decimal constant so large?

decimal constant so large?

[a.biardi]

Hi everybody,

Can anyone tell me why the constant -2147483648, according to what g++ reports 
as a warning for the code below, is considered to be "so large that it is 
unsigned"?

How can it be unsigned when it starts with '-'?

I am compiling on an x86, and the constant seems to be a perfectly good 32 bit 
signed value to me. Where am I wrong?

I've tried it with three different compilers, and they all behave in the same 
funny way: Borland C++ compiler 5.7, as well as g++ 3.2.3 report a warning, 
Intel compiler 8.1 says nothing.

==============================================================================
#include <iostream>

using namespace std;

int main() {

  cout << (int)-2147483648 << endl; // ok
  cout <<      -2147483648 << endl; // prints 2147483648 instead

}
==============================================================================

Re: decimal constant so large?

[Glynn Clements]

a.biardi@tiscali.it wrote:

> Can anyone tell me why the constant -2147483648,

-2147483648 isn't a constant, it's an expression consisting of the
negation operator "-" applied to the decimal constant "2147483648".

The decimal constant 2147483648 is too large to fit into a signed
32-bit integer, so it is treated as unsigned.

If you want the most negative signed 32-bit integer, you have to write
it as (-2147483647 - 1).

-- 
Glynn Clements <glynn@gclements.plus.com>

Training

[Darren Sessions]

I am looking for a training establishment in the South East (United  
States) that offers beginners to advanced training in C with a bang  
for the buck.

Anyone have any suggestions?

Thanks,

  - Darren