An application built with fewer layers no longer restores, stores or shares a
layer it does not have when a link or saved state from a build that has it
names that layer. LayerStateCoordinator drops unregistered enabled-layer ids
before normalizing, so dependent options (a tracked selection) go with them.
An application with every layer registered is unchanged.
windStreamlines takes alpha and paused; rasterShell replaces detailWindow
with a detail image and rectangle and says how shells stack, blend and
update; FieldHandle.on('drawn') says once per change that it is drawn,
never while hidden. The stub fires it on the render after a change, and a
conformance check covers it. The Cesium adapter has no fields yet.
Both reviews of the timeout found that Cesium cannot cancel a full-detail
read: each one given up kept its nine picks traversing every frame, so
stalled tiles piled up a new set every 30 s, while slow but valid answers
were dropped. Raise-only loaded reads then locked in a wrong-high height.
Back to one full-detail read at a time whose answer always applies, and
loaded reads that may adjust a river height but never take it to the sea;
a coarse probe can again lower it partway until the full read lands.
Keeps the test for another ship selected inside the double-click memory,
and corrects CURRENT-STATE: a failed read keeps the height it had.
Before the first full-detail read lands, a loaded-tile read could still
lower a river height partway (27 m to a 15 m bank). Loaded-tile reads now
only raise it; lowering is the full-detail read's to say.
That makes a full read that never lands (stalled tile requests) hold a
wrong height for the whole bridge, so one still running after 30 s is given
up for a fresh one; its late answer is dropped.
Also tests the other half of the double-click memory fix: another ship
selected within the memory is not entered through the first ship's spot.
Before the first full-detail water read lands (it can take many seconds
from cold tiles), a refresh's loaded-tile read with one low probe still
dropped a river height to the geoid. Before that landing, loaded-tile reads
may now find raised water but only a full-detail read lowers it.
CURRENT-STATE also says what the earlier fixes changed: C stays the CCTV
shortcut when the bridge cannot be entered, voice entry is refused on a
bridge, and the ridden ship's trail is hidden with its sprite.
The first press's picture of the ship survived a change of followed
subject: following an aircraft (or another ship) within 700 ms let a
double-click on the old ship's spot enter the new subject. The memory now
carries the ship's MMSI and is used only while that ship is still the one
followed.
The vessel layer draws other ships at their last AIS fix and does not
advance them, so dead-reckoning radar contacts moved the mismatch from own
ship to everyone else: a 15 kn ferry reported 90 s ago showed ~700 m ahead
of the ship out of the window. Own ship stays centred where the camera
rides; contacts go back to their last fixes, and the query drops the extra
reach and candidates (which could also cut the nearest ship in a crowd).
The bridge tests drive a hand-written layer, so nothing exercised the
vessel layer's own scene-click gate. A wire test now clicks open water with
the bridge holding the ship, then again after release.
The Contacts test that adopts Live AIS Vessels directly is renamed to say
so: it never rides a bridge.
The ship bridge rules were appended to cockpit.css, taking it to 2705
lines; they move unchanged to cockpitBridge.css, imported right after it so
the cascade is the same. cockpitBridge.css and cockpitRadar.css join the
format scope and the styles package.
The first press's picture of the track is there because a click on a
selected ship re-centres the camera. It applied to followed aircraft too,
widening their double-click hit area to wherever the aircraft was drawn
as the gesture began. Aircraft are back to the hit where they are drawn.
The sprite and card were hidden from the bridge but the selected-vessel
trail was not: at sea it ran astern from under the eye, and on a raised
river it lay buried under the water. It is now hidden while the bridge
holds its ship, keeping its history, and shown again on release.
The ground clamp read the floor at the ship's last AIS fix while the camera
rode that fix dead-reckoned up to 180 s ahead, so a quay or bridge deck at
the old position could lift the eye. Vessels now read the floor at the
dead-reckoned target, as the water read already does; aircraft are
unchanged.
The radar centred on the ship's last AIS fix while the camera rode that fix
dead-reckoned up to 180 s ahead: at 20 kn about a nautical mile apart, so a
ship already passed still showed ahead. Own ship now sits where the camera
rides and each contact is dead-reckoned to the same moment; the query reaches
past the range for ships that have since moved into it.
Ships with no MMSI each get a key of their own instead of all sharing an
empty one on the scope.
A full-detail water read that sampled nothing (tiles not loaded) fell back
to the geoid, and while a full read ran, a loaded-tile read with one coarse
low probe did the same: on the Seine either dropped the eye ~27 m under the
water until the next read. An empty read is now no answer, and once a
full-detail read has answered only full-detail reads change the height.
The read sequence numbers go: only one full-detail read runs at a time,
so a late read can never be older than one already applied.
Voice entry tracked an aircraft first, which cleared the vessel selection
and ended the bridge, and then failed because Cockpit was still active:
no bridge, no cockpit, no selected ship, and voice said entry was
unavailable. It now refuses before it touches tracking and says to exit
the bridge first.
With a vessel selected and no aircraft tracked, Cockpit took C and stopped
it even when the bridge gate was shut, so the key did nothing and CCTV no
longer toggled. C is now taken only when the bridge can be entered.
The arrow keys also stay with comboboxes, spinbuttons, trees, grids,
menubars, scrollbars and text boxes on a bridge, which move with them.
The globe port takes the globe's material only when it is free, follows a
terrain change, lets go of a material someone replaced, and takes it again
once it comes free; idle ticks ask for no frame.
A new capability, surfaceIdentity, adds globe.surfaceVersion(): a number
that changes whenever what surface heights answer against changes (base,
terrain or photoreal swapped, the globe shown or hidden) and never on
camera moves or tile loads, so the app can key cached heights by it. The
settle kind 'surface' says whether that surface has loaded what is in
view: the terrain while the globe is shown, otherwise the drawn tiles,
false with neither; a renderer without the capability answers it as base.
The Cesium globe port reads both off the scene, so a base changed around
the port counts too; its versions come from one counter shared by every
port. The stub bumps on real base, terrain, photoreal and visibility
changes. A conformance check and a globe port test cover them.
ShapeItem.lit picks the renderer's fixed light or a flat colour; omitted,
extruded shapes and boxes are lit and the rest flat, as they were. The
Cesium draw port lights unclamped colour and image fills by it; clamped
shapes still drape unlit. A new reference scene, shapes-lit, draws a lit
and a flat box and square.
LineItem.mode now says a renderer that cannot drape at all draws draped
lines and clamped shapes as overlay, and clampState() says overlay. The
Cesium port says overlay when its context has no depth textures, which
only a WebGL 1 context can lack.
The satellite dots, the tracked satellite's dot and the orbit rings now go
through the draw port instead of a raw point collection, an entity and
primitives, so the layer's rendering no longer drives Cesium and leaves the
Cesium allow-list.
- Dots are one points set, patched in place and sent once per frame.
- The tracked dot is its own set, drawn at the sample the follow camera
framed (or after the frame's own sample when no camera read it).
- Rings are geodesic lines with a behind-the-globe colour, turned about the
pole by the sidereal angle in the layer each second, since the port has no
inertial frame. The orbit frame matrix is a plain 16-number array, and
the launches layer now receives plain ECEF orbit positions.
- Clicks match the draw hit against the layer's own sets.
- The focus pass projects every dot in one batched call.
- Depth-fail lines in the Cesium draw port honour pickable: false.
Among sets that skip the depth test, a higher drawOrder draws over a lower
one whatever order they were made in, and picks name them topmost first.
Cesium draws the earlier of two opaque near-plane icons on top and picks
it, so the draw port stacks the sets' hosts with a higher order earlier,
ties in the order the sets were made. Hosts trade places only among the
slots draw-set hosts already hold, so no other primitive moves; nothing
moves while every set is at 0. Translucent items blend
order-independently in Cesium and are not ordered by it.
The stub sorts its hits the same way. A conformance check draws two icon
sets at one spot and swaps their orders.
The query contract now says picks answer against the scene as drawn before
effects, and that a drillPick limit bounds the work as well as the answer:
the renderer stops at `limit` hits and looks through at most
max(limit, 16) things under the pointer. The Cesium query port passes that
depth to Cesium's drill (a pick looking under something unpickable stays
unbounded), a click's drill caches one pass per depth, and the stub counts
unpickable items toward the depth.
CameraPose says fovy and aspectRatio describe the nearest centred
perspective frustum, the exact projection being in matrices().
A changed listener's own threshold is now the capability
cameraChangeThreshold; without it every listener fires at 0.5. The Cesium
and stub cameras have it, measured per listener without touching Cesium's
shared percentageChanged.
Conformance: draw.drillPick checks limit 0, never the surface, nothing over
the bare globe, an unpickable cover, nearest first and a bounded prefix;
query.pickBeforeStyles picks an icon under a mirroring style; the threshold
trace is skipped without its capability.
The site dots, outlines and selected footprint fill are now draw-port sets
(points, draped lines, clamped shapes) instead of Cesium entities. The
selected site's dot is drawn by a set of its own, so selecting a site
redraws no other dot. Selection is registered through plain carrier
objects in the context store, as the local GeoJSON layer does, and the
tracked card reads the site's label and position from the carrier's entity
record.
A double-click on a site dot now follows it through camera.follow (gated on
the stock double-click follow), centred on the surface under the dot when
its height is loaded; one on an outline or fill zooms to it. Hiding the
layer stops a follow it started. The layer no longer lowers the shared
camera change sensitivity: its own camera-port listener carries its
threshold.
The follow offset fitting moves to src/geo/frames.js and is shared with
the local GeoJSON layer. rendering.js leaves the Cesium allow-list
(33 -> 32); the layer's entry joins DRAW_FROM_CALLER.
Each product edition now draws as three renderer sets (pins, lines and
fills) instead of a Cesium data source. Pins keep the depth test off and
the Street Level horizon cull, clicks resolve through the draw hit of the
current edition, and a new edition is drawn before the old one goes. A
viewer with no renderer bound draws nothing and releases the edition's
packed images. The layer leaves the Cesium list; a caller handing it a
viewer of their own loads ./render/cesium/unbound first. The QA script
reads pins from the scene instead of data sources.
The launches layer's sites, markers, reticles, mission lines, live orbit
rings and pad zone are now renderer draw sets. Picks go through the query
port, double-click follow through the camera port (shared with other layers
in src/data/stockFollow.js), and per-frame work through the frame port.
rendering.js, orbitRendering.js and launchPad.js leave the Cesium list.
The Cesium draw port gains what these need: pattern line styles, lines in
the inertial frame turned each frame by the set's motion clock, radial-rim
fills on clamped ellipses with a depth bias, and parking for hidden lines
that cannot settle into a static batch, so a hidden line no longer costs a
draw each frame. A roster writes each set once, and a followed live payload
shares one position sample with its marker each frame.
The stations, the cluster dots and the selected station's dot and bracket
are draw-port point and icon sets. Stations drawn too close together on
screen are grouped by geo/screenCluster.js as the default renderer's
entity clustering grouped them: the same boxes and screen coordinates, the
same reuse of the last pass's clusters when the eye comes lower, a pass
whenever the camera moves enough or a station is shown or hidden, and
stations a pass passes over hidden until the next. A cluster dot names its
seed station, so a click plays it as before. The Radio QA script reads the
drawn sets; the unused cluster-collection internals helper is gone.
The fire glows are one draw-port icon set with horizon cull, drawn on the
renderer's moving-icon path with a hold motion, as the layer's own icon
primitive drew them. Only glows that are new or changed are written, after
the set accepts them; fires a refresh drops are removed. Clicks read the
renderer's hit and match the layer's own set id. The sprite order registers
the set by its id.
The camera icons are now a draw-port icon set kept by an icon book; only
icons a view pass or restyle changes are written, in one write per pass.
Picks read the renderer's hit, the view pass and focus pass project
through the camera pose (geo/frustum.js poseProjector, matching the
renderer's projection and near clip) and the query port, and tile
readiness for ground sampling comes from the globe port.
The frustum wireframes, viewshed volumes and the active camera's monitor
plane are now line, mesh and quad items of the layer's own draw sets
(layers/cctv/drawSets.js), written from plain handles only as they change.
The plane takes its canvas double buffer or its video as a quad source.
Picks of lines and planes are matched by set, item id and the record that
draws them. The activation obstruction probe casts through the query
port's rayCast; its clamp lands a microtask later, rewrites the geometry
and asks again for the footprint ground, and a newer activation, a range
edit, a deactivation or a re-init drops an answer still to come.
The gizmo's rings and move arrows are draw-port lines that dim where
geometry hides them, and its handles and arrow tips are depth-test-free
draw-port points; their item ids keep the cctv-gizmo-<part> scheme the
layer's pick owner matches. Picks, rays and the cursor go through the query
port and the renderer's canvas, and the drag math runs on plain vectors.
The layer reports what the gizmo draws for the QA harness, which reads the
parts there instead of from the viewer's entities.
The Cesium draw port now makes quads sets (a posed, textured plane with its
outline, drawn as a plane entity with a transparent image material draws it)
and meshes sets (triangles in the east-north-up frame at an origin, flat and
two-sided or lit and closed). Quads change in place as they move, resize,
fade or take a new canvas; a video quad keeps frames coming while it plays
and gets a new material when its size or source kind changes. Mesh vertices
are placed in the Earth-fixed frame, so Cesium can project them for its 2D
views. The capabilities list both kinds.
The world overlay host now hands each paint lane the eye's height from
the camera port's pose (cameraHeight) beside its position. Detection reads
its view scale and label budget from it instead of the viewer's Cesium
camera, measures box distances with plain arithmetic, and no longer keeps
the viewer, so detection.js leaves the Cesium allow-list.
The satellites layer publishes its focus rectangle with
publishFocusTargetAt, projecting its cached position through the query
port and measuring from the camera port's pose, as flights and military
already do. The scene-taking publishFocusTargetFromCachedPosition and the
unused advanceProjectedSpriteFocus go, and focusDeemphasis.js leaves the
Cesium allow-list. The rectangle is the same; the target's frameNumber,
which nothing reads, is now -1 as the other layers publish it.
screenProjectedRotation, perspectiveProjectedRotation, horizonOccluder and
cameraPoseSignature read a Cesium scene or camera, so they move unchanged
to src/render/cesium/iconOrientation.js, where the CCTV, radio and transit
rendering (still Cesium-bound) load them. src/data/iconOrientation.js keeps
the label backdrop test and the rotation deadband, and leaves the Cesium
allow-list. Local ADS-B computes its rotation from the camera port's axes
with courseScreenRotation directly, the same answer without a stand-in
scene.
The ./aircraft package entry re-exports src/data/iconOrientation.js, so it
no longer carries the four moved helpers or the PERSPECTIVE_* constants.
iconSet.js and iconSegment.js no longer import Cesium and leave the
allow-list. A segment's base is split into high and low floats, and an
arc's east and north axes are computed, the way the default renderer
did it, so the GPU reads the same bits (tests compare every stored float
against Cesium, poles and the centre included). The set's epoch is a
geo/julian date; colours, distance curves and cull points come back as
plain objects. Renderer dates and colours passed in still work.
The primitive that draws icon sets in a Cesium scene now lives at
src/render/cesium/iconSetPrimitive.js beside the draw port's GPU icons,
and leaves the Cesium allow-list. Its importers (the draw port, CCTV,
FIRMS and the cyber sonar overlay) load it from there.
The bundled data centres now draw through the draw port, so there are no
Cesium entities to walk and nothing sets the viewer's selected entity: the
ownership check could not find a target at all. It now finds one from the
layer's context records, accepts a draw-port pick by the context id of the
item it carries, and counts gev:entity-selected announcements per click for
that same item, which both the old and the new build send. Because a
selecting click flies the camera away, the camera is put back and the same
item is found again, before and after draw mode opens, so the drawing click
lands on it rather than on plain ground. The drawing click must add a
vertex with no selection announced and no selected entity.
The hidden-Line check failed now and then with as much amber appearing as
going. Two shots of an unchanged scene differed by 30 to 194 amber pixels on
both builds: imagery was still refining after the fly-to, and the green
area's pulsing fill moves borderline roofs across the amber threshold. Each
shot now waits for the globe's tiles to load, and the other marks are set
aside while the Line is shot and checked back afterwards. The thresholds are
unchanged.
Traffic now hears the camera through the camera port, measured on the
frame at its own 5% threshold, so the raised camera.changed and moveEnd
events no longer reach it. Each move is now setView plus rendered frames,
the way a user move reaches the port, and the setup render records the
baseline pose the first measured move is compared with.
Keyless roads come from OpenFreeMap tiles, so the degraded leg holds those
requests as well as the Overpass proxy's. The interceptor counts road
requests per leg: the degraded leg must hold at least one and the cold city
must reach the network, since requestRoads can be answered from decoded
tiles. The notes say what Fetch->response and response.json measure for
each road source.