Should Diff return opposite orientation for newly created interior holes in a polygon?

[mglasgow]

Using Diff between two polygons, the second polygon of which is contained completely within the first, will result in an interior hole in the original polygon.

Does Diff currently give any consideration to the orientation of that newly created interior hole? To be a valid polygon, the interior ring should have opposite orientation to the outer ring.

I have created a file in the "genericTestCase" format which shows what I am trying to achieve:
issue129.geojson.txt

A screenshot of the polygon is here:

Both polygons are sent to the Diff operation with counter-clockwise orientation. Diff returns the outer and inner rings both in counter-clockwise orientation. However, to be a valid polygon, this interior ring should actually be clockwise orientation.

[rowanwins]

The orientation issue depends on which data spec you are following (eg for geojson your comment is accurate). But this repo doesn't particularly enforce a particular data spec.

There is a fairly simple algorithm that we could add called the shoelace formula, see this reference from TurfJS.
https://github.com/Turfjs/turf/blob/f62ba326cba17d9ebfb29ba8671107f937a89442/packages/turf-boolean-clockwise/index.ts#L19-L35
It's fairly minimal in terms of code but would add some performance overhead. Perhaps we could add an options object argument with this sort of thing?

Alternatively we could ask users to pass their results through the turf.rewind module

[w8r]

Yes, maybe an option would make sense to see how much it is used.