Subj : Re: enumerated variables - how to coerce them to anything else than an Int ?
To   : borland.public.cpp.borlandcpp
From : John F\(dot\)
Date : Tue Oct 11 2005 02:08 am


Hi Rudy!

before you try any hints, read till the end and decide then what you want to 
try... :-)

"R.Wieser" wrote:
> John F(dot) wrote:
> [snip]
>> Try including a value larger than maximum int.
>
> I just tried it.  Guess what : the compiler complained about the value not
> fitting in an Int. :-\

Then you are out of luck with that compiler... not possible at all.

>> OK so far. what are the values for def, y, n ?
>
> My intention (with the declaration I showed) was to give them the values 
> 0,
> 1 and 2 (although I did not yet check if that really happens).

It does happen. One usually writes

enum {
    first_element = 0, //explicitly
    second, //first_element+1 = 1
    third  //second +1 = 2
};
so that it is clear where it starts...

Usually 0 is considered false and anything different from 0 is considered 
true. If you like to do some "if()s" and "||"s or "&&"s you should use 0 for 
_no_.
default should be 0,1 or 2? then it can not be part of the enum. enums are 
fixed.

>> and why weren't they designed to be enums then? :-)
>
> They probably where ....  But the structure I've currently got is only a
> recreation, not the origional.  I made it by looking how changes settings 
> in
> a program changed the stored data.  Yep, a bit of reverse engeneering :-)
>
> [snip]
>> Are you allowed to change the structure??? How comes that?
>
> It's an old program for which I want to be able to create different 
> setups,
> without manually having to select them in the program itself.   A
> template-generator if you will.

Preprocessing?

#define PORT_DEFAULT     -1
#define PORT_COM1            0
#define PORT_COM2            1
#define PORT_LPT1               2
....
#define PORT_USED             (PORT_DEFAULT)
....

>> Do you still need to store arbitrary longs into these member-variables?
>
> Alas, yes.  I can't change anything about the program itself.

OK.

> The only thing I *could* do is to split a number of fields of the type 
> Long
> into two fields of the type Int.  That would most-of-the-time work, as 
> most
> enumerations are done with small positive numbers.

No good idea. No portability - little/big endian are looking over the 
shoulder here...

> But I would like to find a method that is of more generic use (for Ints as
> wel as other types).

Unions?

union {
long MyLong,
enum E MyEnum
}X;

X.E=default;
X.MyLong=1234567;

>> Could you give a short snippet of the code you need to be done?
>
> Well, currently it's just a structure with some fields.  And some of those
> fields only may have (the numerical equivalent) of certain settings (like
> "default", "yes" or "no").
>
> struct {
>  UseLpt=default; -- default, com1, com2, lpt1, lpt2, lpt3
>  ComHandshake=none; -- none, hardware, xonxoff
>  Baudrate=9600; -- 50, 75, 110, 150, 1200, 2400, 9600
> -- etc.
> } Settings;

that is the job for plain enums... or preprocessing. there will hardly ever 
be 2^32 ports... (as long as i'll be still alive...)
Baudrate could be long (115000 e.g.)

In that case I'd use preprocessing...

set_vals.hxx:
#define PORT_LPT1 1
....

settings.hxx:
#include set_vals.hxx
#define PORT_USED (PORT_LPT1)
....

program.cxx
struct ...
UseLPT = PORT_USED; //settings.h
....

it doesn't perform typechecking but it works...

regards
John

.