Follow up of FS#2685
Tested with 3.23.4 trial on Win 7 but probably not related to OS

With the example file attached to the related forum topic p48576
D2 fails for the lower left ellipse in magenta indicating the segment near (11.72, 2.58).

ShapeAlgorithms.getIntersectingShapes() locates 8 intersecting shapes ...
... That is not the issue here.
ShapeAlgorithms.autoSplit() for the given position returns 3 new segments:
(address may vary)

- [0] REllipse(RShape(address: "0x204229d0"), startPoint: "RVector(11.675918, 2.746654, 0.000000, 1)", endPoint: "RVector(11.977686, 2.428095, 0.000000, 1)", center: "RVector(11.852689, 2.644595, 0.000000, 1)", majorPoint: "RVector(-0.124996, 0.216500, 0.000000, 1)", majorRadius: 0.249993, minorRadius: 0.144334, ratio: 0.57735, startAngle: 45, endAngle: 180, full: false, clockwise: true)
- [1] Undefined
- [2] REllipse(RShape(address: "0x204239c0"), startPoint: "RVector(11.675918, 2.746654, 0.000000, 1)", endPoint: "RVector(11.977686, 2.428095, 0.000000, 1)", center: "RVector(11.852689, 2.644595, 0.000000, 1)", majorPoint: "RVector(-0.124996, 0.216500, 0.000000, 1)", majorRadius: 0.249993, minorRadius: 0.144334, ratio: 0.57735, startAngle: 45, endAngle: 180, full: false, clockwise: false)

The issue is that the new entities are only (re-)generated when longer than RS.PointTolerance (Good practice).

The approximated Simpson-length for the ellipse-arcs are:

- [0 CW] **NaN**   (instead about 0.683800238925154)
- [1] Not an RShape
- [2 CCW] 0.5773502691896258

Ellipse-arcs [0] starts at 45°, passes 0° in CW sense and ends at (-)180°.

The full ellipse is removed but not replaced by new segment [0] or [1].
Segment [2] is the segment that is broken out.

Similar new segment lengths for the segment near (12.02, 2.49):

- [0 CW] **NaN**    (instead about 1.12960650667804213)
- [1] Not an RShape
- [2 CCW] 0.13154400143673767

Regards,
CVH