Sicily 1000. 词法分析程序设计 **

Description

设一语言的关键词、运算符、分界符的个数与单词如下： 
struct { int number; string str[10]; } keywords={3,"int","main","return"} ; //关键词
struct { int number; string str[10]; } operators ={5,"+","*","=","+=","*="}; //运算符
struct { int number; string str[10]; } boundaries ={6,"(",")","{","}",",",";"} ; //分界符
struct { int number; string str[100];} identifieres={0}; //标识符
struct { int number; string str[100];} Unsigned_integer={0}; //无符号整数
以上类号分别为1~5，序号从0开始；
标识符是字母开头的字母数字串；常量为无符号整数；
用C++设计一程序实现词法分析。

Input

输入一程序，结束符用”#”；

Output

输出单词数对：<类号，序号>。 输出标识符表，用空格分隔； 输出无符号整数表，用空格分隔；

Sample Input

 Copy sample input to clipboard 

main()
{ int a=2,b=3;
  return 2*b+a;
}#

Sample Output

<1,1><3,0><3,1><3,2><1,0><4,0><2,2><5,0><3,4><4,1><2,2><5,1><3,5><1,2><5,0><2,1>
<4,1><2,0><4,0><3,5><3,3>
identifieres:a b
Unsigned_integer:2 3

解：

用cin>>str读入输入，会以换行和空格分隔，读入一个字符串后，先将其切分为若干个子字符串（切分规则：连续的数字和字母不切开，其他的字符分开），再分别将子字符串转化为单词对，标识符和无符号整数第一次出现时放入表中，代码：

#include<string>
#include<iostream>
#include<vector>
#include<sstream>
using namespace std;

struct { int number; string str[10]; } keywords={3,"int","main","return"} ; //关键词
struct { int number; string str[10]; } operators ={5,"+","*","=","+=","*="}; //运算符
struct { int number; string str[10]; } boundaries ={6,"(",")","{","}",",",";"} ; //分界符
struct { int number; string str[100];} identifieres={0}; //标识符
struct { int number; string str[100];} Unsigned_integer={0}; //无符号整数


bool isLetter(char temp) {
	if ('a' <= temp && 'z' >= temp) return true;
	if ('A' <= temp && 'Z' >= temp) return true;
	return false;
}

bool isDigit(char temp) {
	if (temp >= '0' && temp <= '9') return true;
	return false;
}

vector<pair<int, string> > split(string str) {
	vector<pair<int, string> > temp;
	int len = str.length();
	string letter = "";
	string digit = "";
	for (int i = 0; i < len; i++) {
		if (isLetter(str[i])) {
			letter += str[i];
			if (i+1 == len || !isLetter(str[i+1])) {
				temp.push_back(make_pair(1,letter));
				letter = "";
			}
		} else if (isDigit(str[i])) {
			digit += str[i];
			if (i+1 == len || !isDigit(str[i+1])) {
				temp.push_back(make_pair(2,digit));
				digit = "";
			}
		} else {
			string s;
			stringstream stream;
			stream << str[i];
			s = stream.str();
			temp.push_back(make_pair(3, s));
		}
	}
	return temp;
}

void dealLetters(string str) {
	bool temp = true;
	for (int i = 0; i < 3; i++)
    	if (str == keywords.str[i]) {
    		cout << "<1," << i << ">";
    		temp = false;
    		break;
    	}
	if (temp) {
	    bool temp1 = true;
	    for (int i = 0; i < identifieres.number; i++) {
	    	if (identifieres.str[i] == str) {
	    		cout << "<4," << i << ">";
	    		temp1 = false;
	    		break;
	    	}
	    }
	    if (temp1) {
	    	cout << "<4," << identifieres.number << ">";
	    	identifieres.str[identifieres.number++] = str;
	    }
	}
}

void dealDigits(string str) {
	bool temp = true;
    for (int i = 0; i < Unsigned_integer.number; i++) {
    	if (Unsigned_integer.str[i] == str) {
    		cout << "<5," << i << ">";
    		temp = false;
    		break;
    	}
    }
    if (temp) {
    	cout << "<5," << Unsigned_integer.number << ">";
    	Unsigned_integer.str[Unsigned_integer.number++] = str;
    }
}

bool dealOthers(string str) {
	if (str == "#") return true;
	for (int i = 0; i < 5; i++) {
		if (operators.str[i] == str) {
			cout << "<2," << i << ">";
			return false;
		}
	}
	for (int i = 0; i < 6; i++) {
		if (boundaries.str[i] == str) {
			cout << "<3," << i << ">";
			return false;
		}
	}
	return false;
}

int main() {
	string str;
	while(cin >> str) {
		bool end = false;
	    vector<pair<int, string> > temp = split(str);
	    for (int i = 0; i < temp.size(); i++) {
	    	if (temp[i].first == 1) dealLetters(temp[i].second);
	    	if (temp[i].first == 2) dealDigits(temp[i].second);
	    	else end = dealOthers(temp[i].second);
	    }
	    if (end == true) break;
	}
	cout << "\nidentifieres:";
	for (int i = 0; i< identifieres.number; i++) cout << identifieres.str[i] << " ";
	cout << "\nUnsigned_integer:";
	for (int i = 0; i< Unsigned_integer.number; i++) cout << Unsigned_integer.str[i] << " ";
}

测试：