WinMain()

[Hunderpanzer]

Hey guys. I'm trying to learn some windows programming, but i'm getting some errors with the code:

#include <windows.h>    // include the basic windows header file
 
// the entry point for any Windows program
int WINAPI WinMain(HINSTANCE hInstance,
                   HINSTANCE hPrevInstance,
                   LPSTR lpCmdLine,
                   int nShowCmd)
{
    // create a "Hello World" message box using MessageBox()
    MessageBox(NULL,
               L"Hello World!",
               L"Just another Hello World program!");
               MB_ICONEXCLAMATION | MB_OK);
 
    // return 0 to Windows
    return 0;
}
 

The errors I get (using Dev-C++) are


In function `int WinMain(HINSTANCE__*, HINSTANCE__*, CHAR*, int)':

cannot convert `const wchar_t*' to `const CHAR*' for argument `2' to `int MessageBoxA(HWND__*, const CHAR*, const CHAR*, UINT)'



Alright, I've been using c++ in DOS (i think) so I need some guidance here.

Thanks in advance.

[Cucumber]

When you use the L literal before a string in your code, the compiler understands you are willing to create an Unicode string (wchar_t *) rather than an ANSI string (char *).

The function MessageBox, which is secretly defined as MessageBoxA inside the windows headers in this case, must receive ANSI strings as parameters, not Unicode strings. Thats why the compiler complains.

If you want your code to compile either remove the L literals to use ANSI strings or tell the compiler you are creating an Unicode Win32 application using

#define UNICODE 1
before including the windows.h header.

When defining such preprocessing variable, the windows header will secretly define the function MessageBox as MessageBoxW, such function receive Unicode strings as parameters.

[weaknessforcats]

#define UNICODE 1
before including the windows.h header.

You should not do this. To undefine it, you have to change your code. Instead there is a character set property setting for your project. Set that propertry to the correct value.

When defining such preprocessing variable, the windows header will secretly define the function MessageBox as MessageBoxW, such function receive Unicode strings as parameters.

There is no secret definition. MessageBox is a macro. Based on the character set, the macro resolves to MessageBoxA or MessageBoxW. Your code with the L in front of the literal will not work with MessageBoxA. Further, the L specifes a string of wchar_t and these characters are not Unicode characters unless they contain Unicode values.

Microsoft provides a series of mappings called the TCHAR mappings that you should be using. These macros correctly switch between ASCII and Unicode without your having to change your code.

You #include <tchar.h>

The message bis code should be:

 MessageBox(NULL,
               TEXT("Hello World!"),
               TEXT("Just another Hello World program!"),
               MB_ICONEXCLAMATION | MB_OK);
 

Read this. Be sure to follow the links.

[Cucumber]

I know its not a secret definition, just wanted to keep the explanation clear without going into the macro declaration details.

[Hunderpanzer]

Arghh.


I still can't get it to work.


I reading an online DirectX tutorial, but nothing makes sense or works when I type it in.


WHY is DirectX sooo complicated ?!

[oler1s]

DirectX uses COM. If you aren’t familiar with the paradigms COM introduces, either because you simply haven’t seen COM in any form, or just don’t have that much experience, then DirectX will be overwhelming initially.

[Hunderpanzer]

DirectX uses COM. If you aren’t familiar with the paradigms COM introduces, either because you simply haven’t seen COM in any form, or just don’t have that much experience, then DirectX will be overwhelming initially.

It's extremely overwhelming. And yeah, I haven't seen COM before.

How do I get more experience with COM ? Online tutorials ? Books ?

[Cucumber]

COM+ Programming: A Practical Guide Using Visual C++ and ATL
Pradeep Tapadiya
Publisher: Prentice Hall PTR

That book is the best COM book I have come across.
Yes, I know the tittle says "COM+", but the whole first part of the book is dedicated to explain the COM basis (what is it and why was designed the way it is).

[Hunderpanzer]

COM+ Programming: A Practical Guide Using Visual C++ and ATL
Pradeep Tapadiya
Publisher: Prentice Hall PTR

That book is the best COM book I have come across.
Yes, I know the tittle says "COM+", but the whole first part of the book is dedicated to explain the COM basis (what is it and why was designed the way it is).

I see. Thanks a lot for your help.

[Ninjaamos]

I don't know if it's proper because I am fairly new to C++ as well as COM but the method I used was simply to remove the L (literal?) declaration.
So instead of:
MessageBox(NULL,
L"Hello World!",
L"Just another Hello World program!");
MB_ICONEXCLAMATION | MB_OK);
Just use:
MessageBox(NULL,
"Hello World!",
"Just another Hello World program!");
MB_ICONEXCLAMATION | MB_OK);
Once again, not sure if it's proper but it is functional.

[Ninjaamos]

As a side note I was having the same problem exactly until I read some of the replies here, so thank you all.

[weaknessforcats]

No one so far on this thread has mentioned TCHAR.

TCHAR is a macro that switches betrween char and wchar_t based on the character set used by the compiler. When you set Visual Studio.NET to "Use Multi-byte character set", TCHAR is a char. When you set Visual Studio.NET to "Use Unicode character set", TCHAR is a wchar_t.

Using an L in front of a literal makes the literal a wchar_t but this does not support Unicode. Use the TEXT or _T macros.

MessageBox is itself a macro that swtiches between MessageBoxA and MessageBoxW.

I really suggest tou research TCHAR in MSDN before writing a lot fo code.