如何确认C中unsigned char的范围？

Question

The unsigned char for my compiler contains 1 byte.

So as per my knowledge the range of unsigned char is from 0 to 2^8 - 1 which is 255.

Now I want to confirm the max integer unsigned char can display is 255.

The following is my code:

#include <stdio.h>
#include <limits.h>
int main(void)
{
    printf("unsigned char has %d bytes.\n", sizeof(unsigned char));

    unsigned char a;

    a = 0;

    printf("%d\n", a - 1); 
    printf("%u\n", a - 1); 
    printf("%lu\n", a - 1); 
    printf("%llu\n", a - 1);   
    printf("%zu\n", a - 1); 

    return 0;
}

The output is

unsigned char has 1 bytes.
-1
4294967295
4294967295
4294967295 
4294967295

4294967295 is 2^32 - 1

Where did I go wrong?

Answer 1

First of all, when you pass an integer argument smaller than int to a variadic function (like printf ) then it's automatically promoted to an int .

Secondly, when you use arithmetic, smaller integer types again are promoted to int .

Read eg this implicit conversion reference for more information.

If you want to print the char (or rather unsigned char ) value you need to use the hh prefix for the format specifier (see eg this printf and family reference ):

printf("%hhu\n", (unsigned char) (a - 1));

Answer 2

a-1

This is arithmetic operation and implicit conversion takes place. Compiler will treat it as integer. Try to assign -1 to unsigned type, then check the value. You should also notice the printf operands, like %d, which for that particular example, expects integer.

Answer 3

Your knowledge is incomplete in subtle ways:

The unsigned char for my compiler contains 1 byte.

Type unsigned char is by definition exactly one byte for all compilers.

So as per my knowledge the range of unsigned char is from 0 to 2^8 - 1 which is 255 .

Not exactly: the minimum number of value bits for unsigned char is 8 and no padding bits are allowed for this type, which means the minimum size for a byte is 8 bits. Indeed most architectures nowadays standardize on 8-bit bytes, but the C Standard allows for other, larger, bit widths, and some micro-controllers use 16-bit bytes (or even 32-bit bytes). On such processors, unsigned char as 16 bits (or even 32 bits).

Now I want to confirm the max integer unsigned char can display is 255 .

This is a genuine concern that can be validated both at compile time and at runtime.

The following is my code:

 #include <stdio.h> #include <limits.h> 

Including <limits.h> is a good approach: UCHAR_MAX is the value you want to print or even test directly in a preprocessing directive.

 int main(void) { printf("unsigned char has %d bytes.\\n", sizeof(unsigned char)); 

Undefined behavior: %d expects an argument of type int , which is different and potentially incompatible with sizeof(unsigned char) that has type size_t . Use %zu or cast the argument as (int)sizeof(unsigned char) . Note however that sizeof(unsigned char) by definition always evaluates to 1 with type size_t .

  unsigned char a; a = 0; printf("%d\\n", a - 1); 

a - 1 always evaluates to -1 as a is promoted to int and 0 - 1 has a value of -1 . You want instead to print (unsigned char)(a - 1) , or simply (unsigned char)-1 .

  printf("%u\\n", a - 1); 

This will print the value of (unsigned int)-1 , which is UINT_MAX , not UCHAR_MAX .

  printf("%lu\\n", a - 1); printf("%llu\\n", a - 1); printf("%zu\\n", a - 1); 

The 3 printf calls above have potential undefined behavior because you pass an int value for an argument whose expected byte is respectively unsigned long , unsigned long long and size_t .

  return 0; } 

Here is a simpler approach:

#include <stdio.h>
#include <limits.h>

int main() {
    printf("UCHAR_MAX = %u\n", UCHAR_MAX);
    printf("(unsigned char)-1 = %u\n", (unsigned char)-1);
    printf("-1 with %%hhu conversion: %hhu\n", -1);
    return 0;
}

Output:

UCHAR_MAX = 255
(unsigned char)-1 = 255
-1 with %hhu conversion: 255

Answer 4

Your code invoked undefined behaviour because You have used wrong format specifier of printf.

GCC Compiler warnings:

prog.c: In function 'main':
prog.c:5:32: warning: format '%d' expects argument of type 'int', but argument 2 has type 'long unsigned int' [-Wformat=]
     printf("unsigned char has %d bytes.\n", sizeof(unsigned char));
                               ~^            ~~~~~~~~~~~~~~~~~~~~~
                               %ld
prog.c:13:15: warning: format '%lu' expects argument of type 'long unsigned int', but argument 2 has type 'int' [-Wformat=]
     printf("%lu\n", a - 1);
             ~~^     ~~~~~
             %u
prog.c:14:16: warning: format '%llu' expects argument of type 'long long unsigned int', but argument 2 has type 'int' [-Wformat=]
     printf("%llu\n", a - 1);
             ~~~^     ~~~~~
             %u
prog.c:15:15: warning: format '%zu' expects argument of type 'size_t', but argument 2 has type 'int' [-Wformat=]
     printf("%zu\n", a - 1);
             ~~^     ~~~~~
             %u

C11 $7.19.6.1 p9 :

[...] If any argument is not the correct type for the corresponding conversion specification, the behavior is undefined.