Prev: passing vector as argument
Next: create vector with index base 1
From: TheOne on 22 Apr 2008 22:05
I want to be confirmed.

I'm writing a dll which includes unicode and non-unicode version of
Func, namely, FuncA and FuncW. The dll should include both of them and
the application side must call these function accroding to their
UNICODE setting.

I guess I should give the app-writer header file which looks like,

dllheader.h
+++
#ifdef UNICODE
#define Func FuncW
#else
#define Func FuncA
#endif
+++

However, if I include the same header, dllheader.h when building the
dll. Then the dll's also have only one implementation according to
UNICODE setting. So I guess I should have different headers to build
the actually dll and to give the user to use to build his application.

Am I right? Or, is there a way in which I can use the same header file
for building dll and building app?

TIA.

--
Daewon YOON
From: Igor Tandetnik on 22 Apr 2008 22:19
"TheOne" <daewon.yoon(a)gmail.com> wrote in message
news:2b6c3c7f-a333-462c-8411-08a3c6f6d21c(a)y21g2000hsf.googlegroups.com
> I'm writing a dll which includes unicode and non-unicode version of
> Func, namely, FuncA and FuncW. The dll should include both of them and
> the application side must call these function accroding to their
> UNICODE setting.
>
> I guess I should give the app-writer header file which looks like,
>
> dllheader.h
> +++
> #ifdef UNICODE
> #define Func FuncW
> #else
> #define Func FuncA
> #endif
> +++
>
> However, if I include the same header, dllheader.h when building the
> dll. Then the dll's also have only one implementation according to
> UNICODE setting.

Simply don't use Func in your DLL code. Write FuncA and FuncW
explicitly.
--
With best wishes,
Igor Tandetnik

With sufficient thrust, pigs fly just fine. However, this is not
necessarily a good idea. It is hard to be sure where they are going to
land, and it could be dangerous sitting under them as they fly
overhead. -- RFC 1925


From: Norman Bullen on 23 Apr 2008 00:04
TheOne wrote:
> I want to be confirmed.
>
> I'm writing a dll which includes unicode and non-unicode version of
> Func, namely, FuncA and FuncW. The dll should include both of them and
> the application side must call these function accroding to their
> UNICODE setting.
>
> I guess I should give the app-writer header file which looks like,
>
> dllheader.h
> +++
> #ifdef UNICODE
> #define Func FuncW
> #else
> #define Func FuncA
> #endif
> +++
>
> However, if I include the same header, dllheader.h when building the
> dll. Then the dll's also have only one implementation according to
> UNICODE setting. So I guess I should have different headers to build
> the actually dll and to give the user to use to build his application.
>
> Am I right? Or, is there a way in which I can use the same header file
> for building dll and building app?
>
> TIA.
>
> --
> Daewon YOON

I make a header like this for the users of the DLL.
MyFunc.h ---------------------------------------
int MyFuncA(LPCHAR strArgument);
int MyFuncW(LPWCHAR strArgument);
#if defined(_UNICODE_)
#define MyFunc(strArgument) MyFuncW(strArgument)
#else
#define MyFunc(strArgument) MyFuncA(strArgument)
#endif
MyFunc.h ---------------------------------------

Then I put all of the code for MyFunc() into another header which only
I, the developer, will use.
MyFuncT.h --------------------------------------
#include <tchar.h>
#include "MyFunc.h"
int MyFunc(LPTSTR strArgument)
{ // the body of the function
}
MyFuncT.h --------------------------------------

Then I create two C or C++ modules like this.
MyFuncA.c --------------------------------------
#undef _UNICODE_
#undef UNICODE
#include "MyFuncT.h"
MyFuncA.c --------------------------------------

MyFuncW.c --------------------------------------
#define _UNICODE_
#define UNICODE
#include "MyFuncT.h"
MyFuncW.c --------------------------------------

Because the modules MyFuncA.c and MyFuncW.c have different names and
define different entry point names, they can both be in the same DLL or
LIB file.

--
Norm

To reply, change domain to an adult feline.

From: Giovanni Dicanio on 23 Apr 2008 05:56

"TheOne" <daewon.yoon(a)gmail.com> ha scritto nel messaggio
news:2b6c3c7f-a333-462c-8411-08a3c6f6d21c(a)y21g2000hsf.googlegroups.com...

> I'm writing a dll which includes unicode and non-unicode version of
> Func, namely, FuncA and FuncW. The dll should include both of them and
> the application side must call these function accroding to their
> UNICODE setting.

I would implement the FuncW (Unicode UTF-16 version), and then implement the
FuncA (ANSI version) just as a thin wrapper of Unicode, i.e. the FuncA
version converts the string(s) from Ansi to Unicode using e.g. CA2W and
simply calls the Unicode version FuncW.

void FuncW( const wchar_t * wsz... )
{
... full implementation (Unicode) ...
...
}

void FuncA( const char * sz ... )
{
// Convert from Ansi to Unicode
// each string passed as input
CA2W wsz( sz );
...CA2W wszAnother( szAnother );
...

// Just call the main Unicode implementation
FuncW( wsz ... );
}

You export both functions (A/W) from the DLL, and then use the classic
"#ifdef UNICODE" preprocessor conditional paradigma in the DLL public header
file:

// dllheader:
#ifdef UNICODE
#define Func FuncW
#else
#define Func FuncA
#endif


HTH,
Giovanni


 | 
Pages: 1
Prev: passing vector as argument
Next: create vector with index base 1