@turf/boolean-contains: allow boundary points in MultiPoint containment checks

[rexxars]

(Lifted from #3010)

Fixes behavior of booleanContains to return true if every point in a MultiPoint is inside/on boundary of some polygon, and at least one is strictly interior. This previously returned false, but differs from booleanWithin and doesn't match DE-9IM semantics. In my eyes this is a bugfix, not a breaking change - but will leave it to the maintainers to decide how to classify it.

An example:

Given the following square polygon:

{
  "type": "Polygon",
  "coordinates": [
    [
      [0, 0],
      [1, 0],
      [1, 1],
      [0, 1],
      [0, 0]
    ]
  ]
}

And the following MultiPoint:

{
  "type": "MultiPoint",
  "coordinates": [
    [0.5, 0.5],
    [1, 1]
  ]
}

In this case, one point is strictly on the interior of the polygon, while one point is on the boundary. This previously returned false, but the way I (and DE-9IM) see it should return true.

Resolves #3025

[mfedderly]

This seems like a pretty reasonable change, and I'm kind of torn on whether correcting behavior to match the spec is a breaking fix or not (I went back and forth on #3070 with the same debate). I think semantically they're bug fixes so 7.x is appropriate, but adding them to the v8 milestone and explicitly documenting the behavior change is probably a safer rollout strategy. Its always hard to tell on these calls until we hear back from users after a release that we broke things for them.

Thoughts?

[mfedderly]

Thank you!

[mfedderly]

I have convinced myself that this is a bug that doesn't need a major version bump to correct. The docs and de-9im all suggest that this is a correction of buggy behavior and people should not have been depending on the broken state.