JavaScript Bitwise Masking

2019-02-25 18:18发布

问题:

This question is similar to this other question; however, I'd like to understand why this is working as it is.

The following code:

console.log((parseInt('0xdeadbeef', 16) & parseInt('0x000000ff', 16)).toString(16));
console.log((parseInt('0xdeadbeef', 16) & parseInt('0x0000ff00', 16)).toString(16));
console.log((parseInt('0xdeadbeef', 16) & parseInt('0x00ff0000', 16)).toString(16));
console.log((parseInt('0xdeadbeef', 16) & parseInt('0xff000000', 16)).toString(16));
console.log((parseInt('0xdeadbeef', 16) & parseInt('0x000000ff', 16)).toString(16));
console.log((parseInt('0xdeadbeef', 16) & parseInt('0x0000ffff', 16)).toString(16));
console.log((parseInt('0xdeadbeef', 16) & parseInt('0x00ffffff', 16)).toString(16));
console.log((parseInt('0xdeadbeef', 16) & parseInt('0xffffffff', 16)).toString(16));

Returns:

ef
be00
ad0000
-22000000
ef
beef
adbeef
-21524111

When what I am expecting from .string(16) would be:

ef
be00
ad0000
de000000
ef
beef
adbeef
deadbeef

What's going on with this?

Thank you in advance for your help.


With thank you to the responders and commenters below, and also to the following sources:

  • http://speakingjs.com/es5/ch11.html
  • JavaScript C Style Type Cast From Signed To Unsigned

Here is a solution that works by providing utility functions to convert a radix-16 32-bit number to and from a signed 32-bit integer:

// Convert 'x' to a signed 32-bit integer treating 'x' as a radix-16 number
// c.f. http://speakingjs.com/es5/ch11.html
function toInt32Radix16(x) {
    return (parseInt(x, 16) | 0);
}

// Convert a signed 32-bit integer 'x' to a radix-16 number
// c.f. https://stackoverflow.com/questions/14890994/javascript-c-style-type-cast-from-signed-to-unsigned
function toRadix16int32(x) {
    return ((x >>> 0).toString(16));
}

console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0x000000ff')));
console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0x0000ff00')));
console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0x00ff0000')));
console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0xff000000')));
console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0x000000ff')));
console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0x0000ffff')));
console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0x00ffffff')));
console.log(toRadix16int32(toInt32Radix16('0xdeadbeef') & toInt32Radix16('0xffffffff')));

Which yields the expected output:

ef
be00
ad0000
de000000
ef
beef
adbeef
deadbeef

As well as some good learning on my part about JavaScript integer behavior.

回答1:

In JavaScript, all bitwise operations (and & among them) return signed 32-bit integer as a result, in the range −231 through 231−1, inclusive. That's why you have that extra bit (0xde000000 is greater than 0x7ffffff) representing a sign now, meaning that you get a negative value instead.

One possible fix:

var r = 0xdeadbeef & 0xff000000;
if (r < 0) {
  r += (1 << 30) * 4;
}
console.log( r.toString(16) ); // 'de000000'


回答2:

Because bitwise operations (like &) are done on signed 32-bit integers in javascript. 0xFFFFFFFF, where all bits are set, is actually -1, and the 0xde000000 you are expecting is actually -570425344.