JavaScript converts operands to 32-bit signed ints before doing bitwise operations. It also does the operation with 32-bit signed ints, meaning that the result is a 32-bit signed int.
Because I'd like to do bitwise operations with 32-bit unsigned ints, I'm wondering if there is a way to use JavaScript's mangled result to find out the intended result.
To demonstrate my idea, for example, in C, which is the reference as to what I'd like,
unsigned int a = 3774191835u;
unsigned int b = a >> 2;
/* b == 943547958 */
In JavaScript,
var a = 3774191835;
var b = a >> 2;
/* b == -130193866 */
Let's try this with a different operation. In C,
unsigned int a = 1986735448u;
unsigned int b = a << 1;
/* b == 3973470896 */
In JavaScript,
var a = 1986735448;
var b = a << 1;
/* b == -321496400 */
Now that JavaScript has evaluated my bitwise operation with the operand as an signed int, we of course, get a different result to what we would in C, where we can properly do bitwise operations on unsigned ints.
I know it's possible, but I'm unsure of a way that I can, essentially, turn JavaScript's result into the intended result.
Zero-fill right shift the result by zero works for the second case only, but not the first.
var a = 3774191835;
var b = (a >> 2) >>> 0;
/* b == 4164773430 */
var a = 1986735448;
var b = (a << 1) >>> 0;
/* b == 3973470896 */
See Question&Answers more detail:
os 与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…