Wchar_t wide character set research and com bstr Variant

 

Wchar_t is equal to Char, that is, wchar_t is not derived from typedef and is a native variable.

In short, it has two bytes, which are the same as the short space.

For example:

String "We \ n"

ANSI hexadecimal: Ce D2 C3 C7 0a 00

Six bytes, including \ 0 at the end of the string

Unicode hexadecimal: 11 62 EC 4E 0a 00 00 00

8 bytes, all characters are 2 bytes, even if the letter and number are all, of course, the line feed \ n is also 0a 00.

 

AverageProgramAdding l "" in front of the string indicates that it is a unicode string.

In Windows, a macro _ T ("") is the same as the one above.

 

1. The first simple question is, how to print out Unicode?

2 bytes, which can be printed by number, but if you want to print by character, it cannot be used as a normal printf.

You can use wprintf to print, that is, add a wide W to the front of a common printf. Similar functions, such as wsprintf.

     Char * Lpsztext = " We \ r \ n  " ; //  ANSI: Ce D2 C3 C7  //  UNICODE: 11 62 EC 4E  //  Press enter \ r 0d \ n 0a Printf ( "  Char * Text: % s 0x % 08x 0x % 08x \ nansi encoding:  " , Lpsztext, lpsztext ,* Lpsztext); print_hex_to_file (stdout ,(  Const Uint8_t *) lpsztext, strlen (lpsztext) + 1 , 16  ); // BSTR bstrtext compiled by this function = _ Com_util: convertstringtobstr (lpsztext); wprintf (L  "  BSTR text: % s 0x % 08x 0x % 08x \ nunicode encoding is:  " , Bstrtext, bstrtext ,* Bstrtext); print_hex_to_file (stdout ,(  Const Uint8_t *) bstrtext, wcslen (bstrtext )* 2 + 2 , 16 );

The result is:

 
Char* Text: We
0x013fbd80 0 xffffffceANSI code: 0x ce D2 C3 C7 0d 0a00BSTR text: We
0x007be5b4 Zero X 00006211Unicode encoding: 0x11 62EC 4E 0d000a00 00 00
 

By the way, at first I used wprintf to print Chinese characters, and then I added the following two sentences.

# Include <locale. h>Setlocale (lc_ctype,"CHS");
 

The source code encoding is ANSI or Unicode, which has no effect on the result.

 

By the way, how can I print a single wchar_t? The above are all pointers and they are all strings. That's good. Single...

Setlocale (lc_ctype,"CHS"); Wchar wstr1; wchar_t wstr2; wstr1= L'Me'; Wstr2= L'Are'; Wprintf (L"Every size in the wide character set (% C, % C) is % d bytes \ n", Wstr1, wstr2,Sizeof(Wstr1 ));
 

Always remember l when assigning values, and the result is normal.

 
Each size of our wide character set (US) is:2Bytes
 

If you assign 'I' to a char type, you can only get the first byte of 'my' ce D2. Is it garbled? .

 
CharSS; SS='Me'; Printf ("Ss = % C \ n", SS); the result is: SS=?
 

And? There is no line feed next to it, because \ n is already integrated with % C? .. It may be strange to print ce.

2. The second simple question is how to convert to the char type

 Int  Convertstringtobstrdemo (){  Char * Lpsztext = "  Test "  ; Printf (  "  Char * Text: % s \ n  "  , Lpsztext); BSTR bstrtext = _ Com_util: convertstringtobstr (lpsztext); wprintf (L  "  BSTR text: % s \ n  "  , Bstrtext );  : Sysfreestring (bstrtext );      Return   0  ;}; Int  Convertbstrtostringdemo () {BSTR bstrtext =: Sysallocstring (L "  Test  "  ); Wprintf (L  "  BSTR text: % s \ n  "  , Bstrtext );  Char * Lpsztext2 = _ Com_util: convertbstrtostring (bstrtext); printf (  "  Char * Text: % s \ n "  , Lpsztext2 );  : Sysfreestring (bstrtext );  Delete [] lpsztext2;  Return   0  ;}; 

This global function sysfreestring () does not seem to have a memory leak if it is not added? (VLD detection)

I rely on it, I know, it may be that VLD has not been reloaded to release the memory allocation in COM, so the comments before sysfreestring are removed.

A memory leak of about MB may occur after 10 000 cycles in the experiment. However, VLD cannot be detected. So be careful!

 

In COM programming, BSTR is actually the wchar_t * type. in BSTR, pointers are allocated. You must release the memory yourself!

The conversion of BSTR and string (char *) is actually the conversion of wchar_t * and char. This is the com method.

 

You can also use the methods in stdlib:

Wcstombs and mbstowcs should be widecstring, but how does MBS indicate ANSI normal character encoding? I don't know the abbreviation.

Wchar_t ws [ 10 ]; //  Sizeof (WS) = 20 bytes Wsprintf (WS, l "  We  "  );  Char CS [ 50  ]; Sprintf (CS,  ""  );  //  Clear data and initialize //  Wchar_t * convert to char *      Int Ret = 0  ; Printf (  "  Before wcstombs: cs = % 4 s Ws = % s \ n  "  , Cs, WS); RET = Wcstombs (CS, WS, Sizeof  (WS); printf (  "  After wcstombs: ret = % d, cs = % 4 s Ws = % s \ n  "  , RET, Cs, WS); wsprintf (WS, l ""  );  //  Clear data and initialize  //  Char * To wchar_t * Wprintf (L "  Before mbstowcs: Ws = % 4 s cs = % s \ n  "  , WS, (CS); RET = Mbstowcs (WS, Cs, Sizeof (WS )* 2  ); Wprintf (L  " After mbstowcs: ret = % d, Ws = % 2 S cs = % s \ n  " , RET, WS, (CS ));

Running result

 
Before wcstombs: cs = Ws = % s =After wcstombs: Ret=4, Cs = US Ws = % s =Before mbstowcs: WS= Cs = % s =After mbstowcs: Ret=2, Ws = US cs = % s = us
 

In Windows, there are APIs of the same meaning.

//Multibytetowidechar

 

Add my favorite print_hex_to_file function.

 Void Print_hex_to_file (File * FP, Const Uint8_t * array, Int Count/*  Size of Aray  */ , Int Linecount /*  The default value is 16.  */  ){  Int  I; fprintf (FP,  "  0x  "  );  For (I = 0 ; I <Count;) {fprintf (FP,  "  % 02x  "  , Array [I]); I ++ ;  If (! (I % linecount) & I < Count) {fprintf (FP,  "  \ N0x  "  ) ;}} Fprintf (FP,  "  \ N  " ) ;};