Tim Channell

> ... ugh :point_up: most of my experience trying to figure out how to apply GH to complex geometries. Do you have a test case that's reproducible so I can...

added test cases for this and #14 here: 052678f

I'm not sure this is fixed (or is a real problem) yet, I'll need to find more test cases

:+1: I was just talking to @morganherlocker about this the other day. I'm also pretty sure that in the current state, cloning is not necessary for most of the operations...

3 is _definitely_ something I've been meaning to do. I'm currently also doing some wasteful / expensive stuff just to accommodate pip / clockwise checking via the turf API, so...

> Supporting holes properly is important though so can't get away with one ring. :) certainly! I was just wondering if you were proposing that the turf wrapper handle dealing...

> As far as I remember, the original algorithm could deal with any kinds of rings at the same time when running the algorithm — is this not the case?...

@bhousel @mourner yea, it looks like that repo is a very clean starting point, but is lacking 2 bits: support for multiple rings (multipoly / holes) in the subject and...

At this point, the algorithm depends on two packages: `turf-is-clockwise` (which takes rings, not geojson, so we're okay there) and `point-in-polygon`. All other dependencies have been stripped out / replaced

@Lknechtli Looks like this is the result of [code in `robust-orientation`](https://github.com/mikolalysenko/robust-orientation/blob/1ae9e492e903f8ef50337678adcb4892a50237da/orientation.js#L87), one of the nested dependencies of turf-convex, using `new Function` to generate functions. Without entirely replacing / reimplementing turf-convex...