Commit Graph
4726 Commits
Author SHA1 Message Date
Sameh Khamis 25a3b5c7e3 Add the changelog entry for feed types 2026-11-17 17:41:17 -08:00
Sameh Khamis f2537743ab fix(flights): class an aircraft by its feed's type from the first fix
A civil aircraft was drawn as an airliner until its type lookup answered,
even when its feed already named the type: a helicopter showed the
airliner icon and model for as long as the lookup took.

- Flights classes every contact, and picks its model, by the feed's own
  ICAO type designator from the first fix. A looked-up type replaces it
  once the lookup answers and outranks the feed's on every later poll;
  the lookup still runs and fills the type name and registration.
- A surface emitter category still keeps only a ground type (ZVEH, GRND)
  from the feed.
- State vectors have no type field, so an /api/flights answer may send
  `typeByIcao` beside them, by ICAO address, as `emergencyByIcao` is sent.
  The regional adsb.lol fallback now sends it from each row's `t`.
2026-11-17 17:37:48 -08:00
Sameh Khamis e4e03eb58f Credit the trail fix's changelog entry 2026-11-17 17:36:59 -08:00
Sameh Khamis f985f7611d Add the changelog entry for the trail fix 2026-11-17 17:36:51 -08:00
Sameh Khamis 914b28822a The followed aircraft's trail runs to where the aircraft is drawn
When an application places live contacts itself, the tracked trail was cut at the layer's own delayed clock, which can run minutes behind the placed aircraft: the trail ended far behind the aircraft, and its head was dropped while that clock warmed up. The trail now reads the aircraft's time off its own reports at the point where it is drawn, for Flights and Military Flights.
2026-11-17 17:36:36 -08:00
Sameh Khamis 2f0c01eaa7 Merge pull request #1794 from bilawalsidhu/fix/models-black
Tracked and emergency aircraft models keep their paint instead of turning black
2026-11-17 17:06:31 -08:00
Sameh Khamis 39f2d0b165 Merge remote-tracking branch 'origin/main' into fix/models-black
# Conflicts:
#	docs/CURRENT-STATE.md
2026-11-17 16:54:43 -08:00
Sameh Khamis e9cfd02121 Merge pull request #1793 from bilawalsidhu/fix/panel-speed-reported
Panels show the reported ground speed while a placement waits
2026-11-17 16:53:20 -08:00
Sameh Khamis db27ac6b55 Add the changelog entry for models keeping their paint 2026-11-17 16:51:20 -08:00
Sameh Khamis e83c1672dd fix(models): tracked and emergency tints mix into aircraft paint instead of multiplying it
Aircraft models multiplied their tint into the shaded paint (HIGHLIGHT).
A saturated tint zeroes every channel it lacks: the tracked cyan (no red)
drew a red helicopter solid black, day or night, and a white Cessna one
flat cyan. Tinted models now mix the colour in at 30%
(AIRCRAFT_MODEL_TINT_MIX), so the paint stays readable and the colour
still marks them; untinted models still draw their own colours, and the
IR boost stays flat white.

Neither the multi-part models nor the part shader were involved: old and
new GLBs, with and without the shader, render the same pixels, and an
untinted model at night is lit by the default ambient term.

The new conformance check draw.models.paint draws a model painted with
COLOR_0 under a fixed sun and reads its colour by day, at night, and
under a tint mixed in. The parts test model can now carry vertex paint.
2026-11-17 16:50:52 -08:00
Sameh Khamis 3c3934582e Merge branch 'main' into fix/panel-speed-reported 2026-11-17 16:42:58 -08:00
Sameh Khamis adbcdb8e5a Merge pull request #1785 from bilawalsidhu/models/three-km-true-size
3D models only within 3 km of the camera, at true size
2026-11-17 16:34:47 -08:00
Sameh Khamis 95a38b56bc Add the changelog entry for the panel speed fix 2026-11-17 16:31:25 -08:00
Sameh Khamis d79494ac89 fix(flights): panels show the reported ground speed while a placement waits
A host-supplied placement's speed is how fast the marker is drawn. While it
waits for its fixes to catch up it stands still at 0, so tracked-aircraft
panels and the cockpit readout showed 0 m/s for an aircraft in flight.

displayedKinematics now takes derivedFromPlacement: when the drawn speed is a
placement's, the reported ground speed is shown, and the placement's only when
nothing is reported. The placement speed still drives motion and dead
reckoning. Civil and military tracked descriptors pass it from the new
_drSpeedPlaced scratch, which placeLiveAircraft sets.
2026-11-17 16:31:03 -08:00
Sameh Khamis ae43284d3f Merge remote-tracking branch 'origin/main' into models/three-km-true-size 2026-11-17 16:22:28 -08:00
Sameh Khamis 953af60e1d Merge pull request #1792 from bilawalsidhu/fix/parts-check-while-spinning
draw.models.parts: wait for frames while a part spins
2026-11-17 16:09:10 -08:00
Sameh Khamis 1f9d88ca1a Merge remote-tracking branch 'origin/main' into models/three-km-true-size 2026-11-17 16:06:36 -08:00
Sameh Khamis d146c4bb5a Add the changelog entry for the parts check fix 2026-11-17 15:59:17 -08:00
Sameh Khamis 7ea0903fe2 fix(conformance): draw.models.parts waits for frames, not a settled scene, while a part spins 2026-11-17 15:59:10 -08:00
Sameh Khamis dbe632b51f Merge pull request #1790 from bilawalsidhu/fix/vessels-toggle-without-feed
Vessels turn on and off again when their feed fails
2026-11-17 15:58:12 -08:00
Sameh Khamis ca0aec4d66 Add the changelog entry for the vessel toggle fix 2026-11-17 15:46:06 -08:00
Sameh Khamis 14aa85e638 fix(vessels): the layer turns on even when its feed fails
Turning vessels on returned what the first live poll answered, so a
failing feed (such as a missing AIS key) made enable() resolve false and
the layer reported it could not be turned on. loadLivePositions again
resolves with nothing once the poll is done; going back to live reads the
poll's answer through its own path.
2026-11-17 15:45:49 -08:00
Sameh Khamis acca6a1dcf Merge pull request #1789 from bilawalsidhu/models/multi-part-v4
Multi-part fleet models: rotors, propellers, gear and doors as named parts
2026-11-17 15:37:16 -08:00
Sameh Khamis b9cbc42f20 Merge remote-tracking branch 'origin/main' into models/three-km-true-size
# Conflicts:
#	docs/CURRENT-STATE.md
2026-11-17 15:29:39 -08:00
Sameh Khamis 3c3893d18a Merge remote-tracking branch 'origin/main' into models/multi-part-v4 2026-11-17 15:27:35 -08:00
Sameh Khamis 798d0631e0 Merge pull request #1788 from bilawalsidhu/sprites/kinds-history-live
Sprites for helicopters, ground vehicles, stations and vessel types; history to live without an empty map; model parts in Cesium
2026-11-17 15:26:06 -08:00
Sameh Khamis 08ababc4fc Add the changelog entry for the multi-part fleet models 2026-11-17 15:17:42 -08:00
Sameh Khamis 0b751c2118 Multi-part fleet models: named parts (rotors, propellers, gear, doors) on every aircraft, helicopter, vessel, vehicle and transit model; rocket type rules 2026-11-17 15:17:34 -08:00
Sameh Khamis a1bcce0f63 Merge remote-tracking branch 'origin/main' into sprites/kinds-history-live 2026-11-17 15:12:06 -08:00
Sameh Khamis 49eb5b53c8 Credit the changelog entry to the author's handle 2026-11-17 15:10:12 -08:00
Sameh Khamis c0b31e7edc Add the changelog entry for sprites by kind and the history-to-live handover 2026-11-17 15:10:06 -08:00
Sameh Khamis 7186f720df Sprites for helicopters, ground vehicles, stations and vessel types; history to live without an empty map; model parts in Cesium
Aircraft icons are one size; helicopters turn a three-frame rotor while
airborne or moving; ADS-B ground vehicles share one amber sprite and
stations/obstacles one mast; vessels get hull sprites by type and one buoy.
Going live keeps contacts until the first live answer and replaces them by
id; fast playback holds before the loaded positions end and prefetches
ahead. Cesium models move their parts from the optional model part state.
2026-11-17 15:09:54 -08:00
German Khokhlov 301dee2174 Register TransLink Cairns through transitFeed and drop tests that restate data 2026-11-18 02:14:59 +04:00
German Khokhlov 02f0063cca Tighten the TransLink and Adelaide Metro docs, comments and release note 2026-11-18 02:11:42 +04:00
German Khokhlov d7b38c8c62 Merge branch 'main' into feat/australia-transit
# Conflicts:
#	docs/CURRENT-STATE.md
2026-11-18 02:10:07 +04:00
German Khokhlov c08bd1103d Merge pull request #1699 from bilawalsidhu/german/zh-hant-generated
i18n: generate zh-Hant from zh-CN
2026-11-18 01:14:16 +04:00
German Khokhlov f6aeb91063 Merge remote-tracking branch 'origin/main' into german/zh-hant-generated 2026-11-18 00:44:56 +04:00
German Khokhlov 7dcdca19d0 Merge pull request #1700 from bilawalsidhu/german/i18n-plurals-context
i18n: plural forms and contexts the catalogs could only work around
2026-11-18 00:32:47 +04:00
Sameh Khamis 28baddb923 Name the changelog entry for its pull request number 2026-11-17 11:56:59 -08:00
Sameh Khamis 463946abfa Bring the tracking harness's cameras within 3 km of the planes it models
The tracked-model checks zoom the follow camera to within 2 km of the plane,
and the grounded-model check parks the camera 2 km above the planes instead
of 8 km, since models now draw only within 3 km.
2026-11-17 11:52:34 -08:00
Sameh Khamis 1048a71d73 Add the changelog entry for 3D models within 3 km 2026-11-17 11:27:34 -08:00
Sameh Khamis b041fea7a6 Deselecting a plane drawn as a model no longer flashes its icon
A selected plane is drawn by its own model, and its fleet model was let go once
that model drew. Deselecting showed the icon at once, and the icon stayed for
about a second while a new fleet model loaded.

The fleet model now stays loaded but hidden while the plane is selected, kept
where the plane is. Deselecting hands the plane back to it in the same frame,
and the icon shows only when there is no loaded fleet model. Military aircraft
share this code. Vessels never swap models on selection, so they had no flash.
2026-11-17 11:27:21 -08:00
Sameh Khamis 36624cd329 3D models only within 3 km of the camera, at true size
An aircraft, vessel, transit or rail vehicle now draws as its 3D model only
when it is within 3 km of the camera, and as its icon farther away. A model
already drawn is kept out to 3.6 km, so a vehicle on the edge does not flicker.
Proximity and All use the same distance and still differ by how many models
they allow (150 and 350). Models are drawn at their true size: the minimum
pixel size that inflated distant models is gone.

The camera altitude above which no model is considered is now the highest
aircraft the layers draw (30 km) plus the 3.6 km, so a camera high above the
ground still models an aircraft close to it. Cockpit's aircraft silhouettes
keep their own 150/185 km and 400/450 km bands, since they are icons.
2026-11-17 11:21:08 -08:00
German Khokhlov dde8b52f42 i18n: regenerate zh-Hant after merging main
Main added zh-Hant entries by hand, in Taiwan wording (搜尋, 資料, 即時,
儲存, 「」). `npm run i18n -- --zh-hant` rewrites the 91 that are not
zh-CN converted; the 3818 keys and their order stay.
2026-11-17 19:45:17 +04:00
German Khokhlov 0f5869fadd Merge remote-tracking branch 'origin/main' into german/zh-hant-generated 2026-11-17 19:44:56 +04:00
German Khokhlov b63205f69c i18n: tighten the zh-Hant check and test its command line
- Without opencc-js the zh-Hant check says to run npm ci, and the other
  checks still run.
- --prune handles zh-CN before zh-Hant: shippedLocales() sorts, since
  git lists untracked catalogs first.
- A key missing from zh-Hant is reported once, with "run npm run i18n --
  --zh-hant" when zh-CN has it; a zh-Hant entry still in English points
  to zh-CN and same-as-english.json.
- 並道 -> 併道 applies only to entries whose English key says merge, so
  并 (and) before 道路 or 道謝 stays 並.
- Two more corrections for strings main has added: 箇中 -> 個中 after a
  count (全部 {n} 個中), and 宣佈 -> 宣布 like 發佈 and 公佈.
- zh-hant-cli.test.mjs runs the command in a throwaway repository, with
  opencc-js linked as a junction, so Windows needs no symlink privilege.

zh-Hant.json does not change.
2026-11-17 19:44:49 +04:00
German Khokhlov 64b9cef14b Merge main into german/i18n-plurals-context 2026-11-17 19:29:53 +04:00
Sameh Khamis 624c229bc2 Merge pull request #1778 from bilawalsidhu/fix/clock-sources-on-destroy
Launches and satellites leave the world clock when destroyed
2026-11-16 21:30:58 -08:00
Sameh Khamis 9db71ad443 Add the changelog entry for the world clock fix 2026-11-16 21:19:09 -08:00
Sameh Khamis ee306a46fc Launches and satellites leave the world clock when destroyed
Both layers registered their time coverage with the page's world clock when
created and never unregistered it, so an application destroyed and started
again kept each old layer, and its scene and tiles, reachable through the clock.
They now unregister on destroy and register again on init, as earthquakes do.
2026-11-16 21:18:58 -08:00