Counters that may only fall, each with a committed baseline and a test
that fails on growth:
- report-engine-coupling.mjs gains layer modules that take the viewer
in attachRenderer (33), layer folders with their own "view being
restored" flag (9), engine.surface calls outside src/globe (41) and
layer ids named in the engine switch (3).
- report-qa-harness.mjs counts QA scripts that launch Chrome themselves
(98), wait with fixed sleeps (117), take the hidden loading screen as
readiness (9), read Cesium-only objects (75) or define their own
waitFor helpers (19).
- withCallLog records every contract member the contract suite calls on
the fake engine; contractCoverage.test.mjs fails when a member outside
contract-uncalled-baseline.json goes uncalled, and when a listed one
is called now, so the list of 12 only shrinks.
Remove code that nothing in the app, the runners or CI uses:
- the one-off migration compare, look and bench scripts and the vector
ground prototype they measured, with the bench tables in the docs;
- the unused Navara-named shared render modules (polygons, ground picks,
mover bank, ground ring);
- the Cesium-shaped aircraft and vessel test renderers, and the trail,
engineFromViewer and horizon-occluder helpers only they kept alive.
Their tests now run on the production displays over the fake engine;
- Navara's unread diagnostics, fallback ports, dead exports and
parameters, stale effect-chain anchors and the unread capability flags.
The QA scripts share one keyless guard and readiness wait, the session
harness reuses the strict harness's browser flags and block list, and the
harness honours QA_ANGLE and QA_RESERVED_PORTS. The bikeshare fixture
keeps only what its scenario reads. The docs name the engine picker where
it is (VISUAL PRESETS) and drop dated run results.
No behaviour changes on either engine. Navara layer renderer lines fall
from 189 to 0 and viewer lines from 891 to 849.
engine.imagery gains what these rasters need, on every engine:
- geographic (EPSG:4326) tile schemes and a transform applied to each
decoded tile (the satellite infrared mask);
- staged source changes, which keep the drawn source until the new one
has loaded (the weather frame swap), and credit-only changes in place;
- raised layers that draw on a shell above terrain and 3D tiles, rising
with the camera, with a detail image inside a rectangle. On Cesium this
is the weather shell primitive, moved into the engine; on Navara the
pixel's ray meets the ellipsoid grown by the height.
Rain radar, satellite clouds, lightning, the wind field and the Bhote
Koshi event and locator each draw through one render/display.js on both
engines; their Navara renderers and Navara's weather surface pass go.
Cesium's pixels are unchanged in every compared scene. On Navara the
Bhote Koshi images are no longer upside down. Two engine-specific parts
stay behind one backend option each: the wind's streamlines and the
Bhote Koshi story entities.
The coupling report now counts defineGroundEffect calls, not its own
declaration. Coupling: Navara renderer lines fall from 2304 to 1695,
Navara ground passes from 3 to 1 (the Global Context pulses), engine
kind checks from 5 to 2. The contract suite passes on both engines
(Cesium 71/71, Navara 74/74).
What only one engine has now travels with the engine instead of a
WeakMap in navaraMode.js: engine.extensions.navara holds Navara's view,
its SDK classes and the daylight reset, engine.extensions.cesium the
Cesium viewer. The contract documents the field as outside the contract,
and the coupling report counts its reads outside src/globe in place of
the navaraViews reads (27, unchanged).
scripts/report-engine-coupling.mjs counts the engine-specific code left
outside src/globe/cesium and src/globe/navara: layer renderer lines per
engine, Cesium imports, `viewer` mentions, engine-name checks, raw Navara
reads, Navara ground passes, Cesium SDK loads and ScreenSpaceEventHandler
users. engineCoupling.test.mjs fails npm test when a count grows past
scripts/engine-coupling-baseline.json, so each step towards drawing
through the globe contract can only lower them.
scripts/qa-contract-suite.mjs runs src/globe/testing/contractSuite.mjs in
headless Chrome on the Cesium and Navara of a keyless dev server, which
until now only ran on the fake engine. The page imports the suite from the
dev server, starts at a fixed view, waits for the camera to rest and lets
the pointer through the panels. Cesium passes 57 of 57 checks, Navara 64
of 67; docs/CURRENT-STATE.md names the three that fail.
German Khokhlov