markup.cpp

一款改进了的查询系统----北京公交车查询系统

// Markup.cpp: implementation of the CMarkup class.
//
// Markup Release 6.3
// Copyright (C) 1999-2002 First Objective Software, Inc. All rights reserved
// Go to www.firstobject.com for the latest CMarkup and EDOM documentation
// Use in commercial applications requires written permission
// This software is provided "as is", with no warranty.

#include "stdafx.h"
#include "afxconv.h"
#include "Markup.h"

#ifdef _DEBUG
#undef THIS_FILE
static char THIS_FILE[]=__FILE__;
#define new DEBUG_NEW
#endif

#ifdef _MBCS
#pragma message( "Note: MBCS build (not UTF-8)" )
// For UTF-8, remove _MBCS from project settings C/C++ preprocessor definitions
#endif

// Defines for Windows CE
#ifndef _tclen
#define _tclen(p) 1
#define _tccpy(p1,p2) *(p1)=*(p2)
#endif


void CMarkup::operator=( const CMarkup& markup )
{
	m_iPosParent = markup.m_iPosParent;
	m_iPos = markup.m_iPos;
	m_iPosChild = markup.m_iPosChild;
	m_iPosFree = markup.m_iPosFree;
	m_nNodeType = markup.m_nNodeType;
	m_aPos.RemoveAll();
	m_aPos.Append( markup.m_aPos );
	m_csDoc = markup.m_csDoc;
	MARKUP_SETDEBUGSTATE;
}

bool CMarkup::SetDoc( LPCTSTR szDoc )
{
	// Reset indexes
	m_iPosFree = 1;
	ResetPos();
	m_mapSavedPos.RemoveAll();

	// Set document text
	if ( szDoc )
		m_csDoc = szDoc;
	else
		m_csDoc.Empty();

	// Starting size of position array: 1 element per 64 bytes of document
	// Tight fit when parsing small doc, only 0 to 2 reallocs when parsing large doc
	// Start at 8 when creating new document
	int nStartSize = m_csDoc.GetLength() / 64 + 8;
	if ( m_aPos.GetSize() < nStartSize )
		m_aPos.SetSize( nStartSize );

	// Parse document
	bool bWellFormed = false;
	if ( m_csDoc.GetLength() )
	{
		m_aPos[0].Clear();
		int iPos = x_ParseElem( 0 );
		if ( iPos > 0 )
		{
			m_aPos[0].iElemChild = iPos;
			bWellFormed = true;
		}
	}

	// Clear indexes if parse failed or empty document
	if ( ! bWellFormed )
	{
		m_aPos[0].Clear();
		m_iPosFree = 1;
	}

	ResetPos();
	return bWellFormed;
};

bool CMarkup::IsWellFormed()
{
	if ( m_aPos.GetSize() && m_aPos[0].iElemChild )
		return true;
	return false;
}

bool CMarkup::Load( LPCTSTR szFileName )
{
	CString csDoc;
	CFile file;
	if ( ! file.Open(szFileName,CFile::modeRead) )
		return false;
	int nLength = file.GetLength();

#if defined(_UNICODE)
	// Allocate Buffer for UTF-8 file data
	unsigned char* pBuffer = new unsigned char[nLength + 1];
	nLength = file.Read( pBuffer, nLength );
	pBuffer[nLength] = '\0';

	// Convert file from UTF-8 to Windows UNICODE (AKA UCS-2)
	int nWideLength = MultiByteToWideChar(CP_UTF8,0,(const char*)pBuffer,nLength,NULL,0);
	nLength = MultiByteToWideChar(CP_UTF8,0,(const char*)pBuffer,nLength,
		csDoc.GetBuffer(nWideLength),nWideLength);
	ASSERT( nLength == nWideLength );
	delete [] pBuffer;
#else
	nLength = file.Read( csDoc.GetBuffer(nLength), nLength );
#endif
	csDoc.ReleaseBuffer(nLength);
	file.Close();

	return SetDoc( csDoc );
}

bool CMarkup::Save( LPCTSTR szFileName )
{
	int nLength = m_csDoc.GetLength();
	CFile file;
	if ( ! file.Open(szFileName,CFile::modeWrite|CFile::modeCreate) )
		return false;
#if defined( _UNICODE )
	int nUTF8Len = WideCharToMultiByte(CP_UTF8,0,m_csDoc,nLength,NULL,0,NULL,NULL);
	char* pBuffer = new char[nUTF8Len+1];
	nLength = WideCharToMultiByte(CP_UTF8,0,m_csDoc,nLength,pBuffer,nUTF8Len+1,NULL,NULL);
	file.Write( pBuffer, nLength );
	delete pBuffer;
#else
	file.Write( (LPCTSTR)m_csDoc, nLength );
#endif
	file.Close();
	return true;
}

bool CMarkup::FindElem( LPCTSTR szName )
{
	// Change current position only if found
	//
	if ( m_aPos.GetSize() )
	{
		int iPos = x_FindElem( m_iPosParent, m_iPos, szName );
		if ( iPos )
		{
			// Assign new position
			x_SetPos( m_aPos[iPos].iElemParent, iPos, 0 );
			return true;
		}
	}
	return false;
}

bool CMarkup::FindChildElem( LPCTSTR szName )
{
	// Change current child position only if found
	//
	// Shorthand: call this with no current main position
	// means find child under root element
	if ( ! m_iPos )
		FindElem();

	int iPosChild = x_FindElem( m_iPos, m_iPosChild, szName );
	if ( iPosChild )
	{
		// Assign new position
		int iPos = m_aPos[iPosChild].iElemParent;
		x_SetPos( m_aPos[iPos].iElemParent, iPos, iPosChild );
		return true;
	}

	return false;
}

CString CMarkup::GetTagName() const
{
	// Return the tag name at the current main position
	CString csTagName;

	if ( m_iPos )
		csTagName = x_GetTagName( m_iPos );
	return csTagName;
}

bool CMarkup::IntoElem()
{
	// If there is no child position and IntoElem is called it will succeed in release 6.3
	// (A subsequent call to FindElem will find the first element)
	// The following short-hand behavior was never part of EDOM and was misleading
	// It would find a child element if there was no current child element position and go into it
	// It is removed in release 6.3, this change is NOT backwards compatible!
	// if ( ! m_iPosChild )
	//	FindChildElem();

	if ( m_iPos && m_nNodeType == MNT_ELEMENT )
	{
		x_SetPos( m_iPos, m_iPosChild, 0 );
		return true;
	}
	return false;
}

bool CMarkup::OutOfElem()
{
	// Go to parent element
	if ( m_iPosParent )
	{
		x_SetPos( m_aPos[m_iPosParent].iElemParent, m_iPosParent, m_iPos );
		return true;
	}
	return false;
}

CString CMarkup::GetAttrName( int n ) const
{
	// Return nth Attrute name of main position
	if ( ! m_iPos || m_nNodeType != MNT_ELEMENT )
		return _T("");

	TokenPos token( m_csDoc );
	token.nNext = m_aPos[m_iPos].nStartL + 1;
	for ( int nAttr=0; nAttr<=n; ++nAttr )
		if ( ! x_FindAttr(token) )
			return _T("");

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

bool CMarkup::SavePos( LPCTSTR szPosName )
{
	// Save current element position in saved position map
	if ( szPosName )
	{
		SavedPos savedpos;
		savedpos.iPosParent = m_iPosParent;
		savedpos.iPos = m_iPos;
		savedpos.iPosChild = m_iPosChild;
		m_mapSavedPos.SetAt( szPosName, savedpos );
		return true;
	}
	return false;
}

bool CMarkup::RestorePos( LPCTSTR szPosName )
{
	// Restore element position if found in saved position map
	SavedPos savedpos;
	if ( szPosName && m_mapSavedPos.Lookup( szPosName, savedpos ) )
	{
		x_SetPos( savedpos.iPosParent, savedpos.iPos, savedpos.iPosChild );
		return true;
	}
	return false;
}

bool CMarkup::GetOffsets( int& nStart, int& nEnd ) const
{
	// Return document offsets of current main position element
	// This is not part of EDOM but is used by the Markup project
	if ( m_iPos )
	{
		nStart = m_aPos[m_iPos].nStartL;
		nEnd = m_aPos[m_iPos].nEndR;
		return true;
	}
	return false;
}

CString CMarkup::GetChildSubDoc() const
{
	if ( m_iPosChild )
	{
		int nL = m_aPos[m_iPosChild].nStartL;
		int nR = m_aPos[m_iPosChild].nEndR + 1;
		TokenPos token( m_csDoc );
		token.nNext = nR;
		if ( ! x_FindToken(token) || m_csDoc[token.nL] == _T('<') )
			nR = token.nL;
		return m_csDoc.Mid( nL, nR - nL );
	}
	return _T("");
}

bool CMarkup::RemoveElem()
{
	// Remove current main position element
	if ( m_iPos && m_nNodeType == MNT_ELEMENT )
	{
		int iPos = x_RemoveElem( m_iPos );
		x_SetPos( m_iPosParent, iPos, 0 );
		return true;
	}
	return false;
}

bool CMarkup::RemoveChildElem()
{
	// Remove current child position element
	if ( m_iPosChild )
	{
		int iPosChild = x_RemoveElem( m_iPosChild );
		x_SetPos( m_iPosParent, m_iPos, iPosChild );
		return true;
	}
	return false;
}

//////////////////////////////////////////////////////////////////////
// Private Methods
//////////////////////////////////////////////////////////////////////

int CMarkup::x_GetFreePos()
{
	//
	// This returns the index of the next unused ElemPos in the array
	//
	if ( m_iPosFree == m_aPos.GetSize() )
		m_aPos.SetSize( m_iPosFree + m_iPosFree / 2 );
	++m_iPosFree;
	return m_iPosFree - 1;
}

int CMarkup::x_ReleasePos()
{
	//
	// This decrements the index of the next unused ElemPos in the array
	// allowing the element index returned by GetFreePos() to be reused
	//
	--m_iPosFree;
	return 0;
}

int CMarkup::x_ParseError( LPCTSTR szError, LPCTSTR szName )
{
	if ( szName )
		m_csError.Format( szError, szName );
	else
		m_csError = szError;
	x_ReleasePos();
	return -1;
}

int CMarkup::x_ParseElem( int iPosParent )
{
	// This is either called by SetDoc, x_AddSubDoc, or itself recursively
	// m_aPos[iPosParent].nEndL is where to start parsing for the child element
	// This returns the new position if a tag is found, otherwise zero
	// In all cases we need to get a new ElemPos, but release it if unused
	//
	int iPos = x_GetFreePos();
	m_aPos[iPos].nStartL = m_aPos[iPosParent].nEndL;
	m_aPos[iPos].iElemParent = iPosParent;
	m_aPos[iPos].iElemChild = 0;
	m_aPos[iPos].iElemNext = 0;

	// Start Tag
	// A loop is used to ignore all remarks tags and special tags
	// i.e. <?xml version="1.0"?>, and <!-- comment here -->
	// So any tag beginning with ? or ! is ignored
	// Loop past ignored tags
	TokenPos token( m_csDoc );
	token.nNext = m_aPos[iPosParent].nEndL;
	CString csName;
	while ( csName.IsEmpty() )
	{
		// Look for left angle bracket of start tag
		m_aPos[iPos].nStartL = token.nNext;
		if ( ! x_FindChar( token.szDoc, m_aPos[iPos].nStartL, _T('<') ) )
			return x_ParseError( _T("Element tag not found") );

		// Set parent's End tag to start looking from here (or later)
		m_aPos[iPosParent].nEndL = m_aPos[iPos].nStartL;

		// Determine whether this is an element, or bypass other type of node
		token.nNext = m_aPos[iPos].nStartL + 1;
		if ( x_FindToken( token ) )
		{
			if ( token.bIsString )
				return x_ParseError( _T("Tag starts with quote") );
			_TCHAR cFirstChar = m_csDoc[token.nL];
			if ( cFirstChar == _T('?') || cFirstChar == _T('!') )
			{
				token.nNext = m_aPos[iPos].nStartL;
				if ( ! x_ParseNode(token) )
					return x_ParseError( _T("Invalid node") );
			}
			else if ( cFirstChar != _T('/') )
			{
				csName = x_GetToken( token );
				// Look for end of tag
				if ( ! x_FindChar(token.szDoc, token.nNext, _T('>')) )
					return x_ParseError( _T("End of tag not found") );
			}
			else
				return x_ReleasePos(); // probably end tag of parent
		}
		else
			return x_ParseError( _T("Abrupt end within tag") );
	}
	m_aPos[iPos].nStartR = token.nNext;

	// Is ending mark within start tag, i.e. empty element?
	if ( m_csDoc[m_aPos[iPos].nStartR-1] == _T('/') )
	{
		// Empty element
		// Close tag left is set to ending mark, and right to open tag right
		m_aPos[iPos].nEndL = m_aPos[iPos].nStartR-1;
		m_aPos[iPos].nEndR = m_aPos[iPos].nStartR;
	}
	else // look for end tag
	{
		// Element probably has contents
		// Determine where to start looking for left angle bracket of end tag
		// This is done by recursively parsing the contents of this element
		int iInner, iInnerPrev = 0;
		m_aPos[iPos].nEndL = m_aPos[iPos].nStartR + 1;
		while ( (iInner = x_ParseElem( iPos )) > 0 )
		{
			// Set links to iInner
			if ( iInnerPrev )
				m_aPos[iInnerPrev].iElemNext = iInner;
			else
				m_aPos[iPos].iElemChild = iInner;
			iInnerPrev = iInner;

			// Set offset to reflect child
			m_aPos[iPos].nEndL = m_aPos[iInner].nEndR + 1;
		}
		if ( iInner == -1 )
			return -1;

		// Look for left angle bracket of end tag
		if ( ! x_FindChar( token.szDoc, m_aPos[iPos].nEndL, _T('<') ) )
			return x_ParseError( _T("End tag of %s element not found"), csName );

		// Look through tokens of end tag
		token.nNext = m_aPos[iPos].nEndL + 1;
		int nTokenCount = 0;
		while ( x_FindToken( token ) )
		{
			++nTokenCount;
			if ( ! token.bIsString )
			{
				// Is first token not an end slash mark?
				if ( nTokenCount == 1 && m_csDoc[token.nL] != _T('/') )
					return x_ParseError( _T("Expecting end tag of element %s"), csName );

				else if ( nTokenCount == 2 && ! token.Match(csName) )
					return x_ParseError( _T("End tag does not correspond to %s"), csName );

				// Else is it a right angle bracket?
				else if ( m_csDoc[token.nL] == _T('>') )
					break;
			}
		}

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

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

bool CMarkup::x_FindChar( LPCTSTR szDoc, int& nChar, _TCHAR c )
{
	// static function
	LPCTSTR pChar = &szDoc[nChar];