Defning SIZE_MAX

[ozbear]

In another thread related to size_t and SIZE_MAX, a #define
was suggested for SIZE_MAX where is it not predefined.

The #define was:

#define MY_SIZE_MAX ((size_t)-1)

I have tried this on my system in a printf and it yields a
reasonable value but I do not fully understand how it it
evaluated. size_t is a /type/, defined on my system as

typedef unsigned int size_t;

The part I do not understand is the the meaning of subtracting 1
from a /type/ rather than an expression? How does the compiler
evaluate this?

Oz

 

[Ian Collins]

In another thread related to size_t and SIZE_MAX, a #define
was suggested for SIZE_MAX where is it not predefined.

The #define was:

#define MY_SIZE_MAX ((size_t)-1)

I have tried this on my system in a printf and it yields a
reasonable value but I do not fully understand how it it
evaluated. size_t is a /type/, defined on my system as

typedef unsigned int size_t;

The part I do not understand is the the meaning of subtracting 1
from a /type/ rather than an expression? How does the compiler
evaluate this?

It's not a subtraction, it's a cast!

 

[Keith Thompson]

In another thread related to size_t and SIZE_MAX, a #define
was suggested for SIZE_MAX where is it not predefined.

The #define was:

#define MY_SIZE_MAX ((size_t)-1)

I have tried this on my system in a printf and it yields a
reasonable value but I do not fully understand how it it
evaluated. size_t is a /type/, defined on my system as

typedef unsigned int size_t;

The part I do not understand is the the meaning of subtracting 1
from a /type/ rather than an expression? How does the compiler
evaluate this?

As Ian Collins already posted, it's a cast, not a subtraction.

Looking at the expression in painful detail from the inside out:

``1'' is an integer constant.

``-1'' is an expression consisting of a unary "-" applied to ``1''.
It's value is -1, and its type is int.

``(size_t)-1'' is a cast expression, specifying that the result of the
expression ``-1'' is to be converted to type size_t. Given the the
rules for converting a signed int value to an unsigned type, the
result is the maximum value of type size_t.

Finally, we put parentheses around the whole thing because it's a
macro definition.

 

[ozbear]

D'OH!
Now I feel like an idiot. Of course it's a cast....
A perfect example of not being able to see what something
really is after you've made up your mind what it is.

Many thanks to all that replied.

Oz