Unicode LPCTSTR Problem

[Ocrana007]

Hello,

I have an Unicode application and need to call an DLL with Multibyte Interface. The Interface have an Function:

ThisFunction(LPCTSTR name)

How I can now give an Multibyte LPCTSTR from an Unicode application to that DLL. Everything I tried gave me the linker error that there is no Function available with wchar_t *. This is right cause the DLL expect const char *

So how to call it outside the Unicode application with an call like:

ThisFunction("Dialog");

With _T("Dialog") and L"Dialog it do not work (linker problem).

Ocrana

[googler]

If the parameter is a constant then just call it with

ThisFunction("Dialog");

If the parameter is not a constant, use WideCharToMultiByte() to convert your Unicode string to a multibyte string (of type char*) and pass it to ThisFunction().

[Siddhartha]

ThisFunction(LPCTSTR name)

How I can now give an Multibyte LPCTSTR from an Unicode application to that DLL.

This should be called as follows -

ThisFunction (_T ("A TCHAR Constant String"));

And...

Everything I tried gave me the linker error that there is no Function available with wchar_t *. This is right cause the DLL expect const char *

You mean compiler error (and not linker error)?

So how to call it outside the Unicode application with an call like:
With _T("Dialog") and L"Dialog it do not work (linker problem).

Post the actual error you see.

[Siddhartha]

How I can now give an Multibyte LPCTSTR from an Unicode application to that DLL.

BTW, there is no such thing as a Multibyte LPCTSTR.

LPCTSTR is a typedef defined as -

typedef const TCHAR* LPCTSTR;

Where TCHAR is defined as -

#ifdef _UNICODE
typedef wchar_t TCHAR;
 #else
 typedef char_t TCHAR;

...put simply.

So, as you can see, depending on the definition of the UNICODE macro, LPCTSTR can be a wchar_t* or a char*

MBCS (Multi-Byte Character String) is basically a char* string that allows a simple character string to hold international characters by allowing a variable width of characters per symbol (alphabet).

So, a LPCTSTR can be a Wide Character String, or a Simple Character string depending on the definition of the UNICODE macro. If the latter, it can optionally be a MBCS string depending on the definition of the MBCS macro (that redirects all _t* string functions to mb* functions). So, a LPCTSTR being multi-byte is a special case of a LPCTSTR, and one must not pass a pure-MBCS string as a LPCTSTR - it will not compile for UNICODE settings.

Essentially, if your function accepts a LPCTSTR, supply a LPCTSTR string wrapped in the _T macro, and let the pre-processor definition convert the string for you - i.e. _T gets mapped as L when _UNICODE is defined (making the string literal a wide-character one), and ignored otherwise (making it a character string).

[eero_p]

Have you compiled the DLL with unicode -option as well? If not, you have to convert the unicode string passed to DLL-function to multibyte string. Since they are separate binaries, they don't automatically adapt to each other's settings.