Development/Tutorials/Programming Tutorial KDE 4/How to write an XML parser: Difference between revisions

    From KDE TechBase
    mNo edit summary
    (highlight the xml snippets)
    Line 115: Line 115:
    HTML parsing only works for "legal" html documents.  
    HTML parsing only works for "legal" html documents.  
    For example, look at this code:
    For example, look at this code:
    <code>
    <code xml>
    <pre>
    <html>
    <html>
       <body>
       <body>
    Line 124: Line 123:
       </body>
       </body>
    </html>
    </html>
    </pre>
    </code>
    </code>
    This code contains a &amp; and will bring your parser to an error.
    This code contains a &amp; and will bring your parser to an error.


    See here:
    See here:
    <code>
    <code xml>
    <pre>
    <html>
    <html>
       <body>
       <body>
    Line 138: Line 135:
       </body>
       </body>
    </html>
    </html>
    </pre>
    </code>
    </code>
    This code will throw an error because of the '''nowrap''' that is not xml-conform.
    This code will throw an error because of the '''nowrap''' that is not xml-conform.

    Revision as of 11:24, 30 May 2007

    Warning
    This section needs improvements: Please help us to

    cleanup confusing sections and fix sections which contain a todo


    A parser is used to distinguish between formal language and bulk data of a given grammar. See http://en.wikipedia.org/wiki/Parser for more information. There are two ways to write a parser: to split up the content of a file into an object as known from object-oriented programming (the DOM approach) or to trigger a function everytime a reader occurs a given syntax tag (the QXML approach).

    The QXML approach

    parser.h: /*

    parser.h - demonstration of a parser in C++
    
    • /
    1. ifndef PARSER_H
    2. define PARSER_H
    1. include <qstring.h>
    2. include <QtXml/QXmlDefaultHandler>
    3. include <QtXml/QXmlAttributes>

    class Parser : public QXmlDefaultHandler { public:

     Parser();
    
     /** given by the framework from qxml. Called when parsing the xml-document starts.          */
     bool startDocument();
    
     /** given by the framework from qxml. Called when the reader occurs an open tag (e.g. \<b\> ) */
     bool startElement( const QString&, const QString&, const QString& qName, const QXmlAttributes& att );
    

    };


    1. endif

    parser.cpp: /*

    parser.cpp - demonstration of a parser in C++
    
    • /
    1. include "parser.h"
    2. include <kdebug.h>
     Parser::Parser()
     {
     }
     
     bool Parser::startDocument()
     {
       kDebug() << "Searching document for tags" << endl;
       return true;
     }
     
     bool Parser::startElement( const QString&, const QString&, const QString& qName, const QXmlAttributes& att )
     {
       kDebug() << "Found Element " << qName << endl;
       return true;
    }
    

    hello.cpp: /* hello.cpp compile it with g++ -I. -I/home/kde-devel/kde/include -I/home/kde-devel/qt-unstable/include/Qt -I/home/kde-devel/qt-unstable/include /home/kde-devel/qt-unstable/include/QtXml parser.h parser.cpp hello.cpp -L/home/kde-devel/kde/lib -L/home/kde-devel/qt-unstable/lib -lQtCore_debug -lQtXml_debug -lkdeui

    • /


    1. include <qstring.h>
    2. include <QXmlInputSource>
    3. include <qfile.h>
    4. include <parser.h>

    int main() {

     Parser* handler=new Parser();
     QXmlInputSource* source=new QXmlInputSource(new QFile("hello.htm"));
     QXmlSimpleReader reader;
     reader.setContentHandler( handler );
     reader.parse( source );
    

    }

    The DOM approach

    /*

      dom.cpp
      A demonstration how to use the dom parsing framework.
      Prints the first subnode of an HTML file, i.e. typically 
      "head" or "body".
      compile it like this:
      g++ -I. -I/opt/kde3/include -I/usr/lib/qt3/include dom.cpp \
      -L/opt/kde3/lib -L/usr/lib/qt3/lib -lqt-mt -lkdeui   
    
    • /
    1. include <qdom.h>
    2. include <qfile.h>
    3. include <kdebug.h>

    int main() {

     QDomDocument doc( "myDocument" );
     QFile qf("hello.htm");
     doc.setContent( &qf );
     QDomElement docElement = doc.documentElement(); 
     QDomNode node;
     node = docElement.firstChild();
     kdDebug() << node.nodeName() << endl;
    

    }

    Drawbacks

    HTML parsing only works for "legal" html documents. For example, look at this code: <html>

     <body>
         <a href="http://www.kde.org/"></a>
         <a href="/index.php?title=Special:User&returnto=Main_Page">Log in</a>
         <a href="http://www.gmx.de"></a>
     </body>
    

    </html> This code contains a & and will bring your parser to an error.

    See here: <html>

     <body>
         <a href="http://www.kde.org/"></a>
         <a href="/index.php" nowrap>Log in</a>
         <a href="http://www.gmx.de"></a>
     </body>
    

    </html> This code will throw an error because of the nowrap that is not xml-conform.