Correct binary representation for PHP_INT_MAX + 1 with whom make bitwise operations in PHP?

Question

According to this post -> How to get binary representation of floating-point number in PHP?

the function floatToBinStr($value) is used to represent floats/doubles in binary in PHP. If I call it with this parameter:

floatToBinStr(PHP_INT_MAX + 1);

I get this binary representation:

0100001111100000000000000000000000000000000000000000000000000000

64 bits, the left-most bit is the 63-rd bit, as the 64-th bit is used for the negative floats, everything seem to work fine.

I have a function that returns the position of the left-most bit set to one given a number as parameter, and I noticed that if I pass values bigger than PHP_INT_MAX (e.g. )

 public function getPositionOfLeftMostBitSetTo1($n) {
    $masks = array(0x1, 0x2, 0xC, 0xF0, 0xFF00, 0xFFFF0000, 0xFFFFFFFF00000000);
    $positions = array(1, 2, 2, 4, 8, 16, 32);
    $r = ($n < 0) ? -1 : 0);
    for ($i = PHP_INT_SIZE == 8 ? 6 : 5; $i >= 0; $i--)
    {
      if ($n & $masks[$i]) {
            $n >>= $positions[$i];
            $r += $positions[$i];
        }
    }
    return $r;
 }
 echo getPositionOfLeftMostBitSetTo1(PHP_INT_MAX + 1);

I get 65 as output. If I try this:

echo decbin(PHP_INT_MAX + 1); // PHP_INT_MAX + 1 = 9.22337203685E+18

I get this

1000000000000000000000000000000000000000000000000000000000000000

So it looks like the left-most bit is the 64-th, though this is wrong cause this is not a negative float and my function doesn't return properly and instead returns 65 because it acts like it is right-shifting on a negative int...

How can I make PHP recognize the floats/doubles bigger than PHP_INT_MAX as they should and perform safely bitwise operations on them?

Thanks for attention!

score 0 · Accepted Answer · answered Nov 10 '14 at 14:12

I resolved by converting the float into binary with the floatToBinStr($value) function and then I used bindec to convert that binary representation into an integer on which I can shift without any problem, in the end the purpose of getPositionOfLeftMostBitSetTo1() is to return the position of the leftmost bit set to one, nothing else, so converting that binary float into whatever integer it would be did the job.

Correct binary representation for PHP_INT_MAX + 1 with whom make bitwise operations in PHP?

1 Answers1