VC字符集,Visual c++ character sets:Unicode, _MBCS

 

在VC6中创建一个MFC Dialog工程，下面是创建Unicode版本的操作步骤：
1.Build-->Configurations-->Add，添加一个Unicode Debug配置；
2.Build-->Set Active Configuration，选择Win32 Unicode Debug作为当前活动配置；
3.在Project-->Settings，在C/C++属性页中选择Preprocessor条目，在Preprocessor definitions中添加编译项 _UNICODE；
如果此时就编译，就会立生错误：
msvcrtd.lib(crtexew.obj) : error LNK2001: unresolved external symbol _WinMain@16
还要进行下面操作：
4.在上一步中，在Link属性页中选择Output条目，在Entry-point symbol中填入 wWinMainCRTStartup
5.OK，重新编译，通过。

Microsoft run-time library provides Microsoft-specific generic-text mappings for many data types, routines and other objects, mappings that are defined in TCHAR.h. There are three supported character sets:

1. [*]ASCII (single-byte character set – SBCS)
2. [*]MBCS (multi-byte character set)
3. [*]Unicode

The use of one or another character set is controlled by two pre-processor directives:

• _UNICODE: if defined, Unicode is the character set used

• _MBCS: if defined, MBCS is used

• If neither of the above (mutually-exclusive) is defined, ASCII is the character set used

The Windows API provides different version of each function for Unicode and ASCII.

Q: How do I select the character set?
A: You have to go to Project Properties > Configuration Properties > General and change the value of the Character Set property. The three available options are:

• Not Set (neither _UNICODE nor _MBCS are defined)

• Use Multi-byte Character Set (_MBCS is defined)

• Use Unicode Character Set (_UNICODE is defined)

Q: How exactly do the generic-text mapping directives affect the data types and functions that I'm using?
A: C run-time library functions, such as _itot, or Windows API functions, such are MessageBox, aren't functions at all; they are macros.

The C run-time library provides functions for all character sets and a macro to define one or another of these functions depending on the used character set. For instance macro _itot resolves to:

• _itoa, when _UNICODE is not defined

• _itow, when _UNICODE is defined

Similarly, TCHAR resolves:

• char, when _UNICODE is not defined

• wchar_t, when _UNICODE is defined

You can read more about the mappings in MSDN.

On the other hand, the Windows API comes in two versions: for Unicode and for ASCII/Multi-byte. If you read the MSDN page for MessageBox it says:

The MessageBox function creates, displays, and operates a message box. The message box contains an application-defined message and title, plus any combination of predefined icons and push buttons.

int MessageBox(
HWND hWnd,
LPCTSTR lpText,
LPCTSTR lpCaption,
UINT uType);

Actually, MessageBox and LPCTSTR are both macros. You can see how MessageBox it's defined in WinUser.h:

#ifdef UNICODE
#define MessageBox MessageBoxW
#else
#define MessageBox MessageBoxA
#endif // !UNICODE

There are two version of the function, actually: MessageBoxA for ASCII & MBCS and MessageBoxW for Unicode. When UNICODE (which is the same with _UNICODE) is defined then MessageBox resolves to MessageBoxW and LPCTSTR to LPCWSTR (i.e. const whar_t*); otherwise MessageBox resolves to MessageBoxA and LPCTSTR to LPCSTR (i.e. const char*).

Q: How do I write my program so that it builds for any of these character sets without modifying the code when the character set changes?
A: In a single-byte or multi-byte character set the strings and characters are not prefixed my anything ('string', 'c'). However, for Unicode strings and characters required the suffix L, such as L"string" and L'c'. You can use the Microsoft-specific macros _T() or _TEXT(). These macros are removed by the pre-processor when _UNICODE is not defined, and replaced with L when _UNICODE is defined.

Unicode defined:

• no: _T("string") becomes "string" and _T('c') becomes 'c'

• yes: _T("string") becomes L"string" and _T('c') becomes L'c'

Q: How do I fix the mention line of code?
A: It should be clear now:

MessageBox(NULL, _T("Test message"), _T("Title"), MB_OK);