Make gatherWith tail-call optimized

[jfmengels]

This makes gatherWith tail-call optimized. I haven't benchmarked it, but since there is no other significant change in this function, I can only imagine that the performance improved.

Note: I discovered this issue with https://package.elm-lang.org/packages/jfmengels/elm-review-performance/latest/

[Chadtech]

Awesome! I did a benchmark just to check. In small lists, the performance increase appears to be around 5%: https://ellie-app.com/dnMGv4hW5kza1

[jfmengels]

I need to get into the habit of benchmarking, thanks setting it up!

For posterity's sake, here is my benchmark results

• Brave (Version 1.25.68 Chromium: 91.0.4472.77 (Official Build) (64-bit))
• Ubuntu 20.04

So not as good as the 5% you're seeing, but definitely an improvement, and stack safety :+1: