SCJP5 K&B Octal Literals

[Ruwan Scott]

Hi,
question on the following code which appears on Page 177.

  class Octal {
public static void main(String [] args) {
int six = 06; // Equal to decimal 6
int seven = 07; // Equal to decimal 7
int eight = 010; // Equal to decimal 8
int nine = 011; // Equal to decimal 9
System.out.println("Octal 010 = " + eight);
}
}

out put of this comes as 8 ( integer number). Here the 'eight' is a integer variable. How does it know that 010 is in octal form. why doesn't this variable considers 010 as integer 'Ten'. ( This is not about converting octal 010 to 8. I know how it works). Is there a implicit conversion of anything written as 0s and 1s to Integers ?.

Can some one please explain?

Thanks,
Ruwan

[Henry Wong]

When a literal starts with zero, the compiler treats the literal as specified in octal.

How does it know that 010 is in octal form.

And there is no such a thing as an "octal form". An int is a value, it doesn't really have a base. It is the compiler that is taking your source of 010, and interpreting it as octal.

Henry

[Ruwan Scott]

Henry,
thanks for the explanation.
I need another clarification with following, which is on same page.

int z = 0xDeadCafe;

this will give the printing output as z = -559035650. Whats the reason to get a minus (-) sign here?

/Ruwan

[Henry Wong]

int z = 0xDeadCafe;

Well, this is hexidecimal, which I am assuming you know.

this will give the printing output as z = -559035650. Whats the reason to get a minus (-) sign here?

This requires a bit of explanation... With octal, and with hexidecimal, the compiler allows the number to be specified as an unsigned number. This is because octal and hexidecimal numbers are a common technique to specify bit patterns (such as bit masks). In this case, the number 0xDEADCAFE actually overflows the value of a signed int -- but not an unsigned int. It uses the sign bit as the highest order bit in an unsigned int.

Anyway, to understand how a unsigned number maps to it signed counterpart, you need to userstand "twos-complement".

http://en.wikipedia.org/wiki/Two's_complement

Basically, the very big unsigned number is actually a negative number when the bit patterns are treated as a signed number.

Henry

[Ruwan Scott]

Henry,
Thanks. I got it.

/Ruwan