2ff11a385e23a1eebff76c5b88a1aa9c240311d5
8 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 karti
|
2ff11a385e |
fix: the sky meets the sea, and the places rail admits it scrolls
Two defects a photograph found and no gate could. **The sky seam.** A phone held upright drew a hard horizontal step across the top 10.5% of the frame — a flat duller slab above a warm gradient, meeting in one pixel. Painting the sky dome red proved the slab was the dome itself and everything below the line was sea; so this was never missing geometry, it was two halves of one join disagreeing about a colour. The sea is a flat plane, so however wide it is made the camera's own far plane cuts it at an elevation *below* the horizontal — about 12 degrees in a plan view of the state. atmosphere.ts has fogged it to exactly the horizon colour by then, which is the entire mechanism by which water dissolves into sky instead of ending in an edge. The dome at that same angle has to be the same number, and it was not: the below-horizon darkening ramped to its full 0.82 over the first 11.5 degrees. The sea arrived at the horizon colour and met a sky already a fifth darker. Measured 155,107,99 against 127,88,81 across one pixel. Holding the horizon's own colour for the first 30 degrees costs nothing the darkening was for — a floor or an ocean covers that band whenever there is one, and where there is not, matching is the whole point. Measured after: the worst single-row step anywhere in the sky falls from 15/255 to 8, and the 8 is water texture rather than a line. render/skyHorizonSeam.test.ts parses the shader text, which is a weaker instrument than a frame and the strongest one available with no GL context. It asserts the one number the picture proved wrong — how far down the dome still paints the horizon colour — and it caught a real off-by-a-hair while being written: 0.42 is sin(24.8 deg), just inside the 25 deg the sea can reach. It also guards the backtick-in-a-template-literal trap that has now broken this same shader three times. **The places rail.** Its cap is min(22rem, 42vh) and rows are variable height, so it always landed mid-row, and a row sliced through its own text reads as a button that failed to draw rather than as more list. markOverflow writes data-overflow and the stylesheet fades whichever end has list behind it. Every measurement in it is guarded, because ui/mount.ts is deliberately tested with no DOM at all: no geometry must mean no attribute, not data-overflow=NaN. A pixel of slack at each end, because fractional layout reports a fully-scrolled list a fraction short and a fade that never switches off dims a row over nothing. stylesheet.test.ts asserts the two files still speak the same four words — nothing else connects them and the failure is silent. 1,664 + 295 tests, all gates, budget 779,789 tri / 150 draws / p95 16.7 ms. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
karti
|
acf4d1a510 |
feat: the state becomes California, and the port fills with ships
**Stage 1 of one California.** The owner's complaint had two halves and this is
the first: the state board was a CROPPED SLAB. `california.ts` stopped at 38.05 N,
so the board disagreed with its own minimap about the shape of California in a
single frame, and Bug Fire's 93,733 acres burned off-frame while the panel said
all clear. Bounds now run 32.50-42.05 N / -124.50 to -114.0 W — Cape Mendocino,
the ruled Oregon parallel, the 120th-meridian corner into the Nevada diagonal.
**And it got cheaper.** 391,169 triangles to 375,351, while gaining the North
Coast, the Sacramento Valley, the Klamath knot, the Cascade arc, Shasta at 4,320 m
and Lassen at 3,190 m. Extending the bounds alone would have doubled the lattice
to 168,813 points and blown the mobile cap; coarsening cellLat 0.022 -> 0.0312 and
cellLng 0.027 -> 0.0383 holds it at ~83,800. The cell as a FRACTION of the board
moves 0.0030 -> 0.0033 — unchanged in frame — because the camera retreats to frame
whatever it is given. That argument was already written in the pack's own comment.
The second half — three boards becoming one world you zoom through — is NOT here.
Merging at Bay density would be 34.04M triangles, 13x the highest budget, and
merging at SoCal density would downgrade San Francisco from 40 m lots to 164 m.
Both delete the board every marketing still is shot from. `sf.ts` and `socal.ts`
are untouched by design.
**Aerial perspective, which the state board could not have had before.** The old
fog started at 1.15 board spans = 944 km, on a board whose longest diagonal is
820 km — so no pixel could ever be fogged. Fog now responds to camera altitude,
clamped to the authored pair as a ceiling.
`Atmosphere.aerial(env, view)` is a second pure method returning `{ near, far }`
and **deliberately no colour**. That is structural, not stylistic: it is why a
future camera-dependent term cannot reach `environmentKey()`'s colour fingerprint
and start rebuilding the PMREM cubemap on every camera step. Coarsening the
fingerprint instead would have hidden one instance and armed the mechanism. A
mutation-tested seam guard fails if anyone merges the two paths back together.
**The port.** Terminal Island rendered as a bare tan polygon with generic white
blocks while the chapter text called it the busiest port complex in the
hemisphere. Now six container yards drawn as canvas atlases, 56 gantry cranes at
varied boom angles, the 13 km San Pedro breakwater, the dredged channel. Five
buckets merging ACROSS ports the way airports.ts merges across fields, so a
second complex costs no extra draws: +11 draws and +4,377 triangles for all of it.
At vertical exaggeration 3.4 a 130 m gantry is 1.132 units tall against a 400 m
ship's 1.024 long — the crane is the taller object, and it is what makes a port
read as a port from altitude.
**Ships, and the wake carries the information.** Moored hulls have no foam,
verified at three terminals; a tug under way in the Main Channel trails a clean
Kelvin V. One hull geometry, one InstancedMesh, orientation from the BERTH rather
than the wire. The AIS gate strips sog 102.3, heading 511 and cog 360 — all mean
"not available" — with an explicit test that cog 358.7 SURVIVES, because a naive
range check on cog eats real headings near north.
"Empty or full" is not in AIS position reports and is not invented per ship. The
honest answer is at port level and is a better story: 348,691 of 460,467 boxes
left Los Angeles empty in July 2026, corroborated by FBX01 $7,491 inbound against
FBX02 $347 outbound.
**Radar and birds ship dark, and say why.** California is 0.47% wet and migration
is nocturnal and seasonal, so both layers have nothing to say on most days. The
panel reads "No radar feed is configured, so this board draws no weather. That is
a fact about this box, not about the sky."
Also recorded, and it matters beyond this commit: **the GPU on amd-server never
leaves 500 MHz of a possible 2725**, traced across 80 seconds of sustained load.
`bay-area/desktop` is fragment-bound at that clock and sits on the vsync deadline,
so a trivial change in fragment work flips it between 16.8 and 33.3 with geometry
identical to the digit. Every frame-time number measured on this box is a floor.
Two investigations reached two different wrong conclusions from single-run
comparisons before this was traced. Geometry is the gate; frame time is advisory.
No cap was raised.
Tests 1,340 -> 1,540, server 280 -> 295.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
||
|
karti
|
b25f217e3e |
feat: real fire on the boards, the LA office as a twin, and a night sky worth reading
The world stops being a simulation of California and starts being California. **THE PROMOTION GATE WAS THE FIRST COMMIT, BEFORE ANY ORANGE PIXEL EXISTED.** On today's live store the SoCal board contains 22 incidents. Every one has NULL acreage and fifteen are nameless LA County dispatch numbers. Drawn naively that is 22 orange marks over Los Angeles on a day nothing is burning — in a frame that contains no other warm colour, so one glyph would be the most salient object on the board and twenty-two would spend its credibility permanently. `acres >= 10 AND contained < 80 AND type != 'RX' AND last_seen = max(last_seen)` returns 0 on SoCal, exactly 5 on California, 0 on the Bay — same body, same day, three correct answers. The empty board is a deliverable, not a fallback: it says "No active fire on this board — CAL FIRE and WFIGS, just now", states that 21 records were gated and why, lists the largest fires burning OUTSIDE the frame with distances, and counts the hot pixels it is deliberately not drawing. **The privacy leak is structurally impossible rather than carefully avoided.** cloud-1 serves a projection; the four home-relative columns never leave that box. `observations.threat` was the one that nearly got through — it is `(16/distance)^2 x log10(acres) x momentum x containment x wind-alignment`, so with acreage and containment public it inverts to a distance circle around a house and three fires give an intersection. A grep of the built bundle for distance_km, bearing_deg, threat, 7762 and the street name returns nothing. **Deliberately not used, and both would have produced a confident wrong answer:** the store's `air` table retains only the last parameter of each poll, so all 78 rows read "Good" while the live feed reports ozone 101 "Unhealthy for Sensitive Groups" — haze driven off it would clear the sky during a smoke event. And `weather` is written only inside the NWS alerts loop, so a quiet day stores no wind at all. Tera's own per-region NWS wind is already correct and already what the clouds drift on. Satellite detections are drawn as evidence and never as incidents. The permanent industrial heat source 4.7 km from the owner's house is flagged persistent and dropped, asserted by a test that first proves it is present in the fixture. MODIS integer confidence and VIIRS string confidence are branched on `sat`. **The LA office is a twin.** Its entire authored second storey — Model Loft, Model Bay, The Materials Room, 430 lines nobody had ever stood in — is reachable on foot: a walker crosses level-1 to level-2 in 73 fixed steps, floorY 0 to 5, verified against the real pack rather than a synthetic plan. Its two studio devices read real hardware through a field-allowlisted bridge: mute, volume and reachability only. Never level, because there is no passive level upstream and obtaining one would record a room with people in it. Never dB, because upstream is gainPct across four different native scales. The bridge refuses all writes. Fixed at its root: an anonymous visitor was getting permanently at-rest instruments backing off against a 401. The tier moves into `createDeviceSource`, so anon gets the living simulator three file headers already promised. **Item 8 is closed, not fixed, and the correction is the point.** The Bay Area "stutter" was GPU power management — the card sat at 500 MHz of 2725 through every run that reproduced it, 4096/2048/1024/256 shadow maps all render in 1.21-1.31 ms, and two consecutive runs over a byte-identical dist gave 33.4 then 16.7. The allowance is removed and the cell is back to 16.7. Geometry is the gate; frame time is advisory. Item 7 was re-scoped after measuring: 1,069,006 of the Bay Area's 2,265,056 triangles were the second submission of the same buildings into the shadow pass. Mobile now has its own triangle caps and bay-area mobile draws 1,266,096. Also: bridges and the freeway corridor light up at night as emission, not lights — 1,614 deck lamps and 18 tower heads on the Bay in two draw calls. The single change that made US-101 legible was moving its edge lines from the lit material to the unlit one: retroreflective paint, the argument the SFO night frame already makes. California went 21,991 lamps to 4,051, clustered at the 17 town districts, because a rural interurban corridor genuinely is unlit. Tests 1137 -> 1340, server 280. All ten budget cells pass on first attempt with no cap raised. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
karti
|
8fb85cd2e5 |
feat: give the boards a horizon, a sea that reflects, and a state worth flying over
The wide shot, which is what an anonymous visitor actually lands on. **The sea was `MeshLambertMaterial`** — a material with no specular term at all, by construction — on a board where water is half the frame. It is now a low-roughness dielectric that reads `scene.environment`, with a runtime-generated tiling swell normal map sampled twice per fragment at two scales and two headings, so the sun breaks into a moving glitter path instead of a mirror point. An `onBeforeCompile` patch takes the body colour toward the deep value looking straight down and leaves it to the reflection at grazing, and walks roughness up past 1.6 board spans so the far water cannot shimmer. The swell spectrum is 1/k^2 and not 1/k because the first attempt was photographed: at 1/k every component carries the same slope, the shortest wave wins, and the sea renders as hard diagonal corduroy. A test holds it now. **The board no longer ends in a diamond.** The sea plane went from 1.8 board spans to 18, past the fog's far plane from anywhere the orbit reaches, and the sky is a world-space dome rather than a screen-space gradient. That gradient was wrong in a way dusk made obvious: the sunset band was painted along the *bottom* of the picture, under the board, while the true horizon at the top of frame stayed zenith blue. `daylight.ts` pinning the horizon stop to the fog colour to hide the seam was a symptom of it. **Terrain casts shadows.** Left off before because double-sided terrain against a ~16 m-per-texel shadow map gives acne; `shadowSide = BackSide` is the cure, shot at four sun elevations down to +0.0 degrees to confirm no stippling. The caster is a stride-2 decimation appended to the same index buffer and swapped in by `onBeforeShadow`/`onAfterShadow` via `drawRange`: no extra draw call, a quarter of the depth cost, and indistinguishable from the full-resolution caster in a side-by-side crop. Stride 1 was measured at +65,566 triangles and would have missed the budget by ~47,000, so it was not shipped. **California reads as California.** It was a beige kite: the eastern edge one ruled line for five degrees of latitude, the south closing in a diagonal V, the whole south-east a featureless tan wedge. Now the coast runs to the Mexican border with San Diego on it, the eastern edge follows the Colorado and the Nevada diagonal, and the south-east is the Basin and Range — forty parallel desert ridges throwing shadows east, Death Valley as a white pan between the Panamints and the Black Mountains, the Salton Sea the one cool value for two hundred kilometres. The opening pose is retuned to the bigger board; the old 452/392 stand-off left a slab of empty ocean where the state should be. **The aircraft were six pixels.** Measured, by enlarging a screenshot 200% to find one at all — indistinguishable from a dead pixel, on a board whose entire claim is that the sky is live. They are airliners now, with planform and trail, and clicking one raises its card for a signed-out visitor. **The Model X is off the wall.** It stood at floor level outside a studio 188 m up a Transbay tower, reading as a car balanced on a parapet. The apron is now chosen from `site.elevation`, which the pack already carries — not from an office id, which is the bug class this repo already hit once when a door marker gated on `id === "sf"` and would have pinned the Los Angeles building to San Francisco. Also fixed, and nearly shipped: sea z-fighting dithered every flat piece of ground on the Bay Area and SoCal boards. And one test asserted an exact source line for the water material, so the better multi-line implementation failed it — it now asserts the property (dielectric, metalness 0, low roughness) rather than the author's first guess at formatting. Tests 964 -> 1015. California desktop 562/650 draw calls and 728,744/750,000 triangles — 2.8% of triangle headroom left, which is the number the next person should check first. No budget was raised. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
karti
|
51979feea0 |
Shadows land on the building, and the sky layers stop repeating themselves
**Shadows were the right size and pointed at nothing.** Last round fixed the missing `updateProjectionMatrix()`, so the frustum finally became the size every caller asks for — but nothing aimed it, and `sun.target` sits at the world origin. A pack's origin is the **north-west corner of its slab**, so for lumbridge-hq the box was off-centre by half the building: 14.4 m of a 48 m plate, about a third of the floor, fell outside the frustum and neither cast nor received. Invisible while three's broken ±5 default made shadows useless everywhere; obvious the moment they started working. `SceneKitOptions` takes a `shadowTarget` now, both callers pass one, and the light's target is added to the scene — which is the part that actually matters, because `LightShadow.updateMatrices` reads `target.matrixWorld` and an unparented `Object3D` is never reached by the traversal that updates it. The sun is also placed relative to the target rather than the origin, so light-to-target is exactly `sunDistance` for every direction, which is the invariant each caller's `shadowNear`/`shadowFar` were chosen against. **`flights.ts` could not be tested, and that is why it was untested.** It used a TypeScript parameter property — the one piece of TS syntax that *emits code* rather than annotating a type — so Node's type stripping refused the whole module. The bundler never cared, so nobody found out until the first `node --test` file tried to import it. The module carrying the worst bug this project has shipped was, by construction, the one module that could not have a test. It has eleven now, including one that fails if the live-aircraft repeat-skip is removed. **Robots are on the plan panel** — a turned marker with a bow for heading, in the one hue left that is neither the people-blue nor the camera-amber. Review findings cleared across the four new sky/robot modules: a real 24 mm void at the ankle and an 8 mm hole through each forearm, a per-frame allocation in the robot heading picker, a per-frame sort in the starlink ranking, `uTime` growing unbounded until the cloud breath quantises, and `DAY_REFERENCE`'s derivation which did not reproduce. `createStarlinkMeshLayer` now takes a **board** radius — the same unit its sibling takes — instead of a dome radius with nothing in the types to tell them apart. That is the exact confusion that has already caused one real bug here. `DOME_RADIUS_FACTOR` has one owner and is imported, not copied: the points and the meshes must be on the same dome or a satellite that grows geometry also jumps. Several comments were wrong rather than merely stale — a fabricated claim about `Object3D.clone`, a fabricated attribution to `Plan`, an inverted `DoubleSide` argument, a triangle ledger citing a function that no longer exists, and a defensive-call rationale that contradicted the paragraph above it. In a codebase where the comments are the design record, those are defects. Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com> |
||
|
karti
|
06455f7424 |
The office learns where it stands, and the sky stops being a backdrop
**Aircraft actually move now, and the reason they did not is the headline.**
`HttpFlights` holds a frozen snapshot between network refreshes and is
polled at 1 Hz, so a live feed handed the layer the same position five to
fifteen times and then jumped. `span` therefore measured the poll interval
rather than the gap between the two positions that differ, the teleport
test saw an airliner covering 36 units in a "second" against a ceiling of
8, and **every live track's history was wiped on every refresh** — so no
aircraft on the deployed site could ever grow a trail, however long
TRAIL_POINTS was set. Skipping the repeat fixes the motion and the trail
at once. Trails then go to 72 points / 240 s, which is about seventy
seconds of flying.
Three more defects in the same file, found while looking: trail
truncation dropped the segments nearest the aircraft (leaving a streak
with no aeroplane attached), MAX_TRACKS was declared and never enforced,
and one missing target deleted its whole trail. The buffer now uploads
only what it wrote, rather than 46 MB/s of untouched array.
**You can get above the constellation.** Dome to 1.05 board *radii* and
the orbit to 2.0 spans. Radii, not spans: scene space is centred on the
city and the Bay Area board runs forty kilometres down the peninsula, so
the furthest corner is 0.94 spans out where the half-diagonal is 0.65 —
sized off the half-diagonal the dome sat inside its own city. The far
plane goes to 4 spans to stop clipping the sky from off-centre chapters,
and `PointsMaterial` defaults `fog: true`, which was quietly dimming the
whole constellation with the city's haze.
**An office can say where it stands.** `Office.site` — lat, lng, height
above the ground outside, and the compass bearing the pack's −Z points
along — and with one it gets the same sun the city does, a sky, and a
horizon at `-elevation`. CONTRACT §4 reserved this as "a later
refinement"; it is taken up rather than overturned, and `daylight.ts`
computes no light of its own. It does the two things a room needs that a
map does not: turn the sun into the building's frame, and move the fog
outdoors before it greys out the far wall.
Two buildings now, and they are deliberately unalike: Lumbridge HQ 188 m
up a Transbay tower facing 205°, and **Frontier Valley**, a startup in a
hangar at Alameda Point — one room, 54 x 30 m, nine metres to the
trusses, four metres above reclaimed ground.
Floor-to-floor in the reference pack is now 16.8 m: the interstitial is
ten times a real one, so the space between the slabs is somewhere things
can hang. It is frankly not architecture, `PLENUM` is the one number to
change, and the file says so.
Also fixed, all found by review rather than by looking at the screen:
- `sun.shadow.camera.updateProjectionMatrix()` was never called, so
three's default ±5 unit box has been in force this whole time and
every `shadowExtent` this repo passes — including the city's ±752 —
has been silently ignored.
- A missing aircraft was kept alive by the new grace period and *drawn*,
so it froze in mid-air at full opacity for 32 s.
- Frontier Valley's mezzanine was a `Room`, which carries no height: its
slab lay on the concrete, its chairs floated 4.4 m over it, and its
balustrade fenced off a patch of ground floor. It is a `Level`.
- Overlapping floor slabs z-fought. The format permits overlap and
resolves later-first, so `shell.ts` now lifts a slab a hair per
earlier slab it overlaps — and by nothing at all in a pack, like the
reference office, whose rooms only ever abut.
- `switchOffice` bypassed the `entering` guard (leaking a whole scene
per double-click) and tore down the old room before knowing the new
one would load, with no way back.
Known and not fixed: raising MAX_SPAN to 30 s doubles the worst-case
re-base snap when a feed's gap shortens. It is bounded, pre-existing in
kind, and the fix wants carrying the live head into the next leg.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
||
|
karti
|
e41c90fe8d |
Real weather, real aircraft, a heightfield off the main thread, and instruments
Three things that were built and never connected, connected.
**The weather was already there.** `observe()` has always taken a
`WeatherObservation` and `main.ts` has always passed null, so the cloud,
precipitation, visibility and marine-layer paths in atmosphere.ts had never run
outside a test. The server already shipped NWS, met.no and Open-Meteo, all
configured off. What was actually missing was that a single TERA_ORIGIN_LAT/LNG
served one metro and lied to the other — so weather and traffic are per-region
now, derived from the city's own bounds, and the Bay Area gets its fog while
Long Beach gets its own sky. The route takes ?city= or a validated ?lat=&lng=
and refuses to become an open geocoding proxy for the planet.
**The heightfield moved to a Worker.** 2.3 s of blocked main thread at boot, and
another ~950 ms of point-in-polygon on top of it: the park mask is filled in the
worker now, and block placement samples four corners and only runs the exact
test on a cell that straddles an edge — 8 buildings differ out of 185,036.
createScene is async and takes a Stage as a consequence, and there is a
main-thread fallback because "clone it and it works" has no exception clause.
**Spaces is a chunk you fetch when you reach for the door**, not one everybody
downloads. Same for the godmode tools. The entry chunk is 722 kB rather than
772; three.js is most of what is left and splitting it is a different job.
**Godmode is an instrument panel now** rather than one slider: the date and the
season, not just the hour, so the Meeus moon and the sun's seasonal arc become
visible instead of merely correct; a weather override that says on screen when
it is lying; a frame-time and draw-call readout; and a pose editor that emits a
paste-ready Chapter block, which is the thing that makes adding New York cheap.
Two blockers the review caught:
- Every city switch leaked 8 GPU textures — one of them a 2048x2048 shadow map
— and ~10.5 shader programs, and deleteTexture had never been called once in
the app's lifetime. The renderer was being built per scene; it belongs to the
canvas, for the life of the page.
- An upstream fetch that threw rather than returning null skipped the cache
stamp, so the TTL — the only rate limit on outbound calls — collapsed to one
upstream request per inbound request, and the caller got a 500.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
|
||
 Karti Tripathi
|
d464459838 |
Spaces: the inside of the world, and a sun that is actually where it should be
Ten agents wrote this in parallel against CONTRACT.md, which exists because the five design agents before them collided on fifteen blocking points — four files specified twice with incompatible contents, three separate backends for one box, and `Environment` exported twice meaning different things. What landed: a Stage owning only the renderer and the loop, with the city and an office as two scenes over it. They cannot share one — San Francisco is ~94 m per scene unit with 3.6x vertical exaggeration and an office is 1 unit = 1 m — and the city is paused rather than disposed on the way in, because rebuilding its 336,864-point heightfield costs about a second on the way back out. Offices are data. `src/offices/lumbridge-hq.ts` is fifteen rooms and seventy-six seats, and it is the file a self-hoster copies. Walls are a segment list with 1-D openings, so doors and windows are holes punched in a wall rather than placed objects, and the pass that splits a wall around its openings hands the walk-mode collider its segments for free. The sun is real. `solar.ts` is a NOAA/Meeus implementation with no imports at all — not even three.js — so time of day keeps working on a laptop in a field. Verified against known values: 75.45 degrees at the June solstice in SF, 28.79 at December, sunset at 03:15Z. The first screenshot after wiring it was a black rectangle, which turned out to be correct: it was midnight in San Francisco. Presence binds to a seat id and never to a coordinate. The pack knows where `eng-04` is; who is sitting in it is private data behind an API. Same shape as the marker rule, one level in. Two corrections to ARCHITECTURE.md are in here. Containment does not discharge ODbL — publishing OSM-derived coordinates is Public Use of a Derivative Database wherever the rows live, so the rule is about the geocoder (US Census, public domain) and not the storage. And a person at a desk is not a Marker; markers are geographic. One contract gap surfaced only in a screenshot: two agents read `height` on a viewpoint differently, so the establishing shot aimed at empty air fourteen metres above the roof. It now means what the same field means for a city. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> |