Miss Dada  🏳️‍⚧️

I think the JS one has a specifiable depth but I can never remember how it works across different languages and data structures; I just think it's an over-complicated function

Apparently it removes "array holes" which I did not know was a thing until right now. Apparently `[1, , 2]` is a thing (and is not the same as `[1, undefined, 2]`)

yyeeeeaaahhh

the example I gave when this was being added to the lang is, what happens if you flatten [1, new Set([2, [3]]), 2, null, "hello"] I don't believe there's a well-defined answer to this question and I get tired just thinking about it

also I never actually want arbitrary flattening - if my data is arbitrarily deep, I want an interpreter or visitor - if my data is one level deep, I want flatMap - if I don't know how deep it is, I have a design problem

I'm with you that flatten should always be `flatMap(id)`. FWIW, `Array.prototype.flat` lets you pass a depth, but it defaults to 1 so other than the hole thing it is identical to `flatMap(id)`

Have you seen how an array’s length property is defined? https://www.ecma-international.org/ecma-262/5.1/#sec-15.4 … tl;dr it looks at properties on the object, determines which are array indexes, adds one to the largest, and bam, length.