Spurious points in the corner polygon

[Fil]

This is not a graphic error, as the points are aligned, but topologically it could be nice (?) to avoid them, as we did in #83 to avoid repeated points.

In the following example the cell for point 0 has 6 different points, when 4 would be enough (discounting the last one which closes the path).

https://observablehq.com/d/5c76273688ee3fc9

[mbostock]

So, eliminate successive collinear points, the same way we eliminate coincident points to fix #42? My guess is that this probably isn’t worth the additional computational effort.

[mourner]

Filtering out collinear points should be very fast so it likely won't add much overhead. And it only happens on cells near the border right?

[Fil]

I seem to find them only at the corners.

[Fil]

Here's an explorable notebook that seems to confirm it's only the corners:
https://observablehq.com/d/390c5610c938f86f

The filter I have works for what I was looking for, but maybe the situation can be prevented earlier.

[mourner]

@Fil if it's only for the corners, you could replace orientation test with a much simpler check since edges are axis-aligned (x0 === x1 === x2 or y0 === y1 === y2).