problem understanding a paragraph in c++ primer

[nirbh]

Unlike the other integer types, there are three distinct basic character types:

char,signed char, and unsigned char. In particular, char is not the same type

as signed char. Although there are three character types, there are only two

representations: signed and unsigned. The (plain) char type uses one of these

representations. Which of the other two character representations is equivalent

to char depends on the compiler.

can someone please explain to me the meaning of this paragraph?

thank you very much in advance!

[tpb]

Firstly, remember that chars are just little integers, usually byte-sized. So on my system (and probably yours, too), these are the sizes of the basic integer types:

                       guaranteed by the
                        standard to be
         bytes  bits        at least
char       1      8            1
short      2     16            2
int        4     32            2
long       8     64            4
long long  8     64            8

All the basic integer types except char are assumed to be signed if you don't say "signed" or "unsigned". E.g., signed int is exactly the same as int (so there is never really a reason to say signed int, and signed int is not a distince type, just a synonym for int). unsigned int is the unsigned version of int.

However, for char it's different. Because of possible differences between different computer systems, it may be best to have char (the plain char type, without "signed" or "unsigned" in front of it) as equivalent to an unsigned char instead of signed char. Even if char happens to be signed on your system (as it is on mine), to write portable code you can't assume that it's signed.

Therefore there is a unique meaning to saying "signed char" (unlike saying "signed int" which is identical to saying "int"), so chars actually have 3 distinct types while short, int, long, long long only have 2 distinct types. That is, the type "char" is different from the type "signed char", even if it's actually the same!

If you convert the char to an int you can see whether char is signed or unsigned on your system:

#include <iostream>
int main() {
    char ch = 0x7f; // highest 7-bit value 0111 111 (which is highest positive value if char is signed)
    char ch2 = ch;

    ++ch2; // overflow the positive value of signed char (if char is signed char)
    // Technically the above is "undefined behavior" so according to the standard
    // we can't rely on what happens. But the usual thing is for the value
    // to "wrap around" to the most negative value.

    // if ch2 is negative then it overflowed and char is signed char
    std::cout << int(ch) << ", " << int(ch2) << '\n';
}

[JLBorges]

1. These are three different types: char, signed char, unsigned char

2. In a C++ implementation, the representation of objects of type char would be identical to one of
either signed char or unsigned char,
ie. in some implementations, the representation of objects of type char would be identical to that of signed char
in others, the representation of objects of type char would be identical to that of unsigned char

What this implies is that in our code, we always use char to represent a character.

We would resort to either signed char or unsigned char only when we want to interoperate with another program (or with hardware); say, we want to send character objects from our program to another program running on a different machine.

[nirbh]

Thank you tpb.
Thank you JLBorges.