Add tests for Array#splice when only the first argument is provided. - #385

I am actually wondering what the correct results should be as I get different answers from multiple environments.

My initial expectation was what I got from Google Chrome 48.0.2564.97 and from FireFox v43.0.3
IE9 & IE10 also give

[]
[0]
[0, 1]

But all node versions in the travis tests seem to give

[0, 1, 2]
[1, 2]
[2]

IE 11 gives me

undefined
[0]
[0,1]

IE8 gives

[0, 1, 2]
[0, 1, 2]
[0, 1, 2]

ES5.1 spec http://www.ecma-international.org/ecma-262/5.1/#sec-15.4.4.12

Array.prototype.splice (start, deleteCount [ , item1 [ , item2 [ , … ] ] ] )

9 Else if the number of actual arguments is 1, then

a Let insertCount be 0.

b Let actualDeleteCount be len – actualStart.

ES6 spec http://www.ecma-international.org/ecma-262/6.0/#sec-array.prototype.splice

Array.prototype.splice (start, deleteCount , ...items )

9 Else if the number of actual arguments is 1, then

a Let insertCount be 0.

b Let actualDeleteCount be len – actualStart.

I'm going to have to run it through the polyfil and check the result.

This gives

[]
[]
[]

I believe this is actually a change from ES5 to ES6, which would mean that es5-shim would not check or test for deviations from ES6 semantics. I'll leave this open to verify later.

Ie11 is clearly in the wrong whichever :P
But an interesting mix of results.

In latest Safari and Chrome and Firefox, these tests fail natively, so I think these are invalid tests.

However, when I update your tests so they're not assuming that splice returns the receiver - since it in fact returns the removed elements - they all pass, including in IE 8. I'll update this and land.

@Xotic750 oops, i can't; can you please check the "allow edits" box on the RHS of the PR?