min_endcap_enclosure

Like min_enclosure, but the margin is only required on at least one side rather than all of them — the wire endcap case (e.g. a via sitting at the corner of its metal, where the metal only needs to run past it on one side).

Semantics

Shares min_enclosure’s engine (same containment test), but instead of the facing edge-pair scan it uses the enclosing region’s bounding-box margin on each side of the enclosed shape and takes the best (largest) one — a region violates only if every side is short. This bounding-box reduction is deliberate: an edge-to-contour distance is corner-limited and would understate a long endcap run.

In practice this rule is paired with an ordinary min_enclosure entry at a much smaller value covering the rest of the shape’s sides (e.g. IHP’s V1.c at 0.01 µm alongside V1.c1 at 0.05 µm) — together they express “mostly flush is fine, but at least one side needs real margin.”

Layers

Two layers, positional — same roles as min_enclosure:

1. layers[0] — the enclosing region.

2. layers[1] — the enclosed shapes being measured.

Parameters

None beyond layers and value (µm, the minimum required margin on the best side).

Violation markers

One edge marker per offending shape, along its first contour edge (a location marker — the endcap reduction doesn’t identify a single “worst wall” the way the facing-pair scan does, since it’s driven by the bounding box).

KLayout equivalent

No single built-in KLayout primitive for “at least one side” enclosure; it’s the same enclosed family, restricted here to a bounding-box side-margin comparison instead of per-edge facing pairs.

Example

# V1.c allows the via to be nearly flush with Metal1 on most sides (0.01 µm); V1.c1
# additionally requires a real endcap (0.05 µm) on at least one side.
- id: V1.c
  check: min_enclosure
  layers: [Metal1, Via1NoSealring]
  value: 0.01
- id: V1.c1
  check: min_endcap_enclosure
  layers: [Metal1, Via1NoSealring]
  value: 0.05