/LiveWriterBackupWPF/reg2xml/xmlReader.cs

using System;

using System.IO;

using System.Windows.Forms;

using System.Web;



public enum NODETYPE

{

	NODETYPE_NA = 0,

	NODETYPE_BEGINELEMENT = 1,

	NODETYPE_ENDELEMENT = 2,

	NODETYPE_ATTRIB = 3,

	NODETYPE_CONTENT = 4,

	NODETYPE_PI = 5,

	NODETYPE_COMMENT = 6,

	NODETYPE_CDATA = 7

};



namespace Cll

{

	/// <summary>

	/// xmlReader reads and parses an XML file

	/// </summary>

	public class xmlReader

	{

		public const string IDS_EMPTYELEMENTNAME = "Empty element name";

		public const string IDS_BADBEGINNODESYMBOL = "Bad '<' symbol";

		public const string IDS_NOBEGINNODESYMBOLINEOL = "'<' symbol not allowed at the end of a line";

		public const string IDS_GENERICSYNTAXERROR = "Syntax error";

		public const string IDS_MISSINGATTRIBNAME = "Missing attribute name before '='";

		public const string IDS_MISSINGEQUALSYMBOL = "Missing '=' after attribute name";

		public const string IDS_NOEQUALSYMBOLINEOL = "There should not be a '=' symbol at the end of a line";

		public const string IDS_BADATTRIBUTEVALUESYNTAX = "There should not be a quote char at the end of a line";



		StreamReader _sr; // general file members

		bool		_bFileOpen;

		string		_strFilename;

		bool		_bShowMsgBoxOnError; // true if message boxes are allowed to display while parsing

		string		_strLastError; // filled with error description if ParseContent() returns false



		//

		string		_strContent; // internal use : parser buffer

		int		_nCursor, _nbLines; // internal cursors (horizontal and vertical directions)

		NODETYPE	_nCurNodeType; // returns where is the parser on at the moment

		string		_strCurNodeName, _strCurNodeContent; // returns the current node value, and the current node content value

		string		_strCurAttribName, _strCurAttribValue; // returns the current attrib name/value pair

		string		_strCurPInstruction; // returns the current PInstruction (for instance ?xml, !DOCTYPE, ...)

		bool		_bCurInsideComment; // true if the parser is inside a comment ( <!-- ... -->)

		bool		_bCurInsideCDATA; // true if hte parser is inside a CDATA secrtion ( <![CDATA[[ ... ]]> )

		bool		_bCurInsideNode; // true if the parser is inside a node begin tag

		bool		_bCurInsideAttrib; // true if _strCurAttribName is valid and _strCurAttribValue is pending

		bool		_bCurInsideContent; // true if the parser is inside content



		public xmlReader()

		{

			init();

		}



		void init()

		{

			_bFileOpen = false;

			showMsgBoxOnError(false);

		}



		public void showMsgBoxOnError(bool bShow)

		{

			_bShowMsgBoxOnError = bShow;

		}



		public bool open(string strFilename)

		{

			if (_bFileOpen)

				return true;



			_strFilename = strFilename;



			return true;

		}



		public bool readString()

		{

			if (!_bFileOpen) // open file for reading

			{

				try 

				{

					_sr = File.OpenText(_strFilename);

					_bFileOpen = _sr != null;

					_nCursor = -1;

					_nbLines = 0;

					_bCurInsideNode = _bCurInsideComment = _bCurInsideAttrib = _bCurInsideContent = _bCurInsideCDATA = false;

					_strLastError = "";

					_strCurNodeName = "";

					_strCurNodeContent = "";

					_strCurAttribName = "";

					_strCurAttribValue = "";

					_strCurPInstruction = "";

				}

				catch (Exception ) 

				{

					_bFileOpen = false;

				}

			}



			if (!_bFileOpen)

				return false;



			bool bResult = true;



			if (_nCursor==-1)

			{

				_strContent = _sr.ReadLine();

				_nCursor = 0;

				_nbLines++;

			}



			if (_strContent==null)

			{

				bResult = false;

			}



			if (_strContent!=null && !parseContent())

			{

				string s;

				s = "Parse error in line "+_nbLines+" : " + _strLastError;

				_strLastError = s;



				if (_bShowMsgBoxOnError)

				{

					MessageBox.Show (s, "Office Recovery Manager", 

						MessageBoxButtons.OK, MessageBoxIcon.Exclamation);

				}



				bResult = false;

			}



			return bResult;

		}

	

		public bool close()

		{

			if (_bFileOpen)

				_sr.Close();

			

			init();



			return true;

		}



		public NODETYPE getNodeType()

		{

			return _nCurNodeType;

		}



		public string getNodeName()

		{

			return _strCurNodeName;

		}



		public string getAttribName()

		{

			return _strCurAttribName;

		}



		public string getAttribValue()

		{

			return _strCurAttribValue;

		}

		

		string getNodeContent()

		{

			return _strCurNodeContent;

		}





		string getPInstruction()

		{

			return _strCurPInstruction;

		}



		public int getCurrentLine()

		{

			return _nbLines;

		}





		public string getLastError() // if any

		{

			return _strLastError;

		}



		public bool parseContent()

		{

			_nCurNodeType = NODETYPE.NODETYPE_NA;



			int i = (int)_nCursor;

			int imax = _strContent.Length-1;



			if (i>imax)

			{

				_nCursor = -1; // force next string to be read from file

				return true;

			}



			string strTemp = _strContent + i;



			// pass spaces if we are inside a <...> and not yet processing an attribute value

			while ( (i<=imax) && 

				(_bCurInsideNode && !_bCurInsideAttrib) && 

				(_strContent[i]==' ' || _strContent[i]==0x0A || _strContent[i]==0x0D) )

					i++;



			if (i>imax)

			{

				_nCursor = -1; // force next string to be read from file

				return true;

			}



			// are we inside a comment ?

			if (_bCurInsideComment)

			{

				while ( (i<=imax-2) && 

				!(_strContent[i]=='-' && _strContent[i+1]=='-' && _strContent[i+2]=='>') )

				i++;



				if (i<=imax-2) // found an end-comment

				{

					_nCurNodeType = NODETYPE.NODETYPE_NA; // tell user we have nothing to provide him with

					_nCursor = i+2+1;



					// after '-->' we are automatically within a content

					_bCurInsideNode = _bCurInsideAttrib = _bCurInsideComment = _bCurInsideCDATA = false;

					_bCurInsideContent = true;

					_strCurNodeContent = "";



					return true;

				}

				else // we still are inside an comment

				{

					_nCurNodeType = NODETYPE.NODETYPE_COMMENT;

					_nCursor = imax+1; // force next string to be read

					return true;

				}

			}





			// are we inside a CDATA section ?

			if (_bCurInsideCDATA)

			{

				while ( (i<=imax-2) && 

				!(_strContent[i]==']' && _strContent[i+1]==']' && _strContent[i+2]=='>') )

				i++;



				if (i<=imax-2) // found an end-comment

				{

					_nCurNodeType = NODETYPE.NODETYPE_NA; // tell user we have nothing to provide him with

					_nCursor = i+2+1;



					// after ']]>' we are automatically within a content

					_bCurInsideNode = _bCurInsideAttrib = _bCurInsideComment = _bCurInsideCDATA = false;

					_bCurInsideContent = true;

					_strCurNodeContent = "";



					return true;

				}

				else // we still are inside an CDATA section

				{

					_nCurNodeType = NODETYPE.NODETYPE_CDATA;

					_nCursor = imax+1; // force next string to be read

					return true;

				}

			}





			if (_bCurInsideAttrib) // extracting the attrib value, possibly in multiple passes

			{

				if ( _strCurAttribValue.Length==0 )

				{

					// pass EOL

					while ( (i<=imax) && (_strContent[i]==' ' || _strContent[i]==0x0A || _strContent[i]==0x0D) )

					i++;



					if (i>imax)

					{

						_nCurNodeType = NODETYPE.NODETYPE_NA;

						_nCursor = i;

						return true;

					}



					char quotechar = _strContent[i++];



					_strCurAttribValue += quotechar; // start with something whatsoever!

					// in fact, we don't check the quotechar is an actual quotechar, ie " or '



					_nCurNodeType = NODETYPE.NODETYPE_NA;

					_nCursor = i;

					return true;

				}

				else

				{

					long ibegin = i;



					// pass until we find spaces or EOL or >

					while ( (i<=imax) 

                    && _strContent[i]!='\"'

					//&& _strContent[i]!='\''

					&& _strContent[i]!=0x0A

					&& _strContent[i]!=0x0D

					&& _strContent[i]!='>') 

					i++;



					// TODO : properly manage the case of a multiple-line attrib-value

					// (we should in this case return a N/A nodetype as long as we haven't

					// encountered the ending quotechar, while buffering all the chars in

					// the strAttribValue member).



					long iend = i;



					_strCurAttribValue += _strContent.Substring((int)ibegin, (int)(iend-ibegin));

                    _strCurAttribValue = HttpUtility.HtmlDecode(_strCurAttribValue); // Work Item #1781 http://www.codeplex.com/WLWBackup/WorkItem/View.aspx?WorkItemId=1781



					if (i>imax)

					{ // don't forget to add the EOL as well

						_strCurAttribValue += "\r\n";



						_nCurNodeType = NODETYPE.NODETYPE_NA;

						_nCursor = i;

						return true;

					}



					// and remove the prefixed quote char

					while ( _strCurAttribValue.Length!=0 && 

					(_strCurAttribValue[0]=='\"' || _strCurAttribValue[0]=='\'') )

					{

						_strCurAttribValue = _strCurAttribValue.Substring(1);

					}



					_nCurNodeType = NODETYPE.NODETYPE_ATTRIB;

					_bCurInsideAttrib = false;



					if ( _strContent[i]!='>' )

						i++; // pass ending quote char



					_nCursor = i;

					return true;

				}



			} // end if _bCurInsideAttrib==true





			if (_bCurInsideContent)

			{

				long ibegin = i;



				// pass until we find spaces or EOL or >

				while ( (i<=imax) && _strContent[i]!=0x0A

					&& _strContent[i]!=0x0D

					&& _strContent[i]!='<')

						i++;



				long iend = i;



				if ( (i<=imax) && _strContent[i]=='<')

					_bCurInsideContent = false;



				_strCurNodeContent = _strContent.Substring((int)ibegin, (int)(iend-ibegin));

				if (_strCurNodeContent.Length==0)

					_nCurNodeType = NODETYPE.NODETYPE_NA;	

				else

					_nCurNodeType = NODETYPE.NODETYPE_CONTENT;



				_nCursor = i;

				return true;

			} // end if (_bCurInsideContent)



			//

			char c = _strContent[i];



			// a node ?

			if (c=='<')

			{

				if (_bCurInsideNode) // error, we were already inside one

				{

					_strLastError = IDS_BADBEGINNODESYMBOL;

					return false;

				}



				_bCurInsideNode = true;

				_bCurInsideAttrib = _bCurInsideContent = _bCurInsideComment = _bCurInsideCDATA = false;



				i++;



				// pass spaces

				while ( _strContent[i]==' ' || _strContent[i]==0x0A || _strContent[i]==0x0D)

					i++;



				if (i>imax)

				{

					_strLastError = IDS_NOBEGINNODESYMBOLINEOL;

					return false;

				}



				// here we have either a node name, a PI, or a begin comment

				if (imax-i>=2) // is it a begin comment ? ( <!-- )

				{

					if ( _strContent[i+0]=='!' &&

					_strContent[i+1]=='-' &&

					_strContent[i+2]=='-')

					{

						_nCurNodeType = NODETYPE.NODETYPE_COMMENT;

						_bCurInsideComment = true;



						i+=3; // go to actual comment content



						_nCursor = i;

						return true;

					}

				}



				if (imax-i>=7) // is it a begin cdatasection ? ( <![CDATA[ )

				{

					if ( _strContent[i+0]=='!' &&

					_strContent[i+1]=='[' &&

					_strContent[i+2]=='C' &&

					_strContent[i+3]=='D' &&

					_strContent[i+4]=='A' &&

					_strContent[i+5]=='T' &&

					_strContent[i+6]=='A' &&

					_strContent[i+7]=='[')

					{

						_nCurNodeType = NODETYPE.NODETYPE_CDATA;

						_bCurInsideCDATA = true;



						i+=8; // go to actual cdata section content



						_nCursor = i;

						return true;

					}

				}





				// the node name begins at position i

				long ibegin = i;



				// pass until we find spaces or EOL or >

				while ( (i<=imax) && _strContent[i]!=' ' 

				&& _strContent[i]!=0x0A

				&& _strContent[i]!=0x0D

				&& (_strContent[i]!='/' || (i==ibegin)) // don't forget empty elements (for instance <br/>)

				&& _strContent[i]!='>')

				i++;



				long iend = i;

				

				_strCurNodeName = _strContent.Substring((int)ibegin, (int)(iend-ibegin));

				if (_strCurNodeName.Length==0)

				{

					_strLastError = IDS_EMPTYELEMENTNAME;

					return false;

				}





				if (_strCurNodeName[0]=='?' || _strCurNodeName[0]=='!')

				{

					_nCurNodeType = NODETYPE.NODETYPE_PI;

					_strCurPInstruction = _strCurNodeName;

					_strCurNodeName = ""; // erase the PI instruction so it does not appear as a node name

				}

				else if (_strCurNodeName[0]=='/')

				{

					_nCurNodeType = NODETYPE.NODETYPE_ENDELEMENT;

					_strCurNodeName = _strCurNodeName.Substring(1); // remove /

				}

				else

					_nCurNodeType = NODETYPE.NODETYPE_BEGINELEMENT;



				_nCursor = i;

				return true;

			}

			else // >, or ?, or content or attribute

			{

				if (c=='?')

				{

					_nCurNodeType = NODETYPE.NODETYPE_NA;

					_nCursor = i+1;

					return true;

				}

				else if (c=='/')

				{

					i++;



					// pass node name

					long ibegin = i;



					// pass until we find spaces or EOL or >

					while ( (i<=imax) && _strContent[i]!=' ' 

						&& _strContent[i]!=0x0A

						&& _strContent[i]!=0x0D

						&& _strContent[i]!='>')

							i++;



					long iend = i;

						

					_nCurNodeType = NODETYPE.NODETYPE_ENDELEMENT;



					_nCursor = i;

					return true;

				}

				else if (c=='>')

				{

					_bCurInsideNode = _bCurInsideAttrib = false;

					_bCurInsideContent = true;

					_strCurNodeContent = "";

					_nCurNodeType = NODETYPE.NODETYPE_NA;

					_nCursor = i+1;

					return true;

				}



				if (_bCurInsideNode) // attributes

				{



					if (!_bCurInsideAttrib)

					{

						if (c=='=')

						{

							_nCurNodeType = NODETYPE.NODETYPE_NA;

							_bCurInsideAttrib = true; // enable extraction of the associated attribute value



							i++; // pass '=' symbol



							_nCursor = i;

							return true;

						}



						// get attribute name

						long ibegin = i;



						// pass until we find spaces or EOL or >

						while ( (i<=imax) && _strContent[i]!=' '

							&& _strContent[i]!=0x0A

							&& _strContent[i]!=0x0D

							&& _strContent[i]!='='

							&& _strContent[i]!='>') // check against > is just for safety

								i++;



						long iend = i;



						_strCurAttribName = _strContent.Substring((int)ibegin, (int)(iend-ibegin));

						if (_strCurAttribName.Length==0)

						{

							_strLastError = IDS_MISSINGATTRIBNAME;

							return false;

						}



						_strCurAttribValue = ""; // make sure the attrib value is empty for the moment

						_nCurNodeType = NODETYPE.NODETYPE_NA;



						_nCursor = i;

						return true;

					}

				}

			}



			// this code never executes

			_strLastError = IDS_GENERICSYNTAXERROR;

			return false;

		}

	}

}