C / C++ / MFC

21-Mar-09 7:48

p_1960 wrote:
Hi,
Really im getting confused in using unicode in my project.....When im writing the code using ANSI it is working fine but when im writing code in Unicode im getting lot of errors....im confused with LPCWSTR,LPCSTR,LPWSTR,LPBYTE.....some functions take some of these as parameters but im really confused of using them, can i get any basic article to start with.....bcos lot of time wasted by just not knowing how to use these things....
suppose

See here and here.

p_1960 wrote:
CString csPath="D:\\test";
::CreateDirectory(csPath+_T("\\tester"),NULL);

Change to

CString csPath = _T("D:\\test");
csPath += _T("\\tester");
::CreateDirectory(csPath , NULL);

The fix occurs int the first line, the middle one is just a matter of style.
Smile | :)

If the Lord God Almighty had consulted me before embarking upon the Creation, I would have recommended something simpler.
-- Alfonso the Wise, 13th Century King of Castile.

This is going on my arrogant assumptions. You may have a superb reason why I'm completely wrong.
-- Iain Clarke

[My articles]

Re: Unicode..

Jörgen Sigvardsson21-Mar-09 22:43

Jörgen Sigvardsson

21-Mar-09 22:43

There's two basic defines in the windows header files: CHAR and WCHAR. CHAR is defined as an ordinary char, and WCHAR is defined as a 16 bit WORD. CHAR is used for ANSI/MBCS (Multibyte character sets - pre-unicode solutions for "exotic" languages). WCHAR is used for UTF16/Unicode.

To assign an ANSI character, you do as you are used to. CHAR c = 'A'; To assign a Unicode character, you do it like this: CHAR u = L'A'; The L prefix tells the compiler that the character is a Unicode character. The same pattern applies for strings.

Strings are defined like this in the header files:

typedef LPSTR char*;
typedef LPWSTR char*;

Here you can also see a pattern emerge in the string defines. Whenever you see a W in it somewhere, it's a Unicode string. If it's got a C in it (like LPCSTR), it's a const string.

To assign an Ansi string, you do something like this: LPCSTR str = "Hello world!"; Similarly in Unicode: LPCWSTR str = L"Hello world!";

With me so far?

Right. There's a third define in the windows header file which is a bit special. It's called TCHAR. TCHAR is defined like this:

#ifdef _UNICODE
typedef TCHAR WCHAR;
#else
typedef TCHAR CHAR;
#endif

The meaning of TCHAR is dependent on the _UNICODE symbol. When you specify in the project properties that you want to be using Unicode, the compiler will define this symbol for you.

Anyway, this is ideal if you want to support both ANSI and Unicode builds of the same application. There's just one problem though. This line: LPCTSTR str = "Hello world!"; will only compile in ANSI builds, but not in Unicode build. This is because the string in the example is of type LPCSTR (const char*), while str is of type LPCWSTR (const WORD*)¹. The remedy are the macros _T() and TEXT(). _T and TEXT are equivalent. I prefer typing _T because it's less typing.
_T (and TEXT) are define like this:

#ifdef _UNICODE
#define _T(x) L##x
#else
#define _T(x) x
#endif

So, the previous code LPCTSTR str = "Hello world!" should be rewritten as LPCTSTR str = _T("Hello world!");. This will be expanded to the correct version, depending on wheter you've specified _UNICODE or not.

You will find that the Microsoft APIs have support for both Ansi and Unicode in parallel. In most cases (such as the Win32 API), the header files will define functions like this:

BOOL CreateWindowA(...);
BOOL CreateWindowW(...);
#ifdef _UNICODE
#define CreateWindow CreateWindowW
#else
#define CreateWindow CreateWindowA
#endif

This way, as long as you are consistent in the use of either Ansi or Unicode strings, you won't have to adapt much. Other APIs such as the C runtime library are a bit different. There you have the good old functions such as strlen left intact - they only work on ANSI strings. In addition to strlen, there is also wcslen (Unicode), and _tcslen. the _t-functions are defined differently depending on the _UNICODE symbol, much like the Win32 functions mentioned earlier.

The quick answer to your question is: Use _T() for all your strings and characters, and use _t-functions from the C runtime library, and you will be pretty much fine. Do remember though, that when you write text to sockets, files or via some other medium, that you might have to reencode your Unicode strings. That, however, is a bit larger topic...

Personally, I think it's best to just pick Unicode from scratch. The world is becoming more and more globalized, and different cultures meet more often, so in a business application, odds are that there will be "trouble" down the line. Windows is also a bit faster at handling Unicode strings, because starting with NT4 (or was it Windows 2000? can't remember), the Ansi-versions of the API functions are just shallow wrappers that converts the strings into Unicode and pass them onto the Unicode implementations.

¹ The compiler in these errors complain that the types const char* isn't compatible with the type const unsigned short*.

convert

durban221-Mar-09 5:07

durban2

21-Mar-09 5:07

MFC, STUDIO 2008,
Project use UNICODE
CString m_sTest;
m_sTest = L"Байкал";
char *m_pchar3;
m_pchar3 = new char[70];
How todo m_sTest into m_pchar3?
//- m_pchar3=m_sTest; // error C2440: '=' : cannot convert from 'CString' to 'char *'
//- strcpy(m_pchar3, m_sTest); // error C2664: 'strcpy' : cannot convert parameter 2 from 'CString' to 'const char *'
//- strcpy(m_pchar3, (TCHAR)m_sTest); // error C2440: 'type cast' : cannot convert from 'CString' to 'TCHAR'
//- _tcscpy( m_pchar3, (TCHAR*)(LPCTSTR)m_sTest ); // error C2664: 'wcscpy' : cannot convert parameter 1 from 'char *' to 'wchar_t *'

CPallini21-Mar-09 5:13

21-Mar-09 5:13

Why do you need to do that?
Smile | :)

Dominik Reichl21-Mar-09 5:14

Dominik Reichl

21-Mar-09 5:14

CHIVOTIOV wrote:
char *m_pchar3;
m_pchar3 = new char[70];

Use TCHAR instead of char.

Too many passwords to remember? Try KeePass Password Safe!

Akt_4_U21-Mar-09 5:52

Akt_4_U

21-Mar-09 5:52

The m_sTest is a UNICODE string. So if you want to convert this UNICODE string to char* use WideCharToMultiBye API or use CRT function wcstombs.

prvn

how to write list control data in text file in MFC

Eytukan21-Mar-09 22:02

Eytukan

21-Mar-09 22:02

CHIVOTIOV wrote:
/ error C2664: 'wcscpy' : cannot convert parameter 1 from 'char *' to 'wchar_t *'

Just a padding to the other replies here. As the above lines says, You are trying to squeeze a 2 Byte string (UNICODE) into a single byte string(ANSI).

char , std::string , etc by default belong to the single byte representation.
wchar_t, std::wstring, etc represent UNICODE representation.

TCHAR is "neutral" based on the scheme used, it switches type accordingly. If you define # UNICODE in the setting, the string becomes 2 byte reps. If you don't define it, it becomes single byte. Also when you use TCHARs, you should use strings apis specific to TCHARs. (like _tcscpy) that assures complete independence to your character-set implementations.

Similarly, CString switches types based on the project character-set settings.(UNICODE/Single-byte).

Now you have defined UNICODE in your project settings but trying a cross-copy operation. So just figure out what needs to be done.. Smile | :)

He never answers anyone who replies to him. I've taken to calling him a retard, which is not fair to retards everywhere.-Christian Graus

amit14aug21-Mar-09 4:38

amit14aug

21-Mar-09 4:38

hi,
Can Anybody tell me how to write a list control data in text file in MFC?

Re: how to write list control data in text file in MFC

Rajesh R Subramanian21-Mar-09 4:41

Rajesh R Subramanian

21-Mar-09 4:41

CListCtrl has member functions like GetItem(), GetItemData(), etc., and there's CFile that you can use to write to file. Where exactly are you stuck?

It is a crappy thing, but it's life -^ Carlo Pallini

pointer-to-array

korte2521-Mar-09 3:49

21-Mar-09 3:49

Hi!

I have an array of pointers, which point to classes derived from the same class. I also have a pointer, that points to the array. Is there a simple way to find out to which element of the array this pointer points to, or shall i use a variable in the base class for this purpose?
Thanks

CPallini21-Mar-09 4:09

21-Mar-09 4:09

You may use RTTI [^].
Smile | :)

korte2521-Mar-09 6:09

21-Mar-09 6:09

Actually, this solves another problem of mine, so thanks a lot Big Grin | :-D

However, I'd like to know, whether the pointer points to the 2nd, 5th, nth element of the array. If I understand the description correctly, type_info only gives me the name of the class my pointer-in-the-array points to (right?). First I thought of using sizof(), but my derived classes are of different sizes.

CPallini21-Mar-09 6:25

21-Mar-09 6:25

Please post the relevant code (I need to know the pointer relationship with the array).
Smile | :)

korte2521-Mar-09 6:57

21-Mar-09 6:57

I have a base class, let's say Car, and derived classes BMW, Mercedes, Renault, etc.
The array is: Car* p_Cars[2][16];
The pointer: Car** pCar;

for (int i = 0; i < 2; i++)
{
    for (int j = 0; j < 16; j++)
    {
        if (j==2 || j==3)
            p_Cars[i][j] = new BMW(i, j);
        else if (j==4 || j==5)
            p_Cars[i][j] = new Mercedes(i, j);
        .
                    .
                    .
    }
}

So I'd like to know, whether I can determine the element from its address.
I hope you understand my problem, i tried my best Sniff | :^)

CPallini21-Mar-09 7:41

21-Mar-09 7:41

A given value of pCar refers to the element p_Cars[i][j] (i.e. pCar == &p_Cars[i][j] holds true), provided i,j are computed as follows:

int k = (int) (pCar - &p_Cars[0][0]);
int i = k / 16;
int j = k % 16;

korte2521-Mar-09 7:44

21-Mar-09 7:44

thanks

CPallini21-Mar-09 7:48

21-Mar-09 7:48

You are welcome.
Smile | :)

Stuart Dootson21-Mar-09 12:29

Stuart Dootson

21-Mar-09 12:29

So....you have:

CBase* pArray[] = { whatever... };
CBase* pItem = address_of_some_item_in_the_array;

?

In that case:

const size_t index = pItem - pArray;

Java, Basic, who cares - it's all a bunch of tree-hugging hippy cr*p

korte2522-Mar-09 2:30