C/C++ 字元編碼的轉換(ut8、gb2312)

有三分代碼

strCoding.h

strCoding.cpp

test.cpp

*******************************************************************************************************************************

[cpp] view plaincopyprint?

1. //這是個類strCoding (strCoding.h檔案)   
2. #pragma once   
3. #include <iostream>   
4. #include <string>   
5. #include <windows.h>   
6. using namespace std;  
7. class strCoding  
8. {  
9. public:  
10.     strCoding(void);  
11.     ~strCoding(void);  
12.     void UTF_8ToGB2312(string &pOut, char *pText, int pLen);//utf_8轉為gb2312   
13.     void GB2312ToUTF_8(string& pOut,char *pText, int pLen); //gb2312 轉utf_8   
14.     string UrlGB2312(char * str);                           //urlgb2312編碼   
15.     string UrlUTF8(char * str);                             //urlutf8 編碼   
16.     string UrlUTF8Decode(string str);                  //urlutf8解碼   
17.     string UrlGB2312Decode(string str);                //urlgb2312解碼   
18. private:  
19.     void Gb2312ToUnicode(WCHAR* pOut,char *gbBuffer);  
20.     void UTF_8ToUnicode(WCHAR* pOut,char *pText);  
21.     void UnicodeToUTF_8(char* pOut,WCHAR* pText);  
22.     void UnicodeToGB2312(char* pOut,WCHAR uData);  
23.     char  CharToInt(char ch);  
24.     char StrToBin(char *str);  
25. };  

//這是個類strCoding (strCoding.h檔案)
#pragma once
#include <iostream>
#include <string>
#include <windows.h>
using namespace std;

class strCoding
{
public:
    strCoding(void);
    ~strCoding(void);

    void UTF_8ToGB2312(string &pOut, char *pText, int pLen);//utf_8轉為gb2312
    void GB2312ToUTF_8(string& pOut,char *pText, int pLen); //gb2312 轉utf_8
    string UrlGB2312(char * str);                           //urlgb2312編碼
    string UrlUTF8(char * str);                             //urlutf8 編碼
    string UrlUTF8Decode(string str);                  //urlutf8解碼
    string UrlGB2312Decode(string str);                //urlgb2312解碼

private:
    void Gb2312ToUnicode(WCHAR* pOut,char *gbBuffer);
    void UTF_8ToUnicode(WCHAR* pOut,char *pText);
    void UnicodeToUTF_8(char* pOut,WCHAR* pText);
    void UnicodeToGB2312(char* pOut,WCHAR uData);
    char  CharToInt(char ch);
    char StrToBin(char *str);

};      

1. //這是個類strCoding (strCoding.cpp檔案)   
2. #include "StdAfx.h"   
3. #include ".\strcoding.h"   
4. strCoding::strCoding(void)  
6. }  
7. strCoding::~strCoding(void)  
10. void strCoding::Gb2312ToUnicode(WCHAR* pOut,char *gbBuffer)  
12.     ::MultiByteToWideChar(CP_ACP,MB_PRECOMPOSED,gbBuffer,2,pOut,1);  
13.     return;  
15. void strCoding::UTF_8ToUnicode(WCHAR* pOut,char *pText)  
17.     char* uchar = (char *)pOut;  
18.     uchar[1] = ((pText[0] & 0x0F) << 4) + ((pText[1] >> 2) & 0x0F);  
19.     uchar[0] = ((pText[1] & 0x03) << 6) + (pText[2] & 0x3F);  
22. void strCoding::UnicodeToUTF_8(char* pOut,WCHAR* pText)  
24.     // 注意 WCHAR高低字的順序,低位元組在前，高位元組在後   
25.     char* pchar = (char *)pText;  
26.     pOut[0] = (0xE0 | ((pchar[1] & 0xF0) >> 4));  
27.     pOut[1] = (0x80 | ((pchar[1] & 0x0F) << 2)) + ((pchar[0] & 0xC0) >> 6);  
28.     pOut[2] = (0x80 | (pchar[0] & 0x3F));  
31. void strCoding::UnicodeToGB2312(char* pOut,WCHAR uData)  
33.     WideCharToMultiByte(CP_ACP,NULL,&uData,1,pOut,sizeof(WCHAR),NULL,NULL);  
36. //做為解Url使用   
37. char strCoding:: CharToInt(char ch){  
38.         if(ch>='0' && ch<='9')return (char)(ch-'0');  
39.         if(ch>='a' && ch<='f')return (char)(ch-'a'+10);  
40.         if(ch>='A' && ch<='F')return (char)(ch-'A'+10);  
41.         return -1;  
43. char strCoding::StrToBin(char *str){  
44.         char tempWord[2];  
45.         char chn;  
46.         tempWord[0] = CharToInt(str[0]);                         //make the B to 11 -- 00001011   
47.         tempWord[1] = CharToInt(str[1]);                         //make the 0 to 0  -- 00000000   
48.         chn = (tempWord[0] << 4) | tempWord[1];                //to change the BO to 10110000   
49.         return chn;  
51. //UTF_8 轉gb2312   
52. void strCoding::UTF_8ToGB2312(string &pOut, char *pText, int pLen)  
54.      char buf[4];  
55.      char* rst = new char[pLen + (pLen >> 2) + 2];  
56.     memset(buf,0,4);  
57.     memset(rst,0,pLen + (pLen >> 2) + 2);  
58.     int i =0;  
59.     int j = 0;  
60.     while(i < pLen)  
61.     {  
62.         if(*(pText + i) >= 0)  
63.         {  
64.             rst[j++] = pText[i++];  
65.         }  
66.         else                   
68.             WCHAR Wtemp;  
69.             UTF_8ToUnicode(&Wtemp,pText + i);  
70.             UnicodeToGB2312(buf,Wtemp);  
71.             unsigned short int tmp = 0;  
72.             tmp = rst[j] = buf[0];  
73.             tmp = rst[j+1] = buf[1];  
74.             tmp = rst[j+2] = buf[2];  
75.             //newBuf[j] = Ctemp[0];   
76.             //newBuf[j + 1] = Ctemp[1];   
77.             i += 3;      
78.             j += 2;     
80.   }  
81.     rst[j]='\0';  
82.    pOut = rst;   
83.     delete []rst;  
85. //GB2312 轉為 UTF-8   
86. void strCoding::GB2312ToUTF_8(string& pOut,char *pText, int pLen)  
88.     char buf[4];  
90.     pOut.clear();  
91.     int i = 0;  
94.         //如果是英文直接複制就可以   
95.         if( pText[i] >= 0)  
97.             char asciistr[2]={0};  
98.             asciistr[0] = (pText[i++]);  
99.             pOut.append(asciistr);  
101.         else  
103.             WCHAR pbuffer;  
104.             Gb2312ToUnicode(&pbuffer,pText+i);  
105.             UnicodeToUTF_8(buf,&pbuffer);  
106.             pOut.append(buf);  
107.             i += 2;  
109.     }  
112. //把str編碼為網頁中的 GB2312 url encode ,英文不變，漢字雙位元組  如%3D%AE%88   
113. string strCoding::UrlGB2312(char * str)  
115.     string dd;  
116.     size_t len = strlen(str);  
117.     for (size_t i=0;i<len;i++)  
119.         if(isalnum((BYTE)str[i]))  
121.             char tempbuff[2];  
122.             sprintf(tempbuff,"%c",str[i]);  
123.             dd.append(tempbuff);  
125.         else if (isspace((BYTE)str[i]))  
127.             dd.append("+");  
131.             char tempbuff[4];  
132.             sprintf(tempbuff,"%%%X%X",((BYTE*)str)[i] >>4,((BYTE*)str)[i] %16);  
136.     return dd;  
138. //把str編碼為網頁中的 UTF-8 url encode ,英文不變，漢字三位元組  如%3D%AE%88   
139. string strCoding::UrlUTF8(char * str)  
141.     string tt;  
143.     GB2312ToUTF_8(tt,str,(int)strlen(str));  
144.     size_t len=tt.length();  
147.         if(isalnum((BYTE)tt.at(i)))  
149.             char tempbuff[2]={0};  
150.             sprintf(tempbuff,"%c",(BYTE)tt.at(i));  
153.         else if (isspace((BYTE)tt.at(i)))  
160.             sprintf(tempbuff,"%%%X%X",((BYTE)tt.at(i)) >>4,((BYTE)tt.at(i)) %16);  
166. //把url GB2312解碼   
167. string strCoding::UrlGB2312Decode(string str)  
169.    string output="";  
170.         char tmp[2];  
171.         int i=0,idx=0,ndx,len=str.length();  
172.         while(i<len){  
173.                 if(str[i]=='%'){  
174.                         tmp[0]=str[i+1];  
175.                         tmp[1]=str[i+2];  
176.                         output += StrToBin(tmp);  
177.                         i=i+3;  
178.                 }  
179.                 else if(str[i]=='+'){  
180.                         output+=' ';  
181.                         i++;  
183.                 else{  
184.                         output+=str[i];  
188.         return output;  
190. //把url utf8解碼   
191. string strCoding::UrlUTF8Decode(string str)  
193.      string output="";  
194.     string temp =UrlGB2312Decode(str);//   
195.     UTF_8ToGB2312(output,(char *)temp.data(),strlen(temp.data()));  
196.     return output;  

//這是個類strCoding (strCoding.cpp檔案)
#include "StdAfx.h"

#include ".\strcoding.h"



strCoding::strCoding(void)
{
}

strCoding::~strCoding(void)
{
}
void strCoding::Gb2312ToUnicode(WCHAR* pOut,char *gbBuffer)
{
    ::MultiByteToWideChar(CP_ACP,MB_PRECOMPOSED,gbBuffer,2,pOut,1);
    return;
}
void strCoding::UTF_8ToUnicode(WCHAR* pOut,char *pText)
{
    char* uchar = (char *)pOut;

    uchar[1] = ((pText[0] & 0x0F) << 4) + ((pText[1] >> 2) & 0x0F);
    uchar[0] = ((pText[1] & 0x03) << 6) + (pText[2] & 0x3F);

    return;
}

void strCoding::UnicodeToUTF_8(char* pOut,WCHAR* pText)
{
    // 注意 WCHAR高低字的順序,低位元組在前，高位元組在後
    char* pchar = (char *)pText;

    pOut[0] = (0xE0 | ((pchar[1] & 0xF0) >> 4));
    pOut[1] = (0x80 | ((pchar[1] & 0x0F) << 2)) + ((pchar[0] & 0xC0) >> 6);
    pOut[2] = (0x80 | (pchar[0] & 0x3F));

    return;
}
void strCoding::UnicodeToGB2312(char* pOut,WCHAR uData)
{
    WideCharToMultiByte(CP_ACP,NULL,&uData,1,pOut,sizeof(WCHAR),NULL,NULL);
    return;
}

//做為解Url使用
char strCoding:: CharToInt(char ch){
        if(ch>='0' && ch<='9')return (char)(ch-'0');
        if(ch>='a' && ch<='f')return (char)(ch-'a'+10);
        if(ch>='A' && ch<='F')return (char)(ch-'A'+10);
        return -1;
}
char strCoding::StrToBin(char *str){
        char tempWord[2];
        char chn;

        tempWord[0] = CharToInt(str[0]);                         //make the B to 11 -- 00001011
        tempWord[1] = CharToInt(str[1]);                         //make the 0 to 0  -- 00000000

        chn = (tempWord[0] << 4) | tempWord[1];                //to change the BO to 10110000

        return chn;
}


//UTF_8 轉gb2312
void strCoding::UTF_8ToGB2312(string &pOut, char *pText, int pLen)
{
     char buf[4];
     char* rst = new char[pLen + (pLen >> 2) + 2];
    memset(buf,0,4);
    memset(rst,0,pLen + (pLen >> 2) + 2);

    int i =0;
    int j = 0;

    while(i < pLen)
    {
        if(*(pText + i) >= 0)
        {

            rst[j++] = pText[i++];
        }
        else                 
        {
            WCHAR Wtemp;


            UTF_8ToUnicode(&Wtemp,pText + i);

            UnicodeToGB2312(buf,Wtemp);

            unsigned short int tmp = 0;
            tmp = rst[j] = buf[0];
            tmp = rst[j+1] = buf[1];
            tmp = rst[j+2] = buf[2];

            //newBuf[j] = Ctemp[0];
            //newBuf[j + 1] = Ctemp[1];

            i += 3;    
            j += 2;   
        }

  }
    rst[j]='\0';
   pOut = rst; 
    delete []rst;
}

//GB2312 轉為 UTF-8
void strCoding::GB2312ToUTF_8(string& pOut,char *pText, int pLen)
{
    char buf[4];
    memset(buf,0,4);

    pOut.clear();

    int i = 0;
    while(i < pLen)
    {
        //如果是英文直接複制就可以
        if( pText[i] >= 0)
        {
            char asciistr[2]={0};
            asciistr[0] = (pText[i++]);
            pOut.append(asciistr);
        }
        else
        {
            WCHAR pbuffer;
            Gb2312ToUnicode(&pbuffer,pText+i);

            UnicodeToUTF_8(buf,&pbuffer);

            pOut.append(buf);

            i += 2;
        }
    }

    return;
}
//把str編碼為網頁中的 GB2312 url encode ,英文不變，漢字雙位元組  如%3D%AE%88
string strCoding::UrlGB2312(char * str)
{
    string dd;
    size_t len = strlen(str);
    for (size_t i=0;i<len;i++)
    {
        if(isalnum((BYTE)str[i]))
        {
            char tempbuff[2];
            sprintf(tempbuff,"%c",str[i]);
            dd.append(tempbuff);
        }
        else if (isspace((BYTE)str[i]))
        {
            dd.append("+");
        }
        else
        {
            char tempbuff[4];
            sprintf(tempbuff,"%%%X%X",((BYTE*)str)[i] >>4,((BYTE*)str)[i] %16);
            dd.append(tempbuff);
        }

    }
    return dd;
}

//把str編碼為網頁中的 UTF-8 url encode ,英文不變，漢字三位元組  如%3D%AE%88

string strCoding::UrlUTF8(char * str)
{
    string tt;
    string dd;
    GB2312ToUTF_8(tt,str,(int)strlen(str));

    size_t len=tt.length();
    for (size_t i=0;i<len;i++)
    {
        if(isalnum((BYTE)tt.at(i)))
        {
            char tempbuff[2]={0};
            sprintf(tempbuff,"%c",(BYTE)tt.at(i));
            dd.append(tempbuff);
        }
        else if (isspace((BYTE)tt.at(i)))
        {
            dd.append("+");
        }
        else
        {
            char tempbuff[4];
            sprintf(tempbuff,"%%%X%X",((BYTE)tt.at(i)) >>4,((BYTE)tt.at(i)) %16);
            dd.append(tempbuff);
        }

    }
    return dd;
}
//把url GB2312解碼
string strCoding::UrlGB2312Decode(string str)
{
   string output="";
        char tmp[2];
        int i=0,idx=0,ndx,len=str.length();

        while(i<len){
                if(str[i]=='%'){
                        tmp[0]=str[i+1];
                        tmp[1]=str[i+2];
                        output += StrToBin(tmp);
                        i=i+3;
                }
                else if(str[i]=='+'){
                        output+=' ';
                        i++;
                }
                else{
                        output+=str[i];
                        i++;
                }
        }

        return output;
}
//把url utf8解碼
string strCoding::UrlUTF8Decode(string str)
{
     string output="";

    string temp =UrlGB2312Decode(str);//

    UTF_8ToGB2312(output,(char *)temp.data(),strlen(temp.data()));

    return output;

}      

1. //test   
2. #include "stdafx.h"   
3. #include "strCoding.h"   
5. int main()  
6. {   
7.      strCoding cfm;  
8.      string keyword="大家好,歡迎你";  
9.      string Temp="";  
10.      string Output="";  
11.      //把關鍵字做url的utf8編碼   
12.      Temp= cfm.UrlUTF8((char *)keyword.data());  
13.      cout<<Temp<<endl;  
14.      //把url的utf8編碼的結果解碼   
15.      Temp =cfm.UrlUTF8Decode(Temp);  
17.      //把關鍵字做url的gb2312編碼   
18.      Temp =cfm.UrlGB2312((char *)keyword.data());  
20.      //把url的gb2312編碼的結果解碼   
21.      Temp =cfm.UrlGB2312Decode(Temp);  
23.      //把關鍵字GB2312轉UTF_8   
24.      cfm.GB2312ToUTF_8(Output,(char *)keyword.data(),strlen(keyword.data()));  
25.      cout<<Output<<endl;  
26.      //把GB2312轉UTF_8轉為中文   
27.      cfm.UTF_8ToGB2312(Temp,(char *)Output.data(),strlen(Output.data()));  
29.     //system("pasue");   
30.      getchar();  
31.     return 0;  
32.     //