cheng.c

本程序是利用c编写的词法分析器

#include <stdio.h>
#include <string.h>
#include <process.h>

void main()
{
     FILE *in,*out;/*存放输入字符串和输出单词串的文件*/
     char arr[1024];/*arr数组存放单词符号*/
     int ch;
     char N;/*currentchar存放当前输入字符*/
     int i=0;/*arr数组的一个指针*/
      /*以读方式打开输入文件*/
     if((in=fopen("infile.txt","r"))==NULL)
	 {
        printf("can not open file\n");
        exit(0);
	 }
      /*以写方式打开输出文件*/
     if((out=fopen("outfile.txt","w"))==NULL)
	 {
        printf("cannout open outfile\n");
        exit(1);
	 }
      //去掉开头空格
     N=fgetc(in);
     while(N==' ')
	 { 
         N=fgetc(in);
	 }



     while(N!='$')
	 {
           /*常数部分判断*/
           if((N>='a'&&N<='z')||(N>='A'&&N<='Z')
              ||(N>='0'&&N<='9')||N=='('||N==')'||N=='{'||N=='}'||N=='>'||N=='<'||N==';'||N==':'||N=='+'||N=='-'
              ||N=='*'||N=='/'||N=='='||N=='!'||N=='&'||(ch=N)=='\n'||N==' '||N=='['||N==']')
		   {
               i=0;//清空arr字符数组
               if(N>='0'&&N<='9')
			   {
                   arr[i++]=N;
                   N=fgetc(in);//再读入下一个字符
                   while(N>='0'&&N<='9')
				   {
                           arr[i++]=N;
                           N=fgetc(in);
				   }
               if(N!='.')//如果当前读入的符号不为小数点，则常数读完毕，输出此常数
			   {
                   arr[i++]='\0';
                   fprintf(out,"%s%d,%c%s%c%s\n","(",15,'"',arr,'"',")");//格式化输出到目标文件，15代表种别码， arr为内码值，规定用数字本身代表
			   }
               else//当前字符为小数点则继续向下读入
			   {
                    arr[i++]=N;
                    N=fgetc(in);
                    while(N>='0'&&N<='9')
					{
                        arr[i++]=N;
                        N=fgetc(in);
					}
                    arr[i++]='\0';
                    fprintf(out,"%s%d,%c%s%c%s\n","(",15,'"',arr,'"',")");
			   }
               while(N==' ')
			   { 
                     N=fgetc(in);
			   }
		}
/*判断标识符和关键字部分*/
if((N>='a'&&N<='z')||(N>='A'&&N<='Z'))
{
i=0;//清空arr字符数组
    while((N>='a'&&N<='z')||(N>='0'&&N<='9')||(N>='A'&&N<='Z'))
 {
    arr[i++]=N;
    N=fgetc(in);
 }
//把字符数组arr和关键字表比较，判断单词串是关键字还是标识符
    arr[i++]='\0';
    if(strcmp(arr,"if")==0)
    fprintf(out,"%s%d,%c%s%c%s\n","(",0,'"',arr,'"',")");
    else if(strcmp(arr,"else")==0)
    fprintf(out,"%s%d,%c%s%c%s\n","(",1,'"',arr,'"',")");
    else if(strcmp(arr,"while")==0)
    fprintf(out,"%s%d,%c%s%c%s\n","(",2,'"',arr,'"',")");
    else if(strcmp(arr,"do")==0)
    fprintf(out,"%s%d,%c%s%c%s\n","(",3,'"',arr,'"',")");
 else
 fprintf(out,"%s%d,%c%s%c%s\n","(",14,'"',arr,'"',")");
}
while(N==' ')
{ 
 N=fgetc(in);
}
//判断标识符包括：+, -, *, /, (, ), {, }, <, >, ;, :, !=, >=, <=, ==, =.
if(N=='+')
{
    fprintf(out,"%s%d,%c%c%c%s\n","(",9,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='-')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",10,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='*')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",11,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='/')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",12,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='=')
{
 i=0;
 arr[i++]=N;
 N=fgetc(in);
 if(N=='=')
 {
  arr[i++]=N;
  arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",13,'"',arr,'"',")");
     N=fgetc(in);
 }
 else
 { arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",17,'"',arr,'"',")");
    }
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N==';')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",4,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='{')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",5,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='}')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",6,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='[')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",20,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N==']')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",21,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='(')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",7,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N==')')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",8,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N==':')
{
 i=0;
 arr[i++]=N;
 N=fgetc(in);
 if(N==':')
 {
  arr[i++]=N;
  arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",18,'"',arr,'"',")");
     N=fgetc(in);
 }
 else
 { arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",16,'"',arr,'"',")");
    }
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='>')
{
 i=0;
 arr[i++]=N;
 N=fgetc(in);
 if(N=='=')
 {
  arr[i++]=N;
  arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",13,'"',arr,'"',")");
     N=fgetc(in);
 }
 else
 { arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",13,'"',arr,'"',")");
    }
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='<')
{
 i=0;
 arr[i++]=N;
 N=fgetc(in);
 if(N=='=')
 {
  arr[i++]=N;
  arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",13,'"',arr,'"',")");
     N=fgetc(in);
 }
 else
 { arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",13,'"',arr,'"',")");
    }
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='!')
{
 i=0;
 arr[i++]=N;
 N=fgetc(in);
 if(N=='=')
 {
  arr[i++]=N;
  arr[i++]='\0';
  fprintf(out,"%s%d,%c%s%c%s\n","(",13,'"',arr,'"',")");
     N=fgetc(in);
 }
}
while(N==' ')
{ 
 N=fgetc(in);
}
if(N=='&')
{
 fprintf(out,"%s%d,%c%c%c%s\n","(",19,'"',N,'"',")");
    N=fgetc(in);
}
while(N==' ')
{ 
 N=fgetc(in);
}
while((ch=N)=='\n')
{
 N=fgetc(in);
}
 while(N==' ')
{
 N=fgetc(in);
}
}
else
{
 fprintf(out,"%s,%c%c%c\n","error",'"',N,'"');
    N=fgetc(in);
}

}//结束开头while语句
/*关闭文件退出*/
fclose(in);
fclose(out);
}