API functions that return BOOL

[sinsi]

Code Select Expand

BOOL GetFileSizeEx(
  [in]  HANDLE         hFile,
  [out] PLARGE_INTEGER lpFileSize
);
When testing the return value (for any BOOL return) I have been using EAX (because "everything is a DWORD" ).
After my playing with SQL and VB.net, I am more aware of data types and sizes.
From what I can find out, a BOOL in C++ is one byte, so should I be using AL instead of EAX?

[Greenhorn]

There is a difference between C/C++ bool and WinAPI BOOL.

In Windows headers BOOL is defined as DWORD, in opposite the C/C++ bool is not guaranteed to be 1 byte in size.

Boolean type

    bool — type, capable of holding one of the two values: true or false. The value of sizeof(bool) is implementation defined and might differ from 1.

Fundamental types

[sinsi]

Makes a difference when you get an idea of where to look.
I didn't think of the windows headers

Just to muddy the waters for non-C people,
 - WinDef.h has typedef int BOOL;
 - WinNT.h  has typedef BYTE BOOLEAN;

Cheers for that 

[TimoVJL]

Also

Type: _VARIANT_BOOL A 16-bit Boolean value. A value of 0xFFFF (all bits 1) indicates true; a value of 0 (all bits 0) indicates false. No other values are valid.

[daydreamer]

Test lowest bit and conditional jumps works independent of bool size
Use bool to Mask dword result ,dword size bool is enough

[sinsi]

Test lowest bit and conditional jumps works independent of bool size

[jj2007]

Test lowest bit and conditional jumps works independent of bool size

If the function succeeds, the return value is nonzero.

M$ does not say "returns TRUE or FALSE". So GetFileSizeEx could return 2 or 100, and your conditional jump would be screwed. In practice, they all return 0 or 1 in eax, so it's ok - but that behaviour is not documented ;-)

One interesting point is this:

Code Select Expand

BOOL GetFileSizeEx(
  [in]  HANDLE        hFile,
  [out] PLARGE_INTEGER lpFileSize
);
So it returns a BOOL. What is a BOOL, and how is it different from bool or BOOLEAN?

Windef.h:

Code Select Expand

typedef int                BOOL;
typedef BOOL near          *PBOOL;
typedef BOOL far            *LPBOOL;
typedef int                INT;
typedef unsigned int        UINT;
Crystal clear, isn't it?

P.S., just in case you are not yet thoroughly confused:

bool (C++)

A variable of this type can have values true and false. Conditional expressions have the type bool and so have values of type bool. For example, i != 0 now has true or false depending on the value of i.

Visual Studio 2017 version 15.3 and later (Available with /std:c++17 and later): The operand of a postfix or prefix increment or decrement operator may not be of type bool. In other words, given a variable b of type bool, these expressions are no longer allowed:

b++;
    ++b;
    b--;
    --b;

The values true and false have the following relationship:

!false == true
!true == false

And...

The bool type participates in default integral promotions

C/C++ at its best - you must love it

P.P.S., just for the record: uppercase BOOL is a DWORD, i.e. test eax, eax is perfectly valid

[NoCforMe]

So to sum up, the foolproof way to test a BOOL would be to test for zero ("FALSE"), which is actually easier than the alternative:

Code Select Expand

    TEST    EAX, EAX
    JNZ     isTRUE
Amiright?

(Assuming that BOOL == int, of course)