Internal error on DRC operation

[klayoutmatthias]

The attached test case fails with an internal error with this simple DRC script:

deep
l1 = polygons(1,0)
l2 = polygons(2,0)
out = l1.interacting(l2)

The root cause is instance duplication across hierarchy - e.g. C2 instances from C4 exactly overlap identical instances of C2 in the top cell. This case is currently treated illegal and does not make much sense, but apparently in this situation (fill pattern) it cannot be avoided entirely.

smallest_test.gds.zip

[klayoutmatthias]

Problem solved: was an issue related with (fully) overlapping identical cells and loop shortcut. However, such cells are better handled as individual ones to properly connect nets and avoid dangling nets. This is a change in behaviour but probably makes more sense - as far as fully overlapping cells make sense at all.