C++实现对URL的编码和解码（⽀持ansi和utf8格式）
#include<iostream>
#include<string>
#include<stdio.h>
#include<assert.h>
#include<fstream>
using namespace std;
unsigned char ToHex(unsigned char x)
{
return  x >9? x +55: x +48;
}
unsigned char FromHex(unsigned char x)
{
unsigned char y;
if(x >='A'&& x <='Z') y = x -'A'+10;
else if(x >='a'&& x <='z') y = x -'a'+10;
else if(x >='0'&& x <='9') y = x -'0';
else assert(0);
return y;
}
std::string UrlEncode(const std::string& str)
{
std::string strTemp ="";
size_t length = str.length();
for(size_t i =0; i < length; i++)
{
if(isalnum((unsigned char)str[i])||
(str[i]=='-')||
(str[i]=='_')||
(str[i]=='.')||
(str[i]=='~'))
strTemp += str[i];
else if(str[i]==' ')
strTemp +="+";
else
{
strTemp +='%';
strTemp +=ToHex((unsigned char)str[i]>>4);
strTemp +=ToHex((unsigned char)str[i]%16);
}
}
return strTemp;
}
std::string UrlDecode(const std::string& str)
{
std::string strTemp ="";
size_t length = str.length();
for(size_t i =0; i < length; i++)
{
if(str[i]=='+') strTemp +=' ';
else if(str[i]=='%')
{
assert(i +2< length);
unsigned char high =FromHex((unsigned char)str[++i]);
unsigned char low =FromHex((unsigned char)str[++i]);
strTemp += high *16+ low;
}
else strTemp += str[i];
}
return strTemp;
return strTemp;
}
int main()
{
//进⾏URL编码纯英⽂对于不带中⽂的字符转的码ansi和utf8是⼀样的
string str1 ="=;+/,";
string str1_url =UrlEncode(str1);
cout << str1_url << endl;
//对于带中⽂的ansi 编码和解码
string str_ansi ="还魂草";
string str_ansi_en =UrlEncode(str_ansi);
cout << str_ansi_en << endl;
//ansi的中⽂解码后也是ansi格式的中⽂
string str_ansi_de =UrlDecode(str_ansi_en);
cout << str_ansi_de << endl;
//对于带中⽂的utf8的编码和解码
string str_utf8 = u8"还魂草";
string str_utf8_en =UrlEncode(str_utf8);
//可以看到相同的中⽂ utf8和ansi是不⼀样的
cout << str_utf8_en << endl;
//utf8的中⽂解码后也是utf8的
string str_utf8_de =UrlDecode(str_utf8_en);
cout << str_utf8_de << endl;
ofstream wtxt("");
wtxt << str_utf8_de << endl;
wtxt.close();
getchar();
在线url网址编码解码return0;
}
以上程序执⾏的结果如下：
%3D%3B%2B%2F%2C
%BB%B9%BB%EA%B2%DD
还魂草
%E8%BF%98%E9%AD%82%E8%8D%89
杩橀瓊鑽
因为在C++控制台直接输出的为ansi格式字符串才不会乱码，因此最后⼀⾏的解密出的还魂草为utf8，直接输出为乱码，不过可以从中查看，打开，可以看到解密出的字符串并且编码格式为utf8

版权声明：本站内容均来自互联网，仅供演示用，请勿用于商业和其他非法用途。如果侵犯了您的权益请与我们联系QQ:729038198，我们将在24小时内删除。