CFGReader

SLang Scanner

A. First Edition
This is first edition of my SLang Scanner for Comp442 assignment 1.
B.The problem

In this assignment, you have to design a lexical analyzer for tokens of a programming language SLANG (our Source

LANGuage). In SLANG, the following token types exist: (l represents any letter, d any digit, and c any character).

identifer specified by `l( l + d + _(`l + d))*

numrical constant specifed by dd*

character constant specified by 'c'

//THE FOLLOWING ARE SYMBOL TYPE 18
"(", ")", ";", "+", "-", "*",
"/", ":=", "<", ">", "=", "<=",
">=", "!=", "[", "]", ",",
":",
//THE FOLLOWING ARE RESERVED TYPE 15
"begin", "end", "program", "variables","integer", "array", "char",
"module", "if", "then", "else", "loop", "exit", "read", "write"　

C.The idea of program
　
The previous version of my scanner is actually using a series of transition functions to simulate
　
a DFA. And the transition functions have a uniform shape which takes a char as input output another
　
state. They are placed in an array of function pointers for easy calling.
　
The teacher insisted efficiency strongly and I think I take the suggestion: Let's use table-driven.
　
The job is painful as you have to establish a huge table, say 138 rows x 72 columns as I even want
　
to include all keywords inside the table. So, you can imagine the size increase a lot as each keyword
　
need as many states as the number of its characters. And for this simple language, there are merely
　
15 keywords.
　
The efficiency comes from two factors:
　
1. I use a char table to represent all 256 ASCII chars with 72 different types. If I don't include
　
keywords in my DFA, I would save up to 52 types for 26 letters in both capital and small cases.
　
2. My tokenState 2-dimension array give you instantly the next state without function call. This
　
is the extremely fast compared with any transition function approach. But it is exhausting job to
　
initialize the 138x72 tables.
　
3. I use enum TokenType to represent each legal and all illegal token type in language. And a dirty
　
trick which I hate to use often is that I deliberately place all final states in beginning of
　
tokenState so that the first 38 final states have exact 1-1 sequence with TokenTypes. Therefore,
　
you don't have to verify the type of each token, when final state encounters. Simply use tokenState
　
as if it is tokenType, got it?
　
4. The idea of initialize the state table is to write down the most general cases or default ones.
　
Then overwrite with particular cases. This consists the major job of my scanner.
　
It cost me a whole sleepless night as I debug until 4:00 AM. Accidentally I pushed the repeat button
　
of my DVD player, so for a whole night the only song of Fayer Wang is played again and again---
　
<<The edge of hatred and love>>. 
D.The major functions
E.Further improvement
F.File listing
1. scanner.h
2. errorNo.h  
3. scanner.cpp 
4. slang.cpp (main)
　
　
file name: scanner.h
///////////////////////////////////////////////////////////////////////////////////////////
//Program: SLang Scanner
//Author: Qingzhe Huang
//Date: Jan. 18, 2004
//FileName: scanner.h
//Features:
//	1.	I want to improve efficiency of scanning, so I used table-driven method.
//	2.	I used enum to represent character of all ASCII---CharType---where "space, tab,
//		end of line, end of file are all considered to be White Space.
//	3.	All legal token is represented by enum TokenType.   
//	4.  I defined a huge amount of TokenState which is basically the state of a DFA. As
//		I don't want to search reserved keyword with linear search or whatever, I have 
//		many states for the reserved words.
//	5.  I deliberately make the sequence of first 38 TokenState elements exactly same as
//		all that of TokenType, so that each final state of DFA has a 1-1 correspondence with
//		type of token.
//	6.  I defined a struct of Token which may be used in future parser.
//	7.  I defined an errorNo variable to represent various errors. And a series error string
//		for displaying information.
//	8.  When class Scanner is created, it will initialize the big "state-charType" table.
//	9.  When readFromFile is called, it will first read one char in advance.
//	10. When an error is encountered, the caller of Scanner should understand that no further
//		char is read in. So, stop calling "nextToken()". This is a bit controvercial, and I
//		plan to change it in next version.
//////////////////////////////////////////////////////////////////////////////////////////// 

#ifndef SCANNER_H
#define SCANNER_H
#include <iostream>


using namespace std;

extern enum ErrorCode;

const int TokenStateCount=138;
const int CharTypeCount=72;
const int MaxTokenLength=255;


enum CharType
{
	//all small letters 26
	SMALLA,SMALLB,SMALLC,SMALLD,SMALLE,SMALLF,SMALLG,SMALLH,SMALLI,SMALLJ,SMALLK,SMALLL,
	SMALLM,SMALLN,SMALLO,SMALLP,SMALLQ,SMALLR,SMALLS,SMALLT,SMALLU,SMALLV,SMALLW,SMALLX,
	SMALLY,SMALLZ,
	//all big letters 26
	BIGA,BIGB,BIGC,BIGD,BIGE,BIGF,BIGG,BIGH,BIGI,BIGJ,BIGK,BIGL,BIGM,BIGN,BIGO,BIGP,BIGQ,
	BIGR,BIGS,BIGT,BIGU,BIGV,BIGW,BIGX,BIGY,BIGZ,
	//all digit 1
	DIGIT, 
	//all symbols 16
	QUOTE, OPENPAR, CLOSEPAR, SEMICOLON,PLUS, MINUS, TIMES, SLASH, COLON,
	EQUAL,SMALLER,GREATER,EXCLAIM,OPENBRACKET, CLOSEBRACKET,COMMA,
	//space, tab, end of line are regarded as whitespace, 1
	WHITESPACE,
	//UNDERSCORE IS A SPECIAL SYMBOL 1
	UNDERSCORE,
	//all other ASCII is regarded as illigal 1
	ILLIGAL
};



//TOTAL 38, JUST 1-1 WITH THE FIRST 38 OF TOKENSTATE
enum TokenType 
{
	//GENERAL TYPE 5
	IDTYPE, NUMBERTYPE, CHARCONSTTYPE, COMMENTTYPE, ERRORTYPE,
	//THE FOLLOWING ARE SYMBOL TYPE	18
	OPENPARTYPE, CLOSEPARTYPE, SEMICOLONTYPE, PLUSTYPE, MINUSTYPE, TIMESTYPE, 
    SLASHTYPE, ASSIGNMENTTYPE, SMALLERTYPE, GREATERTYPE, EQUALTYPE, SMALLEREQUALTYPE,
	GREATEREQUALTYPE, NOTEQUALTYPE, OPENBRACKETTYPE, CLOSEBRACKETTYPE, COMMATYPE, 
	COLONTYPE, 
	//THE FOLLOWING ARE RESERVED TYPE 15
	BEGINTYPE, ENDTYPE, PROGRAMTYPE, VARIABLESTYPE,INTEGERTYPE, ARRAYTYPE, CHARTYPE, 
	MODULETYPE, IFTYPE, THENTYPE, ELSETYPE, LOOPTYPE, EXITTYPE, READTYPE, WRITETYPE
};



enum TokenState
{
	//THE FINAL STATE 38, in order to easy initialize "finalState", I put them in beginning
	//5 generals
	IDEND, NUMBEREND, CONSTCHAREND, COMMENTEND, ERROR,
	//18 symbols
	OPENPAREND, CLOSEPAREND, SEMICOLONEND, PLUSEND, MINUSEND, TIMESEND, 
	SLASHEND, ASSIGNMENTEND, SMALLEREND, GREATEREND, EQUALEND, SMALLEREQUALEND, 
	GREATEREQUALEND, NOTEQUALEND, OPENBRACKETEND, CLOSEBRACKETEND, COMMAEND, 
	COLONEND, 
	//15 reserved
	BEGINEND, ENDEND, PROGRAMEND, VARIABLESEND, INTEGEREND, ARRAYEND, CHAREND, 
	MODULEEND, IFEND, THENEND, ELSEEND, LOOPEND, EXITEND, READEND, WRITEEND,
	//THE FOLLOWING ARE ALL NON-FINAL STATES
	//THE very FIRST CHAR 1
	READY, 
	//THE FOLLOWING ARE ALL RESERVED STATE
	//the first char 12
	ARRAY1, BEGIN1, CHAR1, E1, I1, LOOP1, MODULE1, PROGRAM1, READ1, THEN1, VARIABLES1,
	WRITE1,
	//THE SECOND CHAR 15
	ARRAY2, BEGIN2, CHAR2, ELSE2, END2, EXIT2, IF2, INTEGER2, LOOP2, MODULE2, PROGRAM2, 
	READ2, THEN2, VARIABLES2, WRITE2, 
	//THE THIRD CHAR 14
	ARRAY3, BEGIN3, CHAR3, ELSE3, END3, EXIT3, INTEGER3, LOOP3, MODULE3, PROGRAM3, READ3, 
	THEN3, VARIABLES3, WRITE3,
	//THE FOURTH CHAR 13
	ARRAY4, BEGIN4, CHAR4, ELSE4, EXIT4, INTEGER4, LOOP4, MODULE4, PROGRAM4, READ4, THEN4, 
	VARIABLES4, WRITE4,

	//THE FIFTH CHAR 7
	ARRAY5, BEGIN5, INTEGER5, MODULE5, PROGRAM5, VARIABLES5, WRITE5,
	//THE SIXTH CHAR 4
	INTEGER6, MODULE6, PROGRAM6, VARIABLES6, 
	//THE SEVENTH CHAR 3
	INTEGER7, PROGRAM7, VARIABLES7,
	//THE EIGHTH CHAR 1
	VARIABLES8, 
	//THE NINETH CHAR 1
	VARIABLES9,

	//THESE ARE NON-RESERVED
	//THESE ARE GENERAL 9
	IDBEGIN, IDUNDERSCORE, NUMBERBEGIN, CONSTCHARQUOTEBEGIN, CONSTCHARBEGIN, COMMENTSTARBEGIN,
	COMMENTBEGIN, COMMENTSTAREND, COMMENTSLASHBEGIN,  
	//the SINGLE symbols 16
	QUOTEBEGIN, OPENPARBEGIN, CLOSEPARBEGIN, SEMICOLONBEGIN, 
	PLUSBEGIN, MINUSBEGIN, TIMESBEGIN, SLASHBEGIN, COLONBEGIN, SMALLERBEGIN, GREATERBEGIN, 
	EQUALBEGIN, EXCLAIMBEGIN, OPENBRACKETBEGIN, CLOSEBRACKETBEGIN, COMMABEGIN, 
	//MULTI SYMBOL 4
	ASSIGNMENTBEGIN, SMALLEREQUALBEGIN,
	GREATEREQUALBEGIN, NOTEQUALBEGIN 
};



class Scanner
{
	struct Token
	{
		TokenType type;	
		char name[MaxTokenLength+1];
	};
private:		
	unsigned char ch;
	FILE* stream;
	bool nextChar();
	void initialize();
public:
	Scanner();
	ErrorCode errorNo;
	Token token;
	void errorHandle();
	bool readFromFile(const char* fileName);
	const char* getToken(){return token.name;}
	bool nextToken();
};
	
void initialTokenState();

#endif



　




file name: errorNo.h 


#ifndef ERRORNO_H
#define ERRORNO_H

const int ErrorCount=5;

const int TokenTypeCount=38;

enum ErrorCode
{IllegalToken, TokenTooLong, UnexpectedReachEOF, FileEmptyError, CannotOpenFile};



#endif






file name: scanner.cpp 


#include <iostream>
#include "scanner.h"
#include "errorNo.h"

using namespace std;

char* errorStr[ErrorCount]=
{"IllegalToken", "TokenTooLong", "UnexpectedReachEOF", "FileEmptyError", "CannotOpenFile"};

char* tokenTypeStr[TokenTypeCount]=
{
	//GENERAL TYPE 5
	"ID", "NUMBER", "CHARACTER CONSTANT", "COMMENT", "ERROR",
	//THE FOLLOWING ARE SYMBOL TYPE	18
	"(", ")", ";", "+", "-", "*", 
    "/", ":=", "<", ">", "=", "<=",
	">=", "!=", "[", "]", ",", 
	":", 
	//THE FOLLOWING ARE RESERVED TYPE 15
	"begin", "end", "program", "variables","integer", "array", "char", 
	"module", "if", "then", "else", "loop", "exit", "read", "write"
};

CharType charType[256];

TokenState tokenState[TokenStateCount][CharTypeCount];

bool Scanner::nextToken()
{
	TokenState state=READY;
	int count=0;//to count the length of token
	char* ptr=token.name; 
	bool isComment=false;
	do
	{
		//have to output anyway
		state=tokenState[state][charType[ch]];
		//because I put all final state in the first 37 positions
		if (state<38)
		{
			//This is a dirty trick! Because I make the "TokenType" 1-1 with
			//TokenState for the 38 finals.
			*ptr='\0';
			if (state==ERROR)
			{
				errorNo=IllegalToken;
			}
			token.type=(TokenType)(state);
			return true;
		}
		if (state==COMMENTBEGIN)
		{
			isComment=true;
		}
		if (count>=MaxTokenLength)
		{
			errorNo=TokenTooLong;
			token.type=ERRORTYPE;
			return true;
		}
		cout<<ch;
		if (!isComment&&state!=READY)
		{
			*ptr=ch;
			ptr++;
			count++;
		}
	}while (nextChar());
	state=tokenState[state][charType[ch]];
	//even READY IS CONSIDERED TO BE GOOD,
	if (state<38)
	{
		//This is a dirty trick! Because I make the "TokenType" 1-1 with
		//TokenState for the 38 finals.
		*ptr='\0';
		token.type=(TokenType)(state);		
	}
	else
	{
		if (state!=READY)
		{
			//this means the token is incomplete
			errorNo=UnexpectedReachEOF;
			token.type=ERRORTYPE;
		}
	}
	return false;
}

Scanner::Scanner()
{
	initialize();
}

void Scanner::errorHandle()
{
	cout<<"\n"<<errorStr[errorNo]<<endl;
}

void Scanner::initialize()
{
	initialTokenState();
}

bool Scanner::readFromFile(const char* fileName)
{
	if ((stream=fopen(fileName, "r"))==NULL)
	{
		errorNo=CannotOpenFile;
		return false;
	}
	else
	{		
		if (!nextChar())
		{
			errorNo=FileEmptyError;
			return false;
		}
	}
	return true;
}


bool Scanner::nextChar()
{
	ch=fgetc(stream);	
	return ch!=255;
}



　


file name: initialize.cpp 


#include "scanner.h"

extern enum CharType;
extern enum TokenState;
//extern const int TokenStateCount;
//extern const int CharTypeCount;


extern CharType charType[256];
extern TokenState tokenState[TokenStateCount][CharTypeCount];


void finalSymbolToken(TokenState state, TokenState endState);
void finalReservedToken(TokenState state, TokenState endState);
void initialCharType();
void setFinalTokenState();
void initialReserved(TokenState state);
void setRange(TokenState state, CharType start, CharType end, TokenState target);
void setState(TokenState state, TokenState targetState);
void setDefaultState();



void setDefaultState()
{
	for (int i=0; i<TokenStateCount; i++)
	{
		setState((TokenState)i, ERROR);
	}
	
	//the default for all letters are IDBEGIN
	setRange(READY, SMALLA, BIGZ, IDBEGIN);

	//THIS IS  another dirty trick, since I put all reserved states together
	//so you can initialize them together. 
	for (i=ARRAY1; i<=VARIABLES9; i++)
	{
		initialReserved((TokenState)i);
	}
	setFinalTokenState();
}

void setFinalTokenState()
{
	//FOR ID
	finalReservedToken(IDBEGIN, IDEND);
	//for number
	finalReservedToken(NUMBERBEGIN, NUMBEREND);
	//THESE FOR RESERVED WORDS

	finalReservedToken(ARRAY5, ARRAYEND);
	finalReservedToken(BEGIN5, BEGINEND);
	finalReservedToken(CHAR4, CHAREND);
	finalReservedToken(ELSE4, ELSEEND);
	finalReservedToken(END3, ENDEND);
	finalReservedToken(EXIT4, EXITEND);
	finalReservedToken(IF2, IFEND);
	finalReservedToken(INTEGER7, INTEGEREND);
	finalReservedToken(LOOP4, LOOPEND);
	finalReservedToken(MODULE6, MODULEEND);
	finalReservedToken(PROGRAM7, PROGRAMEND);
	finalReservedToken(READ4, READEND);
	finalReservedToken(THEN4, THENEND);
	finalReservedToken(VARIABLES9, VARIABLESEND);
	finalReservedToken(WRITE5, WRITEEND);

	//THESE FOR SYMBOLS


	finalSymbolToken(OPENPARBEGIN, OPENPAREND);
	finalSymbolToken(CLOSEPARBEGIN, CLOSEPAREND);
	finalSymbolToken(SEMICOLONBEGIN, SEMICOLONEND);
	finalSymbolToken(PLUSBEGIN, PLUSEND);
	finalSymbolToken(MINUSBEGIN, MINUSEND);
	finalSymbolToken(TIMESBEGIN, TIMESEND);
	finalSymbolToken(SLASHBEGIN, SLASHEND);
	finalSymbolToken(ASSIGNMENTBEGIN, ASSIGNMENTEND);
	finalSymbolToken(SMALLERBEGIN, SMALLEREND);
	finalSymbolToken(GREATERBEGIN, GREATEREND);
	finalSymbolToken(EQUALBEGIN, EQUALEND);
	finalSymbolToken(SMALLEREQUALBEGIN, SMALLEREQUALEND);
	finalSymbolToken(GREATEREQUALBEGIN, GREATEREQUALEND);
	finalSymbolToken(NOTEQUALBEGIN, NOTEQUALEND);
	finalSymbolToken(OPENBRACKETBEGIN, OPENBRACKETEND);
	finalSymbolToken(CLOSEBRACKETBEGIN, CLOSEBRACKETEND);
	finalSymbolToken(COMMABEGIN, COMMAEND);
	finalSymbolToken(COLONBEGIN, COLONEND);

	//COMMENT
	finalSymbolToken(COMMENTSLASHBEGIN, COMMENTEND);
	//CONSTCHAR
	finalSymbolToken(CONSTCHARQUOTEBEGIN, CONSTCHAREND);

}

void initialTokenState()
{
	//initialize all charType
	initialCharType();
	//default is always error
	setDefaultState();

	//loop
	tokenState[READY][WHITESPACE]=READY;
	//number
	tokenState[READY][DIGIT]=NUMBERBEGIN;
	tokenState[NUMBERBEGIN][DIGIT]=NUMBERBEGIN;//HOW LONG SHOULD NUMBER BE?

	//ID
	//setRange(READY, SMALLA, BIGZ, IDBEGIN); THIS IS IN DEFAULT
	setRange(IDBEGIN, SMALLA, DIGIT, IDBEGIN);
	tokenState[IDBEGIN][UNDERSCORE]=IDUNDERSCORE;
	setRange(IDUNDERSCORE, SMALLA, DIGIT, IDBEGIN);

	//reserved words
	//ARRAY1, BEGIN1, CHAR1, E1, I1, LOOP1, MODULE1, PROGRAM1, READ1, THEN1, WRITE1,
	//VARIABLES1,
	tokenState[READY][SMALLA]=ARRAY1;
	tokenState[READY][SMALLB]=BEGIN1;
	tokenState[READY][SMALLC]=CHAR1;
	tokenState[READY][SMALLE]=E1;
	tokenState[READY][SMALLI]=I1;
	tokenState[READY][SMALLL]=LOOP1;
	tokenState[READY][SMALLM]=MODULE1;
	tokenState[READY][SMALLP]=PROGRAM1;
	tokenState[READY][SMALLR]=READ1;
	tokenState[READY][SMALLT]=THEN1;
	tokenState[READY][SMALLV]=VARIABLES1;
	tokenState[READY][SMALLW]=WRITE1;

	/* RESERVED WORDS
	ARRAY2 */
	tokenState[ARRAY1][SMALLR]=ARRAY2;
	//BEGIN2
	tokenState[BEGIN1][SMALLE]=BEGIN2;
	//CHAR2
	tokenState[CHAR1][SMALLH]=CHAR2;
	//ELSE2,
	tokenState[E1][SMALLL]=ELSE2;
	//EXIT2
	tokenState[E1][SMALLX]=EXIT2;
	//END2
	tokenState[E1][SMALLN]=END2;
	//IF2
	tokenState[I1][SMALLF]=IF2;
	//INTEGER2
	tokenState[I1][SMALLN]=INTEGER2;
	//LOOP2
	tokenState[LOOP1][SMALLO]=LOOP2;
	//MODULE2
	tokenState[MODULE1][SMALLO]=MODULE2;
	//PROGRAM2
	tokenState[PROGRAM1][SMALLR]=PROGRAM2;
	//READ2
	tokenState[READ1][SMALLE]=READ2;
	//THEN2
	tokenState[THEN1][SMALLH]=THEN2;
	//VARIABLES2
	tokenState[VARIABLES1][SMALLA]=VARIABLES2;
	//WRITE2
	tokenState[WRITE1][SMALLR]=WRITE2;

	/* RESERVED WORDS
	ARRAY3 */
	tokenState[ARRAY2][SMALLR]=ARRAY3;
	//BEGIN2
	tokenState[BEGIN2][SMALLG]=BEGIN3;
	//CHAR2
	tokenState[CHAR2][SMALLA]=CHAR3;
	//ELSE2,
	tokenState[ELSE2][SMALLS]=ELSE3;
	//END2
	tokenState[END2][SMALLD]=END3;
	//EXIT2
	tokenState[EXIT2][SMALLI]=EXIT3;
	//INTEGER2
	tokenState[INTEGER2][SMALLT]=INTEGER3;
	//LOOP2
	tokenState[LOOP2][SMALLO]=LOOP3;
	//MODULE2
	tokenState[MODULE2][SMALLD]=MODULE3;
	//PROGRAM2
	tokenState[PROGRAM2][SMALLO]=PROGRAM3;
	//READ2
	tokenState[READ2][SMALLA]=READ3;
	//THEN2
	tokenState[THEN2][SMALLE]=THEN3;
	//VARIABLES2
	tokenState[VARIABLES2][SMALLR]=VARIABLES3;
	//WRITE2
	tokenState[WRITE2][SMALLI]=WRITE3;

	/* RESERVED WORDS
	ARRAY3 */
	tokenState[ARRAY3][SMALLA]=ARRAY4;
	//BEGIN2
	tokenState[BEGIN3][SMALLI]=BEGIN4;
	//CHAR2
	tokenState[CHAR3][SMALLR]=CHAR4;
	//ELSE2,
	tokenState[ELSE3][SMALLE]=ELSE4;
	//EXIT2
	tokenState[EXIT3][SMALLT]=EXIT4;
	//INTEGER2
	tokenState[INTEGER3][SMALLE]=INTEGER4;
	//LOOP2
	tokenState[LOOP3][SMALLP]=LOOP4;
	//MODULE2
	tokenState[MODULE3][SMALLU]=MODULE4;
	//PROGRAM2
	tokenState[PROGRAM3][SMALLG]=PROGRAM4;
	//READ2
	tokenState[READ3][SMALLD]=READ4;
	//THEN2
	tokenState[THEN3][SMALLN]=THEN4;
	//VARIABLES2
	tokenState[VARIABLES3][SMALLI]=VARIABLES4;
	//WRITE2
	tokenState[WRITE3][SMALLT]=WRITE4;

	/* RESERVED WORDS
	ARRAY */
	tokenState[ARRAY4][SMALLY]=ARRAY5;
	//BEGIN2
	tokenState[BEGIN4][SMALLN]=BEGIN5;
	//INTEGER2
	tokenState[INTEGER4][SMALLG]=INTEGER5;
	//MODULE2
	tokenState[MODULE4][SMALLL]=MODULE5;
	//PROGRAM2
	tokenState[PROGRAM4][SMALLR]=PROGRAM5;
	//VARIABLES2
	tokenState[VARIABLES4][SMALLA]=VARIABLES5;
	//WRITE2
	tokenState[WRITE4][SMALLE]=WRITE5;

	// RESERVED WORDS*/
	//INTEGER2
	tokenState[INTEGER5][SMALLE]=INTEGER6;
	//MODULE2
	tokenState[MODULE5][SMALLE]=MODULE6;
	//PROGRAM2
	tokenState[PROGRAM5][SMALLA]=PROGRAM6;
	//VARIABLES2
	tokenState[VARIABLES5][SMALLB]=VARIABLES6;

	// RESERVED WORDS*/
	//INTEGER2
	tokenState[INTEGER6][SMALLR]=INTEGER7;
	//PROGRAM2
	tokenState[PROGRAM6][SMALLM]=PROGRAM7;
	//VARIABLES2
	tokenState[VARIABLES6][SMALLL]=VARIABLES7;
	// RESERVED WORDS*/
	//VARIABLES2
	tokenState[VARIABLES7][SMALLE]=VARIABLES8;
	//VARIABLES2
	tokenState[VARIABLES8][SMALLS]=VARIABLES9;

	/*
	CONSTCHAR, UNDERSCOREBEGIN, ASSIGNMENTBEGIN, SMALLEREQUALBEGIN,
	GREATEREQUALBEGIN, NOTEQUAL, COMMENTBEGIN, IDUNDERSCORE,*/

	//now is the symbols
	//QUOTEBEGIN, OPENPARBEGIN, CLOSEPARBEGIN, SEMICOLONBEGIN, 
	//PLUSBEGIN, MINUSBEGIN, TIMESBEGIN, SLASHBEGIN, COLONBEGIN, SMALLERBEGIN, GREATERBEGIN, 
	//EQUALBEGIN, EXCLAIMBEGIN, OPENBRACKETBEGIN, CLOSEBRACKETBEGIN, COMMABEGIN,
	//'
	tokenState[READY][QUOTE]=QUOTEBEGIN;
	//(
	tokenState[READY][OPENPAR]=OPENPARBEGIN;
	//)
	tokenState[READY][CLOSEPAR]=CLOSEPARBEGIN;
	//;
	tokenState[READY][SEMICOLON]=SEMICOLONBEGIN;
	//+
	tokenState[READY][PLUS]=PLUSBEGIN;
	//-
	tokenState[READY][MINUS]=MINUSBEGIN;
	//*
	tokenState[READY][TIMES]=TIMESBEGIN;
	///
	tokenState[READY][SLASH]=SLASHBEGIN;
	//:
	tokenState[READY][COLON]=COLONBEGIN;
	//<
	tokenState[READY][SMALLER]=SMALLERBEGIN;
	//>
	tokenState[READY][GREATER]=GREATERBEGIN;
	//=
	tokenState[READY][EQUAL]=EQUALBEGIN;
	//!
	tokenState[READY][EXCLAIM]=EXCLAIMBEGIN;
	//[
	tokenState[READY][OPENBRACKET]=OPENBRACKETBEGIN;
	//]
	tokenState[READY][CLOSEBRACKET]=CLOSEBRACKETBEGIN;
	//,
	tokenState[READY][COMMA]=COMMABEGIN;

	//AFTER QUOTE IT CAN BE ANY CHARACTER, INCLUDING ILLIGAL CHAR
	setRange(QUOTEBEGIN, SMALLA, ILLIGAL, CONSTCHARBEGIN);
	//ANY OTHER STATE IS BY DEFAULT ERROR
	tokenState[CONSTCHARBEGIN][QUOTE]=CONSTCHARQUOTEBEGIN;
	//FOR /, DEFAULT IS SLASHEND, EXCEPT * WHICH IS COMMENTSTARBEGIN
	tokenState[SLASHBEGIN][TIMES]= COMMENTSTARBEGIN; 

	//FOR :, DEFAULT IS COLONEND, EXCEPT FOR = WHICH IS ASSIGNMENTBEGIN
	tokenState[COLONBEGIN][EQUAL]= ASSIGNMENTBEGIN; 

	//FOR <, DEFAULT IS SMALLEREND, EXCEPT FOR= WHICH IS SMALLEREQAULBEGIN
	tokenState[SMALLERBEGIN][EQUAL]=SMALLEREQUALBEGIN; 

	//FOR >, DEFAULT IS GREATEREND, EXCEPT FOR= WHICH IS GREATEREQAULBEGIN
	tokenState[GREATERBEGIN][EQUAL]= GREATEREQUALBEGIN; 

	tokenState[EXCLAIMBEGIN][EQUAL]= NOTEQUALBEGIN; 
	//WITHIN COMMENT IT IS A LOOP, EXCEPT FOR * WHICH IS POSSIBLE FOR END OF COMMENT
	setRange(COMMENTSTARBEGIN, SMALLA, ILLIGAL, COMMENTBEGIN);
	tokenState[COMMENTSTARBEGIN][TIMES]=COMMENTSTAREND;
	setRange(COMMENTBEGIN, SMALLA, ILLIGAL, COMMENTBEGIN);
	tokenState[COMMENTBEGIN][TIMES]=COMMENTSTAREND;
	//FROM COMMENTSTARBEGIN, ALL IS BACK TO COMMENTBEGIN, EXCEPT / WHICH IS END OF COMMENT
	setRange(COMMENTSTAREND, SMALLA, ILLIGAL, COMMENTBEGIN);
	tokenState[COMMENTSTAREND][SLASH]=COMMENTSLASHBEGIN;
	//
}


void initialReserved(TokenState state)
{
	setRange(state, SMALLA, DIGIT, IDBEGIN);
	finalReservedToken(state, IDEND);
	tokenState[state][UNDERSCORE]=IDUNDERSCORE;//a_
}

void finalSymbolToken(TokenState state, TokenState endState)
{
	for (int i=SMALLA; i<=WHITESPACE; i++)
	{
		tokenState[state][(CharType)i]=endState;
	}
}

void finalReservedToken(TokenState state, TokenState endState)
{
	//all non-letter, non-digit is regarded to be delimeter
	for (int i=QUOTE; i<=WHITESPACE; i++)
	{
		tokenState[state][(CharType)i]=endState;
	}
}


//the default charType is ILLIGAL
void initialCharType()
{
	int chType;
	//the default charType is ILLIGAL
	for (int i=0; i<256; i++)
	{
		charType[i]=ILLIGAL;
	}
	//chType is SMALLA
	chType=SMALLA;
	for (i='a'; i<='z'; i++)
	{
		charType[i]=(CharType)(chType);
		chType++;
	}
	//chType is now BIGA
	chType=BIGA;//I don't want to rely on the trick.
	for (i='A'; i<='Z'; i++)
	{
		charType[i]=(CharType)(chType);
		chType++;
	}
	chType=DIGIT;
	for (i='0'; i<='9'; i++)
	{
		charType[i]=(CharType)(chType);
	}
	/*
	UNDERSCORE, QUOTE, OPENPAR, CLOSEPAR, SEMICOLON,PLUS, MINUS, TIMES, SLASH, COLON,
	EQUAL,SMALLER,GREATER,EXCLAIM,OPENBRACKET, CLOSEBRACKET,COMMA,
	SPACE,TAB, ENDLINE, ILLIGAL
	*/
	charType['_']=UNDERSCORE;
	charType['\'']=QUOTE;
	charType['(']=OPENPAR;
	charType[')']=CLOSEPAR;
	charType[';']=SEMICOLON;
	charType['+']=PLUS;
	charType['-']=MINUS;
	charType['*']=TIMES;
	charType['/']=SLASH;
	charType[':']=COLON;
	charType['=']=EQUAL;
	charType['<']=SMALLER;
	charType['>']=GREATER;
	charType['!']=EXCLAIM;
	charType['[']=OPENBRACKET;
	charType[']']=CLOSEBRACKET;
	charType[',']=COMMA;
	charType[' ']=WHITESPACE;
	charType['\t']=WHITESPACE;
	charType[10]=WHITESPACE;
	charType[13]=WHITESPACE;
	charType[255]=WHITESPACE;//IT IS A KIND OF DELIMETER
}

void setRange(TokenState state, CharType start, CharType end, TokenState target)
{
	for (int i=start; i<=end; i++)
	{
		tokenState[state][i]=target;
	}
}

void setState(TokenState state, TokenState targetState)
{
	for (int i=0; i<CharTypeCount; i++)
	{
		tokenState[state][i]=targetState;
	}
}


　


file name: slang.cpp (main)



#include <iostream>
#include "scanner.h"
#include "errorno.h"

using namespace std;

extern char* tokenTypeStr[];

int main()
{
	Scanner S;
	if (!S.readFromFile("c:\\scannerSource.txt"))
	{
		S.errorHandle();
	}
	
	while (S.nextToken())
	{
		/*
		if (S.token.type!=COMMENTTYPE)
		{			
			cout<<"\nthe token is:"<<S.getToken()<<endl;
		}
		*/
		if (S.token.type==ERRORTYPE)
		{			
			break;
		}

		cout<<"\nthe type of token is:"<<tokenTypeStr[S.token.type]<<endl;
	}
	if (S.token.type==ERRORTYPE)
	{
		cout<<"\nthe token is:"<<S.getToken()<<endl;
		S.errorHandle();
	}

	return 0;
}



　








Here is the result: The input file is "c:\scannerSource.txt".  



Sorry I don't give you the text file, you can even use the source code of program itself, except you have



to remove some illegal symbol, like ", #, {, }. etc.


















                                                               
                      







                       back.gif (341 bytes)       
up.gif (335 bytes)                                       
next.gif (337 bytes)