Commit Graph
2095 Commits
Author SHA1 Message Date
Sameh Khamis 8ff9a1fa29 Base map through the globe port
Cesium switches the map stacks through its globe port, as any renderer
does, rather than through a map engine of its own. Every route keeps its
behaviour:

- Google 3D turned off is kept hidden, and shown again without loading it
  anew; the key route's tiles take clamped marks, as the token routes' do.
- Esri goes to its deepest level (23); its "Map data not yet available"
  picture is a missing tile, so the parent's imagery stays.
- Cesium ion's world imagery is Bing over HTTPS on any page, with ion's
  credits; a token ion refuses refuses the switch, as before.
- Keyless terrain that does not answer is the flat ellipsoid, tried again
  behind it; world terrain asks ion first, as before.
- window.__godsEyeView.tileset is the Cesium photoreal tileset.
2026-11-16 17:06:35 -08:00
Sameh Khamis b1b2ac91c0 Voice: an app can supply what open_stories does
createApplicationTools took the Stories page as the open_stories action
even when the app's voice options gave its own. voice.openStories now
takes precedence, spoken or typed, with the Stories page as the default.
2026-11-16 16:33:46 -08:00
Sameh KhamisandManjunath Paparaju 3a0842c103 Draw the card previews through the renderer, and show vessel photos too
- The route globe, turning model and cockpit view are the renderer's: the
  Cesium plugin hands them over as its `cardPreviews` application hook
  (src/render/cesium/cardPreviews.js), and the card's tiles keep only the
  words, figures and HUD laid over them. Nothing in the UI reaches Cesium,
  and a renderer without the hook leaves the card its photo and figures.
- The turning model is the fleet model the globe draws for the aircraft's
  type, from the set loaded at startup, so opening a card fetches nothing.
  The flights layer's record now carries the type code and emitter category
  that pick it.
- The photo comes from the item's own layer: the flights layer's
  Planespotters photo for the plane it tracks, and the vessels layer's Open
  Vessel Photos photo for the vessel selected, with its credit. It no longer
  waits for Contacts to follow the same aircraft.
- The satellite models are left out: the card shows a satellite's ground
  view and graphs its altitude and speed.
- The route globe's airport codes are drawn into its own small scene, the
  compass reads its course in the active language, the callouts no longer
  speak to screen readers every six seconds, and figures use the app's
  number format. Esri's tile address and credit live in one place.

Co-authored-by: Manjunath Paparaju <[email protected]>
2026-11-16 15:44:35 -08:00
Sameh KhamisandManjunath Paparaju 7552763ea1 Describe the Context panel in the current state
Co-authored-by: Manjunath Paparaju <[email protected]>
2026-11-16 14:59:18 -08:00
Sameh Khamis 56bd14d5f2 Merge remote-tracking branch 'origin/main' into stories/gev-stories 2026-11-16 14:28:09 -08:00
Sameh Khamis 2788a9f667 Docs: the timeline waits at a shot's end while a restarted clip plays 2026-11-16 14:20:55 -08:00
Sameh Khamis 186f8b76d9 Scenes: a clip restarted without a media gate plays its hold out
A resumed hold waited only its rest unless the layer gave a media gate,
and only trimmed clips have one, so an untrimmed clip started again (as
on Gyirong) was cut short. The layer now says when a run going on from a
seek started its clip again, and the hold then plays out as from its
start: through the media gate when there is one, else for the whole hold.
2026-11-16 14:20:55 -08:00
Sameh Khamis 9e6207d7f2 Stories: the button sits left of elements marked data-stories-corner
The corner controls are the ACTIVE STYLE readout and any element marked
with data-stories-corner, looked for again for 20 s after Stories mounts,
in place of an element id this app never has.
2026-11-16 14:14:27 -08:00
Sameh Khamis 0ed6cde956 Idle frame rate in the app, timed line batching, KEY REQUIRED status
Three fixes to how the globe draws and reports loading.

**The idle frame rate is the app's rule.** The Cesium frame port capped the loop at `IDLE_FRAME_RATE` whenever nothing held it, so the rule lived inside the renderer. The render governor now sets the rate: the display's own while the loop is held, `IDLE_FRAME_RATE` while not. The frame port gains `onHeldChange(cb)`, which says when its first hold is taken and its last released (animated visual styles included), so the governor hears every hold. The quality setting's frame-rate cap is kept apart from the governor's rate, and the loop draws at the lower of the two, so neither overwrites the other.

**Line batching is timed.** A line set waited 30 drawn frames before its unchanged lines settled into a static batch (300 for the long wait), so the batch built twice as fast at 120 Hz as at 60 Hz, and took seconds at the idle rate. It now waits `SETTLE_MS` (500 ms) and `LONG_SETTLE_MS` (5 s), and moved shapes go back into their batches on the same timing.

**KEY REQUIRED.** A layer that cannot load without a provider key ended the global status chip LOAD FAILED. The chip now has a KEY REQUIRED outcome, ranked below a real failure: a layer whose stats say `keyRequired` ends KEY REQUIRED, any other failure still ends LOAD FAILED. Vessels report a missing AISStream key as `keyRequired` and name the key in the panel; a later outage clears it, as it is no longer the key that is missing.

Checked: the full suite (new tests for `onHeldChange`, the quality cap against the governor's rate, timed settling, and the chip with and without a key), format, boundaries, build, the stub renderer boot and the Cesium conformance run; in the browser, the idle and held frame rates; and the first-run check now asserts that with no keys set the chip never says LOAD FAILED.
2026-11-16 14:10:01 -08:00
Sameh Khamis 095e4161d8 Scene document: accept the Stories fields from version 6 again
Early Stories builds wrote stories, storyThumb and storyTags in version 6
projects, so those projects load again; the fields stay refused before
version 6. The version 6 round trip is tested again.
2026-11-16 14:00:30 -08:00
Sameh Khamis 6904ce0e7a Scenes: a hold resumed partway waits for the clip it started again
The resumed hold reads scene media like any hold: a clip still playing
is waited for, and only without one does it wait out the rest of the
hold.
2026-11-16 13:55:07 -08:00
Sameh Khamis b2c7144cce Scenes: a run resumed inside a hold plays its layers on from that moment
After placing the layers at the seek's moment, the hold settles them
without the seek, so the reveal continues from where the seek put it and
a clip the moment falls in starts again from its beginning (a clip
cannot be joined partway); one the moment is past stays finished.
2026-11-16 13:54:40 -08:00
Sameh Khamis a9990018ea Stories: name the Opened in GEV toast for what it is 2026-11-16 13:48:14 -08:00
Sameh Khamis 20f6ceb74c Scene document: accept the Stories fields only from version 7
stories, storyThumb and storyTags arrived with version 7, so they are
gated by version like anchors and dataPacks.
2026-11-16 13:40:17 -08:00
Sameh Khamis 4fa80fb18c Stories: a teardown that times out cannot undo the restored view
directorSettled now tells a timeout from settlement, and on a timeout
revokes the scene's layer cleanup still in flight before Close puts the
view back.
2026-11-16 13:31:51 -08:00
Sameh Khamis 0a39b05cdb Authoring: pick any scene again, and select the first scene on load
Restores the scene picker and its selection on load, after a delete and
after an import, so PLAY and NEXT work after a reload. Open in GEV
selecting a shot and NEW SCENE stay as they are.
2026-11-16 13:30:22 -08:00
Sameh Khamis 8734f5e515 Stories: Play resumes at the exact paused or sought moment
startScene takes fromProgress and starts partway into a shot: the camera
goes on along the rest of the flight, the hold waits only what is left
and the scene clock carries on from the same point.
2026-11-16 13:24:59 -08:00
Manjunath fec84dbf78 Address Stories review feedback 2026-11-16 22:35:07 +05:30
Manjunath 2edcd5bc69 Fix Stories history and paused playback 2026-11-16 21:38:14 +05:30
Manjunath 472cc35c47 Stories: phone entry, centred framing, review follow-ups
- On GEV's phone layout, where neither the style readout nor the account
  avatar shows, the Stories button goes icon-only under the corner-actions
  column instead of covering the top row of icon buttons.
- Explore frames its start view and a picked story in the middle of the
  map it leaves visible (beside the feed, above the phone sheet), with one
  short correction if that area moved during the flight.
- The share icon stays visible in icon-only header buttons.
- A world clock nothing moved while Stories was open (no jump, same mode
  and rate) is left alone on Close.
- Card images use the globe's own Esri tile URL; the credit list keeps the
  application's full-screen placement on phones.
- Tests cover the live and playing clock restores, the clock coming back
  after the share state, the tile-settling hold being released (settled,
  10 s, Close), the visible-map aim, and the near-end Play fix in the
  director's real order. One changelog fragment covers the PR.
2026-11-16 21:08:23 +05:30
Kevin Rajan 8a0953c35d fix(stories): retain the last shot when replaying near the end (#1724)
Focused follow-up to manjunath22466's Stories contribution.
A paused share link in the final 250 ms is not a completed scene.
Add node:test coverage for near-end Play and true-end replay; update
current-state documentation and the #1724 changelog fragment.
2026-11-16 20:45:25 +05:30
Manjunath b8e622aac8 Stories: reach the globe through the renderer's ports
Main now keeps Cesium behind the renderer contract, so Stories stops
reaching Cesium and the viewer directly.

- Camera: flights, the -60 degree story framing, cancel, the pose saved
  on open and put back on Close go through the camera port; a follow is
  released with stopFollowing.
- Pins: placed in the frame port's postRender from the query port's
  projection and horizon test, still written straight to each pin's
  transform on every drawn frame.
- Tile settling: a continuous-render hold until the globe port reports
  the base settled, for at most 10 s.
- Clock: Close and Back to Explore put back the page's world clock (live,
  or the time it showed, playing on), and leave a clock that is still
  where it should be alone.
- Card images: drawn from Esri World Imagery's own Web Mercator tiles
  while the Esri map stack is active, read as images; no Cesium imagery
  provider is touched.
- THUMBNAIL: the director hands captureStoryThumb the frame port and the
  renderer's canvas, so the scene document's module stays renderer-free.
- Credits: Stories places and keeps visible the application's credit
  line (#gev-credits) and its list as well as the renderer's own display.
- Tests fake the ports instead of a Cesium viewer, and cover pins per
  frame, the flights, the settling bound and the clock restore; the
  share link tests Stories added build their manager over a camera port.
2026-11-16 20:43:53 +05:30
Manjunath 9e9ce1db22 Fix Stories navigation, sharing, and voice entry 2026-11-16 20:12:10 +05:30
Manjunath 5418438339 Stories: GEV scenes as stories on the live globe
A Stories button (top right) opens Explore: a feed of GEV scenes over
GEV's own Cesium globe, with a pin per story. Explore opens at country
level over the first story and, while idle, flies on from story to story
for a minute. A picked story is framed in the middle of the view; a pin
plays its story on a second click. The Story page plays the scene through
SceneDirector with a timeline, chapters, key moments, sources and
background, then Close (or Back) puts the person's camera, clock, layers,
style and AUTHORING selection back. Open in GEV hands the scene to the
Scenes panel at its current shot.

Stories lists the scenes GEV ships with a camera path (while they are in
the project), the scenes saved with SAVE TO STORIES in AUTHORING, and the
person's other scenes as drafts. The person's own stories live in the
project on this device, so they have no Share; a link to a story that is
not here opens Explore and says so. Each shipped story's
context comes from what GEV knows: what it shows, the place each shot goes,
and each layer's data sources as DATA_SOURCES.md lists them; the Nepal
story carries the event pack's facts and CC BY-NC sources. Card images are
a captured THUMBNAIL or drawn from the keyless Esri World Imagery.

AUTHORING has no scene picker: it starts with no scene, NEW SCENE (or
CAPTURE SHOT) makes one, and Stories' Open in GEV selects one. Before SAVE
TO STORIES it offers topic tags and THUMBNAIL. A scene's stories flag,
storyThumb and storyTags are saved with the project and kept through
export and import.

Only the closed state loads with the page; the screens, styles, story data
and controller load on first open. Stories keeps its place in the query
(one history entry per screen) and holds the share link's hash writes
while open. Keyboard input and voice tools stay with Stories while it is
open. Explore's pins move on every drawn frame without asking for frames;
the idle tour and tile settling are bounded.

All strings are translated in the 13 locales.
2026-11-16 20:12:10 +05:30
Sameh Khamis 256e992372 The contract text speaks of the renderer 2026-11-15 19:20:48 -08:00
Sameh Khamis 9862b2a165 Contract and docs text describe the ports and the Cesium renderer only 2026-11-15 19:08:46 -08:00
Sameh Khamis 4b0958b9e8 Merge main (87ab9d2b1, #1758 and the #1757 revert) into the renderer branch
Grounded aircraft turn 3D on the first settled frame: streaming tiles are not a miss. groundSnap's tests keep main's cases against the branch's query port; trackedModelForm.test.mjs (the branch's form of trackedModelRegime.test.mjs) takes main's edits.
2026-11-15 18:42:06 -08:00
Sameh Khamis 2ac157aca2 CONTACTS takes the selected vessel as its subject instead of following the nearest flight
Turning CONTACTS on with a vessel selected made the nearest flight to the view the subject and followed it. The selected vessel is now adopted as if it had been selected while CONTACTS was on; a followed flight still comes first, and with nothing selected the nearest flight is still focused.
2026-11-15 18:33:29 -08:00
Sameh Khamis 8dc2371ed8 Merge branch 'main' into flights/ground-models-at-once 2026-11-15 17:53:45 -08:00
Sameh Khamis d232d9afe4 flights: grounded aircraft turn 3D as soon as the tiles settle
Tiles still streaming no longer count as a failed ground sample: they leave
the retry backoff alone and return the last measured ground, if any, so the
first tick after the tiles settle samples and hands the icon to the model.
2026-11-15 17:41:37 -08:00
Sameh Khamis 0ab31d4fb8 The renderer seam: clicks go to the newest tool, recession and display groups through the contract, drawing at parity with Cesium
- Clicks: when two tools want the globe's clicks, the newest takes them and the older is told and turns itself off; a click is decided when it arrives, and ambient selection fires only when no tool holds the pointer. The "map is busy" paths and their strings are removed.
- Recession: the limb fade is shared (src/render/recession.js) and reaches icon and model sets through SetSpec.recede; tuning applies on the next frame.
- Display panel groups are published through src/ui/displayGroups.js; the Weather group and src/viewWeather.js are removed.
- Draw port: a set whose icons only move is moved in place, and models that did not change are not restyled each frame, so moving fleets cost what they did when drawn directly; an untinted, fully opaque model is drawn without a colour.
- Conformance: input.lease.lost checks the new lease inside onLost and reports a throwing onLost; draw.sets.recede covers model sets.
- QA scripts read the renderer host instead of internals the seam replaced.
2026-11-15 17:03:09 -08:00
Sameh Khamis a0d299523d Merge main (663b56f62, #1755) into the renderer branch 2026-11-15 16:34:05 -08:00
Sameh Khamis 1f6e8914f2 vessels: move drawn vessels following the clock once they would move half a pixel
A profile of about 9,000 vessels following a playing world clock spent
~13 ms of every frame in the clock pass, moving every drawn vessel a
fraction of a pixel, and ~4 ms more re-uploading their billboards.

- Drawn vessels move again once the fastest would have moved half a pixel
  on screen at the speed each last moved (measured from the world time it
  was drawn at), at least once a second, and at once when the camera moves
  or changes height by a tenth of its height or the view zooms. A pass that
  holds them back asks for a frame, so a paused clock settles on the time.
- Hidden vessels move a share of them each pass, each within 250 ms, rather
  than all of them in one frame; the visibility pass moves one coming into
  view to the time first.
- A vessel's three positions are written from one surface point and normal
  (Cesium's own formula) instead of three fromDegrees calls; the unused
  normal is dropped. The great-circle step builds no arrays.

The pass falls from 4.6 ms to 0.37 ms a frame on average in a benchmark of
9,000 vessels (3,000 drawn) at 240 frames a second.
2026-11-15 13:35:11 -08:00
Bilawal SidhuandClaude 04d08d1c7e README: lead with the official site, the site's clips and copy; fold reference sections; drop launch-era GIFs
The top of the README now matches godseyeview.ai: official-site line, hero clip, a twelve-tile feature
grid and the video series. Docker, Codespaces, the layer table, Street Level and the voice provider
notes are folded into collapsed sections with their text unchanged. The Field Missions and Cockpit
showcase sections and the 17 launch-era GIFs are removed (68 MB of media down to under 8 MB).
package.json homepage points at godseyeview.ai.

Co-Authored-By: Claude <[email protected]>
2026-11-15 19:41:33 +00:00
Sameh Khamis d49552af57 Merge remote-tracking branch 'origin/main' into flights/course-from-fixes
# Conflicts:
#	docs/CURRENT-STATE.md
#	scripts/package-boundaries.json
2026-11-15 11:26:28 -08:00
Sameh Khamis fc4db81d0b Course from fixes only moves forward in time, from an anchor fix, and not for a standing contact; military contacts too
A fix that is not a second newer (a repeat, or an older position from another
feed) never gives a course, so a source switch cannot turn an aircraft
around. The course runs from an anchor fix that moves only when a course is
read, so frequent short steps add up, and a contact reported under 2 m/s
reads none. The helper moves to src/data/courseFromFixes.js and the military
records use it as well.
2026-11-15 11:19:58 -08:00
Sameh Khamis 537a5d6e0f fix(flights): an aircraft with no track or true heading faces the way it moved, not north
Some feed rows carry neither a ground track nor a true heading for an
aircraft on the ground. Its course then fell back to 0, and an aircraft
taxiing south was drawn facing north, moving backwards or sideways. With
neither reported, the course is now read from its last fix to this one when
they are at least 30 m apart; nearer fixes are position noise and keep the
last course.
2026-11-15 11:11:35 -08:00
Sameh Khamis 4b47d5e058 Merge main (8d9c9071, #1750) into the renderer branch
Main brings satellites and vessels moving every frame, smooth rewind for
contacts, global clouds drawn to the level a feed names, and search for
degrees-minutes-seconds, UTM, MGRS and plus codes. They keep this branch's
renderer work: satellites write their reckoned positions into the points set
once a frame; vessels move every frame and ask for frames from the camera pose
and canvas, and a vessel not there yet is hidden through its icon; flights and
military keep the branch's rendering, with main's timeline fixes, trail top-up
and grounded lift; the cloud tiles keep their infrared transfer. The timeline
history module uses the geo helpers instead of Cesium, and main's new tests
use the draw port's items.
2026-11-15 11:00:24 -08:00
Sameh Khamis be77434ddd Merge branch 'main' into render/uncap-frame-rate 2026-11-15 10:43:12 -08:00
Sameh Khamis 9f4254e962 perf: draw at the display's refresh rate while anything moves
Remove the fixed 60 FPS cap. The render governor caps a parked scene at
60 ticks a second, since Cesium still runs the clock, data sources and update
listeners on every animation frame when it skips the draw, and holds
continuous render while the camera moves (moveStart to moveEnd, with pose
sampling for cameras that drift by rounding errors and never raise moveEnd).
Waits that counted frames (a tracked contact's first position, a shared
follow view, a trail's temporary copy) now count drawn time through
src/time/frameWait.js, so they last the same at any refresh rate.
2026-11-15 10:41:36 -08:00
Sameh Khamis 317d06adf6 GNSS Integrity on H3 resolution-4 hexagons, as gpsjam.org
Aircraft are binned into H3 resolution-4 cells (h3-js) and each cell is
drawn as its hexagon instead of a 0.5 degree square. As on gpsjam.org,
every hexagon with an aircraft is shown: the three-aircraft minimum is
gone, and discounting one bad aircraft is the only small-sample guard.
Cells are keyed by their H3 index and carry their centre and boundary.
2026-11-15 10:19:59 -08:00
Sameh Khamis 2b5fb51684 Contacts move back as smoothly as they move on
A snapshot timeline may declare spanMs: each record of a snapshot at a
time then carries its own fixes from spanMs before the time to spanMs
after it. Flights, military flights and vessels following the clock are
placed on those fixes, so a seek or a move back within the fixes loaded
is no jump: nothing is cleared and play does not wait. The step there is
asked for once and stands in for every contact; those it has keep their
place. Fixes are merged across snapshots at least 20 s apart, at most
240 a contact, and a contact is not drawn more than a minute before its
first fix.
2026-11-15 09:27:31 -08:00
Sameh Khamis 97bce938c4 Global clouds drawn from a feed's tiles to the level it names
A weather feed that supplies its own tiles may name, per product, the
deepest level of the geographic tiling its tiles hold detail at
(frameMaximumLevel, 0-12). Global infrared is then drawn from those tiles
to that level, each given the regional infrared display transfer, rather
than from one 2048x1024 image cut to level 3. Other tiled products take
the named level in place of 6. A feed naming no usable level keeps the
layer's own detail.
2026-11-15 09:14:33 -08:00
Sameh Khamis c2fa15b205 Merge branch 'main' into motion/vessels-frame-cadence 2026-11-14 23:37:10 -08:00
Sameh Khamis 52695daf5c fix(vessels): world-clock motion moves only drawn vessels, once per world time
Vessels out of view move on every 250 ms so one coming into view is current;
vessels at the end of their fixes ask for no frames; nothing moves while the
layer is off; 2D views judge the half-pixel step across the view's width; a
re-floor redraws vessels where they are on the next pass.
2026-11-14 23:30:38 -08:00
Sameh Khamis 57bacd98a8 Move vessels every frame and ask for frames as they move
Vessels between two fixes, and live vessels a host placement places, were
redrawn at most every 250 ms, and a frame was asked for only once a second.
With nothing else asking for frames the scene drew vessels at about 1 to
2 fps live, and none at all while the world clock played at 1x.

They are now redrawn every frame the scene draws, moved in place without
new allocations. After each pass the layer works out how far the fastest
drawn vessel moved on screen and asks for the next frame for when it will
have moved another half pixel, at least once a second while any drawn
vessel moves. A vessel crossing the view close up therefore glides, and
one far away lets the scene idle. A frame already asked for later is
brought forward when a vessel speeds up.
2026-11-14 23:18:57 -08:00
Sameh Khamis e31de29ccc Move satellites every frame between SGP4 samples
Satellite points were propagated once a second and held still in between,
so while the world clock plays faster than real time they jumped: at
1 min/s, about 45 px once a second at globe scale.

Each SGP4 call now keeps a sample of the satellite's inertial position,
velocity, gravity and its rate of change (src/layers/satellites/motion.js).
Every frame each shown point is reckoned from its sample to the world time
and turned into the Earth-fixed frame by the sidereal angle, so points
glide at any play speed. Core satellites still take a fresh sample every
second, and at once when the world time is 30 s from it (fast play, a
seek), which keeps every point within about 10 m of SGP4. Starlink extras
take their first sample as they load and are resampled round-robin, more
per frame the faster time plays (up to 2,000 a frame); a point more than
two minutes from its sample waits for the next one. Hidden points are
moved when shown again, and the tracked satellite keeps its per-frame path.

Measured at globe scale: at 1 min/s points move on every frame, 0.4 px a
frame and at most 1.4 px (before: once a second, 45 px). With 11,880
satellites (Starlink on) the scene holds 60 fps, frames 16.7 ms median and
19.7 ms p95; reckoning costs 0.02 ms a frame for 1,342 points.
2026-11-14 23:18:28 -08:00
Sameh Khamis 0ecbfbecd0 Merge main (e4708a81, #1745) into the renderer branch
Main brings the reviewed one-time-control and timeline-mark changes. They
keep this branch's renderer work: earthquakes draw through the draw port and
register their strong quakes as marks; the weather clock keeps `follows` beside
`world`; the weather layer measures ages and delays from the timeline time;
the Visual panel keeps its sky and weather groups under Effects. Tests that
moved time through the weather clock now move it on the world clock.
2026-11-14 22:08:39 -08:00
Sameh Khamis 961e6321a5 Take the timeline commits back out before merging their reviewed versions from main
5150279cc, 97bd61711, 900c76425 and 4f23ad330 went to main, reviewed and
changed, as #1745. They are taken out here so main brings them in as
merged, rather than as two copies of the same changes in conflict.
2026-11-14 21:52:58 -08:00
Sameh Khamis ac04812db2 A row plays only with a timeline to play; the speed shown follows the world clock; mark labels merge whole and cut by character
A weather row's play started the world clock even with no time to show,
where there is no timeline; it now does nothing then, as the timeline's
own Play cannot be reached. A weather layer made without a weather clock
does nothing once destroyed, so enabling it after builds no clock.

The timeline's speed follows the world clock's rate whoever set it, and
its Play runs at that rate, so it and a row's play never run at different
rates; a rate not offered shows as the nearest offered speed. Choosing a
speed sets the clock's rate first, also while a driver steps.

Marks merge on their whole label and are cut for showing after, so two
whose labels only differ past the cut stay two, and the cut counts
graphemes (code points where the runtime cannot tell them), so an emoji is
never split.

Tests that compared a DOM element with null built an assertion message
from the element when they failed, which hung the run; they compare plain
values now, so they fail at once. The timeline tests put the clock back to
real time after each.
2026-11-14 21:12:43 -08:00