v0.2.0

0.2.0 (2026-08-04)

The first release since 0.1.0. klt grew from 5 verbs to 24; the 19
net-new verbs below all existed on main but had never shipped in a
numbered release, so anyone installing from PyPI has until now been getting
a tool bearing little resemblance to the documented one.

Additive at the CLI surface — no existing verb was renamed or removed. One
per-command breaking change is included (klt precheck's own
schema_version moved 1 -> 2, detailed below); per
docs/json-contract.md that is scoped to that
command and does not affect any other verb.

Added — new verbs

• klt precheck — pre-flight layout checks before an expensive run
• klt socket-check — socket/pin-level connectivity checks
• klt ring-check — guard/collector ring integrity checks
• klt layout-metrics — quantitative layout measurements
• klt render — raster/vector rendering of layout views
• klt extract — parasitic/device extraction with provenance hashing
• klt lvs — layout-versus-schematic comparison
• klt gen — parametric analog cell generation
• klt gen-compose — composition and routing of generated blocks
• klt draw — primitive layout drawing
• klt sim — circuit simulation, including the remote AWS backend
• klt report — aggregated run reporting
• klt kb — knowledge-base query/update
• klt trajectory — append-only optimization-run logs (--plot)
• klt synthesize — RTL synthesis (Yosys engine class)
• klt place-and-route — physical implementation (OpenROAD engine class)
• klt functional-verification — cocotb testbenches via Icarus/Verilator
• klt eval — single valid/objective/metrics verdict over gate checks
• klt lef-abstract — LEF abstract generation

Each is documented in docs/cli/.

Changed — behavior detail

The user-visible, additive behavior changes worth calling out explicitly,
because they affect a verb's output. Not an exhaustive commit-by-commit log.

• 2026-08-04 — klt sim: wall-clock budget, orphan safety, and resume for
  corner sweeps (#482). options.wall_clock_budget_s/--budget-s bounds
  the overall sweep (distinct from the existing per-corner
  options.timeout_s); an always-on parent-PID liveness check stops the
  local/local-parallel dispatch loop the instant the launching process
  exits; and options.resume/--resume checkpoints completed corners to
  checkpoint.json and skips them on a later matching invocation. Skipped
  corners are reported as status: "error" with
  budget_exceeded/orphaned diagnostics, and environment.budget/
  environment.orphaned/environment.resume summarise the outcome —
  purely additive.
• 2026-08-04 — klt extract: comma-joined multi-label net collision
  detection (#481, closes #470). Any net whose KLayout-assigned name is a
  comma-joined merge of 2+ distinct labels (e.g. Y,Y2) — a silent signal
  that two differently-named nets were shorted together in the layout, most
  commonly a gen-compose pins[] entry naming a port other connectivity
  already reaches — is now flagged in a new top-level
  merged_net_labels[] array ({"net", "labels"} per affected net) plus a
  matching warnings[] entry. Purely additive:
  net_count/nets[]/pin_count are unchanged.
• 2026-08-04 — klt gen-compose: self-net short detection now reads drawn
  pad geometry, not just port adjacency (#472, closes #469).
  route_two_pin() gains a fourth self-net check that intersects the
  route's own drawn metal against the block's actual drawn shapes on the
  route layer, generalising #467's same-row/same-direction heuristic
  (still kept) to same-facing pin pairs on different rows and wider routes
  that reach an adjacent row's pad — cases missed because a port's
  reported width_um systematically under-states the pad metal drawn
  around it.
• 2026-08-04 — klt gen: res_array can draw sky130's higher-sheet-rho
  precision-resistor flavours (#475, closes #463). A validated flavor
  request param ("generic"/"high"/"xhigh", mirroring mos_array's enum
  pattern) resolves a per-family implant/block layer pair sourced from the
  extraction deck's own requires layers (sky130 res_high_po = psdm +
  rpm; res_xhigh_po = psdm + urpm), so the generated array extracts as
  the requested device class instead of always res_generic_po. Default
  "generic" reproduces prior geometry exactly on both families; an
  unsupported flavour raises a clear GenError.
• 2026-08-04 — klt gen: MOS generators draw a poly gate landing pad past
  the diffusion so gate contacts land legally (#474). mos_array and
  diff_pair previously drew gate poly sharing the diffusion's exact
  extent, leaving no DRC-legal spot for a gate contact. The shared unit
  layout now adds a landing pad extending past the diffusion's gate-side
  edge on the first finger and reports the gate port at the pad centre —
  a JSON-contract-visible change: the gate port's y_um moves into the
  extension and its width_um becomes the pad width, and
  bbox_height_um now spans diffusion plus pad while height_um stays
  the bare diffusion height. A contact placed at the reported gate port is
  now DRC-clean on the curated sky130 deck.
• 2026-08-04 — klt extract: dummy marker-layer suppression (#295)
  extended from MOS gates to resistors and bipolars (#471, closes #462).
  The dummy region is now subtracted from each candidate resistor body
  before device recognition and from bipolar_base before
  emitter/collector are derived, with fully-covered components counted
  into the shared dummy_devices_dropped — so matched resistor/bipolar
  arrays built with dummy edge units no longer inflate device.unmatched
  under klt lvs, and a fully dummy-suppressed poly resistor no longer
  trips a false marked_unrecognised warning.
• 2026-08-04 — klt lvs: new device.combine_incomplete mismatch category
  (#466). Only possible with options.combine_devices: true. KLayout's own
  klayout.db.Netlist.combine_devices() can raise an unhandled
  internal-consistency RuntimeError on a partial-match device group — N
  real (matching-relevant) instances plus M dummy instances that all share
  two of three terminals (e.g. a bipolar device's base and collector, tied
  to a matched array's common well and substrate), but only the N real
  instances additionally share the third (e.g. an emitter bussed to one
  signal net). klt lvs now catches that one error shape per netlist
  (narrowly — only a RuntimeError carrying KLayout's own "...in Netlist.combine_devices" marker text; any other RuntimeError still
  propagates as an application error) instead of letting it abort the whole
  run: whatever combine_devices() already merged stays merged, that
  netlist's remaining devices are left as individual devices, and a
  severity: "warning" entry (never changes status, never breaks
  mismatch_count's error semantics) records that combine did not fully
  apply on that side. Purely additive (no schema_version bump), but adds a
  category_counts["device.combine_incomplete"] entry for any
  combine_devices run that trips the KLayout error — see
  docs/cli/lvs.md's device.combine_incomplete subsection for the full
  trigger conditions.
• 2026-08-04 — klt gen + klt gen-compose: ring routing openings (#434).
  guard_ring, diff_pair (add_guard_ring) and bjt_array
  (add_collector_ring) accept ring_gap_side (""/"N"/"S"/"E"/
  "W"), ring_gap_um and ring_gap_offset_um, cutting one opening
  through the ring's band — on every layer the ring is drawn on, dropping
  any contact it would clip — so the ring stays a single connected C-shaped
  conductor rather than splitting into two arcs. The opening is reported as
  a new GAP_<side> entry in ports[] (width_um = the opening's length
  along that side, direction_deg = the side's outward normal), and
  klt gen-compose uses it to admit a route to a ringed block's non-tap
  port — but only when the drawn backbone actually passes through the
  opening with half the route width plus the block's own
  drc_hints.min_spacing_um of clearance from either cut end; a crossing on
  any other side, a crossing that misses the opening, and a segment laid
  along a ring side are all still reported in unrouted_nets[]. With no
  opening declared, the #199 closed-ring rejection is unchanged (its message
  now also names the new remedy). Wiring or labelling a GAP_* port is an
  application error (exit 1). Before this, a matched analog group could
  keep its default guard/collector ring or be wired into the rest of a
  composed circuit, not both. See docs/cli/gen.md ("Ring routing
  openings") and docs/cli/gen-compose.md.
• 2026-08-03 — Breaking (per-command schema_version bump 1 -> 2): klt precheck: layer_whitelist violations' shapes count is now weighted by
  placement multiplicity across the full cell hierarchy, not summed once per
  cell definition (issue #452). Previously, layout.each_cell() counted
  each cell definition's own (non-recursive) shapes exactly once regardless
  of how many times that cell was actually placed, under-reporting true
  placed-shape prevalence by roughly two orders of magnitude on real
  hierarchical, macro-scale input — e.g. a layer drawn 800 times across 320
  placed sky130_fd_sc_hd instances reported "shapes": 10. The fix
  multiplies each cell definition's own-shape count by its total placement
  multiplicity across all instantiation paths (multiplicities compose
  multiplicatively for nested placement — a cell placed inside a cell that
  is itself placed multiple times gets the product, not the sum), via a
  top-down hierarchy walk (Layout.each_cell_top_down()/Instance.size())
  rather than a full recursive shape flatten. Only klt precheck's own
  schema_version moved to 2; every other command's schema_version is
  unaffected, and a per-command bump never forces one on the package version
  — per-command versioning per docs/json-contract.md. See docs/cli/precheck.md.
• 2026-08-03 — klt eval: synthesize and place-and-route are now
  first-class gate checks (#437, Phase 5 of Epic #391), joining
  drc/lvs/sim/layout-metrics/functional-verification. Both always
  report "status": "ok" (see docs/cli/synthesize.md/
  docs/cli/place-and-route.md — a run either produces its output or
  raises), so — like layout-metrics — a gate naming either one must
  declare an explicit threshold to derive pass/fail; request resolves
  as a path only (never inline JSON/-), matching each verb's own
  load_request. This is what lets a digital candidate's descriptor chain
  synthesize -> functional-verification -> place-and-route ->
  drc/layout-metrics (over the P&R-produced GDS) into the same single
  valid/objective/metrics verdict an analog descriptor already
  produces, and record it as one klt trajectory log entry via the same
  gates/objective shape. No schema change to klt eval's own response
  envelope. See docs/cli/eval.md.
• 2026-08-03 — klt trajectory + klt eval: a scored evaluation can now be
  recorded to a trajectory log without hand-rolling the record (#437, Phase 5
  of Epic #391 — the trajectory-log half of that issue, completing the gate
  wiring above). New klayout_tools.trajectory.record_from_eval/
  append_record build one trajectory record straight from a klt eval
  envelope's own objective/gates (collapsing gates[] to the record
  schema's lighter gate_results) and append it to an append-only JSONL log,
  creating the file and any missing parent directories on first write. Both
  are check-name-agnostic — they read only objective/gates, so an analog
  envelope (drc/lvs/sim/layout-metrics) and a digital one
  (synthesize/functional-verification/place-and-route) log identically.
  klt eval gains --trajectory-log/--turn/--candidate-ref/
  --description/--wall-clock-s, which call that pair as a side effect on
  top of the unchanged response envelope — so one invocation can score a
  candidate and log the turn, matching klt trajectory --plot's existing
  "write a file as a courtesy, independent of --format" precedent.
  --trajectory-log without both --turn and --candidate-ref, and a
  log-append failure, are both application errors (exit 1). See
  docs/cli/eval.md's "Trajectory logging" section and
  docs/cli/trajectory.md's "Building a record from klt eval" section.
• 2026-08-03 — klt functional-verification: new verb (#422, Phase 3 of
  Epic #391). Runs a cocotb testbench against RTL sources through Icarus
  Verilog (default) or Verilator, reporting status
  ("pass"/"fail"), test_count/passed_count/failed_count/
  skipped_count, a per-test tests[] array (with error_type/
  error_message on failures), optional Verilator coverage
  (line_pct/toggle_pct/branch_pct/expr_pct plus an lcov info_path),
  and an environment reproducibility block. Invoked exclusively through
  cocotb 2.0's first-party Python Runner API — never a generated
  Makefile — and the verdict is always derived from the run's own
  results.xml, never from a simulator's exit code (which the Phase 1
  survey observed varying between 0, 1, and 2 for the same failing
  regression). Exit codes reuse klt lvs's 0/1/2/3 trichotomy, so
  status: "pass" → valid: true and status: "fail" → valid: false at
  the klt eval boundary; functional-verification is also now a
  first-class klt eval gate check. cocotb is an optional runtime
  dependency (not pinned in pyproject.toml — cocotb 2.0 caps at Python
  3.13 while klt supports 3.10+). See
  docs/cli/functional-verification.md and
  docs/design/cocotb-verification-spike.md section 7 for the full
  contract.
• 2026-08-03 — klt functional-verification: options.random_seed (issue
  #423). Pinned to Runner.test()'s own seed parameter
  (COCOTB_RANDOM_SEED) when given; the effective seed cocotb actually used
  (pinned or its own generated value) is always echoed back in
  environment.random_seed, read from results.xml's own
  <property name="random_seed"> — the same reproducibility bar klt sim's
  Monte Carlo seeding and klt lvs's environment hashes already set. CI
  now also provisions pinned, checksum-verified Icarus Verilog/Verilator
  builds and the pinned cocotb extra (scripts/install-icarus-verilog.sh,
  scripts/install-verilator.sh, pyproject.toml's
  functional-verification extra), so the GCD worked example
  (docs/design/cocotb-verification-spike.md section 6) runs for real in CI
  against both engine: "icarus" and engine: "verilator" rather than only
  locally.
• 2026-08-03 — klt place-and-route: new verb (#425, Phase 4 of Epic #391).
  Places and routes a gate-level netlist (klt synthesize's own
  netlist_path output) against a resolved sky130 standard-cell LEF/liberty
  deck via OpenROAD's native Tcl API, one subprocess per stage
  (floorplan -> place -> cts -> route), chained via write_db/
  read_db ODB checkpoints, with -metrics <file>.json as the structured
  per-stage metrics channel (confirmed end-to-end against a real
  openroad/orfs run for this issue's own worked example). Supports all
  three non-padframe floorplan methods (utilization/explicit/def);
  target_stage makes a partial run (e.g. "place") a normal, successful
  outcome with def_path/gds_path both null by design. def_path is
  populated once write_def has run; gds_path only once the DEF is
  merged with the resolved standard-cell GDS view via KLayout's pya, in
  -process (never a klayout subprocess) — ported from ORFS's own
  def2stream.py. Adds a LEF resolver (klayout_tools.pdk.lef_files())
  alongside the existing liberty resolution. seed is a required request
  field, echoed unchanged in the response. See docs/cli/place-and-route.md
  and docs/design/digital-flow-contracts-spike.md section 5 for the full
  contract.
• 2026-08-03 — klt synthesize: new verb (#416, Phase 2 of Epic #391).
  Synthesizes RTL sources against a resolved sky130 standard-cell liberty
  via Yosys + bundled ABC (read_verilog -> hierarchy -> synth ->
  dfflibmap -> abc -liberty -> clean -> stat/write_verilog,
  generated into a debuggable .ys script kept alongside the mapped
  netlist), reporting instance_count/area_um2/sequential_area_um2/
  instance_counts_by_type parsed from Yosys's own stat -liberty ... -json output. pdk.cell_library/corner resolve to a liberty file via
  the same find_pdk()/libs_ref discovery klt pdk/klt cells already
  use — no new PDK-fetch mechanism. timing is always null in this
  contract, deferred to a future OpenROAD/OpenSTA place-and-route phase.
  See docs/cli/synthesize.md and
  docs/design/digital-flow-contracts-spike.md section 4 for the full
  contract.
• 2026-08-03 — klt trajectory: new verb (#388). Renders an append-only
  JSONL optimization-trajectory log (one record per evaluation: turn,
  candidate_ref, objective {name, value, polarity}, optional
  gate_results/wall_clock_s) into a markdown milestone table plus a
  self-contained objective-vs-turn SVG plot for a block repo's README.
  Milestones are the turns where the objective improves on the best-prior
  record by more than a configurable --threshold. Operates purely on the
  JSONL file — no live optimizer required — so a hand-written or
  human-curated log renders identically. The record schema mirrors the
  planned klt eval envelope's objective/gate_results shape (#387). See
  docs/cli/trajectory.md.
• 2026-08-02 — klt lvs: new device.body_unverified mismatch category
  (a483ed0, #281/#285). Warns (severity: "warning", never changes
  status) when a MOS body terminal was extracted onto a deck-synthesized
  net rather than a real drawn tap/well-label net — an NMOS entry fires on
  every inline-extraction LVS run with one or more NMOS devices (no curated
  deck draws a distinct NMOS substrate/tap layer), and a PMOS entry
  additionally fires for decks with no distinct well-tap layer (gf180mcu
  today). This is purely additive (no schema_version bump) but changes
  category_counts for any gf180mcu (and, for the NMOS case, sky130)
  inline-extraction fixture that previously reported an empty
  category_counts: {} — see docs/cli/lvs.md's device.body_unverified
  subsection for the full trigger conditions.
• 2026-08-02 — klt lvs: new top-level net_correspondence[] response
  field (#311). Lists every layout↔reference net pairing the comparer
  matched — unambiguous and ambiguously-resolved alike — as {layout, reference, pin} entries, sorted and deduplicated per circuit scope so
  len(net_correspondence) == counts.nets.matched holds even across a
  hierarchy with cross-circuit net-name collisions. Purely additive (no
  schema_version bump) — see docs/cli/lvs.md's net_correspondence[] entries subsection.
• 2026-08-02 — klt extract: the "unmodelled device geometry" diagnostic
  (#288/#299) no longer flags a recognised drawn resistor's own terminal
  head (#324) — a poly component abutting a body region _resolve_resistors
  already recognised is now excluded outright, the same way a real MOS gate
  already was, removing a false positive that previously fired on any
  resistor whose wide terminal head carries an ordinary (2+) contact array.
  New top-level unmodelled_poly[] response field lists the bounding
  box + reason ("unmarked" / "marked_unrecognised") of every shape the
  diagnostic still flags, alongside the existing prose warnings[] strings.
  Purely additive (no schema_version bump) — see docs/cli/extract.md's
  "Known limitation: unmodelled device geometry" subsection. Ordinary poly
  routing tracks sharing the same resistor-body signature remain a known,
  documented false-positive class with a client-side filtering workaround
  via unmodelled_poly[].
• 2026-08-02 — klt sim: new optional request block monte_carlo (#348,
  phase 1 of #344's decomposition) — re-runs each expanded corner point
  n times with a reproducible seed (monte_carlo.seed), standing in for
  the per-instance device variation a mismatch-aware model library's
  behavioral parameters draw on. Adds environment.monte_carlo (echoes
  n/seed/vary) and a per-sample corners[].monte_carlo field
  ({sample_index, seed, process_seed, mismatch_seed}, null for a
  non-sampled corner); a sample's corner_id/artifact path gets a
  /mc<sample_index> suffix so per-sample logs never collide. Ships the
  deterministic negative control two public canary repos already rely on
  in their own hand-rolled orchestration: the seed component for whichever
  axis monte_carlo.vary does not request stays identical across every
  sample of a corner (sigma=0), proving the sampler isn't silently
  injecting or dropping variation. Purely additive (no schema_version
  bump), reuses the existing local/local-parallel/remote backends
  unchanged — see docs/cli/sim.md's "Monte Carlo sampling" section.
  Statistics rollup and limit-window evaluation across a sample set landed
  separately as phase 2 (#349, below).
• 2026-08-02 — klt lvs: new accepted request.engine value "netgen"
  (#343) — a second, independent comparator behind the same request/response
  contract, wrapping the open-flow standard
  netgen as a subprocess
  (netgen -batch lvs). Netlist-vs-netlist only — no magic extraction
  backend; it compares the same layout/reference SPICE netlists the
  "klayout" engine already resolves, so it validates comparator/contract
  independence, not extraction independence (see docs/cli/lvs.md →
  "Engine"). Adds a new mismatches[].details field (object | null) for
  engine-specific data that does not map onto
  category/net/device/property — present and null-valued on every
  "klayout"-engine entry too, so no entry shape changed — plus two
  netgen-only request options (options.netgen_setup,
  options.netgen_timeout_s). environment.engine_version is netgen's own
  banner-reported version for this engine. Known, documented gap:
  counts.*.matched is exact on a "match" verdict and 0 on a
  "mismatch" verdict, and net_correspondence is always [], for
  "engine": "netgen" only. Purely additive (no schema_version bump); the
  default engine is still "klayout" and its output is unchanged apart from
  the new null details key. Findings (netgen invocation quirks, report-
  format stability) are written up in
  docs/design/lvs-extraction-spike.md's 2026-08-02 addendum.
• 2026-08-02 — klt sim: Monte Carlo statistics rollup (#349, phase 2 of
  #344's decomposition). A measurement that ran under monte_carlo now
  carries an additive measurements[].monte_carlo block —
  {n, errored, mean, stddev, min, max, quantiles, sigma_window, by_corner} — so callers no longer reduce the raw per-sample corner list
  themselves. stddev is the sample (n-1) standard deviation and is null
  for n < 2 rather than a fabricated 0.0; quantiles defaults to
  [5, 50, 95] and is configurable via the new monte_carlo.quantiles
  request field. The new monte_carlo.k_sigma (overridable per measurement
  with measurements[].k_sigma) opts into a mean ± k*stddev
  limit-window check; both endpoints are scored through the existing
  _evaluate_limits, so min/max handling and the margin sign convention
  are shared with the path a single deterministic value takes —
  a failing window makes the run fail (exit 3) even when every
  individual sample passed its limits. Without a declared k_sigma,
  pass/fail behavior is unchanged; without monte_carlo, the response
  shape is unchanged (no schema_version bump) — see docs/cli/sim.md's
  "Monte Carlo statistics" section.
• 2026-08-02 — klt sim: new environment.monte_carlo.family_mismatch[]
  response field (85faf9d, #365) — per-device-family mismatch-section
  availability for the selected model library, so a Monte Carlo consumer
  can detect which families actually sampled variation. Purely additive
  (no schema_version bump), but changes environment contents for every
  gf180mcu Monte Carlo run — see docs/cli/sim.md's
  environment.monte_carlo subsection.
• 2026-08-03 — klt lef-abstract: new verb (#438, Epic #393 Phase 2
  Capability A). Emits a LEF abstract (MACRO block with PIN/OBS
  sections) from a block's GDSII/OASIS layout plus its klt socket-check
  descriptor, so OpenROAD can place it as a hard macro. Pin geometry is
  real drawn metal when present, else a synthesized placeholder box
  (reported per pin as geometry_source); obstruction geometry is every
  routing-layer shape not already claimed by a declared pin. Backed by
  klayout_tools.lef_header, a new dependency-free tech-LEF header reader
  (SITE, routing-layer PITCH/OFFSET/DIRECTION, macro PIN
DIRECTION/USE) — the deferred-trigger resolution
  docs/design/sc-leflib-evaluation.md called out. docs/schemas/socket .schema.json gains two new optional, additive pins[] fields
  (direction/use, LEF's own vocabulary) this command consumes when
  present. See docs/cli/lef-abstract.md and
  docs/cli/socket-check.md's new "LEF translation" section.
• 2026-08-03 — klt place-and-route: new request.macros field (#438),
  closing the "macro placement" gap the verb's own v1 docstring originally
  scoped out. Each entry fixes one hard-macro instance (e.g. a klt lef-abstract LEF) at a caller-given location during the floorplan stage
  via OpenROAD's own place_macro -location ... -orientation ... -exact
  (never the automatic macro placer). The DEF→GDS merge tolerates an
  abstract-only macro instance (no declared gds) staying empty instead of
  raising; a declared gds merges the macro's own view in, same as the
  standard-cell GDS merge. Purely additive ([]/omitted unchanged; no
  schema_version bump) — see docs/cli/place-and-route.md's new
  "Hard-macro placement" section.
• 2026-08-04 — Epic #393 Phase 3 (cross-domain signoff, #456): klt drc,
  klt lvs, and klt extract verified against a real mixed
  sky130_fd_sc_hd + analog-macro layout (a klt gen diff_pair block placed
  via klt place-and-route's request.macros alongside real standard
  cells, real Yosys + real OpenROAD). No code changes were needed in any of
  the three verbs — each already spans both domains by construction (klt drc/klt extract's whole-layout flattening has no region concept to
  scope by; klt lvs's comparator has no macro-boundary special case).
  Findings and the injected-violation/corrupted-reference verification
  methodology are recorded in docs/cli/drc.md, docs/cli/lvs.md, and
  docs/cli/extract.md's new "Mixed sky130_fd_sc_hd + analog-macro
  layout"/"...netlist" sections. One real domain-boundary gap was found and
  filed separately (not fixed here, per this phase's own decomposition
  precedent, #451/#452): #464, a klt lef-abstract macro pin with no
  PORT geometry that, if wired into a real net, fails klt place-and-route opaquely (OpenROAD GRT-0029) rather than with a clear
  klt-level error. This is a documentation/verification-only change — no
  schema_version bump for any of the three verbs.
• 2026-08-04 — Fixed #464 (per the #451/#452 precedent of fixing rather
  than only documenting a verb-boundary gap). klt lef-abstract gains a
  new top-level unroutable_pins response field ([{name, layer: [gds_layer, gds_datatype]}, ...]) that promotes the existing per-pin
  geometry_source: "none"/warnings[] signal into a structured,
  programmatically-checkable echo — purely additive, [] when every pin
  resolved a routing-type LEF layer, no schema_version bump. klt place-and-route's _validate_macros now cross-checks each macro's LEF
  PIN blocks (via a new has_port field on klayout_tools.lef_header's
  parsed pins[]) against the netlist's own named port connections for
  that instance, and rejects the request (exit 1, before OpenROAD is ever
  invoked) when a PORT-less pin is actually wired to a real net — turning
  the opaque GRT-0029 global-routing failure into a specific, actionable
  klt-level error. A PORT-less pin the netlist leaves unconnected is
  not rejected; the netlist scan is a best-effort structural-Verilog lookup
  (paren-depth-matched, comment-stripped) that is skipped — never a
  spurious reject — when the macro's own instance can't be confidently
  located in the netlist text. See docs/cli/lef-abstract.md's "Pins" and
  docs/cli/place-and-route.md's "Hard-macro placement" sections.