🔒 Locked

Need help about UNICODE

Started by Mescalito Feb 18, 2003 at 8:10 AM 5 replies 500+ views

Original Post

Feb 18, 2003 at 8:10 AM

Hi everyone, i didn''t knwo where to post a question about UNICODE format so I posted there... So here''s my problem: I need to developp a application using COM interfaces and these are currently using an UNICODE string format (used to give filenames for examples); this format uses 2 bytes for a character (standard string are [char] coded so it''s only 1 byte). I have so troubles using this format as i use standard C functions to managed all my strings (strcpy, strcmp.. aand all str___() functs, fopen, fclose, fprintf, fwrite and fread for files...). How can we "convert" a simple [char*] string to a [unsigned short*] one (the unicode string format) WITHOUT LOSS OF VALIDTY I have tried a (barbarian

) convertion (with casting) but he didn''t keep the filename valid! So can anyone help me about that !!! Thanks for any reply

dalleboy

Feb 18, 2003 at 8:40 AM

The following example will allocate memory on the stack, if you don''t like that you should replace alloca.

  

#include <malloc.h>

const char* sz = "narrow string";
const size_t uiLength = strlen(sz) + 1;
wchar_t* wsz = (wchar_t*) alloca(uiLength * 2);
MultiByteToWideChar(CP_ACP, 0, sz, -1, wsz, uiLength);
// Here you could use your wsz, when it falls out of scope it will be deallocated.

You could hardcode it like this if you need to:

  
const wchar_t* wsz = L"constant wide string";

Fruny

203

Fruny

Feb 18, 2003 at 8:43 AM

First, the wide-character type isn''t unsigned short, but wchar_t. Even though VC6 treats it as a typedef, it really is a basic type, just as much as char, int, short ...

If you''re intent on using C functions, there are wide-character versions available, usually by adding a ''w'' in front of the function name. Whether the encoding is Unicode or something else depends on your locale. I strongly encourage you to read up on locales !

Alternatively, use the C++ wide string type (std::wstring) to store the Unicode text. Assigning to one such should automatically convert from char to wchar_t, using the std::ios::widen function (which is, once again, dependent on locales, read up on them).

Finally, one last option is to download the Boost library and to make use of their lexical_cast function (which uses stringstreams to do the conversion)

Using it, the code ought to look like (assuming the appropriate using statements have been issued) :


MyCApiFunction( lexical_cast<wstring>( mystring ).c_str() );

This ought to work regardless of whether mystring is a C-style NULL-terminated narrow string or a C++ narrow string object.

If not, that should at least have given you some starting points.

Mescalito

Author

Mescalito

Feb 18, 2003 at 9:07 AM

Has unicode strings have an "end" character ???

(every string (char * type) have an ''\0'' char at the end of it)

Fruny

203

Fruny

Feb 18, 2003 at 9:16 AM

quote:
Original post by Mescalito
Has unicode strings have an "end" character ???

(every string (char * type) have an ''\0'' char at the end of it)

Not every string has a ''\0'' at the end, only C-style (a.k.a. ''null-terminated'') strings do. Other schemes involve keeping track of the total string length. That''s how strings work in Pascal, VB, C++ ...

That being said, a null-terminated wide string would obviously end up with a null wide-character...

dalleboy

Feb 18, 2003 at 4:36 PM

quote:
Original post by Mescalito
Has unicode strings have an "end" character ???

(every string (char * type) have an ''\0'' char at the end of it)

A zero terminated wide string will have a zero at the end (or rather a L''\0'').

Update GameDev.net system time campaign - success at last

Topic Locked

This topic has been locked by a moderator. New replies are not allowed.