The upper bound of a signed integer is not 232 – 1, but 231 – 1, since the first bit is the sign bit.
If you make that comparison, you’ll see your test gives the right result.
Be aware that JavaScript uses IEEE-754 floating point representation for numbers, even when they are integers. But the precision of floating point is more than enough to perform exact calculations on 32-bit integers. As you realised, you’ll need to make the necessary test to detect 32-bit overflow.
Some notes about your code: it passes an argument to the Array#reverse method, which is a method that does not take an argument. Here is how I would write it — see comments in code:
// Name argument n instead of x, as that latter is commonly used for decimal numbers
function reverse(n) {
// Array#reverse method takes no argument.
// You can use `Math.abs()` instead of changing the sign if negative.
// Conversion of string to number can be done with unary plus operator.
var reverseN = +String(Math.abs(n)).split('').reverse().join('');
// Use a number constant instead of calculating the power
if (reverseN > 0x7FFFFFFF) {
return 0;
}
// As we did not change the sign, you can do without the boolean isNegative.
// Don't multiply with -1, just use the unary minus operator.
// The ternary operator might interest you as well (you could even use it
// to combine the above return into one return statement)
return n < 0 ? -reverseN : reverseN;
}
console.log(reverse(-123));
console.log(reverse(1563847412));More efficient
Conversion to string, splitting and joining are relatively expensive operations in comparison with simple arithmetic operations. And so it will be more time (and memory) efficient to solve the problem like this:
function reverse(n) {
var reverseN = 0;
var sign = n < 0;
n = Math.abs(n);
while (n) {
reverseN = reverseN*10 + (n % 10);
n = Math.floor(n/10);
}
return reverseN > 0x7FFFFFFF ? 0 : sign ? -reverseN : reverseN;
}
console.log(reverse(-123));
console.log(reverse(1563847412));