markupstl.cpp

一个很好的vc底层代码

		}

		// Was a right angle bracket not found?
		if ( ! token.szDoc[token.nL] || nTokenCount < 2 )
			return x_ParseError( "End tag not completed for element %s", strName.c_str() );
		m_aPos[iPos].nEndR = token.nL;
	}

	// Successfully parsed element (and contained elements)
	return iPos;
}

bool CMarkupSTL::x_FindChar( const char* szDoc, int& nChar, char c )
{
	// static function
	const char* pChar = &szDoc[nChar];
	while ( *pChar && *pChar != c )
		pChar += 1;
	nChar = pChar - szDoc;
	if ( ! *pChar )
		return false;
	/*
	while ( szDoc[nChar] && szDoc[nChar] != c )
		nChar += _tclen( &szDoc[nChar] );
	if ( ! szDoc[nChar] )
		return false;
	*/
	return true;
}

bool CMarkupSTL::x_FindToken( CMarkupSTL::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
	const char* szDoc = token.szDoc;
	int nChar = token.nNext;
	token.bIsString = false;

	// By-pass leading whitespace
	while ( szDoc[nChar] && strchr(" \t\n\r",szDoc[nChar]) )
		++nChar;
	if ( ! 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 == '\"' || cFirstChar == '\'' )
	{
		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 += 1;

		// 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;
}

string CMarkupSTL::x_GetToken( const CMarkupSTL::TokenPos& token ) const
{
	// The token contains indexes into the document identifying a small substring
	// Build the substring from those indexes and return it
	if ( token.nL > token.nR )
		return "";
	return m_strDoc.substr( token.nL,
		token.nR - token.nL + ((token.nR<m_strDoc.size())? 1:0) );
}

int CMarkupSTL::x_FindElem( int iPosParent, int iPos, const char* szPath )
{
	// If szPath is NULL or empty, go to next sibling element
	// Otherwise go to next sibling element with matching path
	//
	if ( iPos )
		iPos = m_aPos[iPos].iElemNext;
	else
		iPos = m_aPos[iPosParent].iElemChild;

	// Finished here if szPath not specified
	if ( szPath == NULL || !szPath[0] )
		return iPos;

	// Search
	TokenPos token( m_strDoc.c_str() );
	while ( iPos )
	{
		// Compare tag name
		token.nNext = m_aPos[iPos].nStartL + 1;
		x_FindToken( token ); // Locate tag name
		if ( token.Match(szPath) )
			return iPos;
		iPos = m_aPos[iPos].iElemNext;
	}
	return 0;
}

int CMarkupSTL::x_ParseNode( CMarkupSTL::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
	// White space between elements is a text node
	int nTypeFound = 0;
	const 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];
		const char* szEndOfNode = NULL;
		if ( cFirstChar == '?' )
		{
			nTypeFound = MNT_PROCESSING_INSTRUCTION; // processing instruction
			szEndOfNode = "?>";
		}
		else if ( cFirstChar == '!' )
		{
			char cSecondChar = szDoc[token.nL+2];
			if ( cSecondChar == '[' )
			{
				nTypeFound = MNT_CDATA_SECTION;
				szEndOfNode = "]]>";
			}
			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 == '>' )
						{
							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 )
		{
			const 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;
		if ( x_FindToken(token) )
		{
			if ( szDoc[token.nL] == '<' )
				token.nNext = token.nL;
			else
			{
				nTypeFound = MNT_TEXT;
				x_FindChar( token.szDoc, token.nNext, '<' );
			}
		}
	}
	return nTypeFound;
}

string CMarkupSTL::x_GetTagName( int iPos ) const
{
	// Return the tag name at specified element
	TokenPos token( m_strDoc.c_str() );
	token.nNext = m_aPos[iPos].nStartL + 1;
	if ( ! iPos || ! x_FindToken( token ) )
		return "";

	// Return substring of document
	return x_GetToken( token );
}

bool CMarkupSTL::x_FindAttrib( CMarkupSTL::TokenPos& token, const char* szAttrib ) const
{
	// If szAttrib is NULL find next attrib, otherwise find named attrib
	// Return true if found
	int nAttrib = 0;
	for ( int nCount = 0; x_FindToken(token); ++nCount )
	{
		if ( ! token.bIsString )
		{
			// Is it the right angle bracket?
			if ( m_strDoc[token.nL] == '>' || m_strDoc[token.nL] == '/' )
				break; // attrib not found

			// Equal sign
			if ( m_strDoc[token.nL] == '=' )
				continue;

			// Potential attribute
			if ( ! nAttrib && nCount )
			{
				// Attribute name search?
				if ( ! szAttrib || ! szAttrib[0] )
					return true; // return with token at attrib name

				// Compare szAttrib
				if ( token.Match(szAttrib) )
					nAttrib = nCount;
			}
		}
		else if ( nAttrib && nCount == nAttrib + 2 )
		{
			return true;
		}
	}

	// Not found
	return false;
}

string CMarkupSTL::x_GetAttrib( int iPos, const char* szAttrib ) const
{
	// Return the value of the attrib at specified element
	if ( ! iPos || m_nNodeType != MNT_ELEMENT )
		return "";

	TokenPos token( m_strDoc.c_str() );
	token.nNext = m_aPos[iPos].nStartL + 1;
	if ( szAttrib && x_FindAttrib( token, szAttrib ) )
		return x_TextFromDoc( token.nL, token.nR - ((token.nR<m_strDoc.size())?0:1) );
	return "";
}

bool CMarkupSTL::x_SetAttrib( int iPos, const char* szAttrib, int nValue )
{
	// Convert integer to string and call SetChildAttrib
	char szVal[25];
	sprintf( szVal, "%d", nValue );
	return x_SetAttrib( iPos, szAttrib, szVal );
}

bool CMarkupSTL::x_SetAttrib( int iPos, const char* szAttrib, const char* szValue )
{
	// Set attribute in iPos element
	if ( ! iPos || m_nNodeType != MNT_ELEMENT )
		return false;

	TokenPos token( m_strDoc.c_str() );
	token.nNext = m_aPos[iPos].nStartL + 1;
	int nInsertAt, nReplace = 0;
	string strInsert;
	if ( x_FindAttrib( token, szAttrib ) )
	{
		// Decision: for empty value leaving attrib="" instead of removing attrib
		// Replace value only
		strInsert = x_TextToDoc( szValue, true );
		nInsertAt = token.nL;
		nReplace = token.nR-token.nL+1;
	}
	else
	{
		// Insert string name value pair
		string strFormat;
		strFormat = " ";
		strFormat += szAttrib;
		strFormat += "=\"";
		strFormat += x_TextToDoc( szValue, true );
		strFormat += "\"";
		strInsert = strFormat;

		// take into account whether it is an empty element
		nInsertAt = m_aPos[iPos].nStartR - (m_aPos[iPos].IsEmptyElement()?1:0);
	}

	x_DocChange( nInsertAt, nReplace, strInsert );
	int nAdjust = strInsert.size() - nReplace;
	m_aPos[iPos].nStartR += nAdjust;
	m_aPos[iPos].AdjustEnd( nAdjust );
	x_Adjust( iPos, nAdjust );
	MARKUP_SETDEBUGSTATE;
	return true;
}

bool CMarkupSTL::x_CreateNode( string& strNode, int nNodeType, const char* szText )
{
	// Set strNode based on nNodeType and szText
	// Return false if szText would jeopardize well-formed document
	//
	switch ( nNodeType )
	{
	case MNT_CDATA_SECTION:
		if ( strstr(szText,"]]>") != NULL )
			return false;
		strNode = "<![CDATA[";
		strNode += szText;
		strNode += "]]>";
		break;
	}
	return true;
}

bool CMarkupSTL::x_SetData( int iPos, const char* szData, int nCDATA )
{
	// Set data at specified position
	// if nCDATA==1, set content of element to a CDATA Section
	string strInsert;

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

	// Build strInsert from szData based on nCDATA
	// If CDATA section not valid, use parsed text (PCDATA) instead
	if ( nCDATA != 0 )
		if ( ! x_CreateNode(strInsert, MNT_CDATA_SECTION, szData) )
			nCDATA = 0;
	if ( nCDATA == 0 )
		strInsert = x_TextToDoc( szData );

	// Decide where to insert
	int nInsertAt, nReplace;
	if ( m_aPos[iPos].IsEmptyElement() )
	{
		nInsertAt = m_aPos[iPos].nEndL;
		nReplace = 1;

		// Pre-adjust since <NAME/> becomes <NAME>data</NAME>
		string strTagName = x_GetTagName( iPos );
		m_aPos[iPos].nStartR -= 1;
		m_aPos[iPos].nEndL -= (1 + strTagName.size());
		string strFormat;
		strFormat = ">";
		strFormat += strInsert;
		strFormat += "</";
		strFormat += strTagName;
		strInsert = strFormat;
	}
	else
	{
		nInsertAt = m_aPos[iPos].nStartR+1;
		nReplace = m_aPos[iPos].nEndL - m_aPos[iPos].nStartR - 1;
	}
	x_DocChange( nInsertAt, nReplace, strInsert );
	int nAdjust = strInsert.size() - nReplace;
	x_Adjust( iPos, nAdjust );
	m_aPos[iPos].AdjustEnd( nAdjust );
	MARKUP_SETDEBUGSTATE;
	return true;
}

string CMarkupSTL::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
		TokenPos token( m_strDoc.c_str() );
		token.nNext = m_aPos[iPos].nStartR+1;
		if ( x_FindToken( token ) && m_strDoc[token.nL] == '<'
				&& token.nL + 11 < m_aPos[iPos].nEndL
				&& strncmp( &token.szDoc[token.nL+1], "![CDATA[", 8 ) == 0 )
		{
			int nEndCDATA = m_strDoc.find( "]]>", token.nNext );
			if ( nEndCDATA != string::npos && nEndCDATA < m_aPos[iPos].nEndL )
			{
				return m_strDoc.substr( token.nL+9, nEndCDATA-token.nL-9 );
			}
		}
		return x_TextFromDoc( m_aPos[iPos].nStartR+1, m_aPos[iPos].nEndL-1 );
	}
	return "";
}

string CMarkupSTL::x_TextToDoc( const char* szText, bool bAttrib ) const
{
	//
	// &lt;   less than
	// &amp;  ampersand
	// &gt;   greater than
	//
	// and for attributes:
	//
	// &apos; apostrophe or single quote
	// &quot; double quote
	//
	static char* szaReplace[] = { "&lt;","&amp;","&gt;","&apos;","&quot;" };
	const char* pFind = bAttrib?"<&>\'\"":"<&>";
	const char* pSource = szText;
	string strResult;
	int nLen = strlen( szText );
	strResult.reserve( nLen + nLen / 10 );
	char cSource = *pSource;
	char* pFound;
	while ( cSource )
	{
		if ( (pFound=strchr(pFind,cSource)) != NULL )
		{
			pFound = szaReplace[pFound-pFind];