编译原理（词法分析程序）

 

#include <stdio.h>

#include <stdlib.h>

#include <string.h>

#include<windows.h>

void Scaner();//词法分析扫描函数

int IsRetainLetters(char *ch);//判断是否为保留字

char * DecimalToBinary(char *buf);//十进制转二进制函数

char GetCh(FILE *fp);//读取文件中的字符,每次只读取一个字符

int  main()

{

       printf(" 编译正在进行....../n");

       Scaner();

       printf("执行已经完成！！请参看工程目录下的out.txt文件/n");

       return 0;

}

void Scaner()//词法分析扫描函数

{

       FILE *fpin;//用于操作输入文件

       FILE *fpout;//用于操作输出文件

       char ch = ' ';//用于存放字符

       int i,k;

       int digit;//数字变量

       char temp[10];//用做中间变量存放字符

       fpin=fopen("in.txt","rb");//以只读的形式打开该文件

       fpout=fopen("out.txt","w+");//若文件不存在则系统创建该文件，否则重写打开的文件

       if(fpin==NULL || fpout==NULL)

       {

              printf("文件不能打开或新建,请重新调试!/n");

              exit (0);

       }

       while(1)

       {  

              i=0;

              int flag;

              while(i++<6) temp[i]='/0'; //初始化temp数组

              i=0;

              if (ch ==' ' ) ch=GetCh(fpin);

        if(ch==EOF)//文件结束

              {

                     fclose(fpin);

                     fclose(fpout);

                     return ;

              }

              //while(ch==' ') ch=GetCh(fpin);

              while(ch == ' ' || ch == 9 || ch == 10 || ch == 13 ) ch=GetCh(fpin);

              // 如果 ch为空格,回车,制表符table时略去！

              if(ch >= 'a' && ch <= 'z' || ch >= 'A' && ch <= 'Z')

              {

                     k=0;

                     while(ch >= 'a' && ch <= 'z' || ch >= 'A' && ch <= 'Z' || ch >='0' && ch <= '9')

                     {//读取的是字母

                            temp[k++]=ch;

                            ch=GetCh(fpin);

                     }

                     flag=IsRetainLetters(temp);

                     if(flag>0||flag<6)//判断是否是保留字

                     {

                            fprintf(fpout,"(%s,%d)/n",temp,flag);

                     }

                     else      //是标识符,用7表示

                            fprintf(fpout,"(%s,%i)/n",temp,7);//写入文件

                     while(i++<10) temp[i]='/0'; i=0;

              }

              else if(ch >= '0' && ch <= '9')

              {//是数字

                     k=0;

                     while(ch >= '0' && ch <= '9'){

                            temp[k++]=ch;

                            ch=GetCh(fpin);

                     }

                     char *binary;

                     digit=atoi(temp);//将字符串转换为数字

                     binary=DecimalToBinary(temp);//将数字转换为二进制

                     fprintf(fpout,"(%d,%s)/n",digit,binary);

                     i=0;

                     while(i++<10) temp[i]='/0';

                     i=0;

                     if(ch==EOF)

                     {

                            fclose(fpin);

                            fclose(fpout);

                            return;

                     }

              }

              else if(ch=='+'||ch=='-'||ch=='*'||ch=='/'||ch==':'||ch==':='||ch=='<'||ch=='<>'

                     ||ch=='<='||ch=='>'||ch=='>='||ch=='='||ch==';'||ch=='('||ch==')'||ch=='#')

              {//为其他字符

                     int result=0;

                     switch(ch)

                     {

                     case '+':  result=13;break;

                     case '-':  result=14;break;

                     case '*':  result=15;break;

                     case '/':  result=16;break;

                     case ':':  result=17;break;

                     case ':=': result=18;break;

                     case '<':  result=19;break;

                     case '<>': result=20;break;

                     case '<=': result=21;break;

                     case '>':  result=22;break;

                     case '>=': result=23;break;

                     case '=':  result=24;break;

                     case ';':  result=25;break;

                     case '(':  result=26;break;

                     case ')':  result=27;break;

                     case '#':  result=28;break;

                     default:

                            break;//不是定义的界符是为reuslt=0;

                     }

            fprintf(fpout,"(%c,%i)/n",ch,result);

                     ch=GetCh(fpin);  

                     continue;

              }

              else if(ch==EOF)

              {

                     while(!feof(fpout))

                            putchar(fgetc(fpout));//将输出的信息打印到屏幕上

                     fclose(fpin);

                     fclose(fpout);

                     return ;

              }//else continue;

else

              {

            MessageBox(NULL,"输入的整数或字母界符不在定义范围之内","Message",MB_RETRYCANCEL);

                     //The message box contains two push buttons: Retry and Cancel.

                     //If 第一个参数 is NULL, the message box has no owner window.

            ch=GetCh(fpin);//读取下一字符;

                     continue;

              }

             

       }//while(1)

       fclose(fpin);

       fclose(fpout);

}

//判断是否为保留字

int IsRetainLetters(char *ch)

{

       int flag=1;

       int i=0;

       char *keyword[6]={"begin","if","then","while","do","end"};

       while(i<6)

       {

              if(strcmp(ch,keyword[i])==0)

              {

                    

                     break;

              }

              flag++;

              i++;

       }

       return flag;

}

//读取文件中的字符

char GetCh(FILE *fp)

{

       char ch;

       if(!feof(fp))

              ch=fgetc(fp);//从fp所指向的文件的当前位置每次读取一个字符

       return ch;

}

 

//十进制转二进制函数

char * DecimalToBinary(char *buf)

{         

       int temp[20];

       char *binary;

       int value=0,i=0,j;

       for(i=0;buf[i]!='/0';i++)

              value=value*10+(buf[i]-48);        /*先将字符转化为十进制数*/

       if(value==0)

       {

              binary=(char *)malloc(2*sizeof(char));

              binary[0]='0';

              binary[1]='/0';

              return(binary);

       }

       i=0;

       while(value!=0)

       {

              temp[i++]=value%2;

              value/=2;

       }

       temp[i]='/0';

       binary=(char *)malloc((i+1)*sizeof(char));

       for(j=0;j<=i-1;j++)

        binary[j]=(char)(temp[i-j-1]+48);

       binary[i]='/0';

       return(binary);

}