Yes, this type of floating point verification is asking for such issues:

Test.assertEquals(+actual.toFixed(6), +expected.toFixed(6));

Instead, a tolerance of 1e-6 should be used.

isn't this possibly an issue?

This is like a 5-6 kyu just to understand it...

Geezus

It's not so much about exactly how many decimals, it's about whether it is a valid solution or not. It simply doesn't seem okay to allow for any solution to sometimes pass a kata.

Oops sorry, I didn't notice I made the same mistake twice in one line... SHOULD be fixed now...

memory[memIndex] == 0 ? leftBrackets.pop() : i = leftBrackets[0]

i = leftBrackets[0] is wrong because, again, the latest bracket is the last item of the array.

I believe it should work now. Thanks for noticing that!

In the memory cells usage checks:

memory[memIndex] == 0 ? leftBrackets.shift() : i = leftBrackets[0]
if (leftBrackets.length == 0) {skip = false}

The first line is wrong as the latest bracket is currently on the last index. The second line is also wrong as when the code is something like this: +[-[-]>+<]>. this will result in 0 instead of 1 since the state skip from inner bracket still holds.

It should be fixed. It wasn't accounting for inputs in the memory check.

My solution uses 3 memory cells and passes all tests.

Fixed!