On Wed, May 23, 2018 at 5:56 PM, Alexander Lichter <e...@lichter.io> wrote: > An optimization would be great because in comparison to the existing concat > method, rest/spread is significantly slower at the moment (see > https://jsperf.com/single-array-composition)
There's a limit to how much optimization can address the fact that `[...original, ...additions]` has to create a new array. And creating the array is significant, as we can see in the difference between `concat` and `push`: https://jsperf.com/concat-vs-push-number-42 I've often wanted an `Array.prototype.append` (often enough I've been known to add it on occasion; obviously only in app/page code, not in lib code). It would be a good addition in my view. I tend to suspect the name isn't web-safe, though. So that would need investigation and design. The polyfill is trivial: ```js Object.defineProperty(Array.prototype, "append", { value: function(...sources) { for (let o = 0, olen = sources.length; o < olen; ++o) { const source = sources[o]; for (let i = 0, ilen = source.length; i < ilen; ++i) { this.push(source[i]); } } return this; }, writable: true, configurable: true }); ``` I've avoided iterators and such as a first pass at optimization (and it seems to do okay: https://jsperf.com/push-vs-append-number-42/1). Didn't use `push.apply` because of the limits it has on it in some implementations. This version assumes all entries are array-like and should have their entries added to the array on which it's called. There are a lot of other designs one could make, though. Just one argument (you can always chain); a flattening version; etc., etc. -- T.J. Crowder
_______________________________________________ es-discuss mailing list es-discuss@mozilla.org https://mail.mozilla.org/listinfo/es-discuss