Stop pointing comments and test headers at notes and QA scripts the repository does not carry

src/bench/w1Feeds.js

@@ -2,9 +2,8 @@
 // states), the /api/military snapshot (readsb rows), 12 000 AIS vessels on /api/vessels (with and without an area) and
 // the local receivers' 1090 MHz and 978 MHz UAT aircraft near JFK, timed
 // relative to one instant, so every run sees the same objects at the same
-// places. Pure: Node and browser, no I/O. Used by the W1 baseline recordings
-// (src/tooling/w1Feeds.js) and the app's QA fixture sources
-// (src/app/layers/qaSources.js).
+// places. Pure: Node and browser, no I/O. Used by the app's QA fixture
+// sources (src/app/layers/qaSources.js).
 
 import { createSyntheticFlights } from './syntheticFlights.js';
 

src/data/bhoteKoshiLocator.edgeCases.test.mjs

@@ -1,7 +1,6 @@
-// Edge cases from docs/migration/planet-separation/layers.md
-// (bhote-koshi-locator): a boundary that resolves late, after its request was
-// cancelled, does not change the live scene; both engines show the same area
-// and the same target.
+// Edge cases for bhote-koshi-locator: a boundary that resolves late, after its
+// request was cancelled, does not change the live scene; both engines show the
+// same area and the same target.
 import assert from 'node:assert/strict';
 import test from 'node:test';
 import * as Cesium from 'cesium';

src/data/militaryFlights.edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge cases for the military layer from docs/migration/planet-separation/layers.md:
+// Edge cases for the military layer:
 // one ICAO present in both feeds yields one target with the military card, and
 // when the owning feed loses the contact the follow camera is released.
 import { test } from 'node:test';

src/layers/awareness/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge cases for military-awareness from docs/migration/planet-separation/layers.md:
+// Edge cases for military-awareness:
 // with no scene the same nearest contacts come back, and NEXT/PREV, a type
 // filter and an exhausted list lead to the same ID whatever scene is attached
 // and whatever order the feeds list equally distant contacts in.

src/layers/cctv/render/navaraMeshes.js

@@ -49,7 +49,7 @@ export function createCctvNavaraMeshes(engine, { view, MeshDesc }) {
         // The opaque scene draws after the scene depth is copied into its
         // target, so walls and terrain hide the frame and volume; Navara's
         // transparent pass runs after the effects without that depth and
-        // painted them through (scripts/qa-edge-cctv-occlusion.mjs).
+        // painted them through.
         getPassKey() {
           return 'opaque';
         }

src/layers/ghostMaps/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge cases from docs/migration/planet-separation/layers.md (s2-ghostmaps):
+// Edge cases for s2-ghostmaps:
 // change the KMZ edition and the engine before parsing finishes; only the
 // selected edition is shown, and a bad archive does not erase the working one.
 import assert from 'node:assert/strict';

src/layers/localAdsb/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge cases for the local-adsb layer from docs/migration/planet-separation/layers.md:
+// Edge cases for the local-adsb layer:
 // a scene switch does not reopen the USB device, decoder stream or tuner
 // permission; an unplugged device ends exactly one session; and no local ID
 // reaches a public share link.

src/layers/satellites/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge case for the satellites layer from docs/migration/planet-separation/layers.md:
+// Edge case for the satellites layer:
 // with the clock stopped at a known epoch, switching the scene while a
 // satellite is followed leaves the pass list and the coordinates unchanged.
 import test from 'node:test';

src/layers/shared/render/navaraSceneDepth.js

@@ -5,7 +5,7 @@
 // its look. This test keeps the mesh in the transparent pass and drops every
 // fragment that lies behind the depth Navara packs after its opaque pass
 // (globe, 3D tiles, MRT and opaque meshes): an unhidden fragment is drawn
-// exactly as before. scripts/qa-edge-occlusion.mjs checks it in a browser.
+// exactly as before.
 import { RGBADepthPacking, Vector2 } from 'three';
 
 const KEY = 'gevSceneDepth';

src/layers/streetLevel/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge cases from docs/migration/planet-separation/layers.md (street-level):
+// Edge cases for street-level:
 // without a paid source, a late answer from a provider the user switched away
 // from is ignored and the lit provider keeps the panel.
 import assert from 'node:assert/strict';

src/layers/transit/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge cases from docs/migration/planet-separation/layers.md (transit): a stop,
+// Edge cases for transit: a stop,
 // a turn, a feed loss and its return keep the heading and the data age
 // continuous, and the selection stays bound to the vehicle/trip identity.
 import test from 'node:test';

src/layers/vessels/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge case for ais-live-vessels from docs/migration/planet-separation/layers.md:
+// Edge case for ais-live-vessels:
 // a vessel crosses into the next area cell while AIS is interrupted. There is
 // never a second request in flight and the card comes back by MMSI.
 import test from 'node:test';

src/layers/wind/edgeCases.test.mjs

@@ -1,4 +1,4 @@
-// Edge cases from docs/migration/planet-separation/layers.md (wind): on a fixed
+// Edge cases for wind: on a fixed
 // grid direction, speed and legend agree whichever way the centre is picked,
 // and an empty field is never presented as calm.
 import test from 'node:test';

src/services/meshFloorSampler.js

@@ -3,8 +3,8 @@ import { engineFromViewer } from '../globe/cesium/engineFromViewer.js';
 import { acceptedSurfaceHeight } from '../globe/surfaceDetail.js';
 
 /** Answers from tiles larger than this are too coarse to latch (m).
- *  Measured on Navara (scripts/qa-surface-heights.mjs): from 800 m the
- *  answering tiles were 112-314 m and within 2 m of Cesium; from 15 km, cold,
+ *  Measured on Navara: from 800 m the answering tiles were 112-314 m and
+ *  within 2 m of Cesium; from 15 km, cold,
  *  3.5-4 km tiles were off by -16.2 to +6.2 m; while streaming a 27.9 km tile
  *  read 143.2 m at Austin where the ground is 122.5 m. A 1.8 km tile at
  *  Austin (the hidden tile under a 600 m camera looking ahead) read the