Andrew,

Breaking out a gap on an Arc is expected to produce 2 Arc segments.
With a Gap larger than zero the total sweep angle would be less than the original.

Found to be only true for Arcs that are not Reversed.
Or when a gap is feasible on the limited Arc.

In dubio on what on what to expect when indicating outside the limited Arc sweep.

A zero or negative Gap size is allowed.
Zero would be similar to DI and when negative the segments would overlap.
For the last case the auxiliary circle shape is invalid with: radius<0.

Results documented in the attached drawing.
Top row are examples.
Second row are the resulting shapes of D3 on duplicates of the top row.
At the right we indicate an endpoint.

---

---

In the JS code I discover that:
- An original Line is modified to the first part. (FS#2668 related when nothing)
- The second part is created from an RShape copying the original attributes.
And included not using current attributes.

- Original Circles and Arcs are deleted.
- Arc entities are created from RShapes NOT copying the original attributes.
Even when included not using current attributes ...
... The resulting new entities are then always on the current active Layer.

Regards,
CVH