Drawn polygons do not have the correct winding number per RFC7946

Nobody has claimed this yet.

[The RFC for GeoJSON](https://www.rfc-editor.org/rfc/rfc7946#section-3.1.6) is clear about the winding order of the points of a polygon: relevant bullet points from the linked section are:
- A linear ring MUST follow the right-hand rule with respect to the area it bounds, i.e., exterior rings are counterclockwise, and holes are clockwise.
- For type "Polygon", the "coordinates" member MUST be an array of linear ring coordinate arrays.

That is, the boundary ring of points for a polygon must be in counter-clockwise order.

However, in version 1.3.0 of mapbox-gl-draw, drawing a triangle in either direction (clockwise or counter-clockwise) causes the feature gotten on the `draw.create` event to have points in the same ordering as drawn, which is illegal GeoJSON, since the RFC mandates the winding order.

A relevant issue is #128, and it's worth noting that the issue predates the RFC linked at the top of this issue. An earlier draft of the RFC ([found here](https://datatracker.ietf.org/doc/id/draft-butler-geojson-00.txt)) does _not_ contain the winding order requirement, which is probably the reason why `mapbox-gl-draw` doesn't conform to it 😆. Some commenters in the linked issue also claim that the winding order does not matter "In the GeoJSON spec", but I don't know which spec they are referring to. I suspect it is the draft, since the RFC explicitly mentions the difference in winding order between the two.

Start by reviewing the draw.create event behavior described here, then compare it with RFC7946 section 3.1.6 and the related issue #128. Confirm how polygon rings are ordered when drawn in either direction. Done means emitted Polygon coordinates follow the required counter-clockwise exterior-ring winding order.

Written by the indexing model from the issue text.