Fix Math.hypot polyfill to call with zero/one params

[teppeis]

A compat-table test fails due to this issue.

• https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/Math/hypot

[MatrixFrog]

lgtm

[MatrixFrog]

Internal review was more involved than I expected. Going to back-burner this for now.

[concavelenz]

As I mentioned in the internal review, I don't think we want to change the type signature but it should be compatible with the actual implementation and I suggest we use V8's. Which looks like:

function MathHypot(x, y) { // Function length is 2.
    var length = arguments.length;
    var args = new Array(length);
    var max = 0;
    for (var i = 0; i < length; i++) {
      var n = arguments[i];
      if (typeof n !== "number") n = Number(n);
      if (n === Infinity || n === -Infinity) return Infinity;
      n = Math.abs(n);
      if (n > max) max = n;
      args[i] = n;
    }

    // Kahan summation to avoid rounding errors.
    // Normalize the numbers to the largest one to avoid overflow.
    if (max === 0) max = 1;
    var sum = 0;
    var compensation = 0;
    for (var i = 0; i < length; i++) {
      var n = args[i] / max;
      var summand = n * n - compensation;
      var preliminary = sum + summand;
      compensation = (preliminary - sum) - summand;
      sum = preliminary;
    }
    return Math.sqrt(sum) * max;
  };

[concavelenz]

I changed my mind, this function isn't used enough for it to valuable to have a tighter signature.

[brad4d]

@MatrixFrog are you proceeding with this?
@concavelenz is there any reason not to proceed. I think you cancelled out your previous objection.

[teppeis]

I do not know if I should fix this PR ...
I will fix If you need.

[MatrixFrog]

Probably not.

[teppeis]

@MatrixFrog @brad4d @concavelenz any updates?

[brad4d]

Created internal Google issue b/121320889

@concavelenz could you determine whether we should accept or reject this PR?

[brad4d]

You need to delete the value1 argument from the parameter list, too.

[teppeis]

@brad4d Nice catch! I've fixed.

[brad4d]

@concavelenz has confirmed to me that he is OK with accepting this PR.
There's one mistake I see in it for which I've added a comment.

[brad4d]

I've imported this PR and sent it for internal testing and review.

[brad4d]

I submitted the internal version of this change a few minutes ago.
Our next daily push to GitHub should include it and close this PR.

[teppeis]

@brad4d thanks!