scanner.cpp

一个能够分析三种整数、标识符、主要运算符和主要关键字的词法分析程序。 ㈠、基本要求： 标识符 <字母>(<字母>|<数字字符>)* 十进制整数 0

// Scanner.cpp : Defines the entry point for the console application.
//

#include "stdafx.h"

#include<stdio.h>
#include<string.h>
#include<conio.h>
#include<stdlib.h>
#include<ctype.h>
#define IDN    100  //标识符
#define INT8   101  //int8
#define INT10  102  //int10
#define INT16  103  //int16
#define REAL8  104  //real8
#define REAL10 105  //real10
#define REAL16 106  //real16

#define ADD    107  //  +
#define SUB    108  //  -
#define MUL    109  //  *
#define DIV    110  //  /
#define GT     111  //  〉
#define LT     112  // 〈
#define EQ     113  //  =
#define SLP    114  //  （
#define SRP    115  //  ）
#define COL    116  //  ；

#define IF     117  //if
#define THEN   118  //then
#define ELSE   119  //else
#define WHILE  120  //while
#define DO     121  //do

#define TRUE 1
#define FALSE 0
int tempt=0;
int length=0;

typedef struct kinds{
    int code;
	//int value;
    char *svalue;
};
kinds link[100];
char word[200];     //存放输入字符串,相当于缓冲区
int base=0;        //单词开始指针
int search=0;      //扫描指针

char* reserveWord[]={"if","then","else", "while", "do"}; //关键字
char* codeWord[]={"IDN","INT8","INT10","INT16","REAL8","REAL10","REAL16",
                   "ADD","SUB","MUL","DIV","GT","LT","EQ","SLP","SRP","COL",
                   "IF","THEN","ELSE","WHILE","DO"};//种类助记符

char getOneChar()  //读取一个字符
{
  char ch;
  ch=word[base];
  base++;
  return ch;
}
void retract()
{
  base--;
}
int isAlpha(char ch)//判断是否为一个字母
{
    if((ch>='A'&&ch<='Z')||(ch>='a'&&ch<='z'))     return TRUE;
	return FALSE;
}
int isDigit(char ch)//判断是否为一个数字
{
   if(ch>='0'&&ch<='9') return TRUE;
   return FALSE;
}
int isKeyword(char *token)//判断是否为一个关键字
{
   int i=0;
   int j=0;
   for(i=0;i<=4;i++)
   {
     j=strcmp(token,reserveWord[i]);//关键字标识符
     if(j==0) return i+IF;
   }
   return IDN;                       //普通标识符
}

kinds scan(void)  
{
   kinds kind;      //最终的token
   int i=0;         //当前字符的下标
   //int number;      //token的属性值

   char ch;         //当前的字符
   char *token;     //当前的字符串
   token=(char *)malloc(20);

   ch=getOneChar();
   //判断出否为是标识符
   if(isAlpha(ch))
   {              
       while(isalpha(ch)||isdigit(ch) )
	   {
          token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
	   }
	   if(ch=='_'||ch=='.')
	   {
	      token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
          while(isalpha(ch)||isdigit(ch) )
		  {
			  token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
		  }  
	   } 	     
       retract();
       kind.code=isKeyword(token); 	   kind.svalue=token;
       //if(kind.code==100)    printf("\nIDN %8s",token);//普通的标识符	   
       //else                  printf("\n%s    _",token);//关键字
	   return kind;
   }   
   //判断出是否为数字
   else if(isDigit(ch))
   {
       if(ch=='0') 
	   {
           token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
		   //判断16进制
           if(ch=='x'||ch=='X')
		   {
	           token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
	           if(ch>='0'&&ch<='9'||ch>='a'&&ch<='f')
			   {
	               while(ch>='0'&&ch<='9'||ch>='a'&&ch<='f') 
				   {   token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();}
	               if(ch=='.')
				   {   
					   token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
				       while(ch>='0'&&ch<='9'||ch>='a'&&ch<='f') 
					   {
	                       token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
					   }
                       retract();
					   kind.code=REAL16;    kind.svalue=token;
                       //printf("\nREAL16     %s",token);
					   return kind;
				   }
				   else
				   {
				       retract();
					   //sscanf(token,"%x",&number); //转化成值
				       //kind.value=number;
                       //printf("\nINT16     %d",number);
					   kind.code=INT16;    kind.svalue=token;
					   //printf("\nINT16     %s",token);	                   
	                   return kind;   //16进制
				   }
			   }
	           else   printf("0x:判断16进制时，发生错误");//报错：错误的16进制
			 
		   }
		   //判断8进制
           else if(ch>='0'&&ch<='7') 
		   {
	           while(ch>='0'&&ch<='7')
			   {
	               token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
			   }
			   if(ch=='.')
			   {
				   token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
                   while(ch>='0'&&ch<='7') 
				   {
	                   token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
				   }
                   retract();
				   kind.code=REAL8;	   kind.svalue=token;
                   //printf("\nREAL8     %s",token);
				   return kind;
			   }	
			   else
			   {
				   retract();
	               //sscanf(token,"%o",&number);//转化成值
                   //kind.value=number;
	               //printf("\nINT8      %d",number);
				   kind.code=INT8;   kind.svalue=token;
				   //printf("\nINT8     %s",token);
	               return kind;   //8进制
			   }

		   }
           else //以0开头的十进制
		   {
			   if(ch=='.')
			   {
				   token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
                   while(ch>='0'&&ch<='9') 
				   {  
					   token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
				   }
                   retract();
				   kind.code=REAL10;   kind.svalue=token;
                   //printf("\nREAL10     %s",token);
				   return kind;
			   }
			   else
			   {
			        retract();
			        //number=0;
			        //kind.value=number;
	                //printf("\nINT10    0");
					kind.code=INT10;kind.svalue=token;
					//printf("\nINT10     %s",token);
	                return kind;   //恰好是0
			   }
		   }
	   }
       //判断非0开头的10进制
       else 
	   {   
           while(ch>='0'&&ch<='9')  
		   {
	           token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
		   }
		   if(ch=='.')
		   {
			   token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
               while(ch>='0'&&ch<='9') 
			   {
	               token[i]=ch;token[i+1]='\0';i++;ch=getOneChar();
			   }
               retract();
			   kind.code=REAL10;   kind.svalue=token;
               //printf("\nREAL10     %s",token);
			   return kind;
		   }	
		   else
		   {
	           retract();
		       //sscanf(token,"%o",&number);//转化成值
               //kind.value=number;
	           //printf("\nINT10      %s",token);
               kind.code=INT10;   kind.svalue=token;
               //printf("\nINT10     %s",token);
	           return kind;   //10进制
		   }
	   } 
   }
   //判断运算符和分隔符
   else if(ch=='+'){
     kind.code=ADD;kind.svalue="+";
     //printf("\n+        _ ");	 
     return kind;
   }
   else if(ch=='-'){
     kind.code=SUB;kind.svalue="-";
     //printf("\n-        _ ");
     return kind;
   }
   else if(ch=='*'){
	 kind.code=MUL;kind.svalue="*";
     //printf("\n*        _ ");
     return kind;
   }
   else if(ch=='/'){
	 kind.code=DIV;kind.svalue="/";
     //printf("\n/        _ ");
     return kind;
   }   
   else if(ch=='>'){
	 kind.code=GT;kind.svalue=">";
     //printf("\n>        _ ");
     return kind;
   }
   else if(ch=='<'){
	 kind.code=LT;kind.svalue="<";
     //printf("\n<        _ ");
     return kind;
   }
   else if(ch=='='){
	 kind.code=EQ;kind.svalue="=";
     //printf("\n=        _ ");
     return kind;
   }
   else if(ch=='('){
	 kind.code=SRP;kind.svalue="(";
     //printf("\n(        _ ");
     return kind;
   }
   else if(ch==')'){
	 kind.code=SLP;kind.svalue=")";
     //printf("\n)        _ ");
     return kind;
   }
   else if(ch==';'){
	 kind.code=COL;kind.svalue=";";
	 //printf("\n;      _ ");
	 return kind;
   }
   else if(ch==' '){
	 kind.code=0;
     return kind;
   }
   else{ //报错
       printf("\nI don't know! ");
   }  
   //kind.value=0;
   //free(token);
   kind.code=0;
   return kind;
}

void main()
{
	int index=0;
   	kinds kind;
    //gets(word);
	FILE *fp;

	if((fp=fopen("xinjian.txt","r"))==NULL)
	{
		printf("打不开文件!\n");
	}
	fgets(word,200,fp);
	printf("%s\n",word);
	fclose(fp);
    
	while(base<strlen(word))
	{
        kind=scan();
		if(kind.code!=0)  link[index++]=kind;
	}
	length=index;base=0;
    for(int m=0;m<index;m++)
		printf("%-8s     %-8s\n",codeWord[link[m].code-100],link[m].svalue);
    printf("词法分析到此结束!\n");
    return;
}