qxml.cpp

奇趣公司比较新的qt/emd版本

/****************************************************************************
**
** Copyright (C) 1992-2007 Trolltech ASA. All rights reserved.
**
** This file is part of the QtXML module of the Qt Toolkit.
**
** This file may be used under the terms of the GNU General Public
** License version 2.0 as published by the Free Software Foundation
** and appearing in the file LICENSE.GPL included in the packaging of
** this file.  Please review the following information to ensure GNU
** General Public Licensing requirements will be met:
** http://trolltech.com/products/qt/licenses/licensing/opensource/
**
** If you are unsure which license is appropriate for your use, please
** review the following information:
** http://trolltech.com/products/qt/licenses/licensing/licensingoverview
** or contact the sales department at [e-mail].
**
** In addition, as a special exception, Trolltech gives you certain
** additional rights. These rights are described in the Trolltech GPL
** Exception version 1.0, which can be found at
** http://www.trolltech.com/products/qt/gplexception/ and in the file
** GPL_EXCEPTION.txt in this package.
**
** In addition, as a special exception, Trolltech, as the sole copyright
** holder for Qt Designer, grants users of the Qt/Eclipse Integration
** plug-in the right for the Qt/Eclipse Integration to link to
** functionality provided by Qt Designer and its related libraries.
**
** Trolltech reserves all rights not expressly granted herein.
**
** This file is provided AS IS with NO WARRANTY OF ANY KIND, INCLUDING THE
** WARRANTY OF DESIGN, MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.
**
****************************************************************************/

#include "qxml.h"
#include "qtextcodec.h"
#include "qbuffer.h"
#include "qregexp.h"
#include "qmap.h"
#include "qstack.h"
#include <qdebug.h>


#ifdef Q_CC_BOR // borland 6 finds bogus warnings when building this file in uic3
#    pragma warn -8080
#endif

//#define QT_QXML_DEBUG

// Error strings for the XML reader
#define XMLERR_OK                         QT_TRANSLATE_NOOP("QXml", "no error occurred")
#define XMLERR_ERRORBYCONSUMER            QT_TRANSLATE_NOOP("QXml", "error triggered by consumer")
#define XMLERR_UNEXPECTEDEOF              QT_TRANSLATE_NOOP("QXml", "unexpected end of file")
#define XMLERR_MORETHANONEDOCTYPE         QT_TRANSLATE_NOOP("QXml", "more than one document type definition")
#define XMLERR_ERRORPARSINGELEMENT        QT_TRANSLATE_NOOP("QXml", "error occurred while parsing element")
#define XMLERR_TAGMISMATCH                QT_TRANSLATE_NOOP("QXml", "tag mismatch")
#define XMLERR_ERRORPARSINGCONTENT        QT_TRANSLATE_NOOP("QXml", "error occurred while parsing content")
#define XMLERR_UNEXPECTEDCHARACTER        QT_TRANSLATE_NOOP("QXml", "unexpected character")
#define XMLERR_INVALIDNAMEFORPI           QT_TRANSLATE_NOOP("QXml", "invalid name for processing instruction")
#define XMLERR_VERSIONEXPECTED            QT_TRANSLATE_NOOP("QXml", "version expected while reading the XML declaration")
#define XMLERR_WRONGVALUEFORSDECL         QT_TRANSLATE_NOOP("QXml", "wrong value for standalone declaration")
#define XMLERR_EDECLORSDDECLEXPECTED      QT_TRANSLATE_NOOP("QXml", "encoding declaration or standalone declaration expected while reading the XML declaration")
#define XMLERR_SDDECLEXPECTED             QT_TRANSLATE_NOOP("QXml", "standalone declaration expected while reading the XML declaration")
#define XMLERR_ERRORPARSINGDOCTYPE        QT_TRANSLATE_NOOP("QXml", "error occurred while parsing document type definition")
#define XMLERR_LETTEREXPECTED             QT_TRANSLATE_NOOP("QXml", "letter is expected")
#define XMLERR_ERRORPARSINGCOMMENT        QT_TRANSLATE_NOOP("QXml", "error occurred while parsing comment")
#define XMLERR_ERRORPARSINGREFERENCE      QT_TRANSLATE_NOOP("QXml", "error occurred while parsing reference")
#define XMLERR_INTERNALGENERALENTITYINDTD QT_TRANSLATE_NOOP("QXml", "internal general entity reference not allowed in DTD")
#define XMLERR_EXTERNALGENERALENTITYINAV  QT_TRANSLATE_NOOP("QXml", "external parsed general entity reference not allowed in attribute value")
#define XMLERR_EXTERNALGENERALENTITYINDTD QT_TRANSLATE_NOOP("QXml", "external parsed general entity reference not allowed in DTD")
#define XMLERR_UNPARSEDENTITYREFERENCE    QT_TRANSLATE_NOOP("QXml", "unparsed entity reference in wrong context")
#define XMLERR_RECURSIVEENTITIES          QT_TRANSLATE_NOOP("QXml", "recursive entities")
#define XMLERR_ERRORINTEXTDECL            QT_TRANSLATE_NOOP("QXml", "error in the text declaration of an external entity")

// the constants for the lookup table
static const signed char cltWS      =  0; // white space
static const signed char cltPer     =  1; // %
static const signed char cltAmp     =  2; // &
static const signed char cltGt      =  3; // >
static const signed char cltLt      =  4; // <
static const signed char cltSlash   =  5; // /
static const signed char cltQm      =  6; // ?
static const signed char cltEm      =  7; // !
static const signed char cltDash    =  8; // -
static const signed char cltCB      =  9; // ]
static const signed char cltOB      = 10; // [
static const signed char cltEq      = 11; // =
static const signed char cltDq      = 12; // "
static const signed char cltSq      = 13; // '
static const signed char cltUnknown = 14;

// Hack for letting QDom know where the skipped entity occurred
bool qt_xml_skipped_entity_in_content;

// character lookup table
static const signed char charLookupTable[256]={
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x00 - 0x07
    cltUnknown, // 0x08
    cltWS,      // 0x09 \t
    cltWS,      // 0x0A \n
    cltUnknown, // 0x0B
    cltUnknown, // 0x0C
    cltWS,      // 0x0D \r
    cltUnknown, // 0x0E
    cltUnknown, // 0x0F
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x17 - 0x16
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x18 - 0x1F
    cltWS,      // 0x20 Space
    cltEm,      // 0x21 !
    cltDq,      // 0x22 "
    cltUnknown, // 0x23
    cltUnknown, // 0x24
    cltPer,     // 0x25 %
    cltAmp,     // 0x26 &
    cltSq,      // 0x27 '
    cltUnknown, // 0x28
    cltUnknown, // 0x29
    cltUnknown, // 0x2A
    cltUnknown, // 0x2B
    cltUnknown, // 0x2C
    cltDash,    // 0x2D -
    cltUnknown, // 0x2E
    cltSlash,   // 0x2F /
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x30 - 0x37
    cltUnknown, // 0x38
    cltUnknown, // 0x39
    cltUnknown, // 0x3A
    cltUnknown, // 0x3B
    cltLt,      // 0x3C <
    cltEq,      // 0x3D =
    cltGt,      // 0x3E >
    cltQm,      // 0x3F ?
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x40 - 0x47
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x48 - 0x4F
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x50 - 0x57
    cltUnknown, // 0x58
    cltUnknown, // 0x59
    cltUnknown, // 0x5A
    cltOB,      // 0x5B [
    cltUnknown, // 0x5C
    cltCB,      // 0x5D]
    cltUnknown, // 0x5E
    cltUnknown, // 0x5F
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x60 - 0x67
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x68 - 0x6F
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x70 - 0x77
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x78 - 0x7F
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x80 - 0x87
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x88 - 0x8F
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x90 - 0x97
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0x98 - 0x9F
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xA0 - 0xA7
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xA8 - 0xAF
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xB0 - 0xB7
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xB8 - 0xBF
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xC0 - 0xC7
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xC8 - 0xCF
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xD0 - 0xD7
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xD8 - 0xDF
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xE0 - 0xE7
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xE8 - 0xEF
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, // 0xF0 - 0xF7
    cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown, cltUnknown  // 0xF8 - 0xFF
};

//
// local helper functions
//

/*
  This function strips the TextDecl [77] ("<?xml ...?>") from the string \a
  str. The stripped version is stored in \a str. If this function finds an
  invalid TextDecl, it returns false, otherwise true.

  This function is used for external entities since those can include an
  TextDecl that must be stripped before inserting the entity.
*/
static bool stripTextDecl(QString& str)
{
    QString textDeclStart(QLatin1String("<?xml"));
    if (str.startsWith(textDeclStart)) {
        QRegExp textDecl(QString::fromLatin1(
            "^<\\?xml\\s+"
            "(version\\s*=\\s*((['\"])[-a-zA-Z0-9_.:]+\\3))?"
            "\\s*"
            "(encoding\\s*=\\s*((['\"])[A-Za-z][-a-zA-Z0-9_.]*\\6))?"
            "\\s*\\?>"
       ));
        QString strTmp = str.replace(textDecl, QLatin1String(""));
        if (strTmp.length() != str.length())
            return false; // external entity has wrong TextDecl
        str = strTmp;
    }
    return true;
}


class QXmlAttributesPrivate
{
};

/* \class QXmlInputSourcePrivate
    \internal

  There's a slight misdesign in this class that can
  be worth to keep in mind: the `str' member is
  a buffer which QXmlInputSource::next() returns from,
  and which is populated from the input device or input
  stream. However, when the input is a QString(the user called
  QXmlInputSource::setData()), `str' has two roles: it's the
  buffer, but also the source. This /seems/ to be no problem
  because in the case of having no device or stream, the QString
  is read in one go.
 */
class QXmlInputSourcePrivate
{
public:
    QIODevice *inputDevice;
    QTextStream *inputStream;

    QString str;
    const QChar *unicode;
    int pos;
    int length;
    bool nextReturnedEndOfData;
#ifndef QT_NO_TEXTCODEC
    QTextDecoder *encMapper;
#endif

    QByteArray encodingDeclBytes;
    QString encodingDeclChars;
    bool lookingForEncodingDecl;
};
class QXmlParseExceptionPrivate
{
public:
    QString msg;
    int column;
    int line;
    QString pub;
    QString sys;

};

class QXmlLocatorPrivate
{
};

class QXmlDefaultHandlerPrivate
{
};

class QXmlSimpleReaderPrivate
{
private:
    // functions
    QXmlSimpleReaderPrivate();
    ~QXmlSimpleReaderPrivate();
    void initIncrementalParsing();

    // used to determine if elements are correctly nested
    QStack<QString> tags;

    // used by parseReference() and parsePEReference()
    enum EntityRecognitionContext { InContent, InAttributeValue, InEntityValue, InDTD };

    // used for entity declarations
    struct ExternParameterEntity
    {
        ExternParameterEntity() {}
        ExternParameterEntity(const QString &p, const QString &s)
            : publicId(p), systemId(s) {}
        QString publicId;
        QString systemId;

        Q_DUMMY_COMPARISON_OPERATOR(ExternParameterEntity)
    };
    struct ExternEntity
    {
        ExternEntity() {}
        ExternEntity(const QString &p, const QString &s, const QString &n)
            : publicId(p), systemId(s), notation(n) {}
        QString publicId;
        QString systemId;
        QString notation;
        Q_DUMMY_COMPARISON_OPERATOR(ExternEntity)
    };
    QMap<QString,ExternParameterEntity> externParameterEntities;
    QMap<QString,QString> parameterEntities;
    QMap<QString,ExternEntity> externEntities;
    QMap<QString,QString> entities;

    // used for parsing of entity references
    struct XmlRef {
        XmlRef(const QString &_name = QString(), const QString &_value = QString())
            : name(_name), value(_value), index(0) {}
        bool isEmpty() const { return index == value.length(); }
        QChar next() { return value.at(index++); }
        QString name;
        QString value;
        int index;
    };
    QStack<XmlRef> xmlRefStack;

    // used for standalone declaration
    enum Standalone { Yes, No, Unknown };

    QString doctype; // only used for the doctype
    QString xmlVersion; // only used to store the version information
    QString encoding; // only used to store the encoding
    Standalone standalone; // used to store the value of the standalone declaration

    QString publicId; // used by parseExternalID() to store the public ID
    QString systemId; // used by parseExternalID() to store the system ID

    // Since publicId/systemId is used as temporary variables by parseExternalID(), it
    // might overwrite the PUBLIC/SYSTEM for the document we're parsing. In effect, we would
    // possibly send off an QXmlParseException that has the PUBLIC/SYSTEM of a entity declaration
    // instead of those of the current document.
    // Hence we have these two variables for storing the document's data.
    QString thisPublicId;
    QString thisSystemId;

    QString attDeclEName; // use by parseAttlistDecl()
    QString attDeclAName; // use by parseAttlistDecl()

    // flags for some features support
    bool useNamespaces;
    bool useNamespacePrefixes;
    bool reportWhitespaceCharData;
    bool reportEntities;

    // used to build the attribute list
    QXmlAttributes attList;

    // used in QXmlSimpleReader::parseContent() to decide whether character
    // data was read
    bool contentCharDataRead;

    // helper classes
    QXmlLocator *locator;
    QXmlNamespaceSupport namespaceSupport;

    // error string
    QString error;

    // arguments for parse functions (this is needed to allow incremental
    // parsing)
    bool parsePI_xmldecl;
    bool parseName_useRef;
    bool parseReference_charDataRead;
    EntityRecognitionContext parseReference_context;
    bool parseExternalID_allowPublicID;
    EntityRecognitionContext parsePEReference_context;
    QString parseString_s;

    // for incremental parsing
    struct ParseState {
        typedef bool (QXmlSimpleReaderPrivate::*ParseFunction)();
        ParseFunction function;
        int state;
    };
    QStack<ParseState> *parseStack;

    // used in parseProlog()
    bool xmldecl_possible;
    bool doctype_read;

    // used in parseDoctype()
    bool startDTDwasReported;

    // used in parseString()
    signed char Done;


    // variables
    QXmlContentHandler *contentHnd;
    QXmlErrorHandler   *errorHnd;
    QXmlDTDHandler     *dtdHnd;
    QXmlEntityResolver *entityRes;
    QXmlLexicalHandler *lexicalHnd;
    QXmlDeclHandler    *declHnd;

    QXmlInputSource *inputSource;

    QChar c; // the character at reading position
    int   lineNr; // number of line
    int   columnNr; // position in line

    QChar   nameArray[256]; // only used for names
    QString nameValue; // only used for names
    int     nameArrayPos;
    int     nameValueLen;
    QChar   refArray[256]; // only used for references
    QString refValue; // only used for references
    int     refArrayPos;
    int     refValueLen;
    QChar   stringArray[256]; // used for any other strings that are parsed
    QString stringValue; // used for any other strings that are parsed
    int     stringArrayPos;
    int     stringValueLen;
    QString emptyStr;

    const QString &string();
    void stringClear();
    void stringAddC(QChar);
    inline void stringAddC() { stringAddC(c); }
    const QString &name();
    void nameClear();
    void nameAddC(QChar);
    inline void nameAddC() { nameAddC(c); }
    const QString &ref();
    void refClear();
    void refAddC(QChar);
    inline void refAddC() { refAddC(c); }