C ++: interchange between UTF-8 and GB2312

Not to mention, I encountered a problem between characters a few days ago, and garbled characters appeared on the webpage. So I also wrote a Baidu experience.

Now is the time to solve this problem. Of course, you can simply convert it with the "Save as" provided by txt, but now we are discussing how to use functions in c ++ to change it.

The following describes two WinAPI functions: WideCharToMultiByte and MultiByteToWideChar.

Function prototype:

Int WideCharToMultiByte (
UINT CodePage, // code page
DWORD dwFlags, // performance and mapping flags
LPCWSTR lpWideCharStr, // wide-character string
Int cchWideChar, // number of chars in string
LPSTR lpMultiByteStr, // buffer for new string
Int cbMultiByte, // size of buffer
Lpstr lpDefaultChar, // default for unmappable chars
LPBOOL lpUsedDefaultChar // set when default char used
); // Converts a width character to Multiple Narrow characters

Int MultiByteToWideChar (
UINT CodePage, // code page
DWORD dwFlags, // character-type options
LPCSTR lpMultiByteStr, // string to map
Int cbMultiByte, // number of bytes in string
LPWSTR lpWideCharStr, // wide-character buffer
Int cchWideChar // size of buffer
); // Converts Multiple Narrow characters into wide characters
The following functions are required:

CString CTest: HexToBin (CString string) // convert a hexadecimal number to a binary number.
{
If (string = "0") return "0000 ";
If (string = "1") return "0001 ";
If (string = "2") return "0010 ";
If (string = "3") return "0011 ";
If (string = "4") return "0100 ";
If (string = "5") return "0101 ";
If (string = "6") return "0110 ";
If (string = "7") return "0111 ";
If (string = "8") return "1000 ";
If (string = "9") return "1001 ";
If (string = "a") return "1010 ";
If (string = "B") return "1011 ";
If (string = "c") return "1100 ";
If (string = "d") return "1101 ";
If (string = "e") return "1110 ";
If (string = "f") return "1111 ";
Return "";
}

CString CTest: BinToHex (CString BinString) // convert binary to hexadecimal
{
If (BinString = "0000") return "0 ";
If (BinString = "0001") return "1 ";
If (BinString = "0010") return "2 ";
If (BinString = "0011") return "3 ";
If (BinString = "0100") return "4 ";
If (BinString = "0101") return "5 ";
If (BinString = "0110") return "6 ";
If (BinString = "0111") return "7 ";
If (BinString = "1000") return "8 ";
If (BinString = "1001") return "9 ";
If (BinString = "1010") return "";
If (BinString = "1011") return "B ";
If (BinString = "1100") return "c ";
If (BinString = "1101") return "d ";
If (BinString = "1110") return "e ";
If (BinString = "1111") return "f ";
Return "";
}

Int CTest: BinToInt (CString string) // convert binary data to a 10th integer
{
Int len = 0;
Int tempInt = 0;
Int strInt = 0;
For (int I = 0; I <string. GetLength (); I ++)
{
TempInt = 1;
StrInt = (int) string. GetAt (I)-48;
For (int k = 0; k <7-i; k ++)
{
TempInt = 2 * tempInt;
}
Len + = tempInt * strInt;
}
Return len;
}
UTF-8 to GB2312 first convert the UTF-8 to Unicode. Then the Unicode through the function WideCharToMultiByte to GB2312

WCHAR * CTest: UTF_8ToUnicode (char * ustart) // convert the UTF-8 to Unicode
{
Char char_one;
Char char_two;
Char char_three;
Int Hchar;
Int Lchar;
Char uchar [2];
WCHAR * unicode;
CString string_one;
CString string_two;
CString string_three;
CString combiString;
Char_one = * ustart;
Char_two = * (ustart + 1 );
Char_three = * (ustart + 2 );
String_one.Format ("% x", char_one );
String_two.Format ("% x", char_two );
String_three.Format ("% x", char_three );
String_three = string_three.Right (2 );
String_two = string_two.Right (2 );
String_one = string_one.Right (2 );
String_three = HexToBin (string_three.Left (1) + HexToBin (string_three.Right (1 ));
String_two = HexToBin (string_two.Left (1) + HexToBin (string_two.Right (1 ));
String_one = HexToBin (string_one.Left (1) + HexToBin (string_one.Right (1 ));
CombiString = string_one + string_two + string_three;
CombiString = combiString. Right (20 );
CombiString. Delete (4, 2 );
CombiString. Delete (10, 2 );
Hchar = BinToInt (combiString. Left (8 ));
Lchar = BinToInt (combiString. Right (8 ));
Uchar [1] = (char) Hchar;
Uchar [0] = (char) Lchar;
Unicode = (WCHAR *) uchar;
Return unicode;
}

Char * CTest: UnicodeToGB2312 (unsigned short uData) // convert Unicode to GB2312
{
Char * buffer;
Buffer = new char [sizeof (WCHAR)];
WideCharToMultiByte (CP_ACP, NULL, & uData, 1, buffer, sizeof (WCHAR), NULL, NULL );
Return buffer;
}
GB2312 to UTF-8: First GB2312 through the MultiByteToWideChar function to convert Unicode. Then the Unicode By disassembling Unicode assembled into a UTF-8.

Author: Li Mu Space"