xmlparser.cpp

xml文件解析器 解析所有的xml文件 支持unicode

                    ss++;
                    while (*ss!=_X(';'))
                    {
                        if ((*ss>=_X('0'))&&(*ss<=_X('9'))) j=(j<<4)+*ss-_X('0');
                        else if ((*ss>=_X('A'))&&(*ss<=_X('F'))) j=(j<<4)+*ss-_X('A')+10;
                        else if ((*ss>=_X('a'))&&(*ss<=_X('f'))) j=(j<<4)+*ss-_X('a')+10;
                        else { free((void*)s); pXML->error=eXMLErrorUnknownCharacterEntity;return NULL;}
                        ss++;
                    }
                } else
                {
                    while (*ss!=_X(';'))
                    {
                        if ((*ss>=_X('0'))&&(*ss<=_X('9'))) j=(j*10)+*ss-_X('0');
                        else { free((void*)s); pXML->error=eXMLErrorUnknownCharacterEntity;return NULL;}
                        ss++;
                    }
                }
                (*d++)=(XMLCHAR)j; ss++;
            } else
            {
                entity=XMLEntities;
                do
                {
                    if (xstrnicmp(ss,entity->s,entity->l)==0) { *(d++)=entity->c; ss+=entity->l; break; }
                    entity++;
                } while(entity->s);
            }
        } else
        {
#ifdef _XMLWIDECHAR
            *(d++)=*(ss++);
#else
            switch(XML_ByteTable[(unsigned char)*ss])
            {
            case 4: *(d++)=*(ss++); ll--;
            case 3: *(d++)=*(ss++); ll--;
            case 2: *(d++)=*(ss++); ll--;
            case 1: *(d++)=*(ss++);
            }
#endif
        }
    }
    *d=0;
    return (XMLSTR)s;
}

#define XML_isSPACECHAR(ch) ((ch==_X('\n'))||(ch==_X(' '))||(ch== _X('\t'))||(ch==_X('\r')))

// private:
char myTagCompare(XMLCSTR cclose, XMLCSTR copen)
// !!!! WARNING strange convention&:
// return 0 if equals
// return 1 if different
{
    if (!cclose) return 1;
    int l=(int)xstrlen(cclose);
    if (xstrnicmp(cclose, copen, l)!=0) return 1;
    const XMLCHAR c=copen[l];
    if (XML_isSPACECHAR(c)||
        (c==_X('/' ))||
        (c==_X('<' ))||
        (c==_X('>' ))||
        (c==_X('=' ))) return 0;
    return 1;
}

// Obtain the next character from the string.
static inline XMLCHAR getNextChar(XML *pXML)
{
    XMLCHAR ch = pXML->lpXML[pXML->nIndex];
#ifdef _XMLWIDECHAR
    if (ch!=0) pXML->nIndex++;
#else
    pXML->nIndex+=XML_ByteTable[(unsigned char)ch];
#endif
    return ch;
}

// Find the next token in a string.
// pcbToken contains the number of characters that have been read.
static NextToken GetNextToken(XML *pXML, int *pcbToken, enum XMLTokenTypeTag *pType)
{
    NextToken        result;
    XMLCHAR            ch;
    XMLCHAR            chTemp;
    int              indexStart,nFoundMatch,nIsText=FALSE;
    result.pClr=NULL; // prevent warning

    // Find next non-white space character
    do { indexStart=pXML->nIndex; ch=getNextChar(pXML); } while XML_isSPACECHAR(ch);

    if (ch)
    {
        // Cache the current string pointer
        result.pStr = &pXML->lpXML[indexStart];

        // First check whether the token is in the clear tag list (meaning it
        // does not need formatting).
        ALLXMLClearTag *ctag=XMLClearTags;
        do
        {
            if (xstrncmp(ctag->lpszOpen, result.pStr, ctag->openTagLen)==0)
            {
                result.pClr=ctag;
                pXML->nIndex+=ctag->openTagLen-1;
                *pType=eTokenClear;
                return result;
            }
            ctag++;
        } while(ctag->lpszOpen);

        // If we didn't find a clear tag then check for standard tokens
        switch(ch)
        {
        // Check for quotes
        case _X('\''):
        case _X('\"'):
            // Type of token
            *pType = eTokenQuotedText;
            chTemp = ch;

            // Set the size
            nFoundMatch = FALSE;

            // Search through the string to find a matching quote
            while((ch = getNextChar(pXML)))
            {
                if (ch==chTemp) { nFoundMatch = TRUE; break; }
                if (ch==_X('<')) break;
            }

            // If we failed to find a matching quote
            if (nFoundMatch == FALSE)
            {
                pXML->nIndex=indexStart+1;
                nIsText=TRUE;
                break;
            }

//  4.02.2002
//            if (FindNonWhiteSpace(pXML)) pXML->nIndex--;

            break;

        // Equals (used with attribute values)
        case _X('='):
            *pType = eTokenEquals;
            break;

        // Close tag
        case _X('>'):
            *pType = eTokenCloseTag;
            break;

        // Check for tag start and tag end
        case _X('<'):

            // Peek at the next character to see if we have an end tag '</',
            // or an xml declaration '<?'
            chTemp = pXML->lpXML[pXML->nIndex];

            // If we have a tag end...
            if (chTemp == _X('/'))
            {
                // Set the type and ensure we point at the next character
                getNextChar(pXML);
                *pType = eTokenTagEnd;
            }

            // If we have an XML declaration tag
            else if (chTemp == _X('?'))
            {

                // Set the type and ensure we point at the next character
                getNextChar(pXML);
                *pType = eTokenDeclaration;
            }

            // Otherwise we must have a start tag
            else
            {
                *pType = eTokenTagStart;
            }
            break;

        // Check to see if we have a short hand type end tag ('/>').
        case _X('/'):

            // Peek at the next character to see if we have a short end tag '/>'
            chTemp = pXML->lpXML[pXML->nIndex];

            // If we have a short hand end tag...
            if (chTemp == _X('>'))
            {
                // Set the type and ensure we point at the next character
                getNextChar(pXML);
                *pType = eTokenShortHandClose;
                break;
            }

            // If we haven't found a short hand closing tag then drop into the
            // text process

        // Other characters
        default:
            nIsText = TRUE;
        }

        // If this is a TEXT node
        if (nIsText)
        {
            // Indicate we are dealing with text
            *pType = eTokenText;
            while((ch = getNextChar(pXML)))
            {
                if XML_isSPACECHAR(ch)
                {
                    indexStart++; break;

                } else if (ch==_X('/'))
                {
                    // If we find a slash then this maybe text or a short hand end tag
                    // Peek at the next character to see it we have short hand end tag
                    ch=pXML->lpXML[pXML->nIndex];
                    // If we found a short hand end tag then we need to exit the loop
                    if (ch==_X('>')) { pXML->nIndex--; break; }

                } else if ((ch==_X('<'))||(ch==_X('>'))||(ch==_X('=')))
                {
                    pXML->nIndex--; break;
                }
            }
        }
        *pcbToken = pXML->nIndex-indexStart;
    } else
    {
        // If we failed to obtain a valid character
        *pcbToken = 0;
        *pType = eTokenError;
        result.pStr=NULL;
    }

    return result;
}

XMLCSTR XMLNode::updateName_WOSD(XMLSTR lpszName)
{
    if (!d) { free(lpszName); return NULL; }
    if (d->lpszName&&(lpszName!=d->lpszName)) free((void*)d->lpszName);
    d->lpszName=lpszName;
    return lpszName;
}

// private:
XMLNode::XMLNode(struct XMLNodeDataTag *p){ d=p; (p->ref_count)++; }
XMLNode::XMLNode(XMLNodeData *pParent, XMLSTR lpszName, char isDeclaration)
{
    d=(XMLNodeData*)malloc(sizeof(XMLNodeData));
    d->ref_count=1;

    d->lpszName=NULL;
    d->nChild= 0;
    d->nText = 0;
    d->nClear = 0;
    d->nAttribute = 0;

    d->isDeclaration = isDeclaration;

    d->pParent = pParent;
    d->pChild= NULL;
    d->pText= NULL;
    d->pClear= NULL;
    d->pAttribute= NULL;
    d->pOrder= NULL;

    updateName_WOSD(lpszName);
}

XMLNode XMLNode::createXMLTopNode_WOSD(XMLSTR lpszName, char isDeclaration) { return XMLNode(NULL,lpszName,isDeclaration); }
XMLNode XMLNode::createXMLTopNode(XMLCSTR lpszName, char isDeclaration) { return XMLNode(NULL,stringDup(lpszName),isDeclaration); }

#define MEMORYINCREASE 50

static inline void myFree(void *p) { if (p) free(p); };
static inline void *myRealloc(void *p, int newsize, int memInc, int sizeofElem)
{
    if (p==NULL) { if (memInc) return malloc(memInc*sizeofElem); return malloc(sizeofElem); }
    if ((memInc==0)||((newsize%memInc)==0)) p=realloc(p,(newsize+memInc)*sizeofElem);
//    if (!p)
//    {
//        printf("XMLParser Error: Not enough memory! Aborting...\n"); exit(220);
//    }
    return p;
}

// private:
int XMLNode::findPosition(XMLNodeData *d, int index, XMLElementType xtype)
{
    if (index<0) return -1;
    int i=0,j=(int)((index<<2)+xtype),*o=d->pOrder; while (o[i]!=j) i++; return i;
}

// private:
// update "order" information when deleting a content of a XMLNode
int XMLNode::removeOrderElement(XMLNodeData *d, XMLElementType t, int index)
{
    int n=d->nChild+d->nText+d->nClear, *o=d->pOrder,i=findPosition(d,index,t);
    memmove(o+i, o+i+1, (n-i)*sizeof(int));
    for (;i<n;i++)
        if ((o[i]&3)==(int)t) o[i]-=4;
    // We should normally do:
    // d->pOrder=(int)realloc(d->pOrder,n*sizeof(int));
    // but we skip reallocation because it's too time consuming.
    // Anyway, at the end, it will be free'd completely at once.
    return i;
}

void *XMLNode::addToOrder(int memoryIncrease,int *_pos, int nc, void *p, int size, XMLElementType xtype)
{
    //  in: *_pos is the position inside d->pOrder ("-1" means "EndOf")
    // out: *_pos is the index inside p
    p=myRealloc(p,(nc+1),memoryIncrease,size);
    int n=d->nChild+d->nText+d->nClear;
    d->pOrder=(int*)myRealloc(d->pOrder,n+1,memoryIncrease*3,sizeof(int));
    int pos=*_pos,*o=d->pOrder;

    if ((pos<0)||(pos>=n)) { *_pos=nc; o[n]=(int)((nc<<2)+xtype); return p; }

    int i=pos;
    memmove(o+i+1, o+i, (n-i)*sizeof(int));

    while ((pos<n)&&((o[pos]&3)!=(int)xtype)) pos++;
    if (pos==n) { *_pos=nc; o[n]=(int)((nc<<2)+xtype); return p; }

    o[i]=o[pos];
    for (i=pos+1;i<=n;i++) if ((o[i]&3)==(int)xtype) o[i]+=4;

    *_pos=pos=o[pos]>>2;
    memmove(((char*)p)+(pos+1)*size,((char*)p)+pos*size,(nc-pos)*size);

    return p;
}

// Add a child node to the given element.
XMLNode XMLNode::addChild_priv(int memoryIncrease, XMLSTR lpszName, char isDeclaration, int pos)
{
    if (!lpszName) return emptyXMLNode;