scanner.c

完成Java的此法分析功能

#include "stdio.h"
#include "string.h"
#include "stdlib.h"
#include "ctype.h"

#define Half 512
#define Y 1
#define N 0
///////////////////////////////////////////////////////////

FILE *fp;    
char FileName[50];              
char Buffer1[Half];         // 半缓冲区一
char Buffer2[Half];         // 版缓冲区二
char WordBuffer[512];            // 单词缓冲区
FILE *fp_read;                  // 读文件指针
FILE *fp_write;                 // 写文件指针
int point_find;                 // 向前搜索指针
int point_begin;                // 单词开始指针
int word_row;                  // 每行单词数
int line;                       // 行数
int word_total;                   // 文件单词总数

////////////////////////////////////////////////////////////

FILE * OpenFile()                // 打开文件，fp是指向文件的指针变量
{
	if( ( fp = fopen( FileName,"r") )==NULL )
	{
		printf("Can not open the file!");
		exit(0);
	}
	else
	{   printf("Open file successfully!\n");
		return fp;
	}
}

CloseFile()                      // 关闭文件
{ 
	fclose(fp);
}

//////////////////////////////////////////////////////////////

PassLetter(FILE *fp) 
{                                               // 把文件读到半缓冲区一
	int i;
	fp_read = fp;
	fp_write = NULL;
	point_find = -1;
	point_begin = 0;
	word_row = 0;
	line = 1;
	word_total = 0;
	for(i = 0; i < Half; i++) 
	{                                           // 清空缓冲区
		Buffer1[i] = '\0';
        Buffer2[i] = '\0';
	}
	fread(Buffer1,sizeof(Buffer1),1,fp_read);   // 读数据块函数
}

/////////////////////////////////////////////////////////////////

NextChar() 
{                                       // 寻找一个字母             
	int i;
    if (point_find == Half - 1) 
	{
		for(i = 0; i < Half; i++) 
		{                              //清空缓冲区
			Buffer2[i] = '\0';
		}
		if(!feof(fp_read)) 
		{                              // 如果没有到达文件尾
			fread(Buffer2,sizeof(Buffer2),1,fp_read);
		}
		else return NULL;
		point_find++;
	}
	else if( point_find >= 2 * Half - 1 )
	{
		for(i = 0; i < Half; i++)
		{                              //清空缓冲区
			Buffer1[i] = '\0';
		}
		if(!feof(fp_read)) 
		{
			fread(Buffer1,sizeof(Buffer1),1,fp_read);
		}
		else return NULL;
		point_find = 0;
	}
	else
	{
		point_find++;
	}

    if(point_find < Half) 
	{
		return Buffer1[point_find];
	}
	else
	{
		return Buffer2[point_find - Half];
	}
}

///////////////////////////////////////////////////////////////////

BackSpace(char *char1 ) 
{                                  // 相会倒退一个字母
	if (point_find <= 0)
		point_find = 2 * Half -1;
	else point_find --;

	if(point_find < Half) 
	{
		*char1 = Buffer1[point_find];
	}
	else 
	{
		*char1 = Buffer2[point_find - Half];
	}
}

///////////////////////////////////////////////////////////////////////

SendWord() 
{                              // 将分析得到的单词存入单词缓冲区
    int i, j = 0, e;
	for (i = 0; i < 512; i++)  // 清空单词缓冲区
		WordBuffer[i] = '\0';
    if (point_begin > point_find)
		e = point_find + 2 * Half;       // e为单词的结束点
	else
		e = point_find;
	i = point_begin;
	while (i <= e) 
	{
		if (i < Half && Buffer1[i] != '\n') 
		{
			WordBuffer[j] = Buffer1[i];
			j++;
		}
		else if (i < 2 * Half && Buffer2[i-Half] != '\n') 
		{
			WordBuffer[j] = Buffer2[i - Half];
			j++;
		}
		else if (Buffer2[i-Half] != '\n') 
		{
			WordBuffer[j] = Buffer1[i - 2 * Half];
			j++;
		}
		i++;
	}
	point_begin = point_find + 1;
}

/////////////////////////////////////////////////////////////////////////////

MakeSure() 
{                             //检查是否为关键字
    char *KeyWord[] = {"abstract","boolean","break","byte","case","catch","char",
					    "class","const","continue","default","do","double","else","extends",
						"final","finally","float","for","goto","if","implements","import",
						"instanceof","int","interface","long","native","new","null","package",
						"private","protected","public","return","short","static","super","switch",
                        "synchronized","this","throw","throws","transient","try","void","volatile","while"};

	int i,value,flag = 0;
	SendWord();                                 // 将分析得到的单词存入单词缓冲区
	for (i = 0 ; i < 48; i++) 
	{
		if (strcmp(WordBuffer,KeyWord[i]) == 0) 
		{                                    // 是关键字 
			value = 0x103;
			flag = 1;
			break;
		}
	}
	if (((strcmp(WordBuffer,"true") == 0) || (strcmp(WordBuffer,"false") == 0)) && flag == 0)
		value =  0x105;                      // 是布尔型  
	else if (flag ==0) value =  0x104;       // 是标识符 
	WriteFile(value);
	word_row++;                             //每行单词数加一
}
////////////////////////////////////////////////////////////////////////////////////////////

WriteFile(int t) 
{                                       // 把属性值写入文件
    if (fp_write == NULL) 
	{
		if((fp_write = fopen("scanner_output.txt","w")) == NULL) 
		{
		    printf("cannot write to the file!\n");
		    exit(0);
		}
	}
	fprintf(fp_write,"< 0x%x , %s >\n",t,WordBuffer);
	printf("< 0x%x , %s >\n",t,WordBuffer);
}

////////////////////////////////////////////////////////////////////////////////////////////

AtLast(int value, char *char1, int option) 
{	                                        // 分析完每个单词后收尾工作
	int i;
	SendWord();
	if (option == 1) 
	{
	    WriteFile(value);                   // 写入文件同时显示在屏幕上
		word_row++;
	    for (i = 0; i < 512; i++)
		WordBuffer[i] = '\0';
	}
	*char1 = NextChar();
}

/////////////////////////////////////////////////////////////////////////////////////////////

WriteLine() 
{                          //每行结束时，输出行号，及该行字符数
    if (fp_write == NULL) 
	{
		if((fp_write = fopen("scanner_output.txt","w")) == NULL) 
		{
		    printf("cannot write to the file!\n");
		    exit(0);
		}
	}
	fprintf(fp_write,"  Line:%d, Sum of words in this line is:%d.\n",line,word_row); //把行号和每行字符数写文件
	fprintf(fp_write,"******************************************************\n");
	printf("  Line:%d, Sum of words in this line is:%d.\n",line,word_row);
	printf("*****************************************************\n");
	word_total += word_row;
	word_row = 0;
}

/////////////////////////////////////////////////////////////////////////////////////////////////

Analysis() {
	char char1;
	int state = 0, value, ct;
	char1 = NextChar();
	while (char1 != NULL)
	{
		switch(state) 
		{
	    case 0:
		    if (isalpha(char1) || char1=='$' || char1=='_') 
			{
			    state = 1;
		    	break;
			}  // 标志符，关键字
		    else if (char1 == '*') 
			{
			    state = 3;
				break;
			}  // *,*=
			else if (char1 == '>') 
			{
				state = 6;
				break;
			}  // >,>=,>>,>>=,>>>,>>>=
			else if (char1 == '%') 
			{
				state = 15;
				break;
			}  // %,%=
			else if (char1 == '<') 
			{
				state = 18;
				break;
			}  // <,<=,<<,<<=
			else if (char1 == '~') 
			{
				state = 24;
				break;
			}  // ~
			else if (char1 == '|') 
			{
				state = 25;
				break;
			}  // |,||,|=
			else if (char1 == '&') 
			{
				state = 29;
				break;
			}  // &,&&,&=
			else if (char1 == '^') 
			{
				state = 33;
				break;
			}  // ^,^=
			else if (char1 == '=') 
			{
				state = 36;
				break;
			}  // =,==
			else if (char1 == '!') 
			{
				state = 39;
				break;
			}  // !,!=
			else if (char1 == '?' || char1 == ':') 
			{
				state = 41;
				break;
			}  // ?:
			else if (char1 == '/') 
			{
				state = 43;
				break;
			}  // /,/=,//注释,/*注释
			else if (char1 == ',') 
			{
				state = 51;
				break;
			}  // ,
			else if (char1 == ';') 
			{
				state = 52;
				break;
			}  // ;
			else if (char1 == '{' || char1 == '}') 
			{
				state = 53;
				break;
			}  // {,}
			else if (char1 == '[' || char1 == ']' || char1 == '(' || char1 == ')') 
			{
				state = 54;
				break;
			}  // [,],(,)
			else if (char1 == '"') 
			{
				state = 55;
				break;
			}  // 字符串
			else if (char1 == '\'') 
			{
				state = 57;
				break;
			}  //字符
			else if (char1 == '+') 
			{
				state = 60;
				break;
			}  //+,++,+=
			else if (char1 == '-') 
			{
				state = 64;
				break;
			}  //-,--,-=
			else if (char1 == '0') 
			{
				state = 68;
				break;
			}  // 整数,浮点数
			else if (char1 >= '1' && char1 <= '9') 
			{
				state = 75;
				break;
			}  // 整数,浮点数
			else if (char1 == '.') 
			{
				state = 83;
				break;
			}  // .,整数,浮点数
			else if (char1 == '\n') 
			{
				if (word_row != 0)
				    WriteLine();
				line++;
				AtLast(0,&char1,0);
				break;
			}  // 换行符
			else if (char1 == ' ' || char1 == '\t') 
			{
				state = 0;
				AtLast(0,&char1,0);
				break;
			}  // 空格/\t
			else {
				state = 100;
				break;
			}  // 错误
		case 1:
			while (isalnum(char1) || char1 == '_' || char1 == '$')  // isalnum检查字符是否为英文字母或数字
				char1 = NextChar();
			state = 2;
			BackSpace(&char1);   //退会一个字符
			break;
		case 2:      // 0x103
			state = 0;
			MakeSure();
			char1 = NextChar();
			break;  // 关键字,标志符
		case 3:
			char1 = NextChar();
			if (char1 == '=')
				state = 5;
			else  
			{
				state = 4;
				BackSpace(&char1);
			}
			break;
        case 4: case 16: case 44:                  // 0x11b
			state = 0;
			value = 0x11b;
			AtLast(value,&char1,1);                 // 写文件
			break; // *,%,/
		case 5: case 11: case 14: case 17: case 23: case 28: case 32: case 35: case 37: case 45: case 63: case 67:   //0x110
			state = 0;
			value = 0x110;
			AtLast(value,&char1,1);
			break; // *=,>>=,>>>=,%=,<<=,|=,&=,^=,=
		case 6:
			char1 = NextChar();
			if (char1 == '=')
				state = 8;
			else if (char1 == '>')
				state = 9;
            else 
			{
				state = 7;
				BackSpace(&char1);
			}
			break;
		case 7: case 8: case 19:case 20:            //0x118
			state = 0;
			value = 0x118;
			AtLast(value,&char1,1);
			break; // >,>=,<,<=
		case 9:
			char1 = NextChar();
            if (char1 == '=')
				state = 11;
			else if (char1 == '>')
				state = 12;
			else 
			{
				state = 10;
				BackSpace(&char1);
			}
			break;
		case 10: case 13: case 22:                //0x119
			state = 0;
			value = 0x119;
			AtLast(value,&char1,1);
			break; // >>,>>>,<<
		case 12:
			char1 = NextChar();
			if (char1 == '=')
				state = 14;
			else 
			{
				state = 13;
				BackSpace(&char1);
			}
			break;
		case 15:
			char1 = NextChar();
			if (char1 == '=')
				state = 17;
			else 
			{
				state = 16;
				BackSpace(&char1);
			}
			break;
		case 18:
			char1 = NextChar();
			if (char1 == '=')
				state = 20;
			else if (char1 == '<')
				state = 21;
			else {