Why main function cannot return a negative number?

Question

This might be a really simple question for some, but I'm new to C++ and hope someone can answer this for me.

I'm using this online C++ compiler. Here's the simple code I'm running in it:

int main()
{
    int x = 1- 2;
    std::cout << x << std::endl;
    return x;
}

The output is:

-1

...Program finished with exit code 255
Press ENTER to exit console.

That really ponders me. Why would the main() function return 255 when the value of x is -1?

Doesn't main() return an int (not an unsigned int), so it should be able to return a negative number, right?

How does -1 get converted to 255? Something to do with an 8-bit variable? But isn't the int type 16-bit?

***but isn't int type 16 bit?*** It's usually larger than that for a modern CPU. Although it may be 16 bit for micro controllers. — drescherjm, Oct 18 '19 at 02:47
You might get your answer here: [Why do I get error 255 when returning -1](https://unix.stackexchange.com/questions/37915/why-do-i-get-error-255-when-returning-1) — Gilles-Philippe Paillé, Oct 18 '19 at 02:54
The return type is indeed `int`, but on unix systems a programs return value is an unsigned 8 bit integer regardless, so 0 through 255. — Jesper Juhl, Oct 18 '19 at 03:03
@phuclv on Windows, a process's exit code is a 32bit unsigned int. See [`ExitProcess()`](https://learn.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-exitprocess) and [`GetExitCodeProcess()`](https://learn.microsoft.com/en-us/windows/win32/api/processthreadsapi/nf-processthreadsapi-getexitcodeprocess). — Remy Lebeau, Oct 18 '19 at 06:30
@RemyLebeau in cmd it's a signed 32-bit int. [`echo %errorlevel%` after `cmd /c "exit 4000000000"`returns -294967296](https://stackoverflow.com/a/5149267/995714#comment51200906_5149267). *"Windows uses 32-bit unsigned integers as exit codes, although the command interpreter treats them as signed"* [Does Windows shell %ERRORLEVEL% variable have a character limit?](https://stackoverflow.com/a/41153463/995714) — phuclv, Oct 18 '19 at 06:39

score 1 · Accepted Answer · answered Oct 18 '19 at 07:11

This is not related to C language really. The operating system, or possibly just the C runtime (the small piece of the code which sets up things for your C program and actually calls your main function) limits exit code of the program to unsigned 8 bit number.

Very nearly all systems today use two's complement representation for negative numbers, and then bit pattern for -1 is having all bits of the number to be 1. Doesn't matter how many bits, they are all set when value is -1.

The simplest way to convert an int to 8 bit number is to just take 8 lowest bits (which are now all 1 as per above), so you end up with binary number:

11111111

If interpreted as unsigned, then in decimal value of this happens to be 255 (as signed 8 bits it is still -1), which you can check with any calculator which supports binary (such as Windows 10 Calculator app when you switch it to Programmer mode).

Looking at this from the opposite direction: When trying to understand funny numbers related to computers or programming, it is often useful to convert them to binary. If you convert 255 to binary, you get 11111111, and then if you know binary numbers, you should realize this is -1 if interpreted as signed 8 bit number.

Why main function cannot return a negative number?

1 Answers1