UT_XMLSaxParser.h

Go to the documentation of this file.

/*
 * PROPRIETARY INFORMATION.  This software is proprietary to
 * Side Effects Software Inc., and is not to be reproduced,
 * transmitted, or disclosed in any way without written permission.
 *
 * NAME:        UT_XMLSaxParser.h (  Library, C++)
 *
 * COMMENTS:
 *  The SAX parser uses the libxml2 API documented at:
 *  http://xmlsoft.org/html/libxml-parser.html
 *  http://xmlsoft.org/html/libxml-tree.html
 *
 */

#ifndef __UT_XMLSaxParser__
#define __UT_XMLSaxParser__

#include "UT_API.h"
#include "UT_Array.h"
#include "UT_String.h"
#include "UT_UniquePtr.h"
#include "UT_XMLParser.h"

#include <utility>

struct _xmlSAXHandler;
typedef struct _xmlSAXHandler *xmlSAXHandlerPtr;

class UT_API UT_XMLSaxParser : public UT_XMLParser
{
public:
    /// Standard constructor.
    UT_XMLSaxParser();

    /// Standard destructor.
    ~UT_XMLSaxParser() override;

    /// Parse a file, handing the callbacks below libxml2's UTF-8 unchanged.
    /// parseFile() instead converts to Latin-1, which silently drops every
    /// code point above U+00FF; that remains the default so existing
    /// subclasses are unaffected. Use this one if your consumers are UTF-8
    /// clean. Returns true on a successful parse.
    bool                     parseFileUTF8(const char *filename);

protected:

    /// An XML element attribute as defined by a name and a value.
    typedef std::pair<UT_String, UT_String>   UT_XMLAttribute;

public:
    /// A list of XML element attributes.
    typedef UT_Array<UT_XMLAttribute *>     UT_XMLAttributeList;

    /// Callback methods that are called at various times
    /// of the parse.
    virtual void             onStartDocument();
    virtual void             onEndDocument();
    virtual void             onStartElement(const UT_String &name,
                                const UT_XMLAttributeList &attribs);
    virtual void             onEndElement(const UT_String &name);
    virtual void             onCharacters(const UT_String &characters);
    virtual void             onCDATABlock(const UT_String &characters);
    virtual void             onComment(const UT_String &text);
    virtual void             onWarning(const UT_String &text);
    virtual void             onError(const UT_String &text);
    virtual void             onFatalError(const UT_String &text);

    /// Parse XML using the current context.
    /// Return true on a successful parse.  Return false otherwise.
    bool                     parseContextSubclass() override;

    /// Utility method for getting the value of an attribute.
    static bool              getAttribute(const UT_XMLAttributeList &attribs,
                                          const char * name, UT_String & value);

public:
    static UT_String getStringAttribute(const UT_XMLAttributeList &attribs,
                                         const char * name)
                      {
                            UT_String tmp;
                            if (getAttribute(attribs, name, tmp))
                                return tmp;
                            return "";
                      }

    static bool hasAttribute(const UT_XMLAttributeList &attribs,
                                         const char * name)
{
                            UT_String tmp;
                            return getAttribute(attribs, name, tmp);
}


    static int getIntegerAttribute(const UT_XMLAttributeList &attribs,
                                         const char * name)
                      {
                            UT_String tmp;
                            if (getAttribute(attribs, name, tmp))
                                return atoi(tmp.c_str());
                            return 0;
                      }

    static double getDoubleAttribute(const UT_XMLAttributeList &attribs,
                                         const char * name, double defValue)
                      {
                            UT_String tmp;
                            if (getAttribute(attribs, name, tmp)) 
                                return atof(tmp.c_str());
                        return defValue;
                      }
private:

    /// SAX parsing callback handler.
    UT_UniquePtr<_xmlSAXHandler> myCallbackHandler;

    /// We need to give the callback helper class
    /// access to the parser's callback functions (i.e. onStartElement).
    friend class utSaxParserCallback;
};

#endif