Unexpected RectClip output

[EternalXY]

I see an unexpected spike when clipping a long and narrow rectangle:

        Clipper2Lib::Rect64 rect(54690, 0, 65628, 6000);
        Clipper2Lib::Paths64 poly{
        {
                {700000, 6000},
                {0, 6000},
                {0, 5925},
                {700000, 5925},
        }
        };
        Clipper2Lib::Paths64 solution = Clipper2Lib::RectClip(rect, poly);

This outputs the following points:

65628, 5925
65628, 6000
54689, 6000
54690, 0
54689, 5925

Input rectangle in red (clipped to fit the screen) and result in black:

The size of the spike seems related to the size of the clipping rectangle somehow, any idea what is going on here?

[EternalXY]

Small update: the spike is gone when rounding in GetIntersectPoint():

      ip.x = std::llround(static_cast<double>(ln1a.x) + t * dx1);
      ip.y = std::llround(static_cast<double>(ln1a.y) + t * dy1);

[EternalXY]

Another update: also fixed by sorting the input points of GetIntersectPoint/SegmentsIntersect since they seem to round differently if segments or segment points are swapped, and RectClip expects the intersect points to be identical.

[AngusJohnson]

Change:

Clipper2/CPP/Clipper2Lib/src/clipper.rectclip.cpp

Lines 477 to 478 in 3867aab

	if (crossing_prev != Location::Inside)
	AddCorner(crossing_prev, loc);

to:

        if (crossing_prev != Location::Inside && crossing_prev != loc)
          AddCorner(crossing_prev, loc);

[AngusJohnson]

Hopefully fixed now.

[EternalXY]

Thanks, works wonderfully now 👍 Also no longer seeing 1-point polygons in the output.