PHP equivalent javascript >>> shift right with zer

问题:

May I know how can I do PHP >>> ? Such operators is not available in PHP, but is available in Javascript.

I just managed to discover a function as follow:

function zeroFill($a, $b) 
{ 
    $z = hexdec(80000000); 
        if ($z & $a) 
        { 
            $a = ($a>>1); 
            $a &= (~$z); 
            $a |= 0x40000000; 
            $a = ($a>>($b-1)); 
        } 
        else 
        { 
            $a = ($a>>$b); 
        } 
        return $a; 
}

but unfortunately, it doesn't work perfectly.

EG: -1149025787 >>> 0 Javascript returns 3145941509 PHP zeroFill() return 0

回答1:

Twice as fast for negative numbers as using the decimal-binary conversions

function zerofill($a,$b) { 
    if($a>=0) return $a>>$b;
    if($b==0) return (($a>>1)&0x7fffffff)*2+(($a>>$b)&1);
    return ((~$a)>>$b)^(0x7fffffff>>($b-1)); 

回答2:

I've researched a lot on this, collected more than 11 versions from StackOverflow and open-source projects, none of them worked. But finally, I found the solution.

For more details, live demo, tests and examples check my question and answer:
Unsigned Right Shift / Zero-fill Right Shift in PHP (Java/JavaScript equivalent)

function unsignedRightShift($a, $b) {
    if ($b >= 32 || $b < -32) {
        $m = (int)($b/32);
        $b = $b-($m*32);
    }

    if ($b < 0) {
        $b = 32 + $b;
    }

    if ($b == 0) {
        return (($a>>1)&0x7fffffff)*2+(($a>>$b)&1);
    }

    if ($a < 0) 
    { 
        $a = ($a >> 1); 
        $a &= 0x7fffffff; 
        $a |= 0x40000000; 
        $a = ($a >> ($b - 1)); 
    } else { 
        $a = ($a >> $b); 
    }

    return $a; 
}

回答3:

/**
 * The >>> javascript operator in php x86_64
 * Usage: -1149025787 >>> 0 ---> rrr(-1149025787, 0) === 3145941509
 * @param int $v
 * @param int $n
 * @return int
 */
function rrr($v, $n)
{
    return ($v & 0xFFFFFFFF) >> ($n & 0x1F);
}

/**
 * The >> javascript operator in php x86_64
 * @param int $v
 * @param int $n
 * @return int
 */
function rr($v, $n)
{
    return ($v & 0x80000000 ? $v | 0xFFFFFFFF00000000 : $v & 0xFFFFFFFF) >> ($n & 0x1F);
}


/**
 * The << javascript operator in php x86_64
 * @param int $v
 * @param int $n
 * @return int
 */
function ll($v, $n)
{
    return ($t = ($v & 0xFFFFFFFF) << ($n & 0x1F)) & 0x80000000 ? $t | 0xFFFFFFFF00000000 : $t & 0xFFFFFFFF;
}

Enjoy it.

回答4:

I studied around the webs and come out with my own zerofill function, base on the explanation given. This method works for my program.

Have a look:

function zeroFill($a,$b) {
    if ($a >= 0) { 
        return bindec(decbin($a>>$b)); //simply right shift for positive number
    }

    $bin = decbin($a>>$b);

    $bin = substr($bin, $b); // zero fill on the left side

    $o = bindec($bin);
    return $o;
}

回答5:

Your function doesn't work because when $b == 0, the expression

$a >> -1

will be evaluated, which returns 0.

Assuming 32-bit machines, you can add a special case:

if ($z & $a) {
  if ($b == 0)
    return $a + 0x100000000;
  else {
    ...

回答6:

Not sure if this works for php, I've manage to get it to work with C#.

int a, b, result;
//Instead of 
result = a >>> b;
//I do
result = (int)((uint)a >> b);

I found that out by debugging into the code that uses >>> comparing it with C# version of the code I converted from javascript. While trying out with b = 0, and using a scientific calculator to see the different hex/dec result of >> and >>> produce by javascript. When a is negative, >>> actually makes aa as unsigned.

Not sure if that works for all scenario, but for my case the >>> is for md5 hashing. Being able to produce similar output, I'm quite satisfied with the result.

Hope that helps

回答7:

For both 32-bit (PHP_INT_SIZE == 4) and 64-bit integers (PHP_INT_SIZE == 8):

function SHR
($x, $c)
{
    $x = intval ($x); // Because 13.5 >> 0 returns 13. We follow.

    $nmaxBits = PHP_INT_SIZE * 8;
    $c %= $nmaxBits;

    if ($c)
        return $x >> $c & ~ (-1 << $nmaxBits - $c);
    else
        return $x;
}

回答8:

This works for me

function RRR($a, $b){
    return (int)((float)$a/pow(2,(int)$b));
}