- The look-at longitude wraps fully, so a tilted camera at a pole gives a
point on the globe. The strict off-globe check on street_level is undone:
points are clamped again, as before.
- A posted view that fails is answered, and the views behind it still apply.
- A posted view waits at most 2 s for the terrain under its camera, then
keeps its own estimate.
- A camera taken during that wait stays where it is: the rest of the view
applies, and only its photos stand down.
- show_street_level keeps a search or a press on the globe made during its
own flight: it adopts only its own navigation stamps on landing, and a
gesture is never forgiven.
Main added a Swedish catalog in #1630, so the catalog test failed on the
strings this PR adds. They use the catalog's terms (Gatunivå, nyckel,
bilder på gatunivå, vyns mitt).
- Releasing a held viewer frees the old source's adapter without resetting
the panel, so a replacement's error, outcome and armed OPEN survive; a
cross-source replacement drops the old photo's details and marker.
- The action holds the open photo only for its own switch-off: a source
switched by hand during the load stops it and reports cancelled.
- A posted view waiting on terrain watches the camera and stands down after a
navigation or a globe press; a flat or straight-down camera asks no terrain.
- An arming link opens the panel without persisting the layout, and a posted
view's street-level step runs as programmatic.
- An off-globe street_level point throws TypeError.
- zh-Hant uses the file's wording for the new strings; README and
CURRENT-STATE count 35 tools.
Translate the 148 strings main added since (mostly the News layer) and
prune the 5 it no longer uses; keep both Hebrew and Swedish in the
language lists.
show_in_gods_eye_view measured a view's altitude from the terrain before
taking the point the camera looks at; a view posted straight into the
embedded app used the ellipsoid altitude, so a camera over high ground
opened photos kilometres ahead. groundedLookAt moves to the view module,
and the embedded app uses it with the app's terrain heights for a posted
street_level that names no point.
A provider that threw while geocoding (offline, an upstream down) escaped
the action as a bare error. show_street_level now reports it as its
location stage ("Could not arrive at …", the cause in detail), and nothing
is switched or looked up; fly_to_location itself is unchanged.
The voice card showed English for show_street_level failures the catalog
did not know: a missing key, a map or source that could not be switched on,
the layer, invalid coordinates, an imagery fallback and the filter clause.
They now translate like the other action errors (exact messages and
patterns in speech.js), in all twelve languages; the spoken line stays
English, and the filter clause names the filters as their chips do.
- The cloud brightness key is a filter, not a brightness scale.
- Counted lines without a plural form read "label: count", so 1 agrees.
- Fire confidence is «tillförlitlighet», not «säkerhet».
- Agreement with «flöde» and «luftfartyg»; course on the tracking line is
«FÄRDRIKTNING», which fits both aircraft and ships.
- Wildlife is «Artobservationer», since the layer also shows plants and fungi.
- Bridge readouts are nouns (BLICK, ÖGONHÖJD), cockpit key hints share one
verb form, tool names in pointer messages are quoted, and «jfr» replaces
«jmf.»; the guide's example follows.
- Smaller wording and consistency fixes across the catalog.
A source's contextGeneration tells the probe the same context was lost and
restored; it then makes its GPU objects again instead of drawing with dead
ones. A depth reader that throws turns the test off for a while instead of
throwing out of add(). Batches whose control point never answers count as
failed ones, so a query path that never answers stands down and is rebuilt
on the retry.
A weather feed may also be a frame source. With an optional
getFrameImage({ product, time, size, bbox, signal }) the layers take
whole-extent images, detail windows and the global mosaic from it; with an
optional getFrameTile({ product, time, z, x, y, size, signal }) regional
tiles and their prefetch come from it, null being a transparent tile. Pixels
may be an ImageBitmap, ImageData, canvas or VideoFrame of exactly the
requested size; the layer draws them once and closes them. Cancellation
reaches the source, and a feed without these methods reads the image and
tile routes as before.
The GPU probe behind query.occlusionTest no longer depends on Cesium: a
renderer hands it the frame's depth through readSource (a DepthSource), and
createDepthOcclusionTest builds the query port's test from it. Besides depth
packed into RGBA8, as Cesium keeps it for picking, it reads a depth texture
(encoding 'depth', on the GPU only), and log depth with no near plane decodes
too. Cesium's sceneOcclusionProbe.js keeps only its scene
reader. Exported as gods-eye-view/render/occlusion.
A renderer plugin may supply applicationHooks(renderer) for what only its
renderer has (map engine, own credits, stale commands, QA handles). What every
renderer should do moves onto the contract: controls.keepEarthFramed (behind
the keepEarthFramed capability), query.occlusionTest (behind occlusionTest),
and panel drawing, which keepPanelRendering now does through resize and
frame.renderNow through the contract. renderNow throws when the frame fails to draw.
The Cesium port installs its zoom-out guard through keepEarthFramed (a setView
no longer cancels its turn) and answers occlusionTest with the scene's probe;
the stub answers it with a horizon test.
Upstream's layers and changes are carried onto the ports, with no Cesium
outside src/render and no reach into the viewer:
- Shipwrecks draw as one icon set, clicked through the input dispatcher.
- News draws through draw sets, the camera port and the input dispatcher;
its beacons pulse each frame through the frame port.
- The scene occlusion probe moves into src/render/cesium and reaches
detection through a new application hook, createOcclusionProbe; its
private Cesium reads go through internals.js.
- Flight and military trails, the stale and SIGNAL LOST cue, coasting and
following the world clock run on the draw and frame ports. GPU icons
follow the scene clock while the world clock is followed, and fade when
stale. A report that holds is not planned again every tick.
- Vessels are keyed by their record id.
The stub boot now enables military, shipwrecks and news as well.
- Every placement of Cesium's credit display (above the command dock, the
Minimal HUD's corner, clean view and recording, the phone tab bar and an
open sheet) has the same rule for the application's line; a test keeps
each stylesheet's pairs together.
- A credit's logo is drawn in place of its text, which names it (the
image's alt), as Cesium shows logo credits; the contract's Credit says
so. Links open with noopener and noreferrer; a list that empties while
focus is in it hands focus back to the line.
- A stack that stays after any failed switch (a tileset or imagery that
fails to load, or a photoreal base that fails to come back) still falls
back on its tile failures: the controller binds its listener again for
that switch where it handles the failure.
An undraped line in the polyline collection that moved was given a new
array of Cartesians, one new object a point. One that moves by as many
points, not geodesic, now has its new positions written into the points
its polyline draws, which the polyline then takes again: the range ring's
fronts, stepping out every interval, make no new points.
Cesium's camera measures each of heading, pitch and roll by moving the
camera into the eye's east-north-up frame and back, recomputing its
matrices twice each time. In 3D the pose now reads that frame once, takes
the eye's direction and up into it as Cesium's inverse transformation
does, and puts them through Cesium's own formulas: the same heading, pitch
and roll to the last bit or so, and the camera is left as it was. Other
modes still ask Cesium. The camera's matrices are copied straight into
their arrays.
The credits port compared every drawn frame's Cesium credits with the last
frame's by listing them: a sorted copy of the frame's credits, a set and a
string of ids, after every frame. It now first compares what that list is
made from, value for value (the logo, the static credits and the frame's
own, each with its count and placement, and which are the page's), and lists
them only when one of those changed. Most frames owe what the last owed.
- On a renderer whose plugin does not show its credits in a display of
its own (the ownCreditDisplay hook; Cesium's does), the application
shows what the credits port lists at the bottom left (#gev-credits):
inline credits on the line, the rest, the layers' data credits among
them, in the Data attribution list it opens. The button and list are
keyboard operable; Escape closes the list and returns focus. Text is set
as text; links and logos keep only safe schemes.
- The panel, Cockpit and overlay layouts keep clear of it as of Cesium's.
- The stub boot check fails when the credits line is not shown or its list
is empty; it keeps its report when the run stops early, counts only the
dev server's providers and other origins as feeds, and a failed default
renderer load is retried without dropping a newer one.
- When showing the photoreal base fails and the stack before it stays,
its tile-failure listener is bound again for the switch that kept it,
so Esri's OSM fallback still runs.
- The Cockpit test of a ride handed to another aircraft checks that the
follow's change adopted it (the Cockpit took the camera).
AIS Vessels are keyed by the id their live source gives, with the MMSI as an
attribute: an MMSI can be shared, spoofed or reused, so two vessels with one
MMSI are now two vessels. Looking a vessel up or selecting it by MMSI still
finds it.
A share link made while a vessel is selected opens that vessel. A UUID vessel
id travels as 16 bytes in a new thing field (13). A shared thing that is not
reporting now says so in a sentence per kind, with its id (ICAO, NORAD, MMSI
or vessel UUID) on a second line.
A live source may now declare a timeline: it can answer a snapshot at a
given time (source.timeline = { coverage(), stepMs, subscribe? }, read by
readSnapshotTimeline). While such a layer is on it follows the world clock
and registers the coverage. At the live edge it polls exactly as before;
away from it the layer asks for getSnapshot({ ...query, atMs }) at the step
at or before the clock time and shows that snapshot instead of polling.
src/time/snapshotTimeline.js owns the following: a seek, time going
backwards or leaving live replaces every contact with the snapshot at the
new time (play waits on it, for at most 8 s); playing on keeps one
snapshot ahead and gates play on it, applying those snapshots as polls so
contacts move smoothly; outside coverage the layer is empty; back at live
it is cleared and polls at once. One load at a time, and a load the
source drops for a newer one is not a failure.
While following, aircraft are drawn at the clock time with no render
delay, records and absences are judged at the clock, track requests carry
atMs, and vessels move between fixes along the great circle. The world
clock counts jumps (seeks and returns to live) in getState().
A source with no timeline never follows the clock, and every layer stays
live, unchanged.
The radio and push-to-talk helpers shared by both voice providers now pass
English catalog text, and the OpenAI controller words its detail when shown,
like the Gemini session. A token failure without a server message carries
its status as sourceParams, so it is worded again with the number. The docs
say sourceDetail is the message as given, catalog text when it is the app's
own.
- The tests that faked a follow by setting the viewer's tracked entity,
or by raising its change event, follow through the camera port
(followEntity, or a port with follow and followChanged), and check
followedOf: camera handoffs, globe rotation, tilt, ground guard,
vessels, ALPR, CCTV, Contacts, radio, orientation, share links and
follow offsets, the tracked readout and the shell's camera listeners.
- The fake Cesium camera port delivers followChanged from the port itself.
- A ride handed to another aircraft is adopted by the follow's change, as
the application does, not by the test calling the adoption.
- The unbound-factory record lists the scene, the map defaults, share
restoration and the World Packet as entries whose caller loads the
factories; the map startup test stubs the default renderer load.
- The map source controller waits for a base that loads as it is shown
(the globe engine's tilesetReady) before it takes the stack before it
down: one that fails throws there, so that stack and its imagery stay
and the error is reported, as from any failed switch. The Cesium map
engine has no tilesetReady, so its switches are unchanged.
- A failed default renderer load is tried again by the next call, and
the standalone page's early load does not report a failure nobody
awaited yet.
- The adapter ratchet's on-demand import rule refuses a default or an
empty destructuring.
- The stub boot check fails a layer whose manager answers false rather
than throwing, or that is not on (or off) afterwards; a field drawn
counts as drawn; the dev server is closed when the browser cannot
start.
The Gemini session translated each status and error message once, when it
happened, so after a language change the voice control kept showing it in
the old language. The state event now also carries the English catalog text
as sourceDetail, and the control words it when shown. The local server's
known token errors stay catalog text until then instead of being translated
on arrival.
- The standalone page starts loading the default renderer as it starts
(loadDefaultRenderer), so Cesium arrives beside the page's own setup,
not after it; the scene takes a plugin or a promise of one.
- Where a follow target's layer has no display position yet, the followed
position is the target's own position at the frame port's scene time.
- The globe engine's photoreal base that fails to come back is off again
and the globe shown; a later switch tries again. The draped-imagery host
asks the map engine whether the globe is shown where the renderer cannot
say. The scene resizes the renderer with its container.
- Cockpit and voice tests follow through the camera port (followEntity),
not by setting the viewer's tracked entity; the fake frame port has a
scene time.
The application scene takes a renderer plugin and no longer refuses one
without a Cesium viewer. It loads the default (Cesium) plugin only when it
is given none, so a page on the stub never loads Cesium.
- Every layer and tool is handed the application's own scene object
(createSceneHost): the container and isDestroyed(), bound to the
renderer, the same on Cesium and the stub.
- What only Cesium needs at startup is done in its plugin's create: the
free-look height limit, and keyboard access to its credits' expand link
(creditKeyboard.js moves into the adapter).
- What a plugin may hand the application beside the contract goes through
provideApplicationHooks, each hook optional: Cesium's map engine, stale
command trimming, inline panel rendering, viewerDrawsEntity, the zoom-out
guard, and the viewer QA scripts read from window.__godsEyeView.
- The map stacks, and the photoreal base the scene loads, switch through
the renderer's own map engine (mapEngineFor), else its globe port
(globeEngine.js). The Cesium engine takes the scene object and changes
the viewer behind it.
- What the camera follows is the camera port's target only: no layer
tracks an entity of its own any more, so followed.js stops reading and
writing viewer.trackedEntity and the viewer's clock.
- Draped imagery's host is the globe port's imagery while the globe is
shown (isGlobeVisible, or no shown base tileset where a renderer cannot
say), else the shown base tileset; weather uses the same resolver.
- The World Packet reads the eye from the camera port's pose, and vessels
ask again for the view on the camera port's moveEnd.
- The stub's canvas takes event listeners, and in a page is a real canvas
filling the container it is given; createStubRendererPlugin makes it a
renderer plugin.
The viewer access record is now empty.
Every file that read the renderer's canvas as rendererOf(viewer)?.canvas
falling back to viewer.canvas, and the share restoration's gesture
listeners on viewer.canvas, now ask canvasOf(viewer). The 20 canvas
reach-throughs leave the viewer access record.
The Gemini token route now answers a missing key with code API_KEY_MISSING,
the session carries it to the error state as missingKey, and the voice
control hides "Check microphone permission" for that error, as it already
does for OpenAI. The OpenAI controller sends the same flag, so the shared
control and the controller agree whichever updates last.
The band is copied out of its radius table as whole 32-bit pixels where
the image's bytes line up with words, a byte at a time where they do
not. Each front's positions come from the unit vector on the sphere
straight to the ellipsoid, without going through longitude and latitude
and back; they move by nanometres.
Each pulse step painted the band with a square root per pixel and built
each front's per-bearing constants anew. The band now keeps which entry
of its radius table each pixel reads while the subject stays put, so a
step only fills the table and copies it out; the fronts keep their
bearings' constants for the centre they were made for and come out the
same bits.
The voice error was matched against the translated key message to decide
whether to show the microphone hint, so switching language brought the hint
back and left the message in the old language. The error now carries a
missing-key flag into the status, which words the message at render time.
A controller test covers the language switch, and CURRENT-STATE.md documents
the keyless voice response.
The decision no longer reports not-entity-context or channel-closed, which
its only caller rules out first, nor pointer-scale-not-useful, which the
crop path already checks and logs. A closed channel returns before any
record, as on main. The crop decision is pointer-crop-eligible, since the
crop checks after it decide whether the pointer is on ground, and a sender
called on a closed channel logs channel-lost.
In a dense downtown view from about 400 m up, the occlusion margin of
max(3 m, 2% of the range) came to 14 to 20 m, so a wall, awning, tree or
lower facade standing 10 to 20 m in front of a street was read as the
road itself: the vehicle's dot was hidden by the scene, but its bracket and
callout were still drawn. Checked against scene.pickPosition in one such
view, 18 of 470 vehicles in frame fell in that band.
The margin now depends on how steeply the camera looks down on the
contact. A point on the line of sight s metres in front of a vehicle
stands s * sin(elevation) above it, so a surface there hides the vehicle
once that rise passes 2 m. The margin stays at least 3 m and at most 2%
of the range, so grazing and street-level views are unchanged. In the
same views no vehicle with the drawn surface within 3 m in front of it
(its own road) is newly hidden.
From the review of the first change:
- A contact back in view after longer than a retest interval is held back
and retested like a new one, instead of being drawn from an answer
given before it left the view. It gets a fresh record, so a test still
in flight from before cannot end the hold-back either.
- The hold-back for a contact's first answer stretches with the retest
pace when a crowded view slows it.
- A batch whose control point failed no longer counts as a first answer.
- An unknown log-depth flag reads as "no depth" rather than as standard
depth, which would have put every surface in front of every contact.
- A shader that fails to build leaves no shader behind and is not built
again on the same context.
The detection overlay only hid a contact behind the Earth's curve, so a
street vehicle behind or under a building, a 3D tile or a hill still had
its bracket and "VEH-nnnn" callout painted over that geometry, while its
own dot was depth-tested and correctly hidden.
A contact a layer marks `depthTested: true` (only Street Traffic does) is
now tested against the depth the scene drew and left out of the overlay
while a drawn surface stands in front of it by more than max(3 m, 2% of
its range). Other layers are untouched: their markers are drawn over the
scene on purpose.
The test runs on the GPU and never waits for it. Each tested contact is
one point drawn inside an occlusion query; the vertex shader looks up the
frame's packed pick depth at the contact's pixel, decodes it and drops the
point when the surface is nearer than the contact's threshold. Query
results are collected on later frames. Reading depth back was measured
and rejected: `scene.pickPosition` per contact costs 0.2 to 1.4 ms each,
and Chromium answers `getBufferSubData` with a round trip to its GPU
process even behind a signalled sync object (1.5 ms on a quiet machine,
15 to 25 ms on a busy one).
- Contacts are retested every 100 to 200 ms, spread across frames, at
most 256 a frame; with more in view the pace stretches so none starves.
- Hiding is immediate. A hidden contact returns 600 ms after its last
"hidden" answer, so one on a silhouette edge does not blink, and a
contact new to the view is held back until its first answer (400 ms at
most), so orbiting does not flash brackets behind buildings.
- A callout whose contact goes behind something is dropped at once, not
carried to the next solve and not faded out on the facade.
- These times run on painted frames, so a hidden tab does not expire them.
- A control point in every batch must draw; if it does not, the batch is
ignored and after three the probe stands down for ten seconds. With no
pick depth (2D, Columbus view, depth picking off) nothing is hidden.
A WebGL 1 context reads depth back synchronously about 8 times a second.
The renderer internals this depends on are read in one function,
`readSceneDepthSource`, with the engine version they were checked against.