Are C Enumerations always Long data type

[Stefan Schnell]

Hello community,

the following listing shows a small C code with an enumeration, an assignment of value in the main procedure and the disassembled code.

//-Begin----------------------------------------------------------------
 
  typedef enum _RFCTYPE
  {
    RFCTYPE_CHAR   = 0,
    RFCTYPE_DATE   = 1,
    RFCTYPE_BCD    = 2,
    RFCTYPE_TIME   = 3,
    RFCTYPE_BYTE   = 4,
    RFCTYPE_TABLE  = 5,
    RFCTYPE_NUM    = 6,
    RFCTYPE_FLOAT  = 7,
    RFCTYPE_INT    = 8,
    RFCTYPE_INT2   = 9,
    RFCTYPE_INT1   = 10,
    RFCTYPE_NULL   = 255
  }RFCTYPE;
 
  int main (void) {
    RFCTYPE i = RFCTYPE_NULL;
  }
 
/*
55                      push ebp
89E5                    mov ebp, esp
81EC04000000            sub esp, 00000004
90                      nop
B8FF000000              mov eax, 000000FF
8945FC                  mov dword[ebp-04], eax
C9                      leave
C3                      ret
*/
 
//-End------------------------------------------------------------------

Are this kind of enumerations always from the data type LONG resp. DWORD? Or is it possible that they may be of a different data type?

Thanks for tips.

Cheers
Stefan

[Kurt Kuzba]

Are this kind of enumerations always from the data type LONG resp. DWORD? Or is it possible that they may be of a different data type?

The C ANSI Standard requires enumerations to be of type INTEGER.

[Michael Mattias]

Enumerations don't have a "data type"; they are representations of a value

How you store any value - the 'data type' in your compiler or interpreter - is a function of that compiler.

Consider this line of PowerBASIC code:

%EQUATE = 2

The value two could be stored in a byte, a 16, 32, or 64-bit integer, or in an IEEE float.

So how is it stored? Yes, the answer is in the help file under "equates."'

Can you, the programmer, change how a value is stored? Yes, and that's explained in the same 'equate' section of the help ... but it is easier to show:

  A! =  2
  B# = 2 
  C& = 2 
 ...

How is the value "two" stored - that is, what is its data type - in this program?

MCM

[Kurt Kuzba]

Enumerations don't have a "data type"; they are representations of a value

How you store any value - the 'data type' in your compiler or interpreter - is a function of that compiler.

And in a C compiler, the ANSI Standard says that you can store an INTEGER.
If the value exceeds that of the type allowed, the result is undefined.

[Stefan Schnell]

Hello Kurt and Michael,

thanks for your replies.

Cheers
Stefan

[Chris Holbrook]

And in a C compiler, the ANSI Standard says that you can store an INTEGER.

er, no.

The ANSI standard for C does not define an INTEGER but does define an int, having the storage defined for an int by the compiler, generally 32 bits wide in 32-bit architectures. Note: PowerBASIC's INTEGER datatype is 16 bits wide.

There are many references on the web, here's one from M$: http://msdn.microsoft.com/en-us/library/whbyts4t.aspx

[Chris Holbrook]

Just re-read the first post and the OP mentions "small C". Did he mean "Small C"? "Small C" is an ANSI C subset for 8-bit machines. I think that this compiler's int datatype is 16 bits wide.

And in "Tiny C" it's 8 bits, as you may have guessed. *** Added *** oh, cancel that statement. That was in 1988. Another Tiny C has sprung up in the intervening years, for X86. I doubt if it uses 8-bit ints! Here's a link: http://bellard.org/tcc/

[Michael Mattias]

Chris' point - that 'int' means something specific to the 'c' language - is a good one.

Not that it's any different from saying 'CUR' or LONG mean something specific to the PowerBASIC dialect of the BASIC language.

Just something to think about when wondering why your mindless verb-for-verb, function-for-function, statement-for-statement conversion from one source code language to another "don't work."