INXString CMarkup::x_GetData( int iPos ) const
{
// Return a string representing data between start and end tag
// Return empty string if there are any children elements
if ( ! m_aPos[iPos].iElemChild && ! m_aPos[iPos].IsEmptyElement() )
{
// See if it is a CDATA section
const char* tempCh = (const char*)m_csDoc;
char* szDoc = (char*)tempCh;
int nChar = m_aPos[iPos].nStartR + 1;
if ( x_FindAny( szDoc, nChar ) && szDoc[nChar] == '<'
&& nChar + 11 < m_aPos[iPos].nEndL
&& strncmp( &szDoc[nChar], "<![CDATA[",9) == 0)
{
nChar += 9;
int nEndCDATA = m_csDoc.Finds("]]>", nChar);
if ( nEndCDATA != -1 && nEndCDATA < m_aPos[iPos].nEndL )
{
return m_csDoc.Mid( nChar, nEndCDATA - nChar );
}
}
return x_TextFromDoc( m_aPos[iPos].nStartR+1, m_aPos[iPos].nEndL-1 );
}
INXString temp("");
return temp;
}
CString CMarkup::x_GetData( int iPos ) const
{
// Return a string representing data between start and end tag
// Return empty string if there are any children elements
if ( ! m_aPos[iPos].iElemChild && ! m_aPos[iPos].IsEmptyElement() )
{
// See if it is a CDATA section
LPCTSTR szDoc = (LPCTSTR)m_csDoc;
int nChar = m_aPos[iPos].nStartR + 1;
if ( x_FindAny( szDoc, nChar ) && szDoc[nChar] == _T('<')
&& nChar + 11 < m_aPos[iPos].nEndL
&& _tcsncmp( &szDoc[nChar], _T("<![CDATA["), 9 ) == 0 )
{
nChar += 9;
int nEndCDATA = m_csDoc.Find( _T("]]>"), nChar );
if ( nEndCDATA != -1 && nEndCDATA < m_aPos[iPos].nEndL )
{
return m_csDoc.Mid( nChar, nEndCDATA - nChar );
}
}
return x_TextFromDoc( m_aPos[iPos].nStartR+1, m_aPos[iPos].nEndL-1 );
}
return _T("");
}
bool CMarkup::x_FindToken( CMarkup::TokenPos& token )
{
// Starting at token.nNext, bypass whitespace and find the next token
// returns true on success, members of token point to token
// returns false on end of document, members point to end of document
char* szDoc = token.szDoc;
int nChar = token.nNext;
token.bIsString = false;
// By-pass leading whitespace
if ( ! x_FindAny(szDoc,nChar) )
{
// No token was found before end of document
token.nL = nChar;
token.nR = nChar;
token.nNext = nChar;
return false;
}
// Is it an opening quote?
char cFirstChar = szDoc[nChar];
if ( cFirstChar == _T('\"') || cFirstChar == _T('\'') )
{
token.bIsString = true;
// Move past opening quote
++nChar;
token.nL = nChar;
// Look for closing quote
x_FindChar( token.szDoc, nChar, cFirstChar );
// Set right to before closing quote
token.nR = nChar - 1;
// Set nChar past closing quote unless at end of document
if ( szDoc[nChar] )
++nChar;
}
else
{
// Go until special char or whitespace
token.nL = nChar;
while ( szDoc[nChar] && ! strchr(" \t\n\r<>=\\/?!",szDoc[nChar]) )
nChar += _tclen(&szDoc[nChar]);
// Adjust end position if it is one special char
if ( nChar == token.nL )
++nChar; // it is a special char
token.nR = nChar - 1;
}
// nNext points to one past last char of token
token.nNext = nChar;
return true;
}
void CMarkup::x_LocateNew( int iPosParent, int& iPosRel, int& nOffset, int nLength, int nFlags )
{
// Determine where to insert new element or node
//
bool bInsert = (nFlags&1)?true:false;
bool bHonorWhitespace = (nFlags&2)?true:false;
int nStartL;
if ( nLength )
{
// Located at a non-element node
if ( bInsert )
nStartL = nOffset;
else
nStartL = nOffset + nLength;
}
else if ( iPosRel )
{
// Located at an element
if ( bInsert ) // precede iPosRel
nStartL = m_aPos[iPosRel].nStartL;
else // follow iPosRel
nStartL = m_aPos[iPosRel].nEndR + 1;
}
else if ( ! iPosParent )
{
// Outside of all elements
if ( bInsert )
nStartL = 0;
else
nStartL = m_csDoc.GetLength();
}
else if ( m_aPos[iPosParent].IsEmptyElement() )
{
// Parent has no separate end tag, so split empty element
nStartL = m_aPos[iPosParent].nStartR;
}
else
{
if ( bInsert ) // after start tag
nStartL = m_aPos[iPosParent].nStartR + 1;
else // before end tag
nStartL = m_aPos[iPosParent].nEndL;
}
// Go up to start of next node, unless its splitting an empty element
if ( ! bHonorWhitespace && ! m_aPos[iPosParent].IsEmptyElement() )
{
char* szDoc = (char*)m_csDoc;
int nChar = nStartL;
if ( ! x_FindAny(szDoc,nChar) || szDoc[nChar] == _T('<') )
nStartL = nChar;
}
// Determine iPosBefore
int iPosBefore = 0;
if ( iPosRel )
{
if ( bInsert )
{
// Is iPosRel past first sibling?
int iPosPrev = m_aPos[iPosParent].iElemChild;
if ( iPosPrev != iPosRel )
{
// Find previous sibling of iPosRel
while ( m_aPos[iPosPrev].iElemNext != iPosRel )
iPosPrev = m_aPos[iPosPrev].iElemNext;
iPosBefore = iPosPrev;
}
}
else
{
iPosBefore = iPosRel;
}
}
else if ( m_aPos[iPosParent].iElemChild )
{
if ( ! bInsert )
{
// Find last element under iPosParent
int iPosLast = m_aPos[iPosParent].iElemChild;
int iPosNext = iPosLast;
while ( iPosNext )
{
iPosLast = iPosNext;
iPosNext = m_aPos[iPosNext].iElemNext;
}
iPosBefore = iPosLast;
}
}
nOffset = nStartL;
iPosRel = iPosBefore;
}
int CMarkup::x_ParseNode( CMarkup::TokenPos& token )
{
// Call this with token.nNext set to the start of the node
// This returns the node type and token.nNext set to the char after the node
// If the node is not found or an element, token.nR is not determined
int nTypeFound = 0;
char* szDoc = token.szDoc;
token.nL = token.nNext;
if ( szDoc[token.nL] == '<' )
{
// Started with <, could be:
// <!--...--> comment
// <!DOCTYPE ...> dtd
// <?target ...?> processing instruction
// <![CDATA[...]]> cdata section
// <NAME ...> element
//
if ( ! szDoc[token.nL+1] || ! szDoc[token.nL+2] )
return 0;
char cFirstChar = szDoc[token.nL+1];
char* szEndOfNode = NULL;
if ( cFirstChar == '?')
{
nTypeFound = MNT_PROCESSING_INSTRUCTION;
szEndOfNode = (char*)("?>");
}
else if ( cFirstChar == '!' )
{
char cSecondChar = szDoc[token.nL+2];
if ( cSecondChar == '[' )
{
nTypeFound = MNT_CDATA_SECTION;
szEndOfNode = (char*)("]]>");
}
else if ( cSecondChar == '-')
{
nTypeFound = MNT_COMMENT;
szEndOfNode = "-->";
}
else
{
// Document type requires tokenizing because of strings and brackets
nTypeFound = 0;
int nBrackets = 0;
while ( x_FindToken(token) )
{
if ( ! token.bIsString )
{
char cChar = szDoc[token.nL];
if ( cChar == '[')
++nBrackets;
else if ( cChar == ']' )
--nBrackets;
else if ( nBrackets == 0 && cChar == _T('>') )
{
nTypeFound = MNT_DOCUMENT_TYPE;
break;
}
}
}
if ( ! nTypeFound )
return 0;
}
}
else if ( cFirstChar == '/')
{
// End tag means no node found within parent element
return 0;
}
else
{
nTypeFound = MNT_ELEMENT;
}
// Search for end of node if not found yet
if ( szEndOfNode )
{
char* pEnd = strstr( &szDoc[token.nNext], szEndOfNode );
if ( ! pEnd )
return 0; // not well-formed
token.nNext = (pEnd - szDoc) + strlen(szEndOfNode);
}
}
else if ( szDoc[token.nL] )
{
// It is text or whitespace because it did not start with <
nTypeFound = MNT_WHITESPACE;
token.nNext = token.nL;
if ( x_FindAny(szDoc,token.nNext) )
{
if ( szDoc[token.nNext] != _T('<') )
{
nTypeFound = MNT_TEXT;
x_FindChar( szDoc, token.nNext, _T('<') );
}
}
}
return nTypeFound;
}
