Thread: Questions regarding usage of #undef and #ifdef

Lets say I've got a three module source (File1.c, File2.c and File3.c) and I use #undef at a global level in one of the modules like shown in Example A. Will this redefinition effect the other modules?

Example A:

Code:

/* File2.c */

typedef signed char MYBYTE;

#undef LOBYTE
#undef HIBYTE

#define LOBYTE(w)  ((MYBYTE)(int)(w))
#define HIBYTE(w)  ((MYBYTE)(((int)(w) >> 8) & 0xFF))

Scenerio two. Lets say I do a #ifdef as shown below in case WM_SIZE under WinMain(). Will these redefinitions be confined to just case WM_SIZE?

Code:

case WM_SIZE:
    {
        #ifdef LOBYTE
        #ifdef HIBYTE

        #undef LOBYTE
        #undef HIBYTE
        #define LOBYTE(w)  ((signed char)((signed) w & 0x00FF)
        #define HIBYTE(w)  ((signed char)(((signed) w & 0xFF00) >> 8))

        /* .... */  // rest of code

        #endif
        #endif
    }
    break;

BTW, These are just made up scenerios.

In Example A i could see that none of the other modules would be effected by the redefinition of LOBYTE and HIBYTE since like you mentioned there all compiled separately. Now, in scenerio two I would expect the redefinition of LOBYTE and HIBYTE not to effect anything else within the module except for case WM_SIZE; correct?

Originally Posted by TAZIN

In Example A i could see that none of the other modules would be effected by the redefinition of LOBYTE and HIBYTE since like you mentioned there all compiled separately. Now, in scenerio two I would expect the redefinition of LOBYTE and HIBYTE not to effect anything else within the module except for case WM_SIZE; correct?

It's still no to both. #define is just search-and-replace. The pre-processor doesn't even know what a brace is, let alone anything about scope.

Originally Posted by TAZIN

In Example A i could see that none of the other modules would be effected by the redefinition of LOBYTE and HIBYTE since like you mentioned there all compiled separately. Now, in scenerio two I would expect the redefinition of LOBYTE and HIBYTE not to effect anything else within the module except for case WM_SIZE; correct?

AdeyBlue has it right... The C preprocessor doesn't know about the delineated scopes in a program, and it doesn't care... at any point after you redefine those macros the preprocessor will see them as changed... even hundreds of function calls later...

It operates a lot like HTML does... it sees a B and until it sees /B it will bold face everything in site... when the preprocessor sees a definition it will hold that definition until you change it.

The preprocessor is not part of C, per se... it's a separate thing with it's own set of rules.

So, in scenerio two the redefinition of LOBYTE and HIBYTE will effect the whole module.

Now, lets say that you wanted to go back to the precompiled header version of LOBYTE and HIBYTE in scenerio two. What's the best way to do that? Would you have to do what's shown below?

Code:

typedef unsigned char BYTE;

#undef LOBYTE
#undef HIBYTE

#define LOBYTE(w)           ((BYTE)(w))
#define HIBYTE(w)           ((BYTE)(((WORD)(w) >> 8) & 0xFF))

No, you'd call your signed version something else so you wouldn't have to do these dumb hacks in the first place. That way the things you know about the preprocessor and the things you don't can't conspire to use one version when you want the other one.

Thanks guys for your input.

Originally Posted by TAZIN

So, in scenerio two the redefinition of LOBYTE and HIBYTE will effect the whole module.

Everything below your definition in the file will be affected.