Method: Fixnum#bit_length

Defined in:

numeric.c

#bit_length ⇒ Integer

Returns the number of bits of the value of int.

“the number of bits” means that the bit position of the highest bit which is different to the sign bit. (The bit position of the bit 2**n is n+1.) If there is no such bit (zero or minus one), zero is returned.

I.e. This method returns ceil(log2(int < 0 ? -int : int+1)).

(-2**12-1).bit_length     #=> 13
(-2**12).bit_length       #=> 12
(-2**12+1).bit_length     #=> 12
-0x101.bit_length         #=> 9
-0x100.bit_length         #=> 8
-0xff.bit_length          #=> 8
-2.bit_length             #=> 1
-1.bit_length             #=> 0
0.bit_length              #=> 0
1.bit_length              #=> 1
0xff.bit_length           #=> 8
0x100.bit_length          #=> 9
(2**12-1).bit_length      #=> 12
(2**12).bit_length        #=> 13
(2**12+1).bit_length      #=> 13

This method can be used to detect overflow in Array#pack as follows.

if n.bit_length < 32
  [n].pack("l") # no overflow
else
  raise "overflow"
end

Returns:

• (Integer)

# File 'numeric.c', line 3815

static VALUE
rb_fix_bit_length(VALUE fix)
{
    long v = FIX2LONG(fix);
    if (v < 0)
        v = ~v;
    return LONG2FIX(bit_length(v));
}