ASCII and printf

I have a little (big, dumb?) question about int and chars in C. I rememeber from my studies that "chars are little integers and viceversa," and that's okay to me. If I need to use small numbers, the best way is to use a char type.

But in a code like this:

#include <stdio.h>
#include <stdlib.h>
int main(int argc, char *argv[]) {
  int i= atoi(argv[1]);
  printf("%d -> %c\n",i,i);
  return 0;
}

I can use as argument every number I want. So with 0-127 I obtain the expected results (the standard ASCII table) but even with bigger or negative numbers it seems to work...

Here is s开发者_C百科ome example:

-181 -> K
-182 -> J
300 -> ,
301 -> -

Why? It seems to me that it's cycling around the ascii table, but I don't understand how.

When you pass an int corresponding to the "%c" conversion specifier, the int is converted to an unsigned char and then written.

The values you pass are being converted to different values when they are outside the range of an unsigned (0 to UCHAR_MAX). The system you are working on probably has UCHAR_MAX == 255.

When converting an int to an unsigned char:

• If the value is larger than UCHAR_MAX, (UCHAR_MAX+1) is subtracted from the value as many times as needed to bring it into the range 0 to UCHAR_MAX.
• Likewise, if the value is less than zero, (UCHAR_MAX+1) is added to the value as many times as needed to bring it into the range 0 to UCHAR_MAX.

Therefore:

(unsigned char)-181 == (-181 + (255+1)) == 75 == 'K'
(unsigned char)-182 == (-182 + (255+1)) == 74 == 'J'
(unsigned char)300  == (300 - (255+1))  == 44 == ','
(unsigned char)301  == (301 - (255+1))  == 45 == '-'

The %c format parameter interprets the corresponding value as a character, not as an integer. However, when you lie to printf and pass an int in what you tell it is a char, its internal manipulation of the value (to get a char back, as a char is normally passed as an int anyway, with varargs) happens to yield the values you see.

My guess is that %c takes the first byte of the value provided and formats that as a character. On a little-endian system such as a PC running Windows, that byte would represent the least-significant byte of any value passed in, so consecutive numbers would always be shown as different characters.

You told it the number is a char, so it's going to try every way it can to treat it as one, despite being far too big. Looking at what you got, since J and K are in that order, I'd say it's using the integer % 128 to make sure it fits in the legal range.

Edit: Please disregard this "answer".

Because you are on a little-endian machine :) Serously, this is an undefined behavior. Try changing the code to printf("%d -> %c, %c\n",i,i,'4'); and see what happens then...

When we use the %c in printf statement, it can access only the first byte of the integer. Hence anything greater than 256 is treated as n % 256.

For example i/p = 321 yields op=A

What atoi does is converting the string to numerical values, so that "1234" gets 1234 and not just a sequence of the ordinal numbers of the string.

Example:

char *x = "1234";  // x[0] = 49, x[1] = 50, x[2] = 51, x[3] = 52 (see the ASCII table)
int y = atoi(x); // y = 1234
int z = (int)x[0];  // z = 49 which is not what one would want