recvfd = fd; // an int file descriptor to read from (0 by default)
ctx->sendfd = fd; // an int file descriptor to write to (1 by default)
ctx->is = cs; // C only: a const char* string to read from (soap->is will advance)
ctx->os = &cs; // C only: pointer to a const char*, will be set to point to the string output

To parse XML from a file descriptor ctx->recvfd into a DOM node graph you can use to soapcpp2 auto-generated soap_read_xsd__anyType(ctx, dom) function:

// create new context and parse XML from 'cin' into 'dom'
struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
xsd__anyType *dom = soap_elt_new(ctx, NULL, NULL);
ctx->recvfd = stdin;
if (soap_read_xsd__anyType(ctx, dom))
... // check for IO end parse errors
...
soap_end(ctx);
soap_free(ctx);

To write a DOM node graph in XML to a file descriptor ctx->sendfd you can use the soapcpp2 auto-generated soap_write_xsd__anyType(ctx, dom) function:

// create new context and create a 'dom', write its XML to 'cout'
struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
xsd__anyType *dom = soap_elt_new(ctx, NULL, "store");
...
ctx->sendfd = stdout;
... // check for IO errors
...
soap_end(ctx);
soap_free(ctx);

You can also parse XML from a string into a DOM node graph by using the ctx->is input string and the soapcpp2 auto-generated soap_read_xsd__anyType(ctx, dom) function:

// create new context and parse XML from a string into 'dom'
struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
const char *in = "<store><book title=\"Sayings of the Century\"/></store>";
xsd__anyType *dom = soap_elt_new(ctx, NULL, NULL);
ctx->is = in;
if (soap_read_xsd__anyType(ctx, dom))
... // check for IO end parse errors
ctx->is = NULL;
...
soap_end(ctx);
soap_free(ctx);

Likewise, you can write a DOM node graph in XML to a string by using the ctx->os output string pointer and the soapcpp2 auto-generated soap_write_xsd__anyType(ctx, dom) function:

// create new context and create a 'dom', write its XML to a string
struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT | SOAP_DOM_TREE);
const char *out = NULL;
xsd__anyType *dom = soap_elt_new(ctx, NULL, "store");
...
ctx->os = &out; // string to set to point to the output
soap_write_xsd__anyType(ctx, dom); // convert DOM to XML string in UTF-8
ctx->os = NULL;
printf("%s\n", out);
...
soap_end(ctx);
soap_free(ctx);

The soap_read_xsd__anyType(ctx, dom) and soap_write_xsd__anyType(ctx, dom) functions return SOAP_OK (zero) or an error code in ctx->error.

It is also possible to send and receive XML over HTTP as REST operations as was illustrated in the introduction. To make REST HTTP POST, GET, OUT, and DELETE calls, use:

XML DOM parsing and display options

The soap context flags that control XML parsing and visual display are set when the context is created or with the soap_set_mode and soap_clr_mode functions:

struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_DOM_TREE);
soap_set_mode(ctx, SOAP_XML_INDENT); // use indent
soap_clr_mode(ctx, SOAP_XML_INDENT); // no indent

Options to control XML parsing, object serialization, and XML rendition are:

Using DOM together with serializeble types in a gSOAP header file

To use DOM nodes with types declared in a gSOAP header file, import dom.h into your header file for processing with soapcpp2:

#import "dom.h"

The import is automatically added by wsdl2h with option -d. The wsdl2h tool binds xsd:anyType, xsd:any, xsd:anyAttribute and xsd:mixed XML content to DOM nodes instead of using XML literal strings.

The #import "dom.h" enables us use the DOM xsd__anyType element node and xsd__anyAttribute attribute node in a gSOAP header file type declarations, such as:

#import "dom.h"
class ns__record
{
public:
@xsd__anyAttribute atts; // list of DOM attributes
...
xsd__anyType __any; // xsd:any, consumes all other XML
};

where __any contains XML stored in a DOM node set and atts is a list of all visibly rendered attributes. The name atts is arbitrary and any name will suffice.

We should place the xsd__anyType members at the end of the struct or class. This ensures that the DOM members are populated last as a "catch all". A member name starting with double underscore is a wildcard member name and matches any XML tag. These members are placed at the end of a struct or class automatically by soapcpp2.

The soapcpp2 tool generates code that "serializes" the DOM nodes as part of the serialization of the C/C++ types in which they are used.

See also
Documentation of XML Data Bindings: DOM anyType and anyAttribute.

Example

This example demonstrates the domcpp tool to generate code for a SOAP request message to send to a calculator service and to generate code to obtain the floating point value from the SOAP response returned. The generated code is copied and modified to create a DOM-based SOAP client application in C++ dom2calc.cpp.

We use a very simple calculator service to demonstrate the C++ DOM API to populate a DOM node graph and to query a DOM node graph. The codes to accomplish both of these tasks can be generated with domcpp and modified as needed for our purposes.

The SOAP calculator service and client on which this example is based are located in gsoap/samples/calc (for C) and gsoap/samples/calc++ (for C++) in the gSOAP package.

We use the SOAP calculator sample SOAP message calc.add.req.xml generated by soapcpp2 in gsoap/samples/calc++. This XML message contains a SOAP request to add two floating point values:

<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns="urn:calc">
<SOAP-ENV:Body SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<ns:add>
<a>0.0</a>
<b>0.0</b>
</ns:add>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>

To generate code that populates this SOAP request message as a DOM node graph, we invoke:

domcpp -M -n -rrequest calc.add.req.xml

We copied the generated code into dom2calc.cpp as shown further below.

A sample SOAP response message calc.add.res.xml from the calculator service is:

<?xml version="1.0" encoding="UTF-8"?>
<SOAP-ENV:Envelope
xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/"
xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/"
xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
xmlns:xsd="http://www.w3.org/2001/XMLSchema"
xmlns:ns="urn:calc">
<SOAP-ENV:Body SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
<ns:addResponse>
<result>0.0</result>
</ns:addResponse>
</SOAP-ENV:Body>
</SOAP-ENV:Envelope>

We want to generate code that inspects the SOAP response message to extract the value of <result>. For this purpose we can use the XPath query /SOAP-ENV:Envelope/SOAP-ENV:Body/ns:addResponse/result and use this XPath with domcpp option -p and option -x to generate code that prints the <result> value:

domcpp -p'/SOAP-ENV:Envelope/SOAP-ENV:Body/ns:addResponse/result' \
  -rresponse -x'std::cout << "Result = " << v << std::endl;'

We copied the generated code into dom2calc.cpp as shown below:

#include "soapH.h"
const char server[] = "http://websrv.cs.fsu.edu/~engelen/calcserver.cgi";
// copied from:
// domcpp -M -n -rrequest calc.add.req.xml
struct Namespace namespaces[] = {
{"SOAP-ENV", "http://schemas.xmlsoap.org/soap/envelope/", "http://www.w3.org/*/soap-envelope", NULL},
{"SOAP-ENC", "http://schemas.xmlsoap.org/soap/encoding/", "http://www.w3.org/*/soap-encoding", NULL},
{"xsi", "http://www.w3.org/2001/XMLSchema-instance", "http://www.w3.org/*/XMLSchema-instance", NULL},
{"xsd", "http://www.w3.org/2001/XMLSchema", "http://www.w3.org/*/XMLSchema", NULL},
{"ns", "urn:calc", NULL, NULL },
{NULL, NULL, NULL, NULL}
};
int main(int argc, char **argv)
{
struct soap *ctx = soap_new1(SOAP_DOM_TREE | SOAP_XML_INDENT);
if (argc <= 3)
{
std::cerr << "Usage: dom2calc [add|sub|mul|div|pow] <num> <num>" << std::endl;
exit(1);
}
// create command tag ns:add, ns:sub, ns:mul, ns:div, or ns:pow
std::string command = std::string("ns:").append(argv[1]);
// copied and modified from:
// domcpp -M -n -rrequest calc.add.req.xml
xsd__anyType request(ctx, "SOAP-ENV:Envelope");
request["SOAP-ENV:Body"].att("SOAP-ENV:encodingStyle") = "http://schemas.xmlsoap.org/soap/encoding/";
request["SOAP-ENV:Body"][command.c_str()]["a"] = strtod(argv[2], NULL);
request["SOAP-ENV:Body"][command.c_str()]["b"] = strtod(argv[3], NULL);
std::cout << "** Request message: " << std::endl << request << std::endl << std::endl;
// create response
xsd__anyType response(ctx);
// invoke server: make POST XML request and receive XML response
if (soap_dom_call(ctx, server, "", request, response))
{
soap_stream_fault(ctx, std::cerr);
}
else
{
std::cout << "** Response message:" << std::endl << response << std::endl << std::endl;
// copied from:
// domcpp -p'/SOAP-ENV:Envelope/SOAP-ENV:Body/ns:addResponse/result' -rresponse -x'std::cout << "Result = " << v << std::endl;'
if (response.match("SOAP-ENV:Envelope"))
{
size_t pos = 1;
for (xsd__anyType *it = response.elt_get("SOAP-ENV:Body"); it; it = it->get_next(), ++pos)
{
xsd__anyType& v = *it;
size_t pos = 1;
for (xsd__anyType *it = v.elt_get("ns:addResponse"); it; it = it->get_next(), ++pos)
{
xsd__anyType& v = *it;
size_t pos = 1;
for (xsd__anyType *it = v.elt_get("result"); it; it = it->get_next(), ++pos)
{
xsd__anyType& v = *it;
std::cout << "Result = " << v.get_double() << std::endl;
}
}
}
}
}
soap_destroy(ctx); // delete objects
soap_end(ctx); // delete temp data
soap_free(ctx); // free context
return 0;
}

To compile dom2calc.cpp we first generate soapC.cpp, soapStub.h and soapH.h from the gSOAP header file dom.h:

soapcpp2 dom.h
c++ -o dom2calc dom2calc.cpp soapC.cpp dom.cpp stdsoap2.cpp

Running the DOM-based calculator client gives:

./dom2calc add 3 4

** Request message: 
<SOAP-ENV:Envelope xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/">
        <SOAP-ENV:Body SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
                <ns:add xmlns:ns="urn:calc">
                        <a>3</a>
                        <b>4</b>
                </ns:add>
        </SOAP-ENV:Body>
</SOAP-ENV:Envelope>

** Response message:
<SOAP-ENV:Envelope xmlns:SOAP-ENV="http://schemas.xmlsoap.org/soap/envelope/" xmlns:SOAP-ENC="http://schemas.xmlsoap.org/soap/encoding/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xmlns:xsd="http://www.w3.org/2001/XMLSchema" xmlns:ns="urn:calc">
        <SOAP-ENV:Body SOAP-ENV:encodingStyle="http://schemas.xmlsoap.org/soap/encoding/">
                <ns:addResponse>
                        <result>7</result>
                </ns:addResponse>
        </SOAP-ENV:Body>
</SOAP-ENV:Envelope>

Result = 7

Another way to extract the floating point value from the SOAP response is to use the xsd__anyType::find() method to search (and iterate) over <result> elements:

for (xsd__anyType::iterator it = response.find("result"); it != response.end(); ++it)
std::cout << std::endl << "Result = " << it->get_double() << std::endl;

where the xsd__anyType::get_double() method is used to cast the XML value to a double.

Miscellaneous

Floating point format

The floating point format used to output values in XML-RPC and JSON is by default ".17lG' to ensure accuracy up to the last digit. The format can be set as follows:

struct soap *ctx = soap_new1(SOAP_C_UTFSTRING | SOAP_XML_INDENT);
ctx->double_format = "%lG";