XML framework overview

Provides a generic, configurable facility for parsing XML and WBXML, with options to combine parsing with validation and auto-correction using a single interface.

The XML framework is based on the SAX (Simple API for XML) specification. It is also designed for use with WBXML (Wireless Binary XML).

The following key concepts refer to XML:

Tag

Angled brackets containing a tag name and, optionally, attributes: for instance <book>, <book author="Jane Austen">

Element

A sequence beginning with an opening tag and ending with a closing tag, possibly containing nested elements: for instance <book>Pride and Prejudice</book>

Content

The text or character data (CDATA) contained inside the tags: for instance "Pride and Prejudice".

Attribute

A name-value pair separated by an equals sign: for instance author="Jane Austen"

Attribute type

One of certain data types defined for attributes: for instance CDATA.

Document type definition (DTD)

A document which defines a particular use of XML (the names, attributes and values permitted)

Well-formed

A well-formed XML document conforms to the general syntax of XML, which is defined in the official XML specification.

Valid

A valid XML document conforms to a DTD.

Processing instruction

A tag specifying that content should be sent to a named content processor, sometimes used to enclose scripts. For instance: <?xml-stylesheet type="text/css" href="test.css?>"

Prefix

Prefixed to a tag to identify the DTD which defines it.

URI

The web address associated with a prefix: for instance http://www.w3.org/XML/1998/namespace.

Namespace

The association between a prefix and a URI which distinguishes the prefix from synonymous prefixes associated with a different URI.

Whitespace

Outside the content, whitespace is generally meaningless and the parser should ignore it.

Encoding

The scheme by which text is encoded as a sequence of chars or integers: for instance "encoding=utf-8".

Prefix mapping

A prefix mapping is a tag which is used to convert the document tag set into a unique namespace: it is implemented as an attribute with a URI as its value. For instance: xmlns="http://www.w3.org/1999/xlink".

The following key concepts refer to WBXML:

String table

A WBXML document is encoded and decoded by means of a table of frequently encountered strings which the body of the document references by index in order to compress the data.

Extension

WBXML extends XML syntax with extension tokens which are used differently by different applications. One use for them is to refer to a string table created specifically for each message and transmitted in the preamble of the message.

The parser framework conforms to the event-based SAX specification. It outputs an event when it starts or finishes reading one of the following:

• a document

• a start tag

• an end tag

• a prefix mapping

• a processing instruction

• character data

• ignorable whitespace

Client programs call a single interface which selects plugins to provide the appropriate parser (based on the type of the document to be parsed) and, if required, sets up a chain of further plugins to perform various validations. The string dictionaries used to parse WBXML are also accessed as plugins.

The XML framework consists of classes which model the main constituents of the architecture: the framework as a whole, the parser plugins and extensions to XML, the content processor chain and the content handler mechanism.

The entire parser framework is represented by the CParser class: a client with an XML document to be parsed creates a CParser object and calls its parse functions. The information needed to choose a plugin is held in the classes CMatchData (data about plugins) and RDocumentParameters (data about the document to be parsed).

Individual parser plugins implement the MParser interface. They are associated, through the TParserInitParams class, with a character set converter (to convert other formats to Unicode), a string dictionary and an element stack. The RElementStack class is the data structure used to store XML elements: tags are represented by the RTagInfo class and attributes by the RAttribute class. The Symbian OS Framework is delivered with two parser plugins, one for XML and one for WBXML. The XML parser consists of the class CXmlParser, which is wrapped around the class CExpat, an implementation of the stream-based Expat parser. The WBXML parser is implemented as the class CWmxmlParser.

The XML framework provides for extensions to XML: at present WBXML is implemented. WBXML requires use of string dictionaries and extension tokens, which are represented by the classes RStringDictionaryCollection, MWbxmlExtensionHandler and TExtensionTokens.

Content processors are plugins which perform further operations on the output of a parser plugin. They implement the interface MContentProcessor and are associated, through the TContentProcessorInitParams class, with a string dictionary and element stack. They are organised into chains by the MContentSource class which directs the output of each plugin to the next plugin inthe chain.

A client application which is designed to react to the output of the XML framework must implement the MContentHandler interface. The functions to be implemented correspond to the SAX specification discussed above.