Lack of clear alternative for reduceRight in unicorn/no-array-reduce

I do not wish to re-litigate #623 but I am here as the rule triggered for Array#reduceRight after updating xo.

For Array#reduce there is a clear alternative which is to use a for (const ... of array) { ... } loop.

However, there is no clean way to iterate over an Array backwards, the best I can come up with is for (let i = array.length - 1; i >= 0; i--) { const ... = array[i]; ... } but in my (admittedly subjective) opinion having to use indicies is worse than the complexity introduced by reverseRight.

Rule: unicorn/no-array-reduce

The code in question looks a bit like:

const listEmpty = null;

const listCons = (element, list) => ({
  head: element,
  tail: list,
});

const listFromArray = (array) =>
  array.reduceRight((list, element) => listCons(element, list), listEmpty);

And I think that (the best alternative that I can think of)

const listFromArray = (array) => {
    let list = listEmpty;
    for (let i = array.length - 1; i >= 0; i--) {
        const element = array[i];
        list = listCons(element, list);
    }
    return list;
}

is worse because of how easy it is to make an off-by-one error with the indexing logic.

---

I am happy to (and have) used a splash of eslint-disable-next-line so happy for this to be closed but I thought it is worth sharing.

Of course it is possible that there is a much nicer way to iterated backwards over an array that I have missed. If there is it would be great to add it to the docs for unicorn/no-array-reduce which currently do not show an alternative for reduceRight.

[papb]

I think its easier to add the eslint-disable than the helper function in my case (my case is unusually function programming style for js so is probably uniquely suited to a reduce) but I agree it is probably the best of both worlds solution. Would a PR adding it to the docs be welcome?

I am not a maintainer of this repo but for what I know of @sindresorhus I’m pretty sure a PR would be very welcome, and I guess @fisker would agree 😃

[harrysarson]

The reverse method is nice but I don’t have the motivation to benchmark it (even v8 isn’t that smart is it?).

I think its easier to add the eslint-disable than the helper function in my case (my case is unusually function programming style for js so is probably uniquely suited to a reduce) but I agree it is probably the best of both worlds solution. Would a PR adding it to the docs be welcome?