markup.cpp

MarkUP ,用于VC环境下

bool CMarkup::x_FindAttrib( CMarkup::TokenPos& token, MCD_PCSZ szAttrib, int n/*=0*/ )
{
	// Return true if found, otherwise false and token.nNext is new insertion point
	// If szAttrib is NULL find attrib n and leave token at attrib name
	// If szAttrib is given, find matching attrib and leave token at value
	// support non-well-formed attributes e.g. href=/advanced_search?hl=en, nowrap
	// token also holds start and length of preceeding whitespace to support remove
	//
	int nPreSpaceStart;
	int nPreSpaceLength;
	int nChar;
	MCD_CHAR cFirstChar;
	MCD_PCSZ szDoc = token.szDoc;
	int nAttrib = -1; // starts at tag name
	int nFoundAttribNameR = 0;
	bool bAfterEqual = false;
	while ( 1 )
	{
		// Starting at token.nNext, bypass whitespace and find the next token
		nChar = token.nNext;
		nPreSpaceStart = nChar;
		if ( ! x_FindAny(szDoc,nChar) )
			break;
		nPreSpaceLength = nChar - nPreSpaceStart;

		// Is it an opening quote?
		cFirstChar = szDoc[nChar];
		if ( cFirstChar == _T('\"') || cFirstChar == _T('\'') )
		{
			token.nTokenFlags |= MNF_QUOTED;

			// Move past opening quote
			++nChar;
			token.nL = nChar;

			// Look for closing quote
			while ( szDoc[nChar] && szDoc[nChar] != cFirstChar )
				nChar += MCD_CLEN( &szDoc[nChar] );

			// 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
		{
			token.nTokenFlags &= ~MNF_QUOTED;

			// Go until special char or whitespace
			token.nL = nChar;
			if ( bAfterEqual )
			{
				while ( szDoc[nChar] && ! MCD_PSZCHR(_T(" \t\n\r>"),szDoc[nChar]) )
					nChar += MCD_CLEN( &szDoc[nChar] );
			}
			else
			{
				while ( szDoc[nChar] && ! MCD_PSZCHR(_T("= \t\n\r>/?"),szDoc[nChar]) )
					nChar += MCD_CLEN( &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;

		if ( ! bAfterEqual && ! (token.nTokenFlags&MNF_QUOTED) )
		{
			// Is it an equal sign?
			MCD_CHAR cChar = szDoc[token.nL];
			if ( cChar == _T('=') )
			{
				bAfterEqual = true;
				continue;
			}

			// Is it the right angle bracket?
			if ( cChar == _T('>') || cChar == _T('/') || cChar == _T('?') )
			{
				token.nNext = nPreSpaceStart;
				break; // attrib not found
			}

			if ( nFoundAttribNameR )
				break;

			// Attribute name
			if ( nAttrib != -1 )
			{
				if ( ! szAttrib )
				{
					if ( nAttrib == n )
						return true; // found by number
				}
				else if ( token.Match(szAttrib) )
				{
					// Matched attrib name, go forward to value
					nFoundAttribNameR = token.nR;
					token.nPreSpaceStart = nPreSpaceStart;
					token.nPreSpaceLength = nPreSpaceLength;
				}
			}
			++nAttrib;
		}
		else if ( nFoundAttribNameR )
			break;
		bAfterEqual = false;
	}

	if ( nFoundAttribNameR )
	{
		if ( ! bAfterEqual )
		{
			// when attribute has no value the value is the attribute name
			token.nL = token.nPreSpaceStart + token.nPreSpaceLength;
			token.nR = nFoundAttribNameR;
			token.nNext = nFoundAttribNameR + 1;
		}
		return true; // found by name
	}
	return false; // not found
}

MCD_STR CMarkup::x_GetAttrib( int iPos, MCD_PCSZ szAttrib ) const
{
	// Return the value of the attrib
	TokenPos token( m_strDoc, m_nFlags );
	if ( iPos && m_nNodeType == MNT_ELEMENT )
		token.nNext = m_aPos[iPos].nStart + 1;
	else if ( iPos == m_iPos && m_nNodeLength && m_nNodeType == MNT_PROCESSING_INSTRUCTION )
		token.nNext = m_nNodeOffset + 2;
	else
		return _T("");

	if ( szAttrib && x_FindAttrib( token, szAttrib ) )
		return UnescapeText( &token.szDoc[token.nL], token.Length() );
	return _T("");
}

bool CMarkup::x_SetAttrib( int iPos, MCD_PCSZ szAttrib, int nValue )
{
	// Convert integer to string
	MCD_CHAR szVal[25];
	MCD_SPRINTF( szVal, _T("%d"), nValue );
	return x_SetAttrib( iPos, szAttrib, szVal );
}

bool CMarkup::x_SetAttrib( int iPos, MCD_PCSZ szAttrib, MCD_PCSZ szValue )
{
	// Set attribute in iPos element
	TokenPos token( m_strDoc, m_nFlags );
	if ( iPos && m_nNodeType == MNT_ELEMENT )
		token.nNext = m_aPos[iPos].nStart + 1;
	else if ( iPos == m_iPos && m_nNodeLength && m_nNodeType == MNT_PROCESSING_INSTRUCTION )
		token.nNext = m_nNodeOffset + 2;
	else
		return false;

	// Create insertion text depending on whether attribute already exists
	// Decision: for empty value leaving attrib="" instead of removing attrib
	int nReplace = 0;
	int nInsertAt;
	MCD_STR strInsert;
	strInsert += x_ATTRIBQUOTE;
	strInsert += EscapeText( szValue, MNF_ESCAPEQUOTES );
	strInsert += x_ATTRIBQUOTE;
	if ( x_FindAttrib( token, szAttrib ) )
	{
		// Replace value
		nInsertAt = token.nL - ((token.nTokenFlags&MNF_QUOTED)?1:0);
		nReplace = token.Length() + ((token.nTokenFlags&MNF_QUOTED)?2:0);
	}
	else
	{
		// Insert string name value pair
		MCD_STR strFormat;
		strFormat = _T(" ");
		strFormat += szAttrib;
		strFormat += _T("=");
		strFormat += strInsert;
		strInsert = strFormat;
		nInsertAt = token.nNext;
	}

	x_DocChange( nInsertAt, nReplace, strInsert );
	int nAdjust = MCD_STRLENGTH(strInsert) - nReplace;
	if ( m_nNodeType == MNT_PROCESSING_INSTRUCTION )
	{
		x_AdjustForNode( m_iPosParent, m_iPos, nAdjust );
		m_nNodeLength += nAdjust;
		MARKUP_SETDEBUGSTATE;
		return true;
	}
	m_aPos[iPos].AdjustStartTagLen( nAdjust );
	m_aPos[iPos].nLength += nAdjust;
	x_Adjust( iPos, nAdjust );
	MARKUP_SETDEBUGSTATE;
	return true;
}


bool CMarkup::x_CreateNode( MCD_STR& strNode, int nNodeType, MCD_PCSZ szText )
{
	// Set strNode based on nNodeType and szData
	// Return false if szData would jeopardize well-formed document
	//
	switch ( nNodeType )
	{
	case MNT_PROCESSING_INSTRUCTION:
		strNode = _T("<?");
		strNode += szText;
		strNode += _T("?>");
		break;
	case MNT_COMMENT:
		strNode = _T("<!--");
		strNode += szText;
		strNode += _T("-->");
		break;
	case MNT_ELEMENT:
		strNode = _T("<");
		strNode += szText;
		strNode += _T("/>");
		break;
	case MNT_TEXT:
	case MNT_WHITESPACE:
		strNode = EscapeText( szText );
		break;
	case MNT_DOCUMENT_TYPE:
		strNode = szText;
		break;
	case MNT_LONE_END_TAG:
		return false;
	case MNT_CDATA_SECTION:
		if ( MCD_PSZSTR(szText,_T("]]>")) != NULL )
			return false;
		strNode = _T("<![CDATA[");
		strNode += szText;
		strNode += _T("]]>");
		break;
	}
	return true;
}

MCD_STR CMarkup::x_EncodeCDATASection( MCD_PCSZ szData )
{
	// Split CDATA Sections if there are any end delimiters
	MCD_STR strData = _T("<![CDATA[");
	MCD_PCSZ pszNextStart = szData;
	MCD_PCSZ pszEnd = MCD_PSZSTR( szData, _T("]]>") );
	while ( pszEnd )
	{
		strData += MCD_STR( pszNextStart, (int)(pszEnd - pszNextStart) );
		strData += _T("]]]]><![CDATA[>");
		pszNextStart = pszEnd + 3;
		pszEnd = MCD_PSZSTR( pszNextStart, _T("]]>") );
	}
	strData += pszNextStart;
	strData += _T("]]>");
	return strData;
}

bool CMarkup::x_SetData( int iPos, int nValue )
{
	// Convert integer to string
	MCD_CHAR szVal[25];
	MCD_SPRINTF( szVal, _T("%d"), nValue );
	return x_SetData( iPos, szVal, 0 );
}

bool CMarkup::x_SetData( int iPos, MCD_PCSZ szData, int nFlags )
{
	// Set data at specified position
	// if nFlags==1, set content of element to a CDATA Section
	MCD_STR strInsert;

	if ( iPos == m_iPos && m_nNodeLength )
	{
		// Not an element
		if ( ! x_CreateNode(strInsert, m_nNodeType, szData) )
			return false;
		x_DocChange( m_nNodeOffset, m_nNodeLength, strInsert );
		x_AdjustForNode( m_iPosParent, iPos, MCD_STRLENGTH(strInsert) - m_nNodeLength );
		m_nNodeLength = MCD_STRLENGTH(strInsert);
		MARKUP_SETDEBUGSTATE;
		return true;
	}

	// Set data in iPos element
	if ( ! iPos || m_aPos[iPos].iElemChild )
		return false;

	// Build strInsert from szData based on nFlags
	if ( nFlags & MNF_WITHCDATA )
		strInsert = x_EncodeCDATASection( szData );
	else
		strInsert = EscapeText( szData, nFlags );

	// Insert
	NodePos node( MNF_WITHNOLINES|MNF_REPLACE );
	node.strMeta = strInsert;
	int iPosBefore = 0;
	int nReplace = x_InsertNew( iPos, iPosBefore, node );
	int nAdjust = MCD_STRLENGTH(node.strMeta) - nReplace;
	x_Adjust( iPos, nAdjust );
	m_aPos[iPos].nLength += nAdjust;
	if ( m_aPos[iPos].nFlags & MNF_ILLDATA )
		m_aPos[iPos].nFlags &= ~MNF_ILLDATA;
	MARKUP_SETDEBUGSTATE;
	return true;
}

MCD_STR CMarkup::x_GetData( int iPos ) const
{
	if ( iPos == m_iPos && m_nNodeLength )
	{
		if ( m_nNodeType == MNT_COMMENT )
			return MCD_STRMID( m_strDoc, m_nNodeOffset+4, m_nNodeLength-7 );
		else if ( m_nNodeType == MNT_PROCESSING_INSTRUCTION )
			return MCD_STRMID( m_strDoc, m_nNodeOffset+2, m_nNodeLength-4 );
		else if ( m_nNodeType == MNT_CDATA_SECTION )
			return MCD_STRMID( m_strDoc, m_nNodeOffset+9, m_nNodeLength-12 );
		else if ( m_nNodeType == MNT_TEXT )
			return UnescapeText( &(MCD_2PCSZ(m_strDoc))[m_nNodeOffset], m_nNodeLength );
		else if ( m_nNodeType == MNT_LONE_END_TAG )
			return MCD_STRMID( m_strDoc, m_nNodeOffset+2, m_nNodeLength-3 );
		else
			return MCD_STRMID( m_strDoc, m_nNodeOffset, m_nNodeLength );
	}

	// Return a string representing data between start and end tag
	// Return empty string if there are any children elements
	MCD_STR strData;
	if ( ! m_aPos[iPos].iElemChild && ! m_aPos[iPos].IsEmptyElement() )
	{
		// Quick scan for any tags inside content
		int nContentLen = m_aPos[iPos].ContentLen();
		int nStartContent = m_aPos[iPos].StartContent();
		MCD_PCSZ pszContent = &(MCD_2PCSZ(m_strDoc))[nStartContent];
		MCD_PCSZ pszTag = MCD_PSZCHR( pszContent, _T('<') );
		if ( pszTag && ((int)(pszTag-pszContent) < nContentLen) )
		{
			// Concatenate all CDATA Sections and text nodes, ignore other nodes
			TokenPos token( m_strDoc, m_nFlags );
			token.nNext = nStartContent;
			NodePos node;
			while ( token.nNext < nStartContent + nContentLen )
			{
				x_ParseNode( token, node );
				if ( node.nNodeType == MNT_TEXT )
					strData += UnescapeText( &token.szDoc[node.nStart], node.nLength );
				else if ( node.nNodeType == MNT_CDATA_SECTION )
					strData += MCD_STRMID( m_strDoc, node.nStart+9, node.nLength-12 );
			}
		}
		else // no tags
			strData = UnescapeText( &(MCD_2PCSZ(m_strDoc))[nStartContent], nContentLen );
	}
	return strData;
}

MCD_STR CMarkup::x_GetElemContent( int iPos ) const
{
	if ( iPos && m_aPos[iPos].ContentLen() )
		return MCD_STRMID( m_strDoc, m_aPos[iPos].StartContent(), m_aPos[iPos].ContentLen() );
	return _T("");
}

bool CMarkup::x_SetElemContent( MCD_PCSZ szContent )
{
	// Set data in iPos element only
	if ( ! m_iPos )
		return false;

	if ( m_nNodeLength )
		return false; // not an element

	// Unlink all children
	int iPos = m_iPos;
	int iPosChild = m_aPos[iPos].iElemChild;
	bool bHadChild = (iPosChild != 0);
	while ( iPosChild )
		iPosChild = x_ReleaseSubDoc( iPosChild );
	if ( bHadChild )
		x_CheckSavedPos();

	// Parse content
	bool bWellFormed = true;
	TokenPos token( szContent, m_nFlags );
	int iPosVirtual = x_GetFreePos();
	m_aPos[iPosVirtual].ClearVirtualParent();
	m_aPos[iPosVirtual].SetLevel( m_aPos[iPos].Level() + 1 );
	iPosChild = x_ParseElem( iPosVirtual, token );
	if ( m_aPos[iPosVirtual].nFlags & MNF_ILLFORMED )
		bWellFormed = false;
	m_aPos[iPos].nFlags = (m_aPos[iPos].nFlags & ~MNF_ILLDATA) | (m_aPos[iPosVirtual].nFlags & MNF_ILLDATA);

	// Prepare insert and adjust offsets
	NodePos node( MNF_WITHNOLINES|MNF_REPLACE );
	node.strMeta = szContent;
	int iPosBefore = 0;
	int nReplace = x_InsertNew( iPos, iPosBefore, node );
	
	// Adjust and link in the inserted elements
	x_Adjust( iPosChild, node.nStart );
	m_aPos[iPosChild].nStart += node.nStart;
	m_aPos[iPos].iElemChild = iPosChild;
	while ( iPosChild )
	{
		m_aPos[iPosChild].iElemParent = iPos;
		iPosChild = m_aPos[iPosChild].iElemNext;
	}
	x_ReleasePos( iPosVirtual );

	int nAdjust = MCD_STRLENGTH(node.strMeta) - nReplace;
	x_Adjust( iPos, nAdjust, true );
	m_aPos[iPos].nLength += nAdjust;

	x_SetPos( m_iPosParent, m_iPos, 0 );
	return bWellFormed;
}

void CMarkup::x_DocChange( int nLeft, int nReplace, const MCD_STR& strInsert )
{
	// Insert strInsert int m_strDoc at nLeft replacing nReplace chars
	// When creating a document, reduce reallocs by reserving string space
	// If realloc needed, allow for 1.5 times the new length
	//
	int nDocLength = MCD_STRLENGTH(m_strDoc);
	int nInsLength = MCD_STRLENGTH(strInsert);
	int nNewLength = nInsLength + nDocLength - nReplace;
	int nAllocLen = MCD_STRCAPACITY(m_strDoc);
#ifdef MCD_STRINSERTREPLACE // (STL)
	if ( nNewLength > nAllocLen )
		MCD_BLDRESERVE( m_strDoc, (nNewLength + nNewLength/2 + 128) );
	MCD_STRINSERTREPLACE( m_strDoc, nLeft, nReplace, strInsert );
#else // (MFC)
	int nBufferLen = nNewLength;
	if ( nNewLength > nAllocLen )
		nBufferLen += nBufferLen/2 + 128;
	MCD_CHAR* pDoc = MCD_GETBUFFER( m_strDoc, nBufferLen );
	if ( nLeft+nReplace < nDocLength )
		memmove( &pDoc[nLeft+nInsLength], &pDoc[nLeft+nReplace], (nDocLength-nLeft-nReplace)*sizeof(MCD_CHAR) );
	memcpy( &pDoc[nLeft], strInsert, nInsLength*sizeof(MCD_CHAR) );
	MCD_RELEASEBUFFER( m_strDoc, pDoc, nNewLength );
#endif

}

void CMarkup::x_Adjust( int iPos, int nShift, bool bAfterPos /*=false*/ )
{
	// Loop through affected elements and adjust indexes
	// Algorithm:
	// 1. update children unless bAfterPos
	//    (if no children or bAfterPos is true, length of iPos not affected)
	// 2. update starts of next siblings and their children
	// 3. go up until there is a next sibling of a pa